APA Handbook of Industrial and Organizational Psychology (3 Volume Set) [1-3] 9781433807275, 1433807270

Citation preview

Copyright American Psychological Association. Not for further distribution.

APA Handbook of

Industrial and Organizational Psychology

APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

Copyright American Psychological Association. Not for further distribution.

American Psychological Association • Handbooks in Psychology

Copyright American Psychological Association. Not for further distribution.

APA Handbook of

Industrial and Organizational Psychology volume 1 Building and Developing the Organization

Sheldon Zedeck Editor-in-Chief

American Psychological Association • Washington, DC

Copyright © 2011 by the American Psychological Association. All rights reserved. Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, including, but not limited to, the process of scanning and digitization, or stored in a database or retrieval system, without the prior written permission of the publisher.

Copyright American Psychological Association. Not for further distribution.

Published by American Psychological Association 750 First Street, NE Washington, DC 20002-4242 www.apa.org To order APA Order Department P.O. Box 92984 Washington, DC 20090-2984 Tel: (800) 374-2721; Direct: (202) 336-5510 Fax: (202) 336-5502; TDD/TTY: (202) 336-6123 Online: www.apa.org/books/ E-mail: [email protected] In the U.K., Europe, Africa, and the Middle East, copies may be ordered from American Psychological Association 3 Henrietta Street Covent Garden, London WC2E 8LU England AMERICAN PSYCHOLOGICAL ASSOCIATION STAFF Gary R. VandenBos, PhD, Publisher Julia Frank-McNeil, Senior Director, APA Books Theodore J. Baroody, Director, Reference, APA Books Shenyun Wu, Project Coordinator, APA Books Typeset in Berkeley by Circle Graphics, Inc., Columbia, MD Printer: Edwards Brothers, Ann Arbor, MI Cover Designer: Naylor Design, Washington, DC Library of Congress Cataloging-in-Publication Data APA handbook of industrial and organizational psychology / Sheldon Zedeck, editor-in-chief. -- 1st ed. p. cm. -- (APA Handbooks in psychology) Includes bibliographical references and index. ISBN-13: 978-1-4338-0727-5 ISBN-10: 1-4338-0727-0 ISBN-13: 978-1-4338-0732-9 (vol. 1) ISBN-10: 1-4338-0732-7 (vol. 1) [etc.] 1. Psychology, Industrial. 2. Organizational behavior. I. Zedeck, Sheldon. II. American Psychological Association. III. Title: Handbook of industrial and organizational psychology. HF5548.8.A684 2011 158.7--dc22 2009048439 British Library Cataloguing-in-Publication Data A CIP record is available from the British Library.

Printed in the United States of America First Edition

To the Zedeck Organization

Copyright American Psychological Association. Not for further distribution.

To my wife, Marti; to my children Cindy, Jason, and Tracy; to my daughter-in-law Stacey Skura Zedeck; to my son-in-law Jason Singer; and to the future Zedeck organization—the grandkids—Molly, Ella, Aidan, and Noah: Thanks for the support, patience, love, and joy that you have always provided me, which were especially appreciated during this endeavor. I look forward to continuing to build, develop, grow, and sustain the Zedeck organization!

Copyright American Psychological Association. Not for further distribution.

Copyright American Psychological Association. Not for further distribution.

Contents

Volume 1: Building and Developing the Organization Editorial Board . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix About the Editor-in-Chief . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Series Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi Part I. Foundational Issues in Industrial and Organizational Psychology . . . . . . . . . . 1 Chapter 1. A Historical Survey of Research and Practice in Industrial and Organizational Psychology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Andrew J. Vinchur and Laura L. Koppes Chapter 2. Research Strategies in Industrial and Organizational Psychology: Nonexperimental, Quasi-Experimental, and Randomized Experimental Research in Special Purpose and Nonspecial Purpose Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Eugene F. Stone-Romero Chapter 3. Qualitative Research Strategies in Industrial and Organizational Psychology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Thomas W. Lee, Terrence R. Mitchell, and Wendy S. Harman Chapter 4. Advances in Analytical Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 David Chan Part II. Perspectives on Designing Organizations and Human Resource Systems . . 115 Chapter 5. Organizations: Theory, Design, Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 George P. Huber Chapter 6. Strategic Decision Making. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Philip Bromiley and Devaki Rau Chapter 7. Leadership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 Julian Barling, Amy Christie, and Colette Hoption Chapter 8. Entrepreneurship: The Genesis of Organizations . . . . . . . . . . . . . . . . . . . . 241 Robert A. Baron and Rebecca A. Henry

vii

Contents

Chapter 9. Deepening Our Understanding of Creativity in the Workplace: A Review of Different Approaches to Creativity Research . . . . . . . . . . . . . 275 Jing Zhou and Christina E. Shalley Chapter 10. Performance Measurement at Work: A Multilevel Perspective. . . . . . . . . 303 Jessica L. Wildman, Wendy L. Bedwell, Eduardo Salas, and Kimberly A. Smith-Jentsch Chapter 11. Strategic Reward and Compensation Plans . . . . . . . . . . . . . . . . . . . . . . . . 343 Joseph J. Martocchio Chapter 12. Perspectives on Organizational Climate and Culture . . . . . . . . . . . . . . . . 373 Benjamin Schneider, Mark G. Ehrhart, and William H. Macey

Copyright American Psychological Association. Not for further distribution.

Part III. Designing Work and Structuring Experiences . . . . . . . . . . . . . . . . . . . . . . 415 Chapter 13. Work Matters: Job Design in Classic and Contemporary Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417 Adam M. Grant, Yitzhak Fried, and Tina Juillerat Chapter 14. Workplace Safety and Accidents: An Industrial and Organizational Psychology Perspective. . . . . . . . . . . . . . . . . . . . . . . . . . . 455 Seth Kaplan and Lois E. Tetrick Chapter 15. Disability and Employment: New Directions for Industrial and Organizational Psychology . . . . . . . . . . . . . . . . . . . . . . . . 473 Adrienne J. Colella and Susanne M. Bruyère Chapter 16. Role Theory in Organizations: A Relational Perspective. . . . . . . . . . . . . . 505 David M. Sluss, Rolf van Dick, and Bryant S. Thompson Chapter 17. Flexible Work Schedules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535 Ellen Ernst Kossek and Jesse S. Michel Chapter 18. Nonstandard Workers: Work Arrangements and Outcomes . . . . . . . . . . 573 Elizabeth George and Carmen Kaman Ng Chapter 19. Team Development and Functioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . 597 Janis A. Cannon-Bowers and Clint Bowers Chapter 20. Work Team Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 651 Susan E. Jackson and Aparna Joshi

viii

Copyright American Psychological Association. Not for further distribution.

Editorial Board

EDITOR-IN-CHIEF Sheldon Zedeck, PhD, Vice Provost for Academic Affairs and Faculty Welfare and Professor of Psychology, University of California, Berkeley ASSOCIATE EDITORS Herman Aguinis, PhD, Dean’s Research Professor, and Professor of Organizational Behavior and Human Resources, Kelley School of Business, Indiana University, Bloomington Wayne F. Cascio, PhD, Robert H. Reynolds Chair in Global Leadership, University of Colorado, Denver Michele J. Gelfand, PhD, Professor of Organizational Psychology, University of Maryland, College Park Kwok Leung, PhD, Chair Professor, Department of Management, City University of Hong Kong, Kowloon, Hong Kong Sharon K. Parker, PhD, Director of the Institute of Work Psychology, and Professor of Organizational Psychology, University of Sheffield, Sheffield, England Jing Zhou, PhD, Houston Endowment Professor of Organizational Behavior, Jesse H. Jones Graduate School of Management, Rice University, Houston, TX

ix

Copyright American Psychological Association. Not for further distribution.

Copyright American Psychological Association. Not for further distribution.

About the Editor-in-Chief

Sheldon Zedeck, PhD, is a professor of psychology in the Department of Psychology at the University of California at Berkeley and vice provost for academic affairs and faculty welfare. He has been at Berkeley since 1969, when he completed his doctoral degree in industrial and organizational psychology at Bowling Green State University in Ohio. He served as chair of the department from 1993 to 1998 (and as interim chair for the 2003–2004 year); prior to this administrative position, Dr. Zedeck was the director of the campus’ Institute of Industrial Relations from 1988 to 1992. Dr. Zedeck is coauthor of four books on various topics: Foundations of Behavioral Science Research in Organizations (1974, with Milton Blood), Measurement Theory for the Behavioral Sciences (1981, with Edwin E. Ghiselli and John Campbell), Performance Measurement and Theory (1983, with Frank Landy and Jan Cleveland), and Data Analysis for Research Designs (1989, with Geoffrey Keppel). In addition, he has edited a volume titled Work, Family, and Organizations (1992), which is part of the Society for Industrial and Organizational Psychology Frontiers series. Dr. Zedeck has served on the editorial boards of the Journal of Applied Psychology (editor, 2002–2008), Contemporary Psychology, and Industrial Relations. He has also served as editor and associate editor of Human Performance, a journal that he and Frank Landy founded in 1988, as well as associate editor of Applied Psychology: An International Review. Dr. Zedeck has been active in the American Psychological Association (APA) Division 14 (Society for Industrial and Organizational Psychology). He has been on the Society’s Educational and Training Committee and its Workshop Committee; has been a member-at-large; has served as editor of the Society’s newsletter, TIP; has served on two ad hoc committees concerned with revising the Society’s “Principles for the Validation and Use of Personnel Selection Procedures”; has represented the Society on the APA Council of Representatives; and in 1986–1987 served as the president of the Society. He has also served on the executive committees for the Academy of Management’s Personnel/ Human Resources Division and for the Society for Organizational Behavior. Dr. Zedeck has written numerous journal articles on the topics of moderator variables, selection and validation, test fairness, banding, performance appraisal, assessment centers, stress, and work and family issues. His most recent research project was a 9-year project on the identification of factors and criteria of lawyering success and the development and validation of tests that can be used as complements to the Law School Admission Test.

xi

Copyright American Psychological Association. Not for further distribution.

Copyright American Psychological Association. Not for further distribution.

Contributors

Herman Aguinis, PhD, Kelley School of Business, Indiana University, Bloomington Karl Aquino, PhD, Sauder School of Business, University of British Columbia, Vancouver, Canada Winfred Arthur Jr., PhD, Department of Psychology, Texas A&M University, College Station Derek R. Avery, PhD, C. T. Bauer College of Business, University of Houston, Houston, TX Laurie J. Barclay, PhD, School of Business and Economics, Wilfrid Laurier University, Waterloo, Ontario, Canada Julian Barling, PhD, School of Business, Queen’s University, Kingston, Ontario, Canada Robert A. Baron, PhD, Spears School of Business, Oklahoma State University, Stillwater Yehuda Baruch, PhD, Norwich Business School, University of East Anglia, Norwich, England Talya N. Bauer, PhD, School of Business Administration, Portland State University, Portland, OR Wendy L. Bedwell, Department of Psychology and Institute for Simulation and Training, University of Central Florida, Orlando Jennifer L. Berdahl, PhD, Joseph L. Rotman School of Management, University of Toronto, Toronto, Ontario, Canada Uta K. Bindl, Institute of Work Psychology, University of Sheffield, Sheffield, England John W. Boudreau, PhD, Marshall School of Business, University of Southern California, Los Angeles Clint Bowers, PhD, Department of Psychology and Institute for Simulation and Training, University of Central Florida, Orlando Nikos Bozionelos, PhD, Durham Business School, University of Durham, Durham, England Philip Bromiley, PhD, Merage School of Business, University of California, Irvine Kenneth G. Brown, PhD, Henry B. Tippie College of Business, The University of Iowa, Iowa City Susanne M. Bruyère, PhD, Employment and Disability Institute, Cornell University, Ithaca, NY Janis A. Cannon-Bowers, PhD, Department of Psychology and Institute for Simulation and Training, University of Central Florida, Orlando Peter Cappelli, PhD, The Wharton School, University of Pennsylvania, Philadelphia Wayne F. Cascio, PhD, Business School, University of Colorado, Denver xiii

Contributors

Copyright American Psychological Association. Not for further distribution.

David Chan, PhD, School of Social Sciences, Singapore Management University Megan M. Chandler, PhD, Department of Psychology, University of Akron, Akron, OH Oleksandr S. Chernyshenko, PhD, Nanyang Business School, Nanyang Technological University, Singapore Amy Christie, PhD, School of Business, Queen’s University, Kingston, Ontario, Canada Sharon Clarke, PhD, Manchester Business School, University of Manchester, Manchester, England Adrienne Colella, PhD, A. B. Freeman School of Business, Tulane University, New Orleans, LA Satoris S. Culbertson, PhD, Psychology Department, Kansas State University, Manhattan Guangrong Dai, PhD, Korn/Ferry Leadership International, Minneapolis, MN Eric Anthony Day, PhD, Department of Psychology, University of Oklahoma, Norman Carsten K. W. De Dreu, PhD, Faculty of Social and Behavioural Sciences, University of Amsterdam, Amsterdam, the Netherlands Kenneth P. De Meuse, PhD, Korn/Ferry International, Minneapolis, MN Angelo S. DeNisi, PhD, A. B. Freeman School of Business, Tulane University, New Orleans, LA James M. Diefendorff, PhD, Department of Psychology, University of Akron, Akron, OH Erich C. Dierdorff, PhD, Kellstadt Graduate School of Business, DePaul University, Chicago, IL Brian R. Dineen, PhD, Management Area, Gatton College of Business and Economics, University of Kentucky, Lexington Fritz Drasgow, PhD, Department of Psychology, University of Illinois at Urbana–Champaign Lillian T. Eby, PhD, Department of Psychology, University of Georgia, Athens Mark G. Ehrhart, PhD, Department of Psychology, San Diego State University, San Diego, CA Berrin Erdogan, PhD, School of Business Administration, Portland State University, Portland, OR Miriam Erez, PhD, Industrial and Engineering Management, Technion—Israel Institute of Technology, Haifa Gerald R. Ferris, PhD, Department of Management, College of Business, Florida State University, Tallahassee Kevin E. Fox, PhD, Department of Psychology, Saint Louis University, Saint Louis, MO Yitzhak Fried, PhD, Whitman School of Management, Syracuse University, Syracuse, NY Ashley Fulmer, PhD, Department of Psychology, University of Maryland, College Park Michele J. Gelfand, PhD, Department of Psychology, University of Maryland, College Park Elizabeth George, PhD, Department of Management, School of Business and Management, Hong Kong University of Science and Technology, Kowloon, Hong Kong Robyn E. Goodwin, PhD, Australian School of Business, The University of New South Wales, Sydney, Australia Adam M. Grant, PhD, The Wharton School, University of Pennsylvania, Philadelphia Jerald Greenberg, PhD, RAND Corporation, Santa Monica, CA Mark A. Griffin, PhD, School of Psychology, University of Western Australia, Crawley, Australia. Markus Groth, PhD, Australian School of Business, The University of New South Wales, Sydney, Australia xiv

Copyright American Psychological Association. Not for further distribution.

Contributors

Russell P. Guay, Henry B. Tippie College of Business, University of Iowa, Iowa City Leslie B. Hammer, PhD, Department of Psychology, Portland State University, Portland, OR S. Duane Hansen, PhD, Krannert School of Management, Purdue University, West Lafayette, IN Wendy S. Harman, PhD, Foster School of Business, University of Washington, Seattle Rebecca A. Henry, PhD, Lally School of Management and Technology, Rensselaer Polytechnic Institute, Troy, NY M. Sandy Hershcovis, PhD, Department of Business Administration, I. H. Asper School of Business, University of Manitoba, Winnipeg, Manitoba, Canada Wayne A. Hochwarter, PhD, Department of Management, College of Business, Florida State University, Tallahassee Joyce Hogan, PhD, Hogan Assessment Systems, Tulsa, OK Robert Hogan, PhD, Hogan Assessment Systems, Tulsa, OK Peter W. Hom, PhD, Department of Management, W. P. Carey School of Business, Arizona State University, Phoenix Colette Hoption, PhD, School of Business, Queen’s University, Kingston, Ontario, Canada Leaetta M. Hough, PhD, The Dunnette Group, Saint Paul, MN George P. Huber, PhD, McCombs School of Business, University of Texas at Austin Allen I. Huffcutt, PhD, Department of Psychology, Bradley University, Peoria, IL Susan E. Jackson, PhD, School of Management and Labor Relations, Rutgers University, Piscataway, NJ Aparna Joshi, PhD, School of Labor and Employment Relations, University of Illinois at Urbana–Champaign Tina Juillerat, Kenan-Flagler Business School, University of North Carolina at Chapel Hill Rob Kaiser, PhD, Kaplan DeVries Inc., Greensboro, NC Seth Kaplan, PhD, Department of Psychology, George Mason University, Fairfax, VA Laura L. Koppes, PhD, Department of Psychology, University of West Florida, Pensacola Ellen Ernst Kossek, PhD, School of Labor and Industrial Relations, Michigan State University, East Lansing Amy Kristof-Brown, PhD, Henry B. Tippie College of Business, University of Iowa, Iowa City Thomas W. Lee, PhD, Michael G. Foster School of Business, University of Washington, Seattle Kwok Leung, PhD, College of Business, Department of Management, City University of Hong Kong, Kowloon, Hong Kong William H. Macey, PhD, Valtera Corporation, Rolling Meadows, IL William I. MacKenzie Jr., PhD, Management Department, Moore School of Business, University of South Carolina, Columbia Mitchell Lee Marks, PhD, College of Business, San Francisco State University, San Francisco, CA Luis L. Martins, PhD, College of Management, Georgia Institute of Technology, Atlanta Joseph J. Martocchio, PhD, Institute of Labor and Industrial Relations, University of Illinois at Urbana–Champaign xv

Contributors

Copyright American Psychological Association. Not for further distribution.

Jesse S. Michel, PhD, Department of Psychology, Florida International University, Miami Terrence R. Mitchell, PhD, Michael G. Foster School of Business, University of Washington, Seattle Frederick P. Morgeson, PhD, Eli Broad Graduate School of Management, Michigan State University, East Lansing Carmen Kaman Ng, PhD, Department of Management, School of Business and Management, Hong Kong University of Science and Technology, Kowloon, Hong Kong Dennis W. Organ, PhD, Department of Management, Kelley School of Business, Indiana University, Bloomington Frederick L. Oswald, PhD, Department of Psychology, Rice University, Houston, TX James L. Outtz, PhD, Outtz and Associates, Washington, DC Sharon K. Parker, PhD, Institute of Work Psychology, University of Sheffield, Sheffield, England David B. Peterson, PhD, PDI Ninth House, San Francisco, CA Mark F. Peterson, PhD, College of Business, Florida Atlantic University, Boca Raton Robert E. Ployhart, PhD, Management Department, Moore School of Business, University of South Carolina, Columbia Nathan P. Podsakoff, PhD, Department of Management and Organizations, Eller College of Management, University of Arizona, Tucson Philip M. Podsakoff, PhD, Department of Management, Kelley School of Business, Indiana University, Bloomington Marshall Scott Poole, PhD, Center for Supercomputing Applications, University of Illinois, Urbana Devaki Rau, PhD, College of Business, Northern Illinois University, DeKalb Jana Raver, PhD, Queen’s School of Business, Queen’s University, Kingston, Ontario, Canada Tara C. Reich, PhD, Department of Business Administration, I. H. Asper School of Business, University of Manitoba, Winnipeg, Manitoba, Canada Denise M. Rousseau, PhD, H. John Heinz II College of Public Policy, Management, and Information, Tepper School of Business, Carnegie Mellon University, Pittsburgh, PA Eduardo Salas, PhD, Department of Psychology and Institute for Simulation and Training, University of Central Florida, Orlando Deidra J. Schleicher, PhD, Krannert School of Management, Purdue University, West Lafayette, IN Neal Schmitt PhD, Department of Psychology, Michigan State University, East Lansing Benjamin Schneider, PhD, Valtera Corporation, Rolling Meadows, IL, and Department of Psychology, University of Maryland, College Park Christina E. Shalley, PhD, College of Management, Georgia Institute of Technology, Atlanta Laura Severance, PhD, Department of Psychology, University of Maryland, College Park Ruchi Sinha, PhD, Department of Psychology, Michigan State University, East Lansing Traci Sitzmann, PhD, University of Colorado, Denver David M. Sluss, PhD, Moore School of Business, University of South Carolina, Columbia Kimberly A. Smith-Jentsch, Department of Psychology, University of Central Florida, Orlando xvi

Copyright American Psychological Association. Not for further distribution.

Contributors

Scott M. Soltis, Management Area, Gatton College of Business and Economics, University of Kentucky, Lexington Shirley Sonesh, PhD, A. B. Freeman School of Business, Tulane University, New Orleans, LA Stephen Stark, PhD, Department of Psychology, University of South Florida, Tampa Eugene F. Stone-Romero, PhD, Department of Management, University of Texas, San Antonio Lois E. Tetrick, PhD, Department of Psychology, George Mason University, Fairfax, VA Bryant S. Thompson, PhD, Management Department, Moore School of Business, University of South Carolina, Columbia Donald M. Truxillo, PhD, Department of Psychology, Portland State University, Portland, OR Rolf van Dick, PhD, Institute of Psychology, Goethe University, Frankfurt, Germany Andrew J. Vinchur, PhD, Department of Psychology, Lafayette College, Easton, PA Jessica L. Wildman, Department of Psychology and Institute for Simulation and Training, University of Central Florida, Orlando Jing Zhou, PhD, Jesse H. Jones Graduate School of Management, Rice University, Houston, TX Kristi Zimmerman, PhD, Department of Psychology, Portland State University, Portland, OR

xvii

Copyright American Psychological Association. Not for further distribution.

Copyright American Psychological Association. Not for further distribution.

Series Preface

The APA Handbook of Industrial and Organizational Psychology is the first publication to be released in the American Psychological Association’s latest reference line, the APA Handbooks of Psychology series. The series will comprise multiple two- and three-volume sets focused on core subfields, and sets will be issued individually over the next several years. Some 20 are currently envisioned, with more than half already commissioned and in various stages of development and completion. Additionally, some single-volume handbooks on highly focused content areas within core subfields will be released in conjunction with the series. Thus, the APA Handbooks in Psychology series now joins APA’s three critically acclaimed, award-winning, and best-selling dictionaries—the APA Dictionary of Psychology (2006), the APA Concise Dictionary of Psychology (2008), and the APA College Dictionary of Psychology (2009)—as part of a growing suite of distinctly reference literature. Specifically, each handbook set is formulated primarily to address the reference interests and needs of researchers, clinicians, and practitioners in psychology and allied behavioral fields. Second, each set will bear strong interest for professionals in pertinent complementary fields (i.e., by content area), be they corporate executives and human resources personnel (as with the set in hand); doctors, psychiatrists, and other health personnel; teachers and school administrators; cultural diversity and pastoral counselors, and so forth. Third—and not least important—the entire series is geared to graduate students in psychology who require well-organized, detailed supplementary texts, not only for “filling in” their own specialty areas but also for gaining sound familiarity with other established specialties and emerging trends across the breadth of psychology. Under the direction of small and select editorial boards consisting of top scholars in the field, with chapters authored by both senior and rising researchers and practitioners, each reference set is committed to a steady focus on best science and best practice. Coverage converges on what is currently known in the particular subject area (including basic historical reviews) and the identification of the most pertinent sources of information in both core and evolving literature. Volumes and chapters alike pinpoint practical issues; probe unresolved and controversial topics; and present future theoretical, research, and practice trends. The editors provide clear guidance to the “dialogue” among chapters, with internal cross-referencing that demonstrates a robust integration of topics that leads the user to a clearer understanding of the complex interrelationships within each field.

xix

Series Preface

With the imprimatur of the largest scientific and professional organization representing psychology in the United States and the largest association of psychologists in the world, and with content edited and authored by some of its most respected members, the APA Handbooks of Psychology series will be the indispensable and authoritative reference resource to turn to for researchers, instructors, practitioners, and field leaders alike.

Copyright American Psychological Association. Not for further distribution.

Gary R. VandenBos, PhD APA Publisher

xx

Copyright American Psychological Association. Not for further distribution.

Introduction

Introductions to handbooks—whether single volumes or large-scale, multivolume reference works like this—often begin with an explanation of why a handbook is needed. This concern is especially germane given that there have been several explorations of the discipline over the past 2 decades. Our (i.e., the editorial board’s) response to this concern is rather brief and to the point. The field of industrial and organizational (I/O) psychology is rapidly evolving and has entered a new frontier: The world of work and workers is changing; approaches to studying issues are becoming more diverse, more multidisciplinary, and more interdisciplinary; and the study has broadened to include individuals, teams, organizations, environments, cultures, and societies. The goal of this handbook is to capture these current changes and the implications that they have for the research and practice of I/O psychology. This handbook presents what is currently known and, perhaps more important, suggests avenues for further pursuit in light of the conditions existing today and aims to educate and inform readers about the field and how it might have an impact on the future. We hope that the approach of this handbook, as detailed in the remainder of this introduction, presents a more elaborate rationale for the need for another handbook. WHAT IS INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY? The discipline of I/O psychology, which is a subfield of psychology in general, is quite broad. Whereas psychology may generally be defined as a discipline that attempts to understand behavior, I/O psychology focuses on understanding the behavior of those who are in a working situation or organization. The latter focus may appear to be a narrow one but, in fact, it is quite the opposite. I/O psychology attempts to study behavior from many different perspectives, within psychology as well as from other scientific disciplines. As this brief definition notes, I/O psychology focuses on three aspects: (a) the person, the worker; (b) the work (tasks) that is (are) being performed; and (c) the context in which the work is performed. I/O psychology studies the work lives of people as well as how they are influenced by other domains of life, such as family and the broader society and culture in which they live and work. Its purpose is to understand the reasons for behavior (performance of tasks) in a work setting; how people can become effective, satisfied, fulfilled, and rewarded; and how these outcomes can be maintained. As for the organizational side of the equation, I/O psychology studies how the organization can be sustained

xxi

Introduction

Copyright American Psychological Association. Not for further distribution.

and developed. The focus of I/O psychology also involves interactions: I/O psychology studies people—the psychological processes—and the settings in which people work and function, to gain a better understanding of behavior and how it can be influenced, changed, and enhanced to benefit the individuals as well as the organization and society as a whole. In brief, I/O psychology is the understanding and application of psychological principles, theory, research, interventions, and practice, with the target being a particular setting, an organization, which is broadly defined (to include society in general). This simple definition of I/O psychology can be enriched by starting at the beginning— understanding behavior. For I/O psychologists, the behavior of interest occurs in a context and is a study of particular individuals—those who work in a setting (e.g., an organization) and have a goal for that organization. This expansion suggests that the domain of study can be at a minimum two people who come together to form an organization of some kind to pursue a goal. The organization can be structured or unstructured, formal or informal, collocated or virtual, profit or nonprofit. Employees can be volunteer or paid, permanent or temporary. The organization can be part of private industry or part of the government. Regardless of the setting, the domain of I/O psychology has as its objective to understand the behavior (antecedents and consequences) of the participants in these settings: the workers, the employers, the owners, the leaders, and the like. Behavior is often defined as “observable actions,” but I/O psychologists study more than observable actions. They also study intentions, attitudes, emotions, habits, motives, values, beliefs, and any other personal construct that can be used to describe people. They study knowledge, abilities, skills, and other characteristics of people that influence what someone does and thinks of in a work setting. Work also needs to be defined and its place in the discipline elaborated on. Whereas work is usually a set of tasks that someone undertakes in an attempt to achieve a goal, I/O psychologists also study why that work is performed (e.g., motivation), how one may have learned about the work (e.g., vocational psychology, training), and why one may be committed to the work, to the organization, or to both. In essence, they study not only the knowledge, skills, abilities, and other characteristics of the people doing the work but also the factors from within and without the organization that bring people to the work, keep them at work, lead them to shape or change the work, and generate allegiance to the work. As noted earlier, the work is studied in a context, often in a social system where there is some structure, assignment, coordination, and control of the operations. Again, the specific form of the organization can be of different types: private, public, solely owned, or family owned. The specific setting interacts with the people in it as well as those served by the organization (clients) and the requisite work activities. The ability to understand one part requires the study and understanding of the other aspects. I/O psychologists study issues within-persons, at the “people level” of individual differences, at the group or team level, and at the broader, organizational level. The organizational level also is influenced by societal culture, government regulation and intervention, economics, politics, and other societal conditions and factors that influence how and why people behave the way they do. This perspective may suggest that the societal conditions and factors are antecedents, but that would be a particularly narrow and misleading view: There are reciprocal relationships, so an entity such as societal context is an antecedent as well as a consequent. In essence, I/O psychology is a multilevel and multiattribute study of behavior. xxii

Introduction

Copyright American Psychological Association. Not for further distribution.

APPLIED VERSUS BASIC RESEARCH I/O psychology seeks to advance theory as well as to produce strategies and systems that are relevant for and applicable to individuals, teams, organizations, and societies. This perspective is counter to what many consider I/O psychology to be: that is, a predominantly “applied” field. This view is fostered, in part, because one of the main outlets for publishing research conducted by I/O psychologists is a journal titled Journal of Applied Psychology. Our view is that an “applied psychology” perspective is not only limiting but also incorrect. What differentiates I/O psychology is its reliance on basic research and theory. What I/O psychology has produced may be seen as purely applied psychology, and in some instances it is very applied. At the same time, however, the field has generated and presented its results on the basis of theoretical considerations. As an example, consider the “two-factor theory of job satisfaction” research program of Herzberg1 in the 1950s and 1960s, in which he proposed the notion that when someone is satisfied with his or her work, it is due to internal factors such as the intrinsic value or nature of the task being performed, but when someone is dissatisfied, it is the result of an external agent or other individual, such as the supervisor or organization. In essence, whereas success or satisfaction is attributed to oneself or to factors within one’s control, failure or dissatisfaction is attributed to others. This explanation of behavior has basically been described as a job satisfaction theory, which confines it to a particular context—the work setting. Close examination of Herzberg’s research, methodology, and theory, however, reveals that it is somewhat parallel to a theoretical development generated in social and personality psychology, attribution theory, which came into dominance in social psychology at about the same time or later in the works of Heider2 and Jones, Kelley, Weiner, and others.3 In contrast to Herzberg’s applied research, this social psychological research is described by many commentators as “basic research” and is applicable to behavior and settings in general. Perhaps this is because of the laboratory settings in which much of the social psychological research had been conducted, but the theory of attributions is essentially the same whether generated or described in I/O psychology or social psychology domains, or generated in the field (e.g., surveys of employees) or in the laboratory. From another perspective of the “applied versus basic” issue, applied issues studied by I/O psychologists have driven research that has produced theoretical bases of knowledge. For example, the Civil Rights Act of 1964 created a need to understand why particular ethnic and gender groups performed differently—whether performance differences were real or perceived—on selection tests or in their jobs. This applied problem generated research that can be characterized as basic on such topics as cognitive biases, stereotypes, first impressions, impression management—topics often seen as basic domains for social, personality, and cognitive psychology. Thus, one can see that applied problems have a direct link with theoretical, basic research advances. Many basic theories of motivation and performance models have been enhanced or sophisticated by being elaborated on for work situations, whereas observations from the work environment have stimulated basic research on the same topics. The point here,

1

cf. Herzberg, F., Mausner, B., & Snyderman, B. S. (1959). The motivation to work. New York: Wiley.

2

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.

3

Jones, E. E., Kanouse, D., Kelley, H. H., Nisbett, R. E., Valins, S., & Weiner, B. (1972). Attribution: Perceiving the causes of behavior. Morristown, NJ: General Learning Press.

xxiii

Introduction

Copyright American Psychological Association. Not for further distribution.

and as developed further in this handbook’s chapters, is that one cannot isolate research as “applied versus basic”; there is a mutual reliance and dependency, which this handbook’s content acknowledges and adopts. The problems studied by I/O psychologists are grounded in a theoretical foundation. That is, the research problem arises in, or in connection with, theoretical questions or previous research. I/O psychologists build on it by testing it in different situations and under different conditions. This expansion of the problem results in a better understanding of the phenomenon. Likewise, fundamental principles are generated from research that addresses an applied problem; these principles can be generalized to other problems and situations, so long as the problem’s connectivity to other factors is studied and understood. Many I/O psychologists are also change agents; that is, they apply interventions in the pursuit of a solution to a problem. Even when an I/O psychologist is asked to implement an intervention, which appears to be applied psychology at its root level, understanding of the context and situation in which the intervention is to be applied (organizational conditions, characteristics of the group to whom the intervention will be applied) requires knowledge of the underpinnings of the intervention as well as the need for fundamental research to measure its impact. BOUNDARIES OF STUDY As will be seen in the chapters in the handbook, I/O psychology draws on the study, understanding, and application of multiple and varied disciplines, in no particular order, such as sociology, political science, communication studies, law, neuroscience, biology, economics, human resource management, strategy, human factors, and organizational behavior. These fields are integrated with social, personality, cognitive, quantitative, biological, developmental, vocational, and clinical psychology subfields. Thus, one can see that the study of problems pertaining to I/O psychology is untethered and without boundaries. In the end, the topics presented in this handbook should be studied from a multilevel, multidisciplinary, and interdisciplinary perspective. I/O psychologists need to identify a problem, study it from applied and basic perspectives, and involve different disciplines in the process. In summary, I/O psychologists study the behavior of people at work. I/O psychologists derive principles of individual, group, and organization behavior through research. They develop scientific knowledge and apply it to the solution of problems in work settings. The applications are science- and research-based. In the end, the I/O psychologist is the model scientist–practitioner, integrating science and practice so that activities in one domain inform activities in the other domain. PURPOSE OF THE HANDBOOK As noted, I/O psychology is a broad, diverse field that is rapidly changing. It is changing as populations change, the characteristics of the workforce change, the nature of work changes, and the conditions under which and where work is performed change. The goal of this three-volume APA Handbook of Industrial and Organizational Psychology is to capture the diversity and full range of the topics and approaches studied by I/O psychologists and to present findings that define the field of I/O psychology where the intended audience is not only primarily researchers and practitioners in psychology but also proxxiv

Introduction

fessionals in complementary fields, such as corporate executives and human resources professionals. The handbook includes broad as well as specialized issues. Some material is introductory and aimed at the novice, whereas other material is more complex and aimed at the experienced reader. We hope that the reader will gain an appreciation for what I/O psychology can contribute to an understanding of people’s behavior and their contributions to society. We also hope the reader will learn how I/O psychology not only reflects wider changes in work and organizations but also shapes and causes positive change in work and organizations.

Copyright American Psychological Association. Not for further distribution.

ORGANIZATION OF THE HANDBOOK As noted, the purpose of this handbook is to present the types of issues that I/O psychologists study, the questions they pursue, the research they conduct, and the interventions they implement. We present these from both theoretical and applied perspectives. We demonstrate how theory is generated and applied to problems that exist in all types of organizations—problems pertaining to understanding, explaining, and predicting behavior in organizations. The intention is to explore how problems are diagnosed, how research is conducted to generate answers to the problems, what has been found, how interventions are implemented, how they are evaluated, what still needs to be studied, and where the field should be going. Our authors’ focus is on what is currently known, including basic historical reviews, as well as the identification of the most pertinent sources of information in both the core and emerging literatures. Finally, the chapters attempt to pinpoint practical issues and probe unresolved and controversial topics, and present future trends in theory, research, and practice. In developing the outline for the handbook, we generated a list of topics that engage industrial psychologists in study and practice. We did so not only on the assumption that the topics would be presented one at a time but also with the recognition that no topic can stand alone. Our dilemma was to determine a reasonable set of chapters (we identified more than 60 topics) within some broad framework that made sense (to the editor-in-chief!). There were several options, of course, from the reasonably simplistic sequencing in alphabetical order (i.e., the encyclopedia approach) to one that provided a perspective on organizations (admittedly there are many perspectives) with respect to their design, development, and maintenance. As noted, people, work, and organizations are complex and interact in complicated ways. If one attempts to understand or affect a single aspect of behavior (e.g., job satisfaction), one needs to understand that it is likely to affect several other areas, such as compensation, supervision, performance reviews, and the like. If one wants to develop a selection system for an organization, one also needs to think in terms of training and compensation. Should one select those individuals who already are proficient on a task and therefore might demand a higher salary, or should one select those who are trainable, and pay less, while the organization does the training and generates commitment from the employee? Selection itself needs to take into account pre- and postissues. How will the recruiting strategy and test-taking program influence who applies for the position? What influence does the organization’s image have on who applies? Once candidates are hired and trained in the position, how does socialization into the organization influence who provides extra effort and greater commitment to the organization? In practice, all of the aspects of the people, the work, and the setting interact in complicated xxv

Introduction

Copyright American Psychological Association. Not for further distribution.

ways; the topics are inextricably interrelated. In these brief examples, we have mentioned topics such as selection, training, compensation, socialization, job satisfaction, and commitment, to name a few. Individual chapters cover each of these topics, but the reader should always bear in mind that the topics are highly interconnected. They are not mutually exclusive and cannot truly be studied and understood in isolation. To some degree, then, the focus on the interrelatedness or linkages among topics drove the final outline for the handbook. The main perspective was: If one is thinking of becoming involved with an organization (e.g., starting one, buying one, building one), how would one go about doing it and what does one need to know? The approach adopted for this handbook was to consider what one needs to know about peoplerelated business issues if one wanted to start an organization, how one would design it, how one would get members into it, how one would treat members once in the organization, and how one would maintain, sustain, nourish, and develop it. Such considerations yielded a three-volume handbook with a sequence within and among volumes to reflect this perspective: The chapter content flows from the issues associated with building and developing the structure of the organization (Volume 1) to selecting and developing members for the organization and having them become part of it (Volume 2). Once the organization is established and has members with roles, we proceed to topics that pertain to maintaining and ensuring the viability of the organization, which includes issues concerned with growing, expanding, and even contracting the organization (Volume 3). Within each of these volumes, several parts form bundles of issues that need to be considered together. The goal of the entire set of volumes is to provide the reader with an understanding of the complexities and intricacies of trying to study and impact behavior within and between organizations. The chapter sequencing that we chose is just one of many equally plausible options. Moreover, as has been noted, we recognize that the chapters are not mutually exclusive and that the reader cannot concentrate on one chapter without taking into account related and linked material from another. In essence, the sequencing of chapters is a convenience, but one that we hope identifies and distinguishes a focus of concern for study and practice. Volume 1 is titled Building and Developing the Organization. The chapters in this volume discuss the foundation for I/O psychology, the field itself, and then engage the issues that one considers when an individual begins to plan for an organization. Part I presents the foundational issues in I/O psychology. Chapter 1 offers a history of the field; this chapter sets the trend for the way in which “problems” have become the domain of I/O psychology. Chapters 2 through 4 present the research issues and dataanalytic strategies that I/O psychologists use to study the domain’s problems and the approaches to looking at problems, whether at the individual, team, or organizational level. Part II focuses on perspectives on designing organizations and human resource systems. Chapters 5 through 12 focus on broad issues: what type of organization should be constructed (Chapter 5) and how decisions should be made (Chapter 6) and by whom (e.g., leaders; Chapter 7). Chapter 8 focuses on a special type or organization start-up—those initiated by entrepreneurs—as well as who these individuals are and how they function. Chapter 9 focuses on creativity and innovation, which are essential for organizations’ creation (and renewal); discussion focuses on the key enablers and contributors to entrepreneurship, growth, and success. Once the organization is “sketched out” and “filled in,” the next two chapters in Part II turn to performance xxvi

Copyright American Psychological Association. Not for further distribution.

Introduction

measurement and how and at what levels the organization will evaluate itself—whether as individuals, as teams, as an organization, or any combination (Chapter 10)—as well as how it will reward and compensate its members (Chapter 11). Finally, Chapter 12 addresses the issue of organizational climate and culture, which include how these concerns define the organization for the people who work in them and their effectiveness. Part III in Volume I focuses on specific issues identified with designing work and structuring experiences for the members of the organization. Chapter 13 addresses the design of jobs and work. Related to such design issues are concerns for the factors that influence safety in the performance of tasks, such as climate, types of work and schedules, leadership, and the like (Chapter 14). As one plans an organization and its tasks, there should be concern for the integration of people with disabilities into the workforce and for accommodations for this group (Chapter 15). Also related to the design of work are the roles one assumes in the organization; roles define how work is designed, accomplished, evaluated, and experienced (Chapter 16). Chapter 17 focuses on the types of schedules and work arrangements that might fit the designed jobs, and Chapter 18 addresses the issue of whether the workers should work a standard 8-to-5 schedule or whether other arrangements might be more appropriate, such as contingent work, parttime work, or shift work. Beyond the individual, the issues become whether work should be performed in teams (Chapter 19), and if so, what the composition should be for that team, particularly its diversity (Chapter 20). Volume 2 is titled Selecting and Developing Members for the Organization. This volume addresses issues that come into play when the organization’s design and plans have been completed and implemented and it becomes necessary to staff the organization. How will members be selected for the organization, and what issues must one consider as these members become part of it? Part I provides the foundations of selection and development. Chapter 1, on work and job analysis, concerns how jobs are analyzed and studied to ensure that they are performed properly and, if necessary, what changes need to be made. This involves primarily the study of incumbents and how they perform their work activities. This undertaking is critical because the basis for selection and development is keyed to understanding how jobs are performed and the requirements for performing the work. Once there is an understanding of what needs to be done, the organization is concerned with selecting members, evaluating their performance, and training employees to perform the tasks and to be part of the organization. Chapter 2 focuses on recruiting candidates to apply for the positions in the organization. Chapter 3 focuses on career issues, including career progression, how forms of employment affect careers, and the changing nature of careers. Issues identified in this chapter have implications for who will apply to the positions and their likelihood of being selected and, subsequently, being successful in the organization. Part II focuses on specific selection strategies and issues that address particular means for selecting members of the organization. Chapters 4 through 8 discuss types of specific selection devices and strategies that the organization may use to obtain members. Cognitive ability and other skills and personal characteristics (e.g., biographical information) are discussed in the chapter on individual differences (Chapter 4); Chapter 5 devotes attention to the assessment of personality, Chapter 6 focuses on the interview as a selection tool, Chapter 7 presents assessment-center methodology, and Chapter 8 describes one particular type of test, the situational judgment test. The “predictors” are presented as separate chapters, but by no means does this suggest that selection should be based on a xxvii

Introduction

Copyright American Psychological Association. Not for further distribution.

single instrument. Instead, combinations of tests should be considered and examined for their effectiveness in identifying successful future employees of the organization. Part III is concerned with evaluating individuals and performance, particularly with specific emphasis on job-related criteria as determined from the work analysis (Chapter 9). (The reader should note that Chapter 9 in Volume 2 relates to Chapter 10 in Volume 1, with the latter being more general and concerned with “planning the organization,” and how effectiveness can be determined at various levels, whereas the former deals primarily with specific issues and strategies of evaluating incumbent workers.) Chapters 10 and 11 expand on Chapter 9 by focusing on one specific criterion that an organization might employ to evaluate members, such as individual factors that involve organizational citizenship (Chapter 10), and Chapter 11 focuses on “organizational exit,” or turnover, which is another criterion used to evaluate the contribution of employees and the success of the organization. Part IV examines issues pertaining to evaluating systems, which is concerned with how well the organization does with respect to its hiring system and processes in general. A fundamental concern is with applicant reactions to the initial introduction to the organization—the selection process—and in particular, to the specific means by which the organization selects employees (Chapter 12). As part of the natural progression, because decisions as to who joins an organization are based on some of the predictors (tests) identified in the previous set of chapters, it is necessary to determine the appropriateness of those tests for the purpose of selecting the members of the organization. Chapter 13 on validation presents the strategies that organizations use to evaluate their selection system and the degree to which performance can be predicted by the tests. Chapter 14 presents a broad concept of utility for analyzing the selection system as well as other business components. Besides internal, organizational evaluations, society also has a role in evaluating organizations, and it does so, in part, by determining whether selection systems conform to legal requirements. Because of the nature of existing employment law, all of the issues described at the outset of the chapters in Parts I through III of Volume 2 (e.g., work analysis, performance evaluation, recruitment, and testing) become targets for litigation and determination of fairness. Chapter 15 presents an overview of legal cases and guidelines that are keys to selection issues in the United States. The evaluation of the selection system may reveal reasonable success in hiring members, but through its evaluation processes, the organization also identifies those who need development to improve their performance. Part V focuses on developing members. In particular, the topic of training (Chapter 16) is highly related to job analysis, performance evaluation, and staffing; training is concerned with ensuring that those who perform the tasks or staff the positions in the organization meet the criteria for success and effectiveness in that organization. This chapter focuses on training for basic skills as well as development skills needed to grow and enhance performance. Chapter 17 covers mentoring, which is cited in many other chapters as a solution for addressing problems employees may face in their daily work lives and in their development. Chapter 18 focuses on the practice (and theory) of coaching members to promote the likelihood of their success. Finally, Chapter 19 addresses proactive work behavior, that is, how employees go about bringing change to the organization or to themselves. Volume 3, titled Maintaining, Expanding, and Contracting the Organization, addresses issues that become prominent after the organization and jobs are designed and after there xxviii

Copyright American Psychological Association. Not for further distribution.

Introduction

is an understanding of who will be in the organization and how they will be trained, evaluated, and rewarded. The topics in this volume need to be understood—and in some cases fostered and in other cases controlled or eliminated—if the organization and its employees are to have the opportunity to be effective and succeed. Part I focuses on issues related to relationships with work. These chapters address issues of how members see themselves as part of the organization, how they relate to the organization, how the organization relates to them, how members relate to each other, and how the relationships are evaluated. Chapter 1 addresses issues of fit between the person hired into the organization and the type of organization in which he or she will work. The issues influence how people perceive the organization and the type of employees who might be attracted in the future. Chapter 2 focuses on the adjustment and socialization of new employees into the organization, and Chapter 3 addresses motivating them to succeed. A prime goal of the organization is to have employees who are satisfied, are committed to it, and have the opportunity to express their values in the performance of their work (Chapter 4). Chapter 5 focuses on the psychological contract that employees and the organization expect from each other, in the hopes that there will be commitment from both parties. Part II moves from a primary concern with the individual employee to more of an organizational perspective and presents chapters dealing with fostering a positive environment and relationships to and at work. The chapters in this part address the group and organizational level and the dynamics that influence how people behave in and react to an organization. The “people” are the employees, managers, and others who are involved with the organization, such as clients. Chapter 6 addresses interpersonal relationships among the members of the organization and the issues they raise, whereas Chapter 7 addresses issues of communication within the organization and between members; both chapters address issues of networking. Chapter 8 takes a number of topics previously addressed, such as performance evaluation, recruitment, selection, rewards, and the like, and discusses the perceptions of employees with respect to fairness within and by the organization and the extent to which these perceptions influence approach or avoidance behavior on the part of the employee. Chapter 9 deals with another perspective on the organization, that of clients and customers. In particular, this chapter on customer service focuses on the environment needed for the management of service employees and their behavior, performance, and attitudes. Chapters 10 (stress and well-being) and 11 (quality of life) explore the potential problems that may exist in relationship to work: stress, health issues, and quality of life and work–life balance concerns. These two chapters concern, in part, how an organization reacts to its employees to facilitate a better work environment. Part III draws attention to the issues that organizations need to address to be sustainable. This part begins with a discussion of organizational politics (Chapter 12) and then presents a series of chapters that focus on problems that may arise at work in the organization. In particular, Chapter 13 focuses on problems such as conflict at work. Chapter 14 explores the issues that arise when employees exchange with and interact in or outside the organization in the pursuit of their goals and business and resolve disputes through negotiation and mediation. Chapter 15, on derailment, presents a view as to why managers, in particular, have failed and what needs to be done to understand and address the problems, whereas Chapter 16 identifies impediments to promotion. Chapters 17 and 18, respectively, cover two particular issues that present real or potential problems for organizations and their employees: workplace aggression and violence xxix

Introduction

Copyright American Psychological Association. Not for further distribution.

and sexual harassment. If these issues are not addressed and controlled, the effectiveness of the organization is threatened. Part IV presents chapters that focus on planning for change and the future. Chapter 19 discusses the issue of succession planning, and Chapters 20 and 21 deal with the broad issues of what happens when plans may not work out. Chapter 20 addresses organization change, and Chapter 21 describes the issues in downsizing, mergers, and other planned changes to the organization. Finally, Part V brings the societal context interface into play. Chapters 22 and 23 tackle the issues of studying organizations in today’s changing and globalized economy. Chapter 22 addresses the issues of a distributed workforce (e.g., call centers for U.S. consumers who receive responses from technicians located in a foreign country). Chapter 23 addresses the broadest issue for organizations and workers today: the impact of culture on effectiveness and performance. The final chapter, chapter 24, addresses an issue— corporate responsibility and ethics—that is seen by many as a macro issue to which I/O psychology has much to contribute; this chapter identifies issues that set the stage for the kind of organization that one would choose to sustain and be part of, and influences the decisions raised in all of the previous chapters. When all is said and done, is the organization sustainable and contributing in a positive manner to the employees, the stakeholders, the clients, the organization in general, and to society as a whole? In summary, the order within and between chapters is one possible perspective that we have chosen. We will not argue against or dispute a different perspective. For us, the order makes sense. We hope that it will help readers gain an appreciation for what I/O psychologists study, how they study it, what they have found, and what needs to be done in the future. THE AUTHORS The authors invited to participate in this handbook represent many of the leading researchers and practitioners in the field of I/O psychology. They represent those who focus on theory, those who focus on practice, and those who focus on both. Their home institutions are academic (mainly psychology or business) or practices that specialize in human resource management issues, and they represent different countries and continents. In short, good diversity in setting, experience, and focus informs this reference resource. The authors were given guidelines for their chapters. The chapters varied in length, as determined by editorial considerations. Authors were asked to achieve the following general goals, where appropriate: ■

■

■

■

xxx

present the types of issues that I/O psychologists study, from both theoretical and applied perspectives; demonstrate how theory is generated and applied to problems that exist in all types of organizations—problems pertaining to understanding, explaining, and predicting behavior in organizations; explore how problems are diagnosed, how research is conducted to generate answers to the problems, how interventions are implemented, and how they are evaluated; focus on what is currently known, but include basic historical reviews and identify the most pertinent sources of information in both the core and emerging literatures;

Introduction

■

■ ■

Copyright American Psychological Association. Not for further distribution.

■

recognize that the world of work, the workforce (influence of gender and diversity), and the nature of work are changing; that there is an emphasis on globalization of the economy; and that there is a need to consider cultural influences on what and how problems and issues are studied; pinpoint practical issues and probe unresolved and controversial topics; present future theoretical, research, and practice trends; and include a focus on the roles and influence of gender and diversity, as well as the role of technology—issues that are pertinent across all substantive topics.

In addition to these content foci, each chapter author was requested, where possible, to identify and acknowledge the links with other chapter topics to facilitate the reader’s understanding of the integrated approach to the field. Our goal was to provide a clearer understanding of the complexities of each topic covered and of the complexities of organizations and people within those organizations. We hope that the reader notices the significant cross-referencing among chapters (i.e., the reader will be advised to “see Vol. X, chap. Y, this handbook”). This cross-referencing suggests other chapters that might be consulted for more information related to the topic at hand. The cross-referencing process also has the explicit purpose of demonstrating to the reader that, in a number of cases, two authors who are discussing the same topic may have different opinions or conclusions about the body of research being described. This should not be a surprise. The variables studied by I/O psychologists are complex; they cannot understand, explain, or predict every aspect of the constructs studied (see the next section, Effect Size). For some, the glass is half empty, and for others the glass is half full. The point is that the glass can be viewed differently and that more research is needed to understand the phenomena. The authors of this three-volume handbook have put together a tour de force of the theory, research, and applications of one of the most important domains of life that transcends individuals, groups, and cultures: the domain of work. We hope that the handbook stimulates further dialogue and debate and that it ultimately plays an important role in guiding the future of industrial and organizational psychology science and practice. EFFECT SIZE A predominant tendency in books and articles that describe and summarize results is to offer comments such as “x is related to y.” The impression given to many readers— novices and even experienced ones—is that the relationship allows one to understand the phenomenon. For better or worse, however, the phenomena in I/O psychology, or psychology in general, are not fully understood. Also, there is the tendency to assume that if there is a statistically significant relationship, there may be “practical significance” and that the answer has been found and the problem solved! Unfortunately, statistical relationships can range anywhere between greater than zero and 1 (positive or negative). Because statistical significance is a function of sample size—smaller relationships are determined to be significant if the sample size on which the analysis was conducted was relatively large—reliance on the statistical significance test itself and the conclusion it produces (such as p < .05) does not inform the reader about how well x explains (or predicts) y, or how much variance in y is explained by x, or how much variance is “overlapping” in x and y. xxxi

Introduction

In light of the interpretative shortcomings of statistical significance testing per se, and to provide the reader with an appreciation for how much variance is explained by the variables studied in I/O psychology, each of the authors was asked to report effectsize measures for the important or key relationships that were described in their chapters. Thus, almost every chapter offers some presentation of how much variance is explained, or how strong of a relationship there is for the key phenomena that are described therein. Authors differed in what effect-size measure they used to present such information, and thus, we provide here a very brief overview of the effect-size statistics presented by many of the authors. (For more detail, see recent psychological statistics textbooks or an article by one of the associate editors4).

Group Differences Copyright American Psychological Association. Not for further distribution.

Results are often presented in terms of the statistical means for two groups and the difference between those means. A statistic referred to as d (or sometimes as Cohen’s d) is obtained by examining the difference between the two group means divided by the common standard deviation for the two groups. Thus, the d represents the difference in means in terms of standard-deviation units. A value of 0.00 means no difference between the two groups. A critical question for researchers reporting results such as the d statistic is its qualitative interpretation: What values of d represent small, medium, or large effects? As can be seen in journals as well as throughout the chapters in this handbook, researchers and authors have a tendency to describe effects and relationships in qualitative terms (e.g., “the difference between males and females was small”). For better or worse, a tradition has evolved to provide some framework for interpretation of the effect. Cohen5 has described d values of .2, .5, and .8 as representing small, medium, and large effects, respectively. However, as noted by Aguinis, Beaty, Boik, and Pierce,6 it is important to emphasize that Cohen’s benchmarks are based on results of a review and content analysis of a particular journal focused on a particular field in a particular year (abnormal and social psychology in 1960). Thus, the qualitative interpretations are based in part on observed effect sizes reported in a specialized journal as well as Cohen’s own subjective opinion. These interpretations have come to be considered “conventional.” This is not the place, perhaps, to present other interpretations or to suggest other “conventions,” but the critical point that we want to emphasize is that convention and conclusions of “practical” significance are based on effect sizes published more than 45 years ago in a journal in a completely different discipline than I/O psychology. Nevertheless, they provide some broad guidance or benchmark to help the reader assess the size of group differences as typically presented by authors of the chapters.

Strength of Relationship or Amount of Variance Explained The correlation coefficient, r, is a statistic, or index, that is used to describe the strength of the relationship between two variables, such as x (a construct that has the potential to explain or predict a criterion) and y (a criterion that we want to explain or predict). 4

Aguinis, H., Werner, S., Abbott, J. L., Angert, C., Park, J. H., & Kohlhausen, D. (in press). Customer-centric science: Reporting research results with rigor, relevance, and practical impact in mind. Organizational Research Methods.

5

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

6

Aguinis, H., Beaty, J. C., Boik, R. J., & Pierce, C. A. (2005). Effect size and power in assessing moderating effects of categorical variables using multiple regression: A 30-year review. Journal of Applied Psychology, 90, 94–107.

xxxii

Copyright American Psychological Association. Not for further distribution.

Introduction

(Chapters include a variety of additional ways of presenting effect sizes such as rho [ρ] or standardized beta coefficients [β].) Values of r range from 0.00 (or no relationship) to 1.00 (plus or minus), or a perfect linear relationship. If the r is squared (r2), then the resulting value is interpreted as “percentage of variance explained or predicted” or “common variance.” For example, an r of .3 between x and y indicates that 9% (.32 = .09) of the variance in y is explained by variance in x, or that x and y share 9% common variance. The concept is similar for multiple correlation, where r represents the relationship between a y and two or more x variables, and r2 represents the proportion of variance in y explained by the set of x variables. Cohen (1988) indicated that r = .10 is a small effect, r = .30 is a medium effect, and r = .50 is a large effect. Note the extrapolation to r2 values of .01 (1%), .09 (9%), and .25 (25%) as small, medium, and large effect sizes. (As noted earlier for d statistics, the qualitative interpretation of the size of the relationship needs to be considered with caution.) Regardless of what convention is adopted for interpreting the relationship between variables, the absolute values provide information to the researcher and practitioner. If an r2 value of, for example, 9% is obtained, it also means that (1 − r2) or 1 − .09 = .91 or 91% of the variance in y is not explained by the x construct. This suggests a great deal of opportunity for additional research: What other variables can be studied and added to one’s model that would increase the ability to explain y? This is why psychologists continue to study the same issues over time. In the chapters in all three volumes, the absolute explained variance values of the reported results are generally less than 25%. This should demonstrate that the glass is less than half full and much research is needed to achieve higher degrees of explanation. The phenomena of interest are not perfectly understood, and behavior cannot be predicted perfectly. More research is needed on every topic discussed in the three volumes of the handbook. This does not suggest failure on the part of I/O psychology but rather highlights the complexity of human behavior in organizations and suggests opportunities, as well as challenges, to develop a broader, richer understanding of the many important issues that I/O psychologists choose to study. ACKNOWLEDGMENTS The development and production of a handbook requires an organization! My experience of editing this handbook benefitted immensely from an incredible group of colleagues and staff who devoted endless hours in consultation; demonstrated patience, fortitude, and energy when “urgencies” arose; and, overall, contributed to any success this handbook might achieve. I start with the associate editors: Herman Aguinis, Wayne Cascio, Michele Gelfand, Kwok Leung, Sharon Parker, and Jing Zhou. This team contributed invaluable wisdom to the generation of the plan, identification of potential authors, and review of outlines and drafts of chapters. The chapter products were heavily influenced by their feedback to me and the authors. On the project genesis and administration side, I owe gratitude and thanks to APA publisher Gary VandenBos, who raised the possibility of my editorship and who provided valuable input at the outset of the project. Another key figure has been APA’s director of reference books, Ted Baroody, who worked with me from Day 1, helping to generate the plan for the handbook, identify potential chapters and authors, review outlines, and provide feedback on any issue requested throughout the project. Two very important resources for the authors and editorial staff were project coordinators Marian Haggard and Shenny Wu; xxxiii

Introduction

they shepherded the authors and their chapters through the electronic system used for submission, review, and revision of the multiple drafts of chapters, and they performed their tasks with grace and support that served to keep the project on track. The APA Handbook of Industrial and Organizational Psychology is a product of a personally rewarding partnership with each of these individuals. All in all, my performance evaluation of the team results in an “excellent” rating, and I want to thank all of you for your support, encouragement, and wisdom.

Copyright American Psychological Association. Not for further distribution.

Sheldon Zedeck, PhD Editor-in-Chief

xxxiv

Copyright American Psychological Association. Not for further distribution.

PART I

FOUNDATIONAL ISSUES IN INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY

Copyright American Psychological Association. Not for further distribution.

CHAPTER 1

A HISTORICAL SURVEY OF RESEARCH AND PRACTICE IN INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY Copyright American Psychological Association. Not for further distribution.

Andrew J. Vinchur and Laura L. Koppes

The understanding of psychology is one of the most important roads to success for the modern business man. Industrial and commercial work are in thousandfold contact with mental life. Salesmanship and advertising, learning and training for technical labor, choosing the right position and selecting the right employe, greatest efficiency of work and avoidance of fatigue, treatment of customers and of partners, securing the most favorable conditions for work and adapting the work to one’s liking, and ever so many other problems stand before the business world and cannot be answered but by psychology (Münsterberg, 1918, p. v). The above quotation by Hugo Münsterberg is typical of the ambition and confidence (some might say overconfidence) of at least some of the founders of industrial–organizational (I/O) psychology. Although the results of more than 100 years of applying psychology to organizations have caused us to be a bit more humble and circumspect in our pronouncements, it is true that the science and practice of I/O psychology have made a fair amount of progress in dealing with Münsterberg’s concerns. The field has had an interesting and eventful history. As I/O psychology has expanded and matured, we have seen

an increased interest in this history. This is most welcome, as it is our belief that knowledge and appreciation of history is essential for deep understanding. Tracing the evolution of ideas in substantive areas of I/O psychology and examining the litany of insights, incremental progress, and missteps that resulted in the current state of the field can lead to a richer appreciation and understanding for researchers and practitioners alike. Previous historical overviews of I/O psychology vary in emphasis, orientation, and detail. Examples include Ferguson (1962–1965),1 who built his history around the Carnegie Institute of Technology’s (CIT’s) Division of Applied Psychology. Baritz (1960), although critical of industrial psychology’s close ties with management, provided a great deal of information about early industrial psychology, as did Napoli (1981) in his history of the psychological profession. A chapter by Hilgard (1987) concentrated on the history of I/O psychology in the United States; Warr (2007) provided an overview of the development of I/O outside of the United States; and McCollom (1968) and Viteles (1932) gave summaries of the early years of the field both inside and outside of America. Katzell and Austin (1992) offered a contextual approach to American I/O psychology history, as did Koppes (2003). A recent book edited by Koppes (2007) took a topical approach to I/O history. Many other excellent histories of I/O psychology have focused on specific content areas, individuals, or time

We thank Bianca Falbo, Shelly Zedeck, and an anonymous reviewer for their helpful comments. All errors are our own. 1

Leonard W. Ferguson completed a series of pamphlets covering 14 chapters (one volume and part of another) of a planned 12-volume history of industrial psychology centered on the accomplishments of individuals associated with the CIT program (Ferguson, 1962).

http://dx.doi.org/10.1037/12169-001 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

3

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

periods. Where relevant, we refer to these sources in this chapter. Katzell and Austin (1992) and Koppes (2007) provided historical sources and resources for interested readers. The purpose of this chapter is to provide an overview of the history of I/O psychology. To some extent, we cover much of the same ground as the histories mentioned in the preceding paragraph. Like Ferguson (1962–1965), we emphasize the importance of the CIT applied psychology program and concentrate on early developments in the field. Unlike Baritz (1960), we offer an insider’s rather than an outsider’s perspective. We attempt a comprehensive history in the manner of Hilgard (1987) and Katzell and Austin (1992); however, like the earlier treatments by Viteles (1932) and McCollom (1968), we also examine historical developments outside the United States. In addition, we fold in a brief history of the related field of organizational behavior (OB). And although we emphasize a number of substantive areas such as employee selection and motivation, we do not take the explicitly topical approach of Koppes (2007). To cover this expansive history in a single chapter, by necessity we had to be selective in deciding what material to include and what aspects to emphasize. It is not possible in a single chapter to do justice to all of the relevant worldwide research, practice, organizations, and contributing individuals, let alone the associated scientific, social, political, and economic contexts that influenced this history. We have therefore limited the scope of the chapter in the following ways. As is appropriate for a handbook chapter, we emphasize broad trends, familiar, agreed-upon historical landmarks, and the synthesis of the large body of existing historical scholarship. Although we do not assume prior knowledge of the field’s history, we do assume the reader has a working knowledge of I/O psychology. Although we believe that this history is best understood as a complex interaction between the efforts and initiatives of individuals and the context in which they lived and worked, because of space limitations we have kept biographical and contextual information to a minimum, referring the reader to other sources for detailed information. As I/O psychology has grown dramatically over the past 100 years, there is correspondingly more content to cover as the 4

field progresses and expands. We have decided to devote more space and detail to origins and early developments than to more recent history for the following reasons. First, because of the expansion of the field, space consideration precludes in-depth coverage of the large amount of relevant later material; we did not want to reduce this content to a “greatest hits” list. Second, a reasoned historical evaluation often requires the passage of time to determine the relative importance of theories and procedures. Third, the genesis and early development of any field, including I/O psychology, is particularly interesting and informative. Determining the initial motivations, influences, problems, and successes of the early pioneers puts into context present-day practice, and also allows us to develop an appreciation for the efforts of the early applied psychologists. And finally, readers interested in in-depth coverage of more recent activity in the field need only turn to the content chapters in this handbook. For both industrial psychology and organizational psychology, we begin with a broad overview and then cover a few selected areas in greater detail, focusing for the most part on the first half of the 20th century but also examining interesting and relevant later developments. Our coverage of industrial psychology begins with late 19th century work on advertising and on fatigue, followed by the contributions of the Division of Applied Psychology at CIT and the impact of World War I on industrial psychology. We then turn to the early history of psychologists in industry and consulting, as well as professional institutes and organizations. Next, we discuss the history of psychology applied to employee selection, performance appraisal, and training. On the organizational psychology and OB side, we look at the evolution of interest in worker welfare, examine the influential Hawthorne studies, and discuss the history of the human relations movement. This is followed by histories of psychology’s efforts in leadership, employee motivation, and job satisfaction. INDUSTRIAL PSYCHOLOGY Although it can be instructive and entertaining to look for precursors to I/O psychology in the writings of philosophers and historians from antiquity through the 18th century (see Kaiser, 1989, for a

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

chronology beginning in 2100 BC), we begin our story in the second half of the 19th century—the moment when psychology emerged as a scientific discipline distinct from philosophy and physiology. The establishment of Wilhelm Wundt’s laboratory at the University of Leipzig in 1879 is often taken as a convenient starting point. As Murphy (1930), Viteles (1932), and others have emphasized, however, the founding of Wundt’s laboratory was really more the accumulation of a series of events moving psychology from philosophy to an empirically based science than the actual beginning of scientific psychology. In addition, as Campbell (2007) noted, because many nonpsychologists made early contributions to the field, it is overly simplistic to view I/O psychology as an outgrowth of Wundt and his students. Nevertheless, it is true that a number of important contributors to industrial psychology, such as James McKeen Cattell, Hugo Münsterberg, and Walter Dill Scott, received their doctorates at Leipzig under Wundt. This new self-consciously scientific psychology soon spread beyond Germany, and although initial efforts were confined to the laboratory, the fledgling discipline was soon branching out with real-world applications, including applications in business and industry (Koppes & Pickren, 2007; Viteles, 1932). In the United States, these early industrial psychologists2 were influenced by functionalism, an orientation that emphasized the consequences or utility of adaptive behavior and individual differences. A number of prominent early industrial psychologists were trained in graduate programs at Columbia University and the University of Chicago that emphasized this functionalist approach (Vinchur, 2007). Also influential was scientific pragmatism, which emphasizes prediction over understanding and focuses on the importance of utilitarian consequences (Austin & Villanova, 1992). This was evident, for example, in employee selection, where the relationship between test scores and success, rather than what the test

measured, was judged most important (e.g., Freyd, 1923–1924). Finally, early industrial psychology was dependent on advances in measurement and statistics, particularly the accurate measurement of individual differences. We illustrate this dependency in our discussion of employee selection. Cultural forces in the late 19th and early 20th century were supportive of psychology’s forays into business and industry. There was a great deal of concern about social problems, government reform, and the power of large corporations (Zickar & Gibby, 2007). Progress and interest in science was on the upswing, and science, including psychology, was viewed as a source for pragmatic solutions (Koppes & Pickren, 2007). Industrialization was on the rise. As organizations increased in size, they became increasingly difficult to manage. In response, a professional manager class emerged, along with reliance on specialized departments to make decisions that were formerly made by first-line supervisors. In addition, concern for the welfare of workers resulted in legislation, welfare programs, and an increase in labor organization (Nelson, 1975). Psychology, in the form of testing and selection, offered a potential method for helping to manage the labor market by use of procedures based on merit (Hale, 1992).3 At roughly the same time that psychologists were beginning to apply psychology outside the laboratory, engineer Frederick W. Taylor was developing a system to improve worker productivity and efficiency known as scientific management or “Taylorism.” Taylor (1911) argued for a twopronged approach: improving machinery (e.g., developing the most efficient shovel) and improving the individual worker by analyzing the job and determining the optimal way to perform that job. Under scientific management, the organization would benefit from improved productivity; the worker would benefit from increased pay via a piece-rate incentive system. Other advocates of this approach included Frank B. Gilbreth and his wife,

2

Psychologists working in industry went by various labels during these early years, many with slightly different connotations. Labels included economic psychologists, business psychologists, consulting psychologists, employment psychologists, psychotechnicians, industrial psychologists, and applied psychologists. Although industrial psychology was not in common usage during psychology’s early years (Viteles, 1932), for consistency’s sake we use this term.

3

Although it is possible to speculate that general social, cultural, economic, and political forces and trends influenced the general development of the field, identifying which specific force affected which specific development, and how that influence took place, is an important, difficult task that is beyond the scope of this short chapter.

5

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

psychologist Lillian M. Gilbreth,4 who developed time-and-motion studies to analyze tasks and time needed to complete them. The Gilbreths recognized that there needed to be more concern for the worker in scientific management, and they worked to eliminate accidents and fatigue (Lane, 2007). The influence of scientific management on industrial psychology was probably more indirect than direct. Although Viteles (1932) noted that scientific management guided the early scope of industrial psychology and established an economic objective for the field (Hilgard, 1987), Viteles also stated that scientific management contributed little to industrial psychology theory and practice. Certainly the early industrial psychologists and advocates of scientific management were aware of each other (Sokal, 1984). Scientific management did influence engineering psychology or human factors (Lane, 2007), and Kanigel (1997) argued that the effect of scientific management on current business practices is much greater than is commonly perceived. Scientific management, however, exhibited a number of serious drawbacks. The major criticism of Taylorism was the dehumanizing effect it had on the worker. Workers were understandably resentful toward outside experts who purported to tell them how to best perform their jobs, and the system did produce labor unrest (Aiken, 1985; Muscio, 1920). Contemporary psychologists were critical of the system’s impact on the worker, see, for example, British psychologist Charles S. Myers (1925), German psychologist Otto Lipmann (Vinchur, 2005), and American psychologists such as Viteles (1932). The initial forays by psychologists into industry were research on fatigue and energy in Europe and studies on advertising in the United States. Early European researchers on the science of work included E. J. Marey of France, who conducted experimental research on fatigue as early as 1878; Angelo Mosso, who in 1888 invented the ergograph5 to measure muscle fatigue and the corresponding reduction of work potential (Fryer & Henry, 1950);

Gustav Fechner of Germany, who used weights to study fatigue (Münsterberg, 1913); and Wundt’s student Emil Kraepelin, who studied physical and cognitive fatigue and developed work curves to demonstrate the reduction of production over time (Koppes & Pickren, 2007). Hugo Münsterberg (e.g., Münsterberg, 1913), first in Germany and later in the United States, focused on improving worker output and decreasing accidents through his research on fatigue, labor, and training (Koppes & Pickren, 2007). Two of Wundt’s students, Edward W. Scripture and Harlow Gale, were most likely the first in America to apply psychology to advertising. Scripture (1895), although not conducting research, discussed psychological issues relevant to advertising and business. Gale (1896) conducted both laboratory and survey studies and may have been the first to use the order-of-merit procedure to rank-order brands based on advertising data (Schumann & Davidson, 2007). Other early researchers of note in this area include Daniel Starch (1910), who published extensively in advertising and who left academia to start a marketing research company in 1932; Harry Levi Hollingworth, who was interested in purchase behavior and constructed the first panel to systematically track consumer behavior (Kuna, 1976, cited in Schumann & Davidson, 2007); and Edward K. Strong, Jr., who at Hollingworth’s suggestion evaluated the relative merits of advertisements in his 1911 Columbia University dissertation (Hansen, 1987). Of particular note was Walter Dill Scott, another Wundt student and a Northwestern University professor, who in 1901 was asked by magazine editor Thomas L. Balmer to give a talk on the usefulness of psychology in advertising (Ferguson, 1962–65).6 Initially reluctant due to the perceived stigma in academia for applied psychology, Scott eventually agreed. A series of magazine articles followed and resulted in two books (Scott, 1903, 1908) on the psychology of advertising. It was Scott’s 1910 book Increasing Human Efficiency in Business, however,

4

Lillian Gilbreth’s 1915 dissertation from Brown University is one of the earliest on an industrial psychology topic. For biographical information on Gilbreth, see Kelly and Kelly (1990); Koppes (1997); Perloff and Naman (1996); and Cheaper by the Dozen (1948), a memoir by her children Frank B. Gilbreth, Jr., and Ernestine Gilbreth Carey.

5

The egrograph is a device for measuring the work capacity of a muscle.

6

Ferguson (1962–1965) is a good source of biographical information about Scott. Jacobson (1951) provided a book-length biography. Brief biographical sketches on Scott and many other early contributors to industrial psychology can be found in Vinchur and Koppes (2007).

6

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

that Hilgard (1987) called the beginning of serious industrial psychology in America, although Münsterberg’s Psychology and Industrial Efficiency (1913; published in Germany in 1912) is another viable contender for this honor. Determining the beginning of an enterprise as multifaceted as industrial psychology is probably an exercise in futility. This is also true for identifying the founder, although Ferguson (1962–1965) made a compelling case that Scott should have this honor. Scott did have a number of firsts to his credit. In 1916 at CIT he was the first in the United States to have the title of professor of applied psychology; he started the first industrial psychology consulting firm, the Scott Company, in 1919; and he was the first and only psychologist awarded the Distinguished Service Medal for his service in World War I. Like Scott, Hugo Münsterberg received his PhD under Wundt at Leipzig. Both men were also elected president of the American Psychological Association (APA): Scott in 1919; Münsterberg in 1898.7 Münsterberg initially was recruited by William James to run the psychology laboratory at Harvard University. He eventually branched out to applied psychology, making contributions to clinical, educational, and forensic psychology along with his work in industrial psychology. Münsterberg’s initial publication on applying psychology to business appeared in McClure’s Magazine in 1909, and resulted in the consulting work that was included in his 1913 book (Benjamin, 2000). This book set the initial agenda for industrial psychology: Topics included employee selection, vocational guidance, training, monotony, attention, fatigue, social and physical influences on work, advertising, selling, and buying. Münsterberg was responsible for organizing the 1904 International Congress of Arts and Sciences as part of the St. Louis World’s Fair of that year. The congress showcased well-known academics from around the world, including a number of prominent psychologists. Visitors also had the opportunity to experience the

new mental and physical tests (Brown, 1992). Although the unpopularity of Münsterberg’s support for Germany in World War I and his premature death in 1916 resulted in a decrease in his influence, in recent years there has been a resurgence of interest in his career (e.g., Benjamin, 2000; Landy 1992, 1997; Spillmann & Spillmann, 1993). Around this time, two events critical for the development of industrial psychology occurred. The first graduate program in industrial psychology at CIT was founded, and World War I provided the opportunity for psychologists to demonstrate the usefulness of their science. THE DIVISION OF APPLIED PSYCHOLOGY The early industrial psychologists were trained in traditional graduate programs that emphasized laboratory research. In 1915 Walter Van Dyke Bingham8 accepted an invitation from the president of CIT to found a program in applied psychology. The Division of Applied Psychology under the direction of Bingham was the first graduate program in industrial psychology in the United States. Supported by a number of wealthy Pittsburgh businessmen, the division, arranged into various bureaus and departments, served as a model for cooperation between business and academia. The Bureau of Salesmanship Research, proposed by insurance executive Edward A. Woods and directed by Walter Dill Scott, was sponsored by pledges from organizations who believed they would benefit from the bureau’s research on sales (Ferguson, 1962–1965). Under the supervision of Scott, the bureau developed materials to aid in the selection of salespersons, including model application blanks, interviewers’ guides, and tests of intelligence, alertness, carefulness, imagination, resourcefulness, and verbal ability (Prien, 1991). Under the direction of James B. Miner and then W. W. Charters, the Bureau of Retail Training began in 1918. This bureau focused on both

7

Biographical information about Münsterberg is available from Benjamin (2000), Bjork (1983), Hale (1980), Landy (1992, 1997), daughter M. Münsterberg (1922), and Spillmann and Spillmann (1993).

8

Walter Van Dyke Bingham had a long and illustrious career as an industrial psychologist. He received his doctorate from the University of Chicago, studied with William James at Harvard, where he became acquainted with Münsterberg, and through his European travels was also acquainted with the Gestalt psychologists and pioneering British industrial psychologists such as C. S. Myers. Sources of autobiographical and biographical information on Bingham include Benjamin and Baker (2003), Bingham (1952), and Ferguson (1962–1965). Ferguson (1962–1965) remains an authoritative source on the history of the Division of Applied Psychology.

7

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

employee and employer attitudes regarding training and organizational research. Vocational guidance, particularly individual interests, was the area researched by the Bureau of Personnel Research. Leadership was provided by Miner, Clarence Yoakum, and Edward K. Strong, who would build on the work started here to produce the Strong Vocational Interest Blank (Strong, 1927). Despite its many successes, the division had a short existence and closed in 1924. CIT graduate Richard S. Uhrbrock attributed the closing to the following combination of factors: Bingham’s new interest in the Personnel Research Federation; the new director Clarence S. Yoakum’s lack of success in retaining corporate sponsorship; and a new, less supportive president at CIT (Hilgard, 1987; see also Ferguson, n.d.) It is hard to overstate the influence of the CIT program on I/O psychology. Students and staff associated with the Division of Applied Psychology made major contributions to industrial psychology and psychometrics and also were instrumental in spreading industrial psychology to academic and applied settings. A few examples will illustrate. Bingham, who along with Scott contributed greatly to Army personnel research in World War I, produced more than 200 books and papers (Zusne, 1984), directed the Personnel Research Federation (described in a later section), edited the Journal of Personnel Research, and was chief psychologist for the Adjutant General in World War II. Scott’s contributions to the United States Army during World War I are described below. The selection procedures developed at CIT by Scott and his students were widely imitated (Ferguson, 1962–1965). Staff member Louis L. Thurstone later made significant contributions to measurement and factor analysis. Marion A. Bills, who in 1951–1952 was the first woman president of APA Division 14 (which evolved into the Society for Industrial and Organizational Psychology [SIOP]), conducted long-term research in the insurance industry considered a model for

9

collaboration between psychology and business (Ferguson, 1952). Other staff members of note include Arthur W. Kornhauser, who did important early work in selection and testing, attitude surveys, labor union relations,9 and employee mental health (Zickar, 2003) and Clarence Stone Yoakum, who was field supervisor of all mental testing for the Army in World War I (Bingham, 1946) and who supervised the doctoral program at CIT. Four students received their doctorates from that program. The first was Bruce V. Moore, whose 1921 PhD is considered the first in the United States from an industrial psychology program (Ferguson, n.d.). Moore was followed by Max Freyd, Grace Manson, and Merrill Ream. A description of their contributions along with others from the program can be found in Ferguson (1962–1965). WORLD WAR I AND INDUSTRIAL PSYCHOLOGY World War I took place from 1914 to 1918; the United States entered the war in 1917. The conflict provided an opportunity for psychologists to contribute to the war effort and to demonstrate the practical value of their new science. APA president Robert Yerkes decided the best use of psychology would be in developing standardized group intelligence tests that could be used in selecting and classifying recruits. Yerkes’s group developed two tests, the Army Alpha for individuals with Englishlanguage skills, and the Army Beta, for those recruits who were illiterate or did not have strong English-language skills. Working under the Surgeon General, Yerkes and his team of 354 examiners tested 1,726, 966 soldiers in 35 camps at a cost of 50 cents per individual (Ferguson, 1962–1965). This effort demonstrated that large groups of individuals could be tested at a reasonable cost. Although the military’s response to this testing program was mixed (Samuelson, 1977), the perceived success of

Although industrial psychology has long been criticized as overly favoring the management perspective (e.g., Baritz, 1960) and indifferent at best to unions, the relationship between industrial psychology and unions is more complex, and there were certainly exceptions to the pro-management perspective; Kornhauser was one of the most prominent (see Zickar, 2003 for biographical information about Kornhauser, and Zickar, 2004 for more information about industrial psychology and unions).

8

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

the group testing effort would fuel a strong interest by industry for psychological testing after the war ended. In keeping with the hereditary bias regarding intelligence in vogue at the time, interpretations of these tests by Brigham (1923)10 and others were used to cast aspersion on the mental capacity of Americans in general and to reinforce racial and ethnic group stereotypes. Walter Dill Scott strongly disagreed with Yerkes’s approach to the war effort. He saw Yerkes as primarily concerned with furthering the interests of psychology, rather than using the tools of industrial psychology to further the war effort (Ferguson, 1962–1965; von Mayrhauser, 1987). Along with Walter V. Bingham, Scott went his own way, establishing the Committee on Classification of Personnel under the aegis of Adjutant General’s Office. This committee applied the techniques developed in the CIT program to solving Army personnel problems. They adapted rating scales for military use in selecting and rating officers; developed trade specifications and indices of occupations, occupational needs, and personnel specifications; and developed standardized trade tests used on approximately 130,000 soldiers (Bingham, 1919; Strong, 1918). More than 3 million soldiers were classified and rated on job qualifications (Sokal, 1981). This committee also laid the foundation for person and job analysis (Katzell & Austin, 1992; see Wilson, 2007, for a history of job analysis). One notable test to emerge from World War I was Woodworth’s (1919) Personal Data Sheet, a precursor to tests of personality. Although he noted pockets of applied psychology activity in Great Britain before World War I, Hearnshaw (1964) claimed the war produced the real beginning of applied psychology in that country. Concern for the health and well-being of munitions workers, who commonly worked 70 to 90 or more hours per week, resulted in the creation of the Health of Munitions Workers Committee in 1915. Members of this committee, including one of Britain’s first industrial psychologists H. M. Vernon (McCollom, 1968), investigated industrial fatigue and accidents, hours of work, ventilation and lighting of factories, and worker efficiency and output. During the war, 10

work was done on employee selection, although not to the degree conducted in the United States. Future industrial psychologists C. S. Myers, T. H. Pear, and others researched selection procedures for submarine detection operators. Myers stated that his initial exposure to industrial psychology was through the writings of Bernard Muscio, one of the first investigators for the Industrial Fatigue Research Board. In 1916, Muscio, an Australian, delivered a series of lectures on industrial psychology in Sydney that were later published in 1917 (Hearnshaw, 1964). In Germany, World War I necessitated an increased need for aptitude testing for positions such as pilots, radio operators, and transport drivers (Sprung & Sprung, 2001). PSYCHOLOGISTS IN CONSULTING AND INDUSTRY After hostilities ceased, Scott and others involved in the war effort opened the first personnel consulting organization, the Scott Company, in February 1919. Although successful, the Scott Company was only in business a few years. Experiencing financial difficulties in 1921, the company laid off several consultants, including one of the first woman consultant psychologists, Mary Holmes Stevens Hayes (Hopkins, 1921; Koppes, 1997). The Scott Company’s demise may have been due to employees leaving for other opportunities (Katzell & Austin, 1992), or perhaps to a recession in 1921–1922 (Ferguson, 1961). Another early consulting firm of note was the Psychological Corporation, founded in 1921 by James McKeen Cattell (Cattell, 1923). This firm, whose primary purpose as conceived by Cattell was to advance psychology through research, operated as a holding company for psychologists. Although the Psychological Corporation was initially unprofitable, a reorganization under Walter V. Bingham and Paul S. Achilles brought eventual success (Sokal, 1981). The early practitioners of industrial psychology were trained in experimental psychology and for the most part maintained academic careers. In time individuals did begin to work full-time in industry (e.g.,

Brigham (1930) later completely recanted this view.

9

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

Henry C. Link in 191711) or for consulting firms. Employment in industry and consulting provided opportunities for psychologists who were unwilling or unable to secure academic employment (e.g., women). Examples include the aforementioned Lillian Gilbreth; Mary Holmes Stevens Hayes, a 1910 University of Chicago PhD who worked for the Scott Company, coauthored a book with Scott (Scott & Hayes, 1921), and became an authority on youth guidance and placement (Koppes, 1997); and Elsie Oschrin Bregman, a 1922 Columbia University PhD, who worked as an applied psychologist at R. H. Macy (Bregman, 1921; Oschrin, 1918) and for the Psychological Corporation, where she revised the Army Alpha test for civilian use (Bregman, 1926). Also noteworthy are Grace Manson, one of four students to receive PhDs from the CIT program, who conducted selection research for the Bureau of Business Research at the University of Michigan (e.g., Cook & Manson, 1925–26; Manson, 1925–1926); the previously mentioned Marion A. Bills, a 1917 Bryn Mawr PhD, a research assistant at CIT who went on to a productive career conducting research in the life insurance industry (Koppes, 1997); Millicent Pond, a 1925 Yale PhD, who directed personnel research at the Scovill Manufacturing Company (e.g., Pond, 1926–1927); and Sadie Myers Shellow, who followed her 1923 Columbia PhD by collaborating with Morris Viteles in conducting research for the Milwaukee Electric Railway and Light Company and who later was a personnel consultant for that city’s police department. Histories of I/O psychology, including this one, emphasize the published record and as such, may not give a complete picture of the activities of practitioners in the field. As discussed by McCollom (1968), Marion Bills (1953) made this point about practitioners more than 50 years ago when she stated that there is a major difference between what psychologists do versus what they report. Bills noted that much of the

11

day-to-day activities of practitioners may rely on practice not considered scientific enough for publication. From industrial psychology’s inception, there has been tension between academic researchers and practitioners, as evidenced by the condemnation directed at the early industrial psychologists by their academic brethren. Each side has criticized the other. Practitioners complain that academic psychology is artificial and lacks relevance for practice; academics criticize practitioners for their perceived compromise of scientific standards. As with most long-standing disputes in a discipline, there is some validity to the concerns of both sides (see Zickar & Gibby, 2007, for a recent discussion). PROFESSIONAL INSTITUTES AND ORGANIZATIONS A number of institutes were founded in the early years of industrial psychology. In Germany they included the Institute for Applied Psychology and General Psychological Research in Neu Babelsberg, founded by Otto Lipmann and William Stern in 1906, and the Institute for Industrial Psychotechnology at the Technical College in Berlin–Charlottenberg founded in 1918 by Walter Moede and Georg Schlesinger. Curt Piorkowski and Lipmann established the Institute for Vocational and Business Psychology in Berlin in 1920, and a year later Hans Rupp and Carl Stumpf founded the Division of Applied Psychology at Berlin University’s Institute of Psychology. These institutes engaged in activities such as aptitude testing, ergonomics, and career counseling (Sprung & Sprung, 2001; see also Viteles, 1932). In addition, by 1922 at least 22 large companies had established their own psychological laboratories (Viteles, 1923) and by 1926 more than 100 firms, including Krupp, the Siemans Company, and the Loewe Company, were using psychological selection methods (Viteles, 1932). In Great Britain, the National Institute of Industrial Psychology was

Ferguson (1962–1965) stated that Link, a Yale PhD, was the first PhD psychologist to work full time in industry. Link began work as director of training and psychological research at the Winchester Repeating Arms Company in 1917. The first masters-level psychologist to work full time in industry may have been Herbert W. Rogers, a Columbia PhD who in 1916 went to work as an applied psychologist for the Charles William Stores in New York City (H. W. Rogers, 1946; Vinchur & Koppes, 2007). Possibly the first psychologist to work full time for the government was L. J. O’Rourke, a George Washington University PhD who became director of Personnel Research for the Civil Service Commission (CSC) in 1922. O’Rourke was recommended to the CSC by Beardsley Ruml, an officer of the Scott Company and former instructor in the Carnegie Tech program (Hilgard, 1987).

10

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

cofounded in 1921 by businessman H. J. Welch and pioneering industrial psychologist Charles S. Myers12 to conduct research and do applied work in industrial psychology (Myers, 1936). Other countries were also establishing institutes dedicated to applying psychology. As early as 1889, L. O. Patrizi founded a Laboratory of Work Psychology in Modena, Italy (deWolff & Shimmin, 1976), and in 1923 Agostino Gemelli founded the Institute of Psychology at the Catholic University in Milan, where some industrial psychology work took place (McCollom, 1968). In Switzerland, Jules Suter established an Institute of Industrial Psychology in Zurich in 1924 (Heller, 1929–1930). The Institute for Psychotechnics in Krakow, Poland, was founded in 1924 (deWolff & Shimmin, 1976). In Japan, Yoichi Uyeno established the Institute of Industrial Efficiency in 1920, primarily to apply the scientific management methods of Taylor and the Gilbreths. The Kurashiki Institute of the Science of Labor was founded a year later by Ohara, with less of a commitment to scientific management (McCollom, 1968). In Russia, the Central Institute of Labor was established in Moscow in 1920 (see Tagg, 1925), and that same year the Psychotechnic Institute and Center for Vocational Counseling was established in Prague, Czechoslovakia (Warr, 2007). The Australian Institute of Industrial Psychology was founded in 1927 by A. H. Martin (Warr, 2007). In 1921 in the United States, the Personnel Research Federation was created to try to coordinate the large number of agencies conducting personnel research. Chaired by Bingham, the Federation published the Journal of Personnel Research (later Personnel Journal), a rich outlet for industrial psychology research. The American Psychological Association was founded in 1892, and in 1894, it adopted its first constitution. The stated objective for APA was “the advancement of Psychology as a science” (Cattell, 1895, cited in Sokal, 1992, p. 115). With this purely scientific goal, APA’s leadership decided that applying psychology outside the university laboratory was

inappropriate (Benjamin, 1997a). Cattell (1946) estimated that as late as 1917, only 17 of the more than 300 members of APA were working in applications of psychology. Between 1916 and 1938, the number of APA members in teaching positions increased fivefold, from 233 to 1,299; however, the number of members in all applied psychology positions grew dramatically from 24 to 694 (Finch & Odoroff, 1939). After numerous attempts by applied psychologists to organize under the umbrella of APA, these psychologists formed their own applied psychology organizations, most of which were bound by state lines (Benjamin, 1997a). In 1921, the largest of these groups was established, The New York State Association of Consulting Organizations. In 1930, under the leadership of Douglas Fryer, a New York University psychologist, the New York Association was renamed as the Association of Consulting Psychologists (ACP) to relinquish state boundaries and establish a national presence. Amidst the continued dissatisfaction over APA’s response to professional psychologists during the 1930s, Fryer was nominated to form a new national organization of applied psychologists. Proposed in 1937, the organization was named the American Association of Applied Psychology (AAAP). Consequently, ACP voted itself out of existence (Benjamin, 1997a). The membership of AAAP was divided into four sections: clinical, consulting, educational, and industrial and business. Section D, Industrial and Business Psychology, was formed for applied psychologists in industry (Benjamin, 1997a). Many early prominent industrial psychologists were members who offered several professional services; including examining the requirements of occupations, placing workers, and conducting training programs (for a detailed description, see Benjamin, 1997a). In 1941, shortly after the National Research Council called on APA, AAAP, and other psychology groups to organize for “the benefit of the national welfare” (Benjamin, 1997a, p. 464), AAAP merged

12

Charles Samuel Myers had both an ScD (1909) and an MD (1901) from Cambridge University, where he had a successful career as an experimental psychologist and director of the psychological laboratory. Myers discovered an affinity for industrial psychology during military service in World War I. Finding Cambridge University unsupportive of his new interests, Myers and businessman H. J. Welch founded the National Institute of Industrial Psychology (Myers, 1936). For more information about Myers, see remembrances by Burt (1947) and Pear (1947), an autobiographical sketch (Myers, 1936), and a short article by Vinchur (2005).

11

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

with APA to form APA Division 14, Industrial and Business Psychology (Benjamin, 1997b), in 1945. Bruce V. Moore was the first president of APA Division 14. In 1962, “business” was dropped from the name of APA Division 14, which became the Division of Industrial Psychology. In 1973, the name changed again because of the evolving nature of the discipline. “Organizational” was added and the division became labeled as APA Division 14, Industrial and Organizational Psychology. To achieve some independence from the APA, the APA Division 14 incorporated as the Society for Industrial and Organizational Psychology in 1982 (Hakel, 1979). Since 1945, the organization has experienced changes in structure, membership, and activities, primarily due to the growth of discipline and the membership (Koppes & Pickren, 2007). Membership has grown from 130 members (fellows and associates) in 1945 (David Nershi, Executive Director, SIOP, personal communication, July 6, 2005) to 4,015 professional members (fellows, members, associates, international affiliates, retired) and 3,682 student members in early 2009. The international presence in SIOP has increased. In 1997, 6.5% of professional SIOP members were living outside the United States; by 2008 that figure had almost doubled to 12.1% (Tracy L. Vanneman, SIOP Membership Services Manager, personal communication, February 17, 2009). Benjamin (1997a, 1997b) provided a more complete historical account of the development of professional organizations in I/O psychology. Professional organizations were established in countries other than the United States. The International Association of Psychotechnics was founded in 1920 by Edouard Claparède and Pierre Bovet (Pickren & Fowler, 2003) during the first international conference of psychotechnics applied to vocational guidance in Geneva. Original members were from Belgium, Bulgaria, France, Germany, Greece, Holland, Italy, Spain, Switzerland, the United Kingdom, and the United States (Warr, 2007). Conferences of the International Congress of Psychotechnics were held in several European cities in the 1920s and 1930s, but the term psycho-

technics eventually fell out of fashion; in 1955, the International Association of Psychotechnics became the International Association of Applied Psychology (IAAP). While Edwin Fleishman was president, IAAP reorganized into a divisional structure in 1982. Division 1 was named the Division of Organizational Psychology under the leadership of Bernard Bass of the United States (Warr, 2007). Since its inception IAAP has been instrumental in the globalization of industrial and organizational psychology (Warr, 2007). The organization has provided numerous opportunities for communication and collaboration among psychologists around the globe (e.g., Fleishman, 1979, 1999). Activities include the IAAP’s journal Applied Psychology: An International Review and meetings every 4 years in different countries. IAAP has facilitated regional meetings and sponsored programs in developing countries (Fleishman, 1979, 1999). For example, Cheung (2009) reported activities of an IAAP task force on developing communications between IAAP and applied psychologists in the Asian region. Recently, IAAP, along with SIOP and the European Association of Work and Organizational Psychology, signed an agreement to strengthen the collaboration among these organizations (IAAP Division 1, n.d.). We now turn to the history of some of the substantive areas of industrial psychology. Although space considerations do not permit broad coverage of the many topic areas involved, we would like to elaborate on three influential areas: employee selection, performance appraisal, and training. EMPLOYEE SELECTION AND TESTING13 The principle activity of the early industrial psychologists in the United States was employee selection. Zickar and Gibby (2007) identified the focus on selection and differential psychology as one of the four themes that characterize the history of American I/O psychology.14 Because accurate measurement of individual differences is a prerequisite for a scientific approach to selection, measurement and empirical

13

The material on employee selection in this chapter draws heavily from Vinchur (2007). See that chapter for an expanded discussion of this topic.

14

Zickar and Gibby’s (2007) other three themes are emphasis on productivity and efficiency, emphasis on quantification, and the interplay between science and practice.

12

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

verification of procedures were central concerns. That is, psychologists needed a way to demonstrate empirically that their predictors, primarily psychological tests, accurately predicted job performance. Early work on the measurement of individual differences focused on physiological and sensory abilities (e.g., Cattell, 1890). These measures proved unsatisfactory to predict academic performance (Wissler, 1901). Wissler’s study was one of the first uses of the correlation coefficient, developed by Francis Galton, William Weldon, and Karl Pearson, among others, to determine the predictive accuracy of a test (von Mayrhauser, 1992). Early efforts at evaluating tests focused on their cost efficiency and on reliability or freedom from measurement errors. As Bingham (1923) noted, simply evaluating a test by its reliability was inadequate. What was needed is a method to determine the relationship between the test and a measure of success, the criterion.15 By around 1910, the cognitive-based tests pioneered by Binet were replacing the anthropomorphic tests of Cattell (Sokal, 1984). The test-criterion method, evaluating a test by its correlation with the criterion, was becoming standard procedure by this time, although the current term to describe this, validity, was not commonly used until the 1920s (T. B. Rogers, 1995). The use of the term criterion to describe a measure of an employee’s success (Bingham, 1926) or job proficiency (Burtt, 1926) also emerged in the 1920s (Austin & Villanova, 1992). This test-criterion method, familiar to present-day I/O psychologists and described in the 1920s by sources such as Freyd (1923–1924), Kornhauser and Kingsbury (1924), and Bingham and Freyd (1926), consisted of the following general steps: (a) A job analysis is conducted to obtain relevant factors necessary for job success; (b) A criterion, or measure of success, is selected. Criteria can be objective (e.g., number of items sold) or subjective (e.g., supervisor rankings); (c) Select or construct a predictor, generally

some type of test; (d) Correlate predictor scores with criterion scores to see whether there is a relationship; and (e) If there is a relationship, use a decision rule, regression for example, to determine acceptance or rejection of individual applicants. Freyd (1923–1924) discussed comparing the new procedure with existing procedures (what today is termed incremental validity) and Bingham and Freyd (1926) included issues of adequate sample size, reliability and validity of criteria, and the necessity of cooperation from the organization. Not all psychologists were comfortable with a strictly quantitative approach. Viteles (1925), for example, argued for an approach more compatible with practice in Europe and advocated a combination of statistical and clinical procedures that give consideration to the well-being and interests of individual workers. This period also saw an anticipation of the concept of utility in Hull’s (1928) Index of Forecasting Efficiency (H. C. Taylor & Russell, 1939). Industrial psychology was not only concerned with finding the best worker for a job, the question of selecting jobs for individuals, what came to be termed vocational psychology,16 was also an issue. Although considered a mainstream activity in industrial psychology during the early years of the field’s development, beginning in the 1930s psychologists interested in career guidance began to split from those interested in industrial applications. By the 1950s this activity was well on its way to becoming a part of counseling psychology (Savickas & Baker, 2005). In addition to the antecedents already mentioned for selection and industrial psychology in general, predecessors to vocational psychology can be found in vocational counseling work after the Civil War by the Young Men’s Christian Associations and in the vocational guidance work of Boston University law professor Frank Parsons (1909), who pioneered the model that vocational adjustment involves a fit between the requirements and routines of a job and the capabilities and characteristics of the

15

An analysis of 170 validity studies from 1906 through 1930 found that subjective ratings and rankings criteria were more common (68.1%) than objective criteria (31.9%). Across these studies, aptitude tests, especially special aptitude tests for special mental functions, were the most popular predictors. The two most common job types were clerical/office and manufacturing (29.5% and 25.5%, respectively). Service jobs constituted only 5.2% of the sample, an illustration of how the job market has changed from then to now (Vinchur, 2007).

16

See Savickas and Baker (2005) for a history of vocational psychology.

13

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

individual17 (Savickas & Baker, 2005). Particularly important for vocational placement was the later development of vocational interest tests pioneered in the CIT Division of Applied Psychology program. Moving from the laboratory to the real world was not without criticism by mainstream academic psychologists (e.g., Moore, 1961, for a personal reminiscence). To establish a professional identity, it was important for applied psychologists to differentiate themselves from nonscientific practitioners. This was particularly true for industrial psychologists involved in selection. Popular pseudoscientific selection procedures such as Katherine Blackford’s character analysis system (Blackford & Newcomb, 1914) for selecting individuals based on physiognomic characteristics such as face shape and hair color were subjected to scientific scrutiny and determined to be without merit (Cleeton & Knight, 1924; Paterson & Ludgate, 1922–1923). Other nonscientific procedures such as phrenology, graphology, palmistry, and mind reading were put to the test and also found wanting (Moore & Hartmann, 1931a). Psychologists were not immune to cloaking preconceptions in scientific garb, for example, Münsterberg’s “group psychology” that assigned mental traits to various ethnic groups (Baritz, 1960). Countries outside the United States were also involved in selection research and practice. For example, Salgado (2001) noted Ugo Pizzoli of Italy was selecting apprentices with tests as early as 1901, and Agostino Gemelli used psychophysical procedures for selecting aviators for the military in 1917 (McCollom, 1968). Jean Marie Lahy of France was using tests for selecting stenographers in 1905 (Viteles, 1923) and streetcar drivers in 1908 (Fryer & Henry, 1950). It was in Germany, however, where the emphasis on selection was comparable with that in the United States (Viteles, 1923). Industrial psychology, referred to as psychotechnics, got an early start in Germany. In 1907, Otto Lipmann cofounded with William Stern the journal Zeitschrift für angewandte Psychologie [Journal for Applied Psychology] (Viteles, 1932). Walter Moede and Curt Piorkowski

were selecting army chauffeurs by 1916. In addition, in 1917, Stern investigated streetcar drivers; a selection laboratory was established by the Saxon Railway Company in Dresden in 1917 (Viteles, 1925–1926); and in 1918 motormen were studied by the Greater Berlin Tramways (Viteles, 1923). The perceived successes of psychology in World War I, particularly psychological testing, contributed to a boom in the field’s popularity in the United States in the 1920s. The country was prosperous and employment levels were high. For industrial psychology and testing, there was initial success, followed by overconfidence and overselling in the mid 1920s, disillusion by business, and a period of decline in the late 1920s (Sokal, 1984). To be fair, individuals other than psychologists were promoting tests, and reputable psychologists (e.g., Kornhauser & Kingsbury, 1924) were careful not to oversell the usefulness of these tests. In addition, there were psychologists who went beyond the simplistic “square peg in a square hole” view of personnel selection. The Scott Company, for example, when staffing personnel departments used a “worker in his work” approach, which viewed the worker and the job as an integral unit, each capable of change in response to the other (Ferguson, 1961). With a worldwide depression and subsequent high level of unemployment, demand for industrial psychology in the United States declined greatly in the 1930s (Hale, 1992). In Germany, while individual psychologists such as Otto Lipmann and William Stern were dismissed from their posts by the Nazi regime, demand by the military for selection contributed to an increased need for industrial psychology. Psychotechnics, once almost exclusively in industry the domain of engineers in the 1920s (albeit engineers trained by psychologists such as Moede), opened up to psychologists in the 1930s (Geuter, 1992).18 In Japan, the 1930s and 1940s saw an expansion of personnel testing and placement in the military, using adapted American tests and ones developed in Japan (Warr, 2007). One notable non-

17

Hugo Münsterberg met with Parsons to discuss putting his vocational guidance system on a scientific basis. Münsterberg believed that individuals were unaware of their own capabilities and would need psychology to resolve this problem (Savickas & Baker, 2005).

18

Industrial psychology had a complicated relationship with the military and industry under the Nazi regime. A book by Geuter (1992) is a good source on this time period; Sprung and Sprung (2001) provide a useful synopsis.

14

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

military development in the 1930s was Morris S. Viteles’s19 publication of his landmark text, Industrial Psychology, in 1932. In addition to a comprehensive treatment of industrial psychology up to that date, Viteles provided information on selection practices both in the United States and abroad. In the United States, World War II provided a major stimulus for selection research and practice. The enormous need for soldiers put a serious strain on existing military personnel systems (Hale, 1992). The old Army Alpha was replaced in 1939 by the Army General Classification Test (Harrell, 1992), and Walter V. Bingham, soon to become the Army’s chief psychologist, chaired the Committee on Classification of Military Personnel. This committee was responsible for developing tests for classifying officers, trade tests, achievement tests, and various aptitude tests (Napoli, 1981). The Navy established the Applied Psychology Panel in 1942, which developed over 250 tests for classification and selection. The Navy’s aviation psychology program was directed by John G. Jenkins, and under the supervision of John C. Flanagan, the Aviation Cadet Qualifying exam was constructed for the Army Air Force (Napoli, 1981). In the Office of Strategic Services, psychologists adapted earlier German and British efforts to develop a program of global assessment for candidates anticipating sensitive assignments. Using multiple methods, most notably situational tests, this program was the forerunner of the assessment center used in industry (Highhouse, 2002). The assessment center model was introduced into Bell System in 1957 by Douglas Bray and associates (Bray & Campbell, 1968; Bray & Grant, 1966). Initial success at Bell resulted in adoption of assessment centers by a large number of organizations (Hale, 1992). H. C. Taylor and Russell (1939) advanced utility analysis research with their tables to predict success based not only on the validity coefficient, but also on the selection ratio (i.e., number of hires over total applicants) and base rate (i.e., percentage of individuals currently successful in the job). Brogden (1946, 1949; Brogden & Taylor, 1950) demonstrated how

the selection ratio and standard deviation of job performance affects the economic utility of a predictor, laying the groundwork for modern utility theory (see Cronbach & Gleser, 1957). The criterion came under increased scrutiny in the 1940s and 1950s, with calls for more research and developments such as Flanagan’s (1954) critical incident technique and Wherry’s rating process model (see Wherry, 1983). Increased criticism of testing, particularly personality testing, was evident in the 1950s and early 1960s both inside (Guion & Gottier, 1965) and outside the field (Baritz, 1960; Gross, 1962; Whyte, 1954). The civil rights movement of the 1950s and subsequent legislation and court decisions brought issues of test fairness and differential validity to the forefront. In particular, Title VII of the Civil Rights Act of 1964 and an interpretation for employers, the 1978 Uniform Guidelines on Employee Selection Procedures (Equal Employment Opportunity Commission, 1978), forced industrial psychologists to evaluate their selection procedures in light of their differential impact on legally protected minority groups. Related to this was Schmidt and Hunter’s (1977) research on validity generalization. Demonstrating that much of the variability in validity coefficients across situations and groups was often due to statistical artifacts, Schmidt and Hunter provided evidence against the prevailing orthodoxy of the situation specificity of validity. They later used this logic to develop a meta-analytic procedure (Hunter & Schmidt, 1990; Hunter, Schmidt, & Jackson, 1982) that has proven to be very influential in selection research. Metaanalysis has become standard operating procedure for evaluating predictors in selection research. Due to this meta-analytic research and advances in personality theory, use of personality measures saw something of a comeback in the late 1980s and 1990s. Before leaving this section we would like to briefly discuss the flip side of selecting the best candidate for the job: best adapting the job to the individual worker. Like vocational psychology, the field of human factors has roots in and was once considered to be part of mainstream industrial psychology. Also like vocational counseling, human factors went

19

Viteles’s undergraduate, masters, and 1921 doctoral degrees were all from the University of Pennsylvania, where he had a long and very productive career as a professor and dean. Biographical and autobiographical information about Viteles can be found in Thompson (1998) and Viteles (1967, 1974).

15

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

its separate way, despite attempts at a rapprochement in recent years (see Lane, 2007). Human factors goes by other names, such as ergonomics, human factors engineering, human engineering, and engineering psychology; variations that have been defined with varying degrees of difference. All are generally concerned with fitting the job or environment to the human operator and as such, have roots in both experimental and applied psychology.20 Experimental psychology, the functionalist worldview, scientific management, and the anthropometry pioneered by Francis Galton and James McKeen Cattell were all influences on human factors (Lane, 2007). One major influence was the scientific management procedures of Taylor, the Gilbreths, and others, with their emphasis on reducing fatigue and errors and increasing efficiency. The demands of World War I provided psychologists opportunities to apply their skills to developing displays and controls for the increasingly complex tanks, planes, and other equipment. Although activity slowed following the war, there was continuing work done in the 1920s (e.g., Ohio State University and the automobile industry) and the 1930s, particularly in aviation psychology (Lane, 2007). World War II brought an increased set of demands to what were now being called engineering psychologists. This war also saw the expansion of human factors to the point where this is now a separate, multidisciplinary field. PERFORMANCE APPRAISAL AND TRAINING As we noted in the section on employee selection, performance ratings were used early on as criteria for test validation. Paterson (1950) credited Karl Pearson with the first psychological rating scale for his 1907 scale used to estimate intelligence. It was Walter Dill Scott and his colleagues at the Division of Applied Psychology at CIT, however, who in 1917 developed the prototype of rating scales subsequently used in I/O psychology. First developed for use in selecting salespeople and later termed the Man-to-Man Rating Scale (Ferguson, 1961), the scale was modified by Scott for use in rating Army 20

officers in World War I. As Ferguson (1962–1965) described in detail, the Army was initially reluctant, but by dogged determination Scott convinced them of the scale’s usefulness and eventually it was used to evaluate more than 180,000 officers. Farr and Levy (2007) noted that although Scott’s scale was popular, it was not without problems. Raters constructed their own master scales, a difficult and time-consuming process that did not permit comparability across raters. Farr and Levy discussed other rating scales available at the time, such as the Specific Instance Scale, a forerunner of Behaviorally Anchored Rating Scales (Smith & Kendall, 1963) in that scale anchors were developed not by the raters, but were performance examples generated by scale developers; and the Descriptive-Term Scale that used descriptive adjectives as anchors and was very popular in the 1920s. It was the graphic rating scale, however, that had the greatest impact. According to Freyd (1923), the graphic rating scale originated in the Scott Company laboratory. To use this popular scale, the rater simply needs to make a mark on a continuum corresponding to where the ratee falls on a particular behavior or trait. Anchors are provided for reference. Psychologists were aware early on that consistent rating errors could occur with this method. Thorndike, for example, described the halo error in a 1920 article. Farr and Levy (2007) pointed out, however, in the 1920s graphic rating scales were scored normatively and not in raw score form, which made rating errors such as leniency and central tendency less of a problem. As was true for their effect on selection research, for performance appraisal the dire economic conditions of the 1930s likely resulted in a suppression of new activity (Katzell & Austin, 1992). Farr and Levy (2007), however, did note some precursors of later developments that occurred in the 1930s. Harry Hepner (1930) was one of the first psychologists to discuss the potential advantage to organizations of using timely ratings feedback to improve individual performance. H. L. Humke (1938–1939) agreed, stating that ratings could be used as a means to identify individual employees for training, termination, or promotion. And Herbert Moore (1939) discussed

Grether (1968), for one, disagrees. He states that engineering psychology originated in World War II and was an outgrowth of experimental, not industrial psychology.

16

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

promotion ratings as a way to foster a perception that decisions were made in a just fashion, rather than based on the subjective judgment of a superior. World War II brought the same large-scale personnel problem as World War I: how to successfully evaluate and then place a large number of individuals into appropriate positions in a short period of time. Because the Army’s existing officer performance rating system was found to lack sensitivity (i.e., it did not discriminate performance levels among candidates), a new forced-choice procedure was developed to provide a sound basis for personnel decisions (Sisson, 1948; for a discussion, see Farr & Levy, 2007). Following the war, the Army funded Robert Wherry’s groundbreaking theory of ratings research. Not published until the 1980s (Wherry & Bartlett, 1982; Wherry, 1983), this theory, although in the testmetaphor tradition, included a number of psychological, situational, and procedural variables as potentially affecting the accuracy of ratings (Farr & Levy, 2007). Farr and Levy (2007) noted that by the end of the 1950s and the early 1960s there were two literatures extant in performance appraisal. The first focused on measurement, accuracy, and rating format, with a move away from trait ratings to behavioral ratings. This period saw the development of Behaviorally Anchored Rating Scales (Smith & Kendall, 1963) and then later its acronym-labeled variants such as Mixed Standard Scales (MSS; Blanz & Ghiselli, 1972) and Behavior Observation Scales (BOS; Latham & Wexley, 1977). The second literature showed an increased emphasis on the effect of performance ratings on employees’ lives and careers and consequent research on feedback. In 1980, Landy and Farr called for a moratorium on format research and an emphasis on researching the cognitive processes underlying appraisal. This reorientation has occurred, with an increased emphasis on cognitive and social factors involved in performance appraisal (Farr & Levy, 2007). The course of training research and practice in I/O psychology was similar to the field as a whole. Influenced by scientific management and a strong focus on efficiency, training methods concentrated on improving productivity and safety through simplification and standardization in the first quarter of the 20th century. This was followed by the recogni-

tion of the importance of employee attitudes and motivation, and, in the latter half of the century, the inclusion of cognitive and systems models of the training process. Prior to 1900, formal training was primarily conducted in apprenticeship programs or perhaps technical or trade schools (Kraiger & Ford, 2007). Psychologists were optimistic that laboratory work on learning could be used to improve training in industry. Münsterberg (1913), for example, found current unsystematic apprenticeship training to be an enormous waste of energy. He advocated an experimental approach to training, using recent studies on reading acquisition, telegraphy, and typewriting as examples. As they grew in size and complexity, businesses began to set up their own factory schools (Kraiger & Ford, 2007). One of the first formal training programs was instituted in 1912 by the American Steel and Wire Company (Baritz, 1960). A common training method was the use of vestibule schools, usually a room containing duplicates of factory or office machinery where new workers can be trained under careful supervision (Burtt, 1929). Moore and Hartmann (1931a) thought vestibule schools best met the training goals of the company, and Burtt (1929) believed this method resulted in wiser placement of the worker, better instruction because the focus is on instruction rather than productivity, and the ability to keep detailed records of a worker’s progress. As it did for selection and performance appraisal, World War I presented serious challenges for training; there was the need to immediately train approximately a half-a-million workers for almost 100 trade jobs. The need for standardization resulted in the “show, tell, do, check” method of the Energy Fleet Corporation of the U.S. Shipping Board, and by the war’s end a number of training principles were posited. These included the view that training should be done on the job by supervisors, who themselves should be instructed on how to train (Kraiger & Ford, 2007). Although after the war there was a dramatic increase in civilian use of psychological testing and selection, training remained an uncommon activity for industrial psychologists. Contemporary textbooks devoted relatively little space to training, and Kraiger and Ford (2007) noted that although after 17

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

1930 there was a shift in training instruction from the supervisor to a training expert, that expert was rarely an industrial psychologist. In line with the increasing interest in worker attitudes and motivations of that time period, the focus of training expanded from skill acquisition and standardization to include consideration of personal concerns that may interfere with training (Kraiger & Ford, 2007). But there had also been recognition of this earlier. Henry C. Link (1923), for example, noted that in addition to developing the ability to do the job, training should arouse interest in the work and should also foster “goodwill” or “institutional spirit” in workers. As noted by Kraiger and Ford (2007), the period between 1930 and 1960 saw an increasing professionalization of the training process. There was also increasing variability of training methods. By 1960, case studies, role playing, human relations training, sensitivity training, and the use of simulations were being used. Needs assessment became a focus, as illustrated by its treatment in McGehee and Thayer (1961), the first I/O psychology book devoted to training. More attention was also being given to training program evaluation. In 1959 Kirkpatrick first discussed his influential four levels of training program criteria: reactions, learning, behavior change, and performance change. By the 1970s, the cognitive revolution and systems approaches to organizations began to work their way into training research; in 1974 Goldstein introduced his instructional systems design model of training (Kraiger & Ford, 2007). ORGANIZATIONAL PSYCHOLOGY AND BEHAVIOR Organizational psychology and its multidisciplinary sibling organizational behavior are younger disciplines than industrial psychology, although as we will see, their historical roots go back far before they were recognized as formal disciplines. The subject matter, research, practice, and history of organizational psychology and OB overlap to a considerable degree. There is no bright-line distinction between the two 21

disciplines, but attempts have been made to differentiate the two and to separate both from the related orientations of organizational theory (OT) and organizational development (OD). For example, Jex and Britt (2008), although conceding their similarity, noted that organizational psychology tends to be more parochial, relying primarily on other subdisciplines in psychology. OB draws more extensively from other fields such as sociology, economics, and anthropology. They also noted that organizational psychology tends to focus on individual behavior to a greater extent than OB, which is more comfortable with multiple levels of analysis. Vecchio (1995) saw OB as taking a theoretical–conceptual orientation at the micro (individual) level of analysis to distinguish it from OT, which also has a theoretical orientation but at a more macro level, and OD, which takes an applied macro perspective. And on a practical level, OB tends to be a staple of business schools; organizational psychology tends to be taught in departments of psychology. These distinctions are probably moot because all of these disciplines share many common pioneers and landmark events and organizational psychology textbooks cover a similar range of topics as OB texts (Highhouse, 2007).21 Not surprisingly for an interdisciplinary field, the origins of OB tend to be viewed through the disciplinary lens of the individual writer. Lawrence (1987) noted that psychologists tend to focus on the mid1940s contributions of Kurt Lewin and his colleagues on group behavior and leadership. Sociologists cite the earlier work of bureaucratic sociologists such as Peter Blau, Alvin Gouldner, Robert Morton, and Philip Selznick. Lawrence, however, argued that a plausible starting point centers around three earlier works from the 1930s. Two focused on the Hawthorne studies: Elton Mayo’s (1933) Human Problems of an Industrial Civilization and Fritz Roethlisberger and William Dickson’s (1939) Management and the Worker. The third was New Jersey Bell executive Chester Barnard’s Functions of an Executive (1938). Miner (2002), in contrast, viewed the Hawthorne work as the narrowly defined human relations approach. He placed the genesis of OB as a

This is also true for recent handbooks in I/O psychology, such as Anderson, Ones, Sinangil, and Viswesvaran (2001); Borman, Ilgen, and Klimoski (2003); and Dunnette and Hough (1992).

18

History of Research and Practice in I/O Psychology

Copyright American Psychological Association. Not for further distribution.

separate entity in the mid-1950s with the migration of social scientists into the business schools. The present survey goes back further than the 1930s and 1940s in examining the evolution of organizational topics. In the initial sections, landmark events such as the iconic Hawthorne studies that were instrumental in shaping the field are discussed. Then we look at the development of the human relations perspective. Following this, we focus on the history of topics of particular interest to I/O psychologists: organizational leadership, motivation and job satisfaction. EARLY INTEREST IN WORKER WELFARE Both organizational psychology and OB grew out of a human relations tradition, a concern for the welfare of the individual employee. Although it is true that the early years of American industrial psychology were characterized by an emphasis on productivity and efficiency, these were not the only concerns of the early industrial psychologists, especially for industrial psychologists outside the United States. In Great Britain, for example, there was considerable early emphasis on the happiness of workers (Myers, 1920), although efficiency was not neglected. Worker concern can be illustrated by the considerable criticism of the work of Taylor and the Gilbreths, although Pear (1948), possibly tongue-in-cheek, did note that at least Frank Gilbreth “injected some milk of human kindness” into Taylor’s “inhuman doctrines” (p. 112). Prior to the Hawthorne studies in the United States, Myers (1920) was speculating on psychological causes for restricted output and employee discontent; and in 1925 he described employee behavior that was very similar to what became known as the Hawthorne effect. Myer’s National Institute of Industrial Psychology, founded in 1921, had as its primary goal to “ease the effort required by the worker and not to endeavor to increase output by increased incentives” (Farmer, 1958, p. 265). As noted in an earlier section of this chapter, criticism by psychologists of Taylor’s lack of

consideration for the attitudes and emotions of the worker also occurred in other countries, including Germany and the United States. German psychologists, while sharing the United States’ emphasis on selection, also expressed concern for worker well-being. Otto Lipmann, for example, viewed industrial psychology as broader than the efficiency-based selection or scientific management approaches then popular. In addition to maximal performance to work, Lipmann discussed willingness-towork, that is, worker motivation and satisfaction. He believed that too much attention was paid to maximal performance through selection and not enough to willingness-to-work (Hausmann, 1931). As described by Viteles (1932), in the 1920s Lipmann directed a study by the Efficiency Committee of the German Industrial Inquiry Board to evaluate the effect on workers of the increasing use of machinery in coal mining. Miners objected to the new machinery, although they did use it. Lipmann took these objections seriously and attributed them to the conservative nature of workers, to the deprivation of the companionship workers shared when mining by hand, and to a fear of job loss. It is interesting to note the parallels between this early study and the later, better-known Tavistock Institute coal mining studies described later. Organizational psychology and OB texts that provide brief histories of the field tend to identify the same 20th century landmarks. OB texts stress the foundations laid by Frederick Taylor and scientific management; Henri Fayol and administrative theory; Max Weber22 and bureaucracy; the writings of Mary Parker Follett23; the human relations movement as exemplified by Abraham Maslow, Douglas McGregor, Rensis Likert, and Chris Argyris; and then progressing to contingency, open systems, and other more modern approaches to the study of organizations. Organizational psychology (e.g., Jex & Britt, 2008) and I/O texts understandably fold in more of the history of psychology, most notably the contributions of Kurt Lewin, but these texts also emphasize scientific management and human relations. We have discussed scientific management previously; we

22

Max Weber wrote on bureaucracy in the early part of the 20th century. Although the German version of his work was available in United States early on (Chester Barnard read it), it was not until the 1940s that translations became widely available (Miner, 2002).

23

Mary P. Follett’s writings and lectures on topics such as conflict, power (“power-over” vs. “power-with”), authority, participative leadership, and group processes anticipated modern management theory and practice and continue to have relevance today. See Tonn (2003) for a recent biography.

19

Vinchur and Koppes

take a brief look at some of the other landmarks shortly. First, however, we need to examine a set of studies central to all of these histories: the Hawthorne studies.

Copyright American Psychological Association. Not for further distribution.

HAWTHORNE STUDIES Just as the activities of the Division of Applied Psychology at CIT can be viewed as foundational to the industrial side of I/O psychology, the research conducted at the Western Electric Plant in Hawthorne, Illinois, has an iconic status in organizational psychology and OB. Hawthorne is seen as triggering a paradigm shift: from a mechanistic, efficiency-oriented approach to a more humane focus on the attitudes, social pressures, and general wellbeing of the individual worker. Roethlisberger (1941), one of the Hawthorne researchers, stated that the Western Electric researches seemed to him to be “the road back to sanity” in employee relations: human problems require human solutions (p. 8). The perception of the Hawthorne studies’ centrality persists, despite a virtual cottage industry over the years of criticism and reassessment (e.g., Highhouse, 2007). That the importance of these studies has endured despite this revisionist history is a testament to the power of the human relations message championed by the researchers, whether or not the data actually supported their conclusions. The basic research conducted and the conclusions drawn by the Hawthorne researcher are familiar to most students of organizations. Those interested in detail beyond this brief summary can find it in primary sources such as Mayo (1923, 1933) and Roethlisberger and Dickson (1939). The initial studies at Hawthorne were conducted between 1924 and 1927 by an independent group of the gas and electric lighting industry, the Committee on Industrial Lighting. Inspired by claims from industrial psychologists such as Münsterberg and Scott, the committee hoped to demonstrate that improved lighting would result in improved productivity and worker satisfaction (Highhouse, 2007). The research was supervised by C. E. Snow, head of the electrical 24

engineering department at the Massachusetts Institute of Technology (MIT; Hilgard, 1987). The researchers found that any variation in illumination intensity, or even no variation at all, resulted in an increase in productivity. In 1927, Hawthorne managers invited a group of researchers from the Harvard Graduate School of Administration to continue the research. Elton Mayo,24 who earlier had some success increasing production and decreasing turnover at a Philadelphia textile mill, emerged as the early spokesperson and chief publicist for the Hawthorne studies (Hilgard, 1987). The first study, conducted on a small group of women assembling telephone relays, examined the effects of varying rest periods, day and week length, and wage incentives on fatigue and monotony. Regardless of the manipulation, productivity increased. This increase in output was deemed by Mayo (1923) to be the result of the increased attention given to the workers by supervisors and to improved group dynamics; this effect was termed the Hawthorne effect. The final study conducted at Hawthorne was the Bank Wiring Room Study, in which the group dynamics of 14 men who wired telephone banks were observed. Despite being paid on an incentive system, the observer noted that the workers restricted output to a group norm. This norm was maintained by the group through group procedures that ranged from minor verbal harassment through physical harassment to, as a last resort, socially ostracizing the offender (Vecchio, 1995). Before the Bank Wiring Room study, a companywide interviewing program was conducted between 1928 and 1931, to allow employees to discuss their working conditions. The final result of the Hawthorne studies was a personnel counseling program begun in 1936. Highhouse (2007) noted that the goals of this program were to increase productivity and decrease dissatisfaction by using counselor concern to create a “positive Hawthorne effect” (p. 335). Management evaluated the program negatively and closed it in 1956 (Hilgard, 1987). As noted earlier, the Hawthorne studies were criticized heavily (see Landsberger, 1958). Alternative

(George) Elton Mayo was born in 1880 in Australia. After failing to obtain an MD degree, he immigrated to the United States in 1922 and obtained positions at the University of Pennsylvania and Harvard Business School (Miner, 2002). See Griffin, Landy, and Mayocchi (2002) for a discussion of Mayo’s concept of revery, a state of consciousness Mayo believed had negative effects of workers and that influenced his work at Hawthorne.

20

History of Research and Practice in I/O Psychology

Copyright American Psychological Association. Not for further distribution.

explanations for the results included financial incentives, managerial discipline, and the fear of layoffs. Roethlisberger and Dickson (1939) were somewhat circumspect in their treatment of the Hawthorne research and they did acknowledge alternative explanations for their findings. It is indisputable, however, that these studies had a major impact on organizational psychology and OB. As Highhouse (2007) put it: “Only the human relations message survived . . . in the oral history of the Hawthorne experiments” (pp. 334–335). HUMAN RELATIONS AND BEYOND Despite the ambiguity regarding their interpretation, the Hawthorne studies ushered in a sea change in the study of organizations. Many topical areas of organizational psychology and OB, such as leadership, motivation, group processes, and job satisfaction, have relevant antecedents in that research. We now turn to post-Hawthorne developments in organizational psychology and OB. The work of Kurt Lewin25 and his colleagues in the 1930s and 1940s deserves special mention. Lewin immigrated to the United States from Germany in 1933, working first at Cornell University and then at the University of Iowa. While at Iowa, Lewin conducted his groundbreaking research contrasting authoritarian, democratic, and laissez-faire leadership styles (Lewin, Lippitt, & White, 1939). Also in 1939, Lewin was invited to the Harwood Manufacturing Company by company vice president Alfred J. Marrow, who held a doctorate in psychology (Highhouse, 2007). The pajama manufacturing plant was experiencing turnover problems, and Marrow thought Lewin could help. Lewin, along with students Alex Bavelas and John R. P. French, Jr., instituted a number of studies that for a time made Harwood almost as well-known as Hawthorne (Hilgard, 1987). Notable was the work of Lester Coch and John R. P. French, Jr. (1948) on the use of participation to reduce resistance to change. Lewin’s work at Harwood set the stage for his action research model that has proven to be very influential in OD practice (W. L. French, 1982). Lewin’s (1951) “unfreezing, change, refreezing” model of change 25

established the conceptual framework for OD (Shafritz & Ott, 1996). Largely through the recruiting efforts of Douglas McGregor, in the early 1940s Lewin moved from Iowa to MIT, where he established the Research Center for Group Dynamics (W. L. French, 1982). In 1946, Lewin, Kenneth Benne, Leland Bradford, and Donald Lippitt conducted a workshop in New Britain, CT, on reducing intergroup tension, including racial tension. This effort was the genesis of the T-group method of attitude change. After Lewin’s death in 1947, the Research Center for Group Dynamics moved to the University of Michigan to join Rensis Likert’s Survey Research Center. These two entities, along with the Center for Utilization of Scientific Knowledge, became the Institute for Social Research (ISR). In 1962, ISR returned to Harwood, which had just acquired the Weldon Company, to explore Likert’s ideas about participative management and group dynamics (Highhouse, 2007; Hilgard, 1987). The 1950s and 1960s saw a continuation and branching out of research and practice in the human relations vein. Douglas McGregor (1957, 1960) introduced his influential conception of managerial beliefs regarding subordinates: Theory X managers believe workers find work aversive and therefore need to be closely controlled; Theory Y managers believe employees can be self-motivated if they find the work intrinsically rewarding. Chris Argyris (1957) discussed how modern organizations are in conflict with the personality of mature adults (Shafritz & Ott, 1996). Work on participative management continued with Likert’s (1961) “linking pin” model of integrating small groups into the organization and his taxonomy of management systems ranging from System 1 (exploitive–authoritarian) to System 4 (participative management; Hilgard, 1987). By the mid-1950s the assumptions of the human relations or OB perspective revolved around the fit between the organization and the individual: organizations exist to serve human needs; organizations and people need each other, and a good fit between the two benefits both (Shafritz & Ott, 1996). Conflict, decision making, and power, topics long of interest in political science, were receiving increased attention in organizations in the 1940s

Marrow (1969) provided a book-length biography of Kurt Lewin.

21

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

and 1950s (Perrow, 1986). Philip Selznick (1949) discussed goal conflict in his study of the Tennessee Valley Authority. In a 1953 address to the Society for the Psychological Study of Social Issues, Dorwin Cartwright argued that leadership, attitude change, and other variables could only be understood through the prism of power (Ott, 1989). J. R. French and Raven (1959) identified five bases of power; that same year Richard Cyert and James March discussed the impact of power and politics on organizational goals (Shafritz & Ott, 1996). The cognitive limits of decision making and the strategy of “satisficing” rather than working toward the optimal decision were explored by Cyert, March, and Herbert Simon (March & Simon, 1958; Simon; 1947). Their work on decision processes has been very influential. As Miner (2006) stated about OB: “No field could have asked for a more valuable send-off than . . . from the theorizing of Simon and March” (p. 57). In Great Britain, Eric Trist emerged as the leading spokesperson for the Tavistock Institute of Human Relations and the sociotechnical approach to organizations. Trist was influenced by the work of Kurt Lewin, by psychoanalysis, and by open systems theory and he took a decidedly humanist approach to his research (Miner, 2002). The sociotechnical approach postulates that changes in technology must take into account the social system of the organization. The theory has origins in the well-known long-wall versus short-wall coal mining studies conducted by Trist and Bamforth (1951). Another influential research study to come out of Great Britain was Joan Woodward’s (1958) classification of organizations based on the type of technology they employed: large-batch versus small-batch versus continuous-process technologies. Woodward found that effective organizational design depended on the type of technology employed. For example, large-batch, or mass production organizations tended to have bureaucratic structures, whereas small-batch organizations had more humanistic or organic structures. This distinction between mechanistic and organic organizations was also explored by Burns and Stalker (1961). Discussion of the history of big-picture topics such as organizational theory and structure, as well as topics of relatively recent interest such as organi22

zational communication and stress, are beyond the scope of this narrative. We now turn to three topics with a long history of interest in I/O psychology: leadership, motivation, and job satisfaction. LEADERSHIP Leadership was not a focus of study by psychologists during the early years of industrial and organizational psychology (Koppes & Pickren, 2007). In a review of industrial psychology texts, Day and Zaccaro (2007) found no mention of leadership in early industrial psychology texts such as Scott (1911) and Münsterberg (1913). They noted that the first obvious mention of leadership issues was in Viteles’ (1932) text in a section titled “The New Leadership in Industry.” Not until Blum’s 1949 text was an entire chapter devoted to leadership. There were earlier books specifically on leadership, such as a popularpress book that focused on leadership from a selfhelp perspective (Kleiser, 1923) and Tralle (1925); Craig and Charters (1925) first empirically based book about personal leadership in organizational settings; and Tead’s (1935) leadership book, which was grounded in industrial psychology (Day & Zaccaro, 2007). Leadership research in the 20th century proceeded through trait approaches, behavioral approaches, and situational or contingency approaches. According to Day and Zaccaro (2007), the “great man” approach served as the foundation for the trait approach to explaining leadership. Several leadership qualities that today would be labeled leadership traits were proposed in the early literature (e.g., Bingham, 1927; Craig & Charters, 1925). Trait studies were conducted during the 1930s and 1940s to discover personal attributes of leaders. These studies frequently compared leaders with nonleaders on personality and ability tests or calculated correlations between test scores and measures of leader effectiveness. Most of this early research was dominated by those with a social psychological perspective (Bird, 1940; Bogardus, 1934; Britt, 1943; Lewin, Lippitt, & White, 1939). Reviewers of this research were disappointed in the results of these early trait studies (Gibb, 1954; Mann, 1959; Stogdill, 1948) because the results were inconsistent and weak

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

(Yukl & Van Fleet, 1992). While these criticisms and public criticism contributed to a 1950s shift in research that focused on leader behaviors, trait research did continue (see Day and Zaccaro, 2007). Notable is the research conducted at AT&T (Bray, Campbell, & Grant, 1974; Howard & Bray, 1988). These researchers were interested in the career advancement of AT&T managers as predicted by diverse trait measures conducted during assessment centers. Dissatisfied with the inconsistent results of trait studies, researchers turned to identifying behaviors for explaining effective leadership. This approach considers what leaders do and how these behaviors relate to leader effectiveness. At the same time, the postwar economy produced an explosion of psychological applications and research opportunities (Koppes, 2003): Military research centers were formed, psychological research organizations were created, consulting firms were established, and research groups were formed within private companies. One university research center that partially devoted its efforts to leadership research was located at Ohio State University (OSU). In the 1950s, research conducted at OSU discovered measures of leadership behaviors through the perceptions of subordinates (Fleishman, 1953; Halpin & Winer, 1957). It was determined that subordinates’ views of leader behavior could fall into two major categories: initiating structure or task-oriented behaviors, and consideration or people-oriented behaviors. Two questionnaires resulted from this research, Leader Behavior Description Questionnaire and the Supervisory Description Questionnaire. These measures prevailed for the next 2 decades for survey research on leadership behavior; many of the behavior studies in the 1950s, 1960s, and 1970s were based on the OSU leadership measures. A similar research program was conducted at the University of Michigan under the direction of Daniel Katz (Katz & Kahn, 1952). That program and the OSU studies “represented perhaps the tipping point in the changing zeitgeist toward more situational models of leader effectiveness” (Day & Zaccaro, 2007, p. 391). In Japan, more than 30 years of research on these two categories revealed that both of these types of behavior are correlated to leader effectiveness (Misumi,

1985; Misumi & Peterson, 1985). Based on early studies by Lewin, Lippitt, and White (1939) and Coch and W. L. French (1948), a specific aspect of leadership behavior, participative leadership, and its consequences was developed. Following the heavy emphasis on behavioral research, the focus shifted to situational or contingency approaches (Day & Zaccaro, 2007). Contingency theories take into account situational factors that moderate leader effectiveness. The first and one of the most influential of these theories is Fiedler’s contingency theory, published in 1967, which is based on the assumption that effective leadership results from leader characteristics and the features of the situation. Fiedler identified two leader characteristics, task-oriented and relationshiporiented, that he measured with the Least Preferred Coworker scale (Fiedler, 1967). Fiedler also introduced the concept of situation favorability for the leader, which is determined by three factors: the relationship between the leader and subordinates (leader–member relations); the degree of structure in a particular task (task–structure); and the formal authority of the leader (position power). As each factor consists of two levels (e.g., high or low position power), eight situations, or octants, are possible. Task-oriented leaders are most effective in situations of either very high or very low favorableness; relationship-oriented leaders are effective in conditions of moderate favorableness. Other theories of note that take the leadership situation into account include the situational theory of Paul Hersey and Kenneth Blanchard (1969) that examined various leadership styles in light of the maturity (i.e., ability to perform the job) level of subordinates and Robert House’s (1971) path–goal theory that evaluates leadership style as a function of the subordinate and the situation. George Graen and his colleagues developed vertical dyad linkage theory, later termed leader–member exchange theory, focusing on the relationship between supervisor and subordinate (Miner, 2005). In the 1980s, charismatic and transformational leadership approaches emerged. Inspired by the work of Max Weber, the premise of charismatic and transformational theories is that leader behaviors and traits influence others as well as inspire them to achieve high goals or better perfor23

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

mance, which in turn lead to improvement in organizational effectiveness (Judge & Piccolo, 2004). As noted by Day and Zaccaro (2007), an interesting historical trend in the study of leadership is the change in research focus with regard to the organizational level of the leader. In early studies, the focus was on lower-level leaders such as foremen, student leaders, incarcerated criminal leaders, and military leaders among the noncommissioned and junior officer corps to name a few (e.g., Moore & Hartman, 1931b; Cowley, 1931; Jenkins 1947). Later, researchers began to focus on middle-to-upper level leaders (e.g., Browne, 1951; Shartle, 1949); however, the current emphasis on middle to upper-level leaders resulted from the AT&T studies (Bray, Campbell, & Grant, 1974). MOTIVATION AND JOB SATISFACTION Before the Industrial Revolution of the mid 1800s, not much systematic attention was given to employee motivation because units of production were small and capital investment was generally minimal. With the increased investment, competition, and emphasis on efficiency and productivity ushered in by the Industrial Revolution, employers focused on improving individual worker productivity, almost exclusively through monetary incentives (Dunnette & Kirchner, 1965). Although Taylor did recognize other incentives (Viteles, 1932), his scientific management approach provides an example of the use of financial incentives tied to performance. As noted previously, this approach was problematic. Individual workers were seen as interchangeable with one another, all employees were treated the same, and all were assumed to be motivated by money. Although the use of financial incentives did have some success, owners reset performance standards, workers caught on, and restriction of output was often the result (Dunnette & Kirchner, 1965). Münsterberg (1913) discussed factors other than money, such as fatigue, working hours, weather, atmospheric conditions, and especially mental monotony, that could potentially influence the worker (he also devoted the better part of a chapter discussing the effects of alcohol on performance). Münsterberg did seem to be moving toward recog24

nition of the importance of social factors in the workplace, using examples of the positive effect of interventions such as introducing a pet cat to a group of workers. For a short period in the early 20th century psychologists flirted with the notion that economic activity is instinctual rather than completely rational and they went about identifying various instincts to account for work behavior. This instinct explanation for motivation proved to be a tautology lacking logical and empirical support (Viteles, 1932). For the most part, however, it was not until the results of the Hawthorne studies were disseminated that American industrial psychologists began to pay serious attention to social factors in the workplace. As previously noted, these Hawthorne studies are key to the human relations movement and to the idea that productivity cannot be viewed separately from the attitudes, motivations, and satisfactions of the worker. In the early part of the 20th century American industrial psychologists were primarily concerned with improving efficiency and productivity, especially through selection, and although there were exceptions, they were less concerned with employee welfare and motivation per se. As described earlier, outside of the United States in Germany and Great Britain, there was more concern with the influence of social factors on motivation and satisfaction. Lipmann (1928–1929), for example, noted his concern that technical innovations were severing the link between work and the worker, thereby decreasing employee satisfaction. Mayo’s (1933) explanation for the Hawthorne effect was that group processes, norms, and communication were the cause of increased production, not the manipulation of environmental variables such as illumination levels and rest pauses (Hilgard, 1987). Recognizing the importance of allowing the workers to express their opinions about their jobs, in 1928 the Western Electric plant instituted a program of individual confidential conferences between employees and trained interviewers. Analysis of the results of these interviews found that many factors other than payment, such as work interest, sanitation, and social contact, were involved in motivating workers. Of these factors, the relationship between the worker and the first-line supervisor was determined to be most important for employee morale, happiness, and

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

efficiency (Viteles, 1932). Kornhauser and Sharp (1931–1932) also found supervisory practices to be an important influence on attitudes; however, in contrast with the Hawthorne researchers, they found no relationship between employee efficiency ratings and attitudes (Viteles, 1932). The development of attitude surveys were central to research and practice on both motivation and job satisfaction. There were early efforts to measure attitudes via interviews or work participation. Whiting Williams (1925), for example, worked alongside miners, railroad workers, and factory workers to gauge their reactions to their jobs. He concluded that although money was important, of equal importance was the social status the job provided. Williams also noted that workers and supervisors were ignorant of each others’ needs and desires, that feelings and experiences trump reason and logic in attitudes, and that one cannot separate the worker from outside roles and responsibilities (Viteles, 1932). It was not until the mid 1920s that standardized attitude questionnaires were developed. Management consultant J. D. Houser (1927) developed a standard set of questions and categorized responses according to positive and negative emotions; his initial scale, however, required one-on-one interviews (Landy, 1989). Building on the scaling methods of former CIT staffer L. L. Thurstone (1927; Thurstone & Chave, 1929), Rensis Likert in a 1932 doctoral dissertation introduced the simpler and now ubiquitous fivechoice format now known as the Likert scale. By the early 1930s Kornhauser (1933; cited in Wright, 2006) was able to identify five methods of measuring attitudes: (a) an impressionistic, informal approach exemplified by the Williams (1925) study discussed above; (b) an unguided interview, in which employees are encouraged to discuss topics important to them; (c) a guided interview; (d) attitudinal questions blanks, such as those used by Houser (1927); and (e) the more psychometrically sound scales such as those developed by Thurstone and later by Likert. Another doctoral dissertation that would prove to be a landmark in attitude and job satisfaction research was conducted by Robert Hoppock during the early 1930s. Using Thurstone’s scaling techniques, Hoppock’s research assistant (his father-inlaw) interviewed the majority of working adults in

New Hope, Pennsylvania (Landy, 1989). Hoppock (1935) found that the majority of workers reported being satisfied overall, only 12% were classified as dissatisfied. And although there was a fair amount of variability within occupational group, Hoppock was able to rank order occupational classifications by mean satisfaction. Professional, managerial, and executive employees reported the highest levels of satisfaction, followed by subprofessional, skilled manual and white collar, semiskilled, and finally unskilled manual laborers. Wright (2006) noted that research on job satisfaction per se was uncommon until the 1950s, which saw a rise in interest that has continued to the present day. For example, between the year of its inception in 1917 and 1946, only two articles in the Journal of Applied Psychology were published with the phrase “job/work satisfaction” in the title. Since the early 1950s thousands of articles have been published, many focusing on the all-important link between job satisfaction and productivity. Examination of this proposed link has a long history in I/O psychology, and it was more or less taken for granted that a “happy worker is a productive worker” by early researchers, including the Hawthorne researchers (Mayo, 1933). Initial empirical evidence for this link between satisfaction and performance was disappointing. Brayfield and Crockett’s (1955) review found little evidence for a relationship and neither did a review by Vroom (1964), who found a median r of only .14. Iaffaldano and Muchinsky’s (1985) meta-analysis found a similarly disappointing mean r = .17. A more recent meta-analysis by Judge, Thoresen, Bono, and Patton (2001) found a more optimistic estimated population correlation of .30 (Wright, 2006). It was in the 1930s that noted humanist psychologist Abraham Maslow began development of his need theory of motivation. Developed during the Great Depression and based on his observations of individuals having personal difficulties (Latham & Budworth, 2007), Maslow’s (1943) hierarchy of needs theory postulates that individuals progressed through five categories of needs: physiological, security, social, esteem, and self-actualization. Once one set of needs is more or less satisfied, the next group of needs becomes operative and motivates the indi25

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

vidual. Although not particularly well-supported, the theory had a large influence on managers. In 1953, Viteles’s book Motivation and Morale in Industry was published. As with his 1932 landmark text, his focus was consistent with the zeitgeist. Interest in motivation and attitudes has continued to be strong ever since. If research on motivation in industry had been largely atheoretical before 1950 (Latham & Budworth, 2007), this was no longer true after that date. Theories of motivation and motivation/satisfaction appeared on a regular basis in the 1950s and 1960s. These theories, familiar to most readers and a staple of most I/O psychology textbooks, are described only briefly here. Need theories were influenced by Freud’s theories of the unconscious and Murray’s (1938) early theory of needs (Miner, 2002). Clayton Alderfer’s (1972) existence, relatedness, and growth (ERG) theory was developed to deal with problems identified in Maslow’s theory. Alderfer postulated three sets of needs (existence, relatedness, and growth) and included a frustration–regression component (if a higher-level need is frustrated, individuals can regress to a lower need). Theories that focus on the personality trait need for achievement were developed by David McClelland, John Atkinson, and Bernard Weiner. McClelland’s approach includes three motives: achievement motivation (which includes fear of failure), power motivation, and affiliation motivation (Miner, 2002). Taking a different tack, Frederick Herzberg focused on characteristics of the job by postulating two sets of needs: hygiene factors and motivators (Herzberg, Mausner, & Snyderman, 1959). In motivation–hygiene or two-factor theory, dissatisfaction results if hygiene factors, such as pay and coworker relations, are substandard, and to truly satisfy and motivate a worker the job must be made intrinsically interesting through a process termed orthodox job enrichment. Herzberg’s view was influential in shifting attention from the physical work environment toward the intrinsic nature of the work (Miner, 2002). A related approach that also focused on job enrichment, job characteristics theory, was developed by Richard Hackman, Edward Lawler, and Greg Oldham (Hackman & Oldham, 1980). In the late 1960s and early 1970s, the operant conditioning 26

model of behaviorist B. F. Skinner received a flurry of attention. Termed organizational behavior modification, or O.B. Mod (Miner, 2002), this approach is fully developed in the work of Fred Luthans (e.g., Luthans & Kreitner, 1975, 1985). Equity theory, developed by J. Stacey Adams, sees both motivation and job satisfaction as resulting from the comparison of a worker’s perceived outcomes and inputs to the outcomes and inputs of a referent other. Adams was influenced in developing this theory by both the cognitive dissonance theory of Leon Festinger and social justice theories such as those of the sociologist George Homans (Miner, 2002). Expectancy theory had its roots in the purposive behaviorism of Edward Tolman and the work of Kurt Lewin. The first research on this approach from an organizational perspective was published by Basil Georgopoulos, Gerald Mahoney, and Nyle Jones of the University of Michigan’s Survey Research Center in 1957 (Miner, 2002). It was Victor Vroom (1964), however, who was responsible for the initial formal theory, with Lyman Porter and Edward Lawler contributing subsequently (Miner, 2002). The first comprehensive cognitive motivational theory developed by an I/O psychologist (Latham & Budworth, 2007), expectancy theory views motivation as a function of an individual’s expectations that effort will lead to performance, performance will lead to outcomes, and the value (valence) of those outcomes. While goal-setting theory has multiple historical antecedents, Kurt Lewin’s work on the determinants of aspiration level was key. Modern goal-setting theory was developed by Edwin Locke, later joined by Gary Latham (Miner, 2002). CONCLUDING REMARKS In surveying the history of a field as diverse as I/O psychology in a single chapter, it was inevitable that detail and nuance would be sacrificed, not to mention that whole content areas would be given limited or no coverage. As frustrating as it may be for a reader to look in vain for a particular topic or important individual, it was equally frustrating for us to have to leave so much out. The temptation was always to expand this section at bit or to shoehorn in this interesting fact or story; however the accumulation of

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

these small additions resulted in a manuscript far beyond the allotted space. We can only emphasize that this is a history, not the history, of I/O psychology and encourage readers to explore other historical perspectives and the primary sources that form the basis for these narratives. What, if any, general conclusions can one draw from this history of more than 100 years of applying psychology to the workplace? Limiting our observations for the most part to I/O psychology and not the broader field of OB, we have seen quantitative growth in number of I/O psychologists, graduate programs, journals, and books. I/O psychology has grown from a handful of pioneers at the turn of the last century to thousands of psychologists in academia and industry. Paradoxically, as the field has grown and prospered, the scope of research and practice has narrowed (Campbell, 2007). Whole areas of interest to early industrial psychologists, such as vocational counseling and human factors, have been co-opted by other disciplines and are no longer considered a part of mainstream I/O psychology. The bulk of current research and practice in recent years has focused on the dependent variables of performance and job satisfaction, with other dependent variables of interest to early industrial psychologists (e.g., fatigue, injury, performance stability) receiving little attention (Campbell, 2007). No enterprise that survives and prospers for more than 100 years does so without sensitivity to the demands of the greater environment. The worldwide origins of I/O psychology were both influenced and attuned to the rise of industrialization, the increasing difficulty of managing the resultant large work organizations, and the zeitgeist of the progressive era with its focus on efficiency and order. Personnel selection, the dominant activity of the early industrial psychologists, provides an illustration of how the expertise of these applied psychologists could be useful in pursuit of these progressive era goals. The perceived successes of applied psychology, particularly in testing, during World War I contributed to the expansion of the field. Throughout the 20th century the field was able to respond to other social, economic, cultural, and military events, notably World War II and the civil rights legislation of the 1960s, and continue to grow and adapt.

There are a number of ways to summarize the history of I/O psychology. One can look for general themes that reoccur throughout this history. Zickar and Gibby (2007), for example, identified four themes in America: the field’s emphasis on productivity and efficiency, the emphasis on quantification, the focus on selection and individual differences, and the interplay between science and practice. Others have taken a lessons-learned approach (e.g., Katzell & Austin, 1992), tracing the progress (or lack thereof) in various areas. One can point to the durability over time of certain practices or procedures, such as the predictor– criterion validation model (Vinchur, 2007) or the rating systems developed by early pioneers Paterson, Bingham, Freyd, and others (Farr & Levy, 2007). Another approach is to view the history of I/O psychology through the examination of various dichotomous tensions or conflicts (e.g., research vs. practice) that have persisted throughout the field’s history. As noted by Campbell (2007), these conflicts should not be resolved; there is merit on both sides. These tensions are useful in that they illustrate central issues and values important to understanding the history of the field. We would like to close this chapter by commenting on a few of these enduring conflicts. First, as discussed by Zickar and Gibby (2007), there is science versus practice, which is related to the academic versus practitioner divide. These tensions have been characterized a number of ways throughout I/O’s history. The early industrial psychologists were all too aware of the stigma attached to applied work and the majority of them held on to the security and prestige of their academic positions. Even Münsterberg was initially hostile to applied work (Benjamin, 2000). Although things have certainly improved over the last century, it is interesting to note that I/O psychology is still perceived as peripheral to mainstream academic psychology. Rozin (2006) noted that, unlike many other disciplines, psychology is organized around processes (e.g., sensation, cognition) and not major life domains, such as work. He speculated that the unintended result is that academic psychology therefore has paid minimal attention to these life domains. There is therefore the perception that I/O psychology is somewhat underappreciated by main27

Copyright American Psychological Association. Not for further distribution.

Vinchur and Koppes

stream academic psychology (Katzell & Austin, 1992). In recent years there has been much discussion of the perceived irrelevance of the published research for practitioners and the corresponding concern by researchers that practitioners are not relying enough on empirically validated procedures. This concern has also been evident from the very beginnings of the field. Although some conflict between the domains are inevitable, we believe the historical record supports the continued use of the scientist–practitioner model; it has served the field well. There are also long-standing tensions within the research and practitioner communities. There are advocates for a hard-headed, quantitative approach versus a more soft-headed, qualitative approach. The former is sometimes loosely more associated with the “I” side of the field, the latter with the “O” or OB perspective. Again, this tension is not new. For example, as far back as 1925 Max Freyd and Morris Viteles debated in the Journal of Applied Psychology the advisability of a strictly statistical approach to employee selection (Freyd, 1925) versus a more clinical, qualitative approach. Practitioners also had a number of issues to contend with, such as the us versus them distinction. Early industrial psychologists had to differentiate themselves from the nonscientific practitioners of character analysis, palmistry, and other pseudosciences who were competing in the workplace. Today this struggle continues, as I/O psychologists still have to convince managers that what they have to offer is different from the mass of management consultants and others offering solutions to organizational problems. As Brown (1992) noted, the professionalization of a field involves a balance between monopolizing the content knowledge developed in that field and popularizing that knowledge. This is an inherently contradictory process.26 Convincing management of the efficacy of I/O psychology’s procedures has been an ongoing task for the practitioner. And there is the question of the actual effectiveness of the various procedures applied to organizations. Although answering this question is beyond the scope of this essay, it is true that that while I/O psychology can 26

point to many successes, even landmark interventions such as the Hawthorne studies, the Harwood studies, the use of sensitivity training at the U.S. State Department, and the application of Theory Y management procedures at Non-Linear Systems, were much less successful than popularly believed (Highhouse, 2007; Miner, 2002). Finally, there is the question of who applied I/O psychology actually serves, the manager versus worker conflict. Bingham (1923) defined applied psychology as “psychology in the service of ends other than its own” (p. 294, original italics). Münsterberg (1913) argued for a neutral, scientific stance: psychologists should concern themselves only with “means” and not “ends.” In the first issue of the Journal of Applied Psychology, Roback (1917) criticized Münsterberg’s injunction of impartiality, noting that the applied psychologist operates more as an agent of whoever is paying his fee, rather than a broker who serves both parties fairly. The title of Baritz’s (1960) book, The Servants of Power: A History of the Use of Social Science in American Industry, left little doubt where he stood on the issue. Baritz stated that “. . . managers, are in business to make money. Only to the extent that industrial social scientists can help in the realization of this goal will management make use of them” (p. 196). Although there is little doubt that the management perspective has been the dominant one in the history of I/O psychology, this does not necessarily mean that I/O psychologists are antiworker. I/O’s history with organized labor, for example, is actually quite nuanced (see Gordon & Burt, 1981; Stagner, 1981; Zickar, 2001, 2003, 2004). In addition to I/O psychologists’ legitimate concern with helping managers improve efficiency, I/O psychologists should recognize they also have an obligation to the worker. As Viteles (1932) noted in the context of reducing worker fatigue, there must be “a willingness to sacrifice economic values when they clearly conflict with human values” (p. 465). The extent to which I/O psychologists have made these sacrifices has been the source of much debate (e.g., Kornhauser, 1947). I/O psychology has made a tremendous amount of progress over the past 100+ years. From a few pioneers at the turn of the last century, the field has

Brown (1992) viewed metaphor as a mechanism to resolve the contradiction. Specifically for applied psychology, she provided evidence that psychology borrowed metaphors from the more established professions of medicine and engineering to legitimize their own nascent profession.

28

Copyright American Psychological Association. Not for further distribution.

History of Research and Practice in I/O Psychology

grown into a vital scientific enterprise that has contributed greatly to our understanding of organizational behavior and to the effectiveness and well-being of organizations and organizational members. Some of those accomplishments are reflected in this chapter; others are detailed in the chapters of this handbook. Arthur and Benjamin (1999) stated that I/O psychology “. . . has an impact in one way or another on every person in the workforce” (p. 115). Despites setbacks and controversies, on balance the historical record shows that this impact has been and continues to be a positive one.

Benjamin, L. T., Jr., & Baker, D. B. (2003). Walter Van Dyke Bingham: Portrait of an industrial psychologist. In G. A. Kimble & M. Wertheimer (Eds.), Portraits of pioneers in psychology (Vol. 5, pp. 141–157). Mahwah, NJ: Erlbaum.

References

Bingham, W. V. (1927). Leadership. In H. C. Metcalf (Ed.), The psychological foundations of management (pp. 244– 260). Chicago: A. W. Shaw.

Aiken, H. G. J. (1985). Scientific management in action: Taylorism at Watertown arsenal, 1908–1915. Princeton, NJ: Princeton University Press. Alderfer, C. P. (1972). Existence, relatedness, and growth: Human needs in organizational settings. New York: Free Press. Anderson, N., Ones, D. S., Sinangil, H. K., & Viswesvaran, C. (Eds.). (2001). Handbook of industrial, work, and organizational psychology (Vol. 2). London: Sage. Argyris, C. (1957). Personality and organization. New York: Harper & Row. Arthur, W., Jr., & Benjamin, L. T. Jr. (1999). Psychology applied to business. In A. M. Stec & D. A. Bernstein (Eds.), Psychology: Fields of application (pp. 98–115). Boston: Houghton Mifflin. Austin, J. T., & Villanova, P. (1992). The criterion problem: 1917–1992. Journal of Applied Psychology, 77, 836–874. Baritz, L. (1960). The servants of power: A history of the use of social science in American industry. Middletown, CT: Wesleyan University Press. Barnard, C. I. (1938). The functions of the executive. Cambridge, MA: Harvard University Press. Benjamin, L. T., Jr. (1997a). Organized industrial psychology before Division 14: The ACP and the AAAP (1930–1945). Journal of Applied Psychology, 82, 459–466. Benjamin, L. T., Jr. (1997b). A history of Division 14 (Society for Industrial and Organizational Psychology). In D. A. Dewsbury (Ed.), Unification through division: Histories of the divisions of the American Psychological Association (Vol. 2, pp. 101–126). Washington, DC: American Psychological Association. Benjamin, L. T., Jr. (2000). Hugo Münsterberg: Portrait of an applied psychologist. In G. A. Kimble & M. Wertheimer (Eds.), Portraits of pioneers in psychology (Vol. 4, pp. 113–129). Washington, DC: American Psychological Association and Mahwah, NJ: Erlbaum.

Bills, M. A. (1953). Our expanding responsibilities. Journal of Applied Psychology, 37, 142–145. Bingham, W. V. (1919). Army personnel work. With some implications for education and industry. Journal of Applied Psychology, 3, 1–12. Bingham, W. V. (1923). On the possibility of an applied psychology. Psychological Review, 30, 289–305. Bingham, W. V. (1926). Measures of occupational success. Harvard Business Review, 5, 1–10.

Bingham, W. V. (1946). Clarence Stone Yoakum 1879–1945. American Psychologist, 1, 26–28. Bingham, W. V. (1952). Walter Van Dyke Bingham. In E. G. Boring, H. S. Langfeld, H. Werner, & R. M. Yerkes (Eds.), A history of psychology in autobiography (Vol. 4, pp. 1–26). New York: Appleton-CenturyCrofts. Bingham, W. V., & Freyd, M. (1926). Procedures in employment psychology: A manual for developing scientific methods of vocational selection. New York: McGraw-Hill. Bird, C. (1940). Social psychology. New York: AppletonCentury. Bjork, D. W. (1983). The compromised scientist: William James in the development of American psychology. New York: Columbia University Press. Blackford, K. M. H., & Newcomb, A. (1914). The job, the man, the boss. Garden City, NY: Doubleday, Page. Blanz, F., & Ghiselli, E. E. (1972). The mixed standard scale: A new rating system. Personnel Psychology, 25, 185–200. Blum, M. L. (1949). Industrial psychology and its social foundation. New York: Harper. Bogardus, E. S. (1934). Leaders and leadership. New York: D. Appleton-Century. Borman, W. C., Ilgen, D. R., Klimoski, R. J. (Vol. Eds.), Weiner, I. B. (Ed.-in-Chief). (2003). Handbook of psychology: Volume 12. Industrial and organizational psychology. Hoboken, NJ: Wiley. Bray, D. W., & Campbell, R. J. (1968). Selection of salesmen by means of an assessment center. Journal of Applied Psychology, 52, 36–41. Bray, D. W., & Grant, D. L. (1966). The assessment center in the measurement of potential for business management. Psychological Monographs: General and Applied, 80. 29

Vinchur and Koppes

Bray, D. W., Campbell, R. J., & Grant, D. L. (1974). Formative years in business: A long-term AT&T study of managerial lives. New York: Wiley.

Cheung, F. M. (2009, January). Report of the Task Force (TF) on regional development in Asia. The IAAP Bulletin, Issue 21-1.

Brayfield, A. H., & Crockett, W. H. (1955). Employee attitudes and employee performance. Psychological Bulletin, 62, 396–424.

Cleeton, G. U., & Knight, F. B. (1924). Validity of character judgment based on external criteria. Journal of Applied Psychology, 8, 215–231.

Bregman, E. O. (1921). A study in industrial psychology— Tests for special abilities. Journal of Applied Psychology, 5, 127–151.

Coch, L., & French, J. R. P., Jr. (1948). Overcoming resistance to change. Human Relations, 1, 512–532.

Bregman, E. O. (1926). Revision of the Army Alpha Examination: Form A. New York: Psychological Corporation.

Copyright American Psychological Association. Not for further distribution.

Brigham, C. C. (1923). A study of American intelligence. Princeton, NJ: Princeton University Press. Brigham, C. C. (1930). Intelligence tests of immigrant groups. Psychological Review, 37, 158–165. Britt, S. H. (1943). Social psychology of modern life. New York: Farrar & Rinehart. Brogden, H. E. (1946). On the interpretation of the correlation coefficient as a measure of predictive efficiency. Journal of Educational Psychology, 37, 64–76. Brogden, H. E. (1949). When testing pays. Personnel Psychology, 2, 171–183. Brogden, H. E., & Taylor, E. K. (1950). The dollar criterion: Applying the cost accounting concept to criterion construction. Personnel Psychology, 3, 133–154. Brown, J. (1992). The definition of a profession: The authority of metaphor in the history of intelligence testing, 1890–1930. Princeton, NJ: Princeton University Press. Browne, C. G. (1951). Study of executive leadership in business. Journal of Applied Psychology, 35, 36–37. Burns, T., & Stalker, G. M. (1961). The management of innovation. London: Tavistock. Burt, C. (1947). Charles Samuel Myers. Occupational Psychology, 21, 1–6. Burtt, H. (1926). Principles of employment psychology. New York: Harper. Burtt, H. (1929). Psychology and industrial efficiency. New York: Appleton. Campbell, J. P. (2007). Profiting from history. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 441–457). Mahwah, NJ: Erlbaum. Cattell, J. M. (1890). Mental tests and measurements. Mind, 15, 373–381. Reprinted in James McKeen Cattell: Man of science, Vol. 1. Psychological Research (1947). Lancaster, PA: Science Press. Cattell, J. M. (1923). The Psychological Corporation. Annals of the Academy of Political and Social Science, 110, 165–171. Cattell, J. M. (1946). Retrospect: Psychology as a profession. Journal of Consulting Psychology, 10, 289–291. 30

Cook, H. E., & Manson, G. E. (1925–26). Abilities necessary in effective retail selling and a method for evaluating them. Journal of Personnel Research, 4, 74–82. Cowley, W. H. (1931). The traits of face-to-face leaders. Journal of Abnormal and Social Psychology, 26, 304–314. Craig, D. R., & Charters, W. W. (1925). Personal leadership in industry. New York: McGraw-Hill. Cronbach, L. J., & Gleser, G. C. (1957). Psychological tests and personnel decisions. Urbana: University of Illinois Press. Day, D., & Zaccaro, S. (2007). Leadership: A critical historical analysis of the influence of leader traits. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 383–405). Mahwah, NJ: Erlbaum. deWolff, C. J., & Shimmin, S. (1976). The psychology of work in Europe: A review of a profession. Personnel Psychology, 29, 175–195. Dunnette, M. D., & Hough, L. M. (Eds.) (1992). Handbook of industrial and organizational psychology (2nd ed., Vol. 3). Palo Alto, CA: Consulting Psychologists Press. Dunnette, M. D., & Kirchner, W. K. (1965). Psychology applied to industry. New York: Appleton-CenturyCrofts. Equal Employment Opportunity Commission, Civil Service Commission, Department of Labor, & Department of Justice. (1978). Uniform guidelines on employee selection procedures. Federal Register, 43, 38290–38313. Farmer, E. (1958). Early days in industrial psychology: An autobiographical note. Occupational Psychology, 32, 264–267. Farr, J. L., & Levy, P. E. (2007). Performance appraisal. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 311–327). Mahwah, NJ: Erlbaum. Ferguson, L. W. (n.d.). A new light of the history of industrial psychology. Unpublished manuscript, Ferguson Collection, Carnegie Mellon University. Ferguson, L. W. (1952). A look across the years 1920–1950. In L. L. Thurstone (Ed.), Applications of psychology: Essays to honor Walter V. Bingham (pp. 7–22). New York: Harper.

History of Research and Practice in I/O Psychology

Ferguson, L. W. (1961). The development of industrial psychology. In B. V. Gilmer (Ed.), Industrial psychology (pp. 18–37). New York: McGraw-Hill. Ferguson, L. W. (September 1, 1962). Flyer announcing the Heritage of Industrial Psychology (in twelve volumes). Ferguson, L. W. (1962–1965). The heritage of industrial psychology [14 pamphlets]. Hartford, CT: Finlay Press. Fiedler, F. E. (1967). A theory of leadership effectiveness. New York: McGraw-Hill.

Copyright American Psychological Association. Not for further distribution.

Finch, F. H., & Odoroff, M. E. (1939). Employment trends in applied psychology. Journal of Consulting Psychology, 3, 118–122.

Gordon, M. E., & Burt, R. E. (1981). A history of industrial psychology’s relationship with American unions: Lessons from the past and directions for the future. International Review of Applied Psychology, 30, 137–156. Grether, W. F. (1968). Engineering psychology in the United States. American Psychologist, 23, 743–751. Griffin, M. A., Landy, F. J., & Mayocchi, L. (2002). Australian influences on Elton Mayo: The construct of revery in industrial society. History of Psychology, 5, 356–375. Gross, M. L. (1962). The brain watchers. New York: Random House.

Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51, 327–358.

Guion, R. M., & Gottier, R. F. (1965). Validity of personnel measures in personnel selection. Personnel Psychology, 18, 135–164.

Fleishman, E. A. (1953). The description of supervisory behavior. Personnel Psychology, 37, 1–6.

Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, MA: Addison-Wesley.

Fleishman, E. A. (1979). The new applied psychology: An international perspective. International Review of Applied Psychology, 28, 67–74.

Hakel, M. (1979). Proposal to incorporate as the Society for Industrial and Organizational Psychology. Industrial-Organizational Psychologist, 16, 4–5.

Fleishman, E. A. (1999). Applied psychology: An international journey. American Psychologist, 54, 1008–1016.

Hale, M., Jr. (1980). Human science and the social order: Hugo Münsterberg and the origins of applied psychology. Philadelphia: Temple University Press.

French, J. R. P., Jr., & Raven, B. (1959). The bases of social power. In D. P. Cartwright (Ed.), Studies in social power (pp. 150–167). Ann Arbor: Institute for Social Research, University of Michigan.

Hale, M. (1992). History of employment testing. In A. Widgor & W. R. Garner (Eds.), Ability testing: Uses, consequences, and controversies (pp. 3–38). Washington, DC: National Academy Press.

French, W. L. (1982). The emergence and early history of organizational development: With reference to influences on and interaction among some of the key actors. Group and Organization Studies, 7, 261–278.

Halpin, A. W., & Winer, B. J. (1957). A factorial study of the leader behavior descriptions. In R. M. Stogdill & A. E. Coons (Eds.), Leader behavior: Its description and measurement. Columbus, OH: Ohio State University.

Freyd, M. (1923). The graphic rating scale. Journal of Educational Psychology, 14, 83–102.

Hansen, J. C. (1987). Edward Kellogg Strong, Jr.: First author of the Strong Interest Inventory. Journal of Counseling and Development, 66, 119–125.

Freyd, M. (1923–24). Measurement in vocational selection. Journal of Personnel Research, 1, 215–249, 268–284, 377–385. Freyd, M. (1925). The statistical viewpoint in vocational selection. Journal of Applied Psychology, 9, 349–356. Fryer, D. H., & Henry E. R. (Eds.) (1950). Handbook of applied psychology (2 vols.). New York: Rinehart.

Harrell, T. W. (1992). Some history of the Army General Classification Test. Journal of Applied Psychology, 77, 875–878. Hausmann, M. F. (1931). Otto Lipmann and industrial psychology in Germany [Review of Grundriss der arbeitswissenschaft und ergebnisse der arbeitswissenschaftlichen statistik]. Personnel Journal, 9, 417–420.

Gale, H. (1896). On the psychology of advertising. In H. Gale (Ed.), Psychological studies (pp. 39–69). Minneapolis: Author.

Hearnshaw, L. S. (1964). A short history of British psychology 1840–1940. Westport, CT: Greenwood Press.

Geuter, U. (1992). The professionalization of psychology in Nazi Germany (R. J. Holmes, Trans.). Cambridge, England: Cambridge University Press.

Heller, W. J. (1929–1930). Industrial psychology and its development in Switzerland. Personnel Journal, 8, 435–441.

Gibb, C. A. (1954). Leadership. In G. Lindzey (Ed.), Handbook of social psychology. Cambridge, MA: Addison-Wesley.

Hersey, P., & Blanchard, K. H. (1969). Life cycle theory of leadership. Training and Development Journal, 23, 26–34.

Gilbreth, Jr., F. B., & Carey, E. G. (1948). Cheaper by the dozen. New York: Crowell.

Herzberg, F., Mausner, B., & Snyderman, B. S. (1959). The motivation to work. New York: Wiley.

Goldstein, I. L. (1974). Training: Program development and evaluation. Monterey, CA: Brooks/Cole.

Hepner, H. W. (1930). Psychology in modern business. New York: Prentice-Hall. 31

Vinchur and Koppes

Copyright American Psychological Association. Not for further distribution.

Highhouse, S. (2002). Assessing the candidate as a whole: A historical and critical analysis of individual psychological assessment. Personnel Psychology, 55, 363–396.

Judge, T. A., Thoresen, C. J., Bono, J. E., & Patton, G. K. (2001). The job satisfaction–job performance relationship: A qualitative and quantitative review. Psychological Bulletin, 127, 376–407.

Highhouse, S. (2007). Applications of organizational psychology: Learning through failure or failure to learn. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 331–352). Mahwah, NJ: Erlbaum.

Kaiser, A. E. (1989). A chronology of the theoretical developments in organizational behavior. In J. S. Ott (Ed.), Classic readings in organizational behavior (pp. 10–26). Pacific Grove, CA: Brooks/Cole.

Hilgard, E. R. (1987). Psychology in America: A historical survey. San Diego, CA: Harcourt, Brace, Jovanovich.

Kanigel, R. (1997). The one best way: Frederick Taylor and the enigma of efficiency. New York: Viking.

Hopkins, L. B. (1921, June 16). [Letter to active members of the Scott Company]. The Scott Company Collection, Box 1, Folder 5. Evanston, IL: Northwestern University Archives.

Katz, D., & Kahn, R. (1952). Some recent findings in human relations research in industry. In E. Swanson, T. Newcomb & E. Hartley (Eds.), Readings in social psychology (2nd ed., pp. 650–665). New York: Holt, Rinehart, & Winston.

Hoppock, R. (1935). Job satisfaction. Oxford, England: Harper. House, R. J. (1971). A path–goal theory of leader effectiveness. Administrative Science Quarterly, 16, 321–338. Houser, J. D. (1927). What the employee thinks. Cambridge, MA: Harvard University Press. Howard, A., & Bray, D. W. (1988). Managerial lives in transition: Advancing age and changing times. New York: Guilford Press. Hull, C. L. (1928). Aptitude testing. Yonkers-on-Hudson, NY: World Book. Humke, H. L. (1938–39). Full use of employee ratings. Personnel Journal, 17, 292–295. Hunter, J. E., & Schmidt, F. L. (1990). Methods of metaanalysis: Correcting error and bias in research findings. Newberry Park, CA: Sage. Hunter, J. E., Schmidt, F. L., & Jackson, G. B. (1982). Meta-analysis: Cumulating research findings across studies. Beverly Hills, CA: Sage. IAAP Division 1 of Work and Organizational Psychology. (n.d.). Invitation to join Division 1: Organizational Psychology. Retrieved March 29, 2009, from http:// iaapsy.org/division1/ Iaffaldano, M. T., & Muchinsky, P. M. (1985). Job satisfaction and job performance: A meta-analysis. Psychological Bulletin, 97, 366–373. Jacobson, J. Z. (1951). Scott of Northwestern. Chicago: Louis Mariano. Jenkins, W. O. (1947). A review of leadership studies with particular reference to military problems. Psychological Bulletin, 44, 54–79. Jex, S. M., & Britt, T. W. (2008). Organizational psychology: A scientist–practitioner approach (2nd ed.). Hoboken, NJ: Wiley. Judge, T. A., & Piccolo, R. F. (2004). Transformational and transactional leadership: A meta-analytic test of their relative validity. Journal of Applied Psychology, 89, 755–768. 32

Katzell, R. A., & Austin, J. T. (1992). From then to now: The development of industrial–organizational psychology in the United States. Journal of Applied Psychology, 77, 803–835. Kelly, R. M., & Kelly, V. P. (1990). Lillian Moller Gilbreth. In A. N. O’Connell & N. F. Russo (Eds.), Women in psychology: A bio-bibliographic sourcebook (pp. 117–124). New York: Greenwood Press. Kirkpatrick, D. L. (1959). Techniques for evaluating training programs. Journal of the ASTD, 13, 3–9. Kleiser, G. (1923). Training for power and leadership. New York: George H. Doran. Koppes, L. L. (1997). American female pioneers of industrial and organizational psychology during the early years. Journal of Applied Psychology, 82, 500–515. Koppes, L. L. (2003). Industrial–organizational psychology. In I. B. Weiner & D. K. Freedheim (Eds.), Comprehensive handbook of psychology (Vol. 1, pp. 376–389). New York: Wiley. Koppes, L. L. (Ed.). (2007). Historical perspectives in industrial and organizational psychology. Mahwah, NJ: Erlbaum. Koppes, L. L., & Pickren, W. (2007). Industrial and organizational psychology: An evolving science and practice. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 3–37). Mahwah, NJ: Erlbaum. Kornhauser, A. W. (1947). Industrial psychology as management technique and as social science. American Psychologist, 11, 224–229. Kornhauser, A. W., & Kingsbury, F. A. (1924). Psychological tests in business. Chicago: University of Chicago Press. Kornhauser, A. W., & Sharp, A. A. (1931–32). Employee attitudes: Suggestions from a study in a factory. Personnel Journal, 10, 393–404. Kraiger, K., & Ford, J. K. (2007). The expanding role of workplace training: Themes and trends influencing training research and practice. In L. L. Koppes

History of Research and Practice in I/O Psychology

(Ed.), Historical perspectives in industrial and organizational psychology (pp. 281–309). Mahwah, NJ: Erlbaum. Landsberger, H. A. (1958). Hawthorne revisited: Management and the Worker, its critics, and developments in human relations in industry. Ithaca, NY: Cornell University Press. Landy, F. J. (1989). Psychology of work behavior (4th ed.). Pacific Grove, CA: Brooks/Cole.

Copyright American Psychological Association. Not for further distribution.

Landy, F. J. (1992). Hugo Münsterberg: Victim or visionary. Journal of Applied Psychology, 77, 787–802. Landy, F. J. (1997). Early influences on the development of industrial and organizational psychology. Journal of Applied Psychology, 82, 467–477. Landy, F. J., & Farr, J. L. (1980). Performance rating. Psychological Bulletin, 87, 72–107. Lane, S. C. (2007). A historical view of human factors in the United States. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 243–263). Mahwah, NJ: Erlbaum. Latham, G. P., & Budworth, M.-H. (2007). The study of work motivation in the 20th century. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 353–381). Mahwah, NJ: Erlbaum. Latham, G. P., & Wexley, K. N. (1977). Behavioral observation scales for performance appraisal purposes. Personnel Psychology, 30, 225–268. Lawrence, P. R. (1987). Historical development of organizational behavior. In J. W. Lorsch (Ed.), Handbook of organizational behavior (pp. 1–9). Englewood Cliffs, NJ: Prentice-Hall. Lewin, K. (1951). Field theory in social psychology. New York: Harper. Lewin, K., Lippitt, R., & White, R. K. (1939). Patterns of aggressive behavior in experimentally created social climates. Journal of Social Psychology, 10, 271–301. Likert, R. (1961). New patterns of management. New York: McGraw-Hill. Link, H. C. (1923). Education and industry. New York: Macmillan. Lipmann, O. (1928–29). The human factor in production. Personnel Journal, 7, 87–95. Luthans, F., & Kreitner, R. (1975). Organizational behavior modification. Glenview, IL: Scott, Foresman. Luthans, F., & Kreitner, R. (1985). Organizational behavior modification and beyond: An operant and social learning approach. Glenview, IL: Scott, Foresman. Mann, R. D. (1959). A review of the relationship between personality and performance in small groups. Psychological Bulletin, 56, 241–270.

Manson, G. E. (1925–26). What can the application blank tell? Evaluation of items in personal history records of four thousand life insurance salesmen. Journal of Personnel Research, 4, 73–99. March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley. Marrow, A. J. (1969). The practical theorist: The life and work of Kurt Lewin. New York: Basic Books. Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50, 370–396. Mayo, E. (1923). The irrational factor in human behavior. Annals of the American Academy of Political and Social Sciences, 110, 117–121. Mayo, E. (1933). The human problems of an industrial civilization. New York: Viking. McCollom, I. N. (1968). Industrial psychology around the world: Part one: America and Western Europe. Part Two: Eastern Europe, Africa, Asia, and Australasia. International Review of Applied Psychology, 17, 3–19; 137–148. McGehee, W., & Thayer, P. (1961). Training in business and industry. New York: McGraw-Hill. McGregor, D. M. (1957). The human side of enterprise. Management Review, 46, 22–28. McGregor, D. M. (1960). The human side of enterprise. New York: McGraw-Hill. Miner, J. B. (2002). Organizational behavior: Foundations, theories, and analyses. New York: Oxford University Press. Miner, J. B. (2005). Organizational behavior 1: Essential theories of motivation and leadership. Armonk, NY: M. E. Sharpe. Miner, J. B. (2006). Organizational behavior 2: Essential theories of process and structure. Armonk, NY: M. E. Sharpe. Misumi, J. (1985). The behavioral science of leadership: An interdisciplinary Japanese research program. Ann Arbor, MI: University of Michigan Press. Misumi, J., & Peterson, M. (1985). The performancemaintenance (PM) theory of leadership: Review of a Japanese research program. Administrative Science Quarterly, 30, 198–223. Moore, B. V. (1961). Some beginnings of industrial psychology. In B. V. Gilmer (Ed.), Walter Van Dyke Bingham: Memorial Program/March 23, 1961 (pp. 1–5). Pittsburgh, PA: Carnegie Institute of Technology. Moore, B. V., & Hartmann, G. W. (Eds.). (1931a). Readings in industrial psychology. New York: Appleton-Century. Moore, B. V., & Hartmann, G. W. (1931b). The key position of foreman. In B. V. Moore & G. W. Hartmann (Eds.), Readings in industrial psychology (pp. 509–510). New York: Appleton-Century. 33

Vinchur and Koppes

Moore, H. (1939). Psychology for business and industry. New York: McGraw-Hill. Münsterberg, H. (1913). Psychology and industrial efficiency. Boston: Houghton Mifflin. Münsterberg, H. (1918). Business psychology. Chicago: LaSalle Extension University. Münsterberg, M. (1922). Hugo Münsterberg: His life and work. New York: Appleton. Murphy, G. (1930). A historical introduction to modern psychology (2nd ed.). London: Harcourt, Brace, & Co.

Copyright American Psychological Association. Not for further distribution.

Murray, H. A. (1938). Explorations in personality: A clinical and experimental study of fifty men of college age. New York: Oxford University Press. Muscio, B. (1920). Lectures on industrial psychology. London: Routledge. Myers, C. S. (1920). Psychology and industry. The British Journal of Psychology, 10, 177–182. Myers, C. S. (1925). Industrial psychology. New York: People’s Institute. Myers, C. S. (1936). Charles Samuel Myers. In C. Murchinson (Ed.), A history of psychology in autobiography (Vol. 3, pp. 215–230). Worcester, MA: Clark University Press. Napoli, D. S. (1981). Architects of adjustment: The history of the psychological profession in the United States. Port Washington, NY: Kennibat Press. Nelson, D. (1975). Managers and workers: Origins of the new factory system in the United States 1880–1920. Madison: University of Wisconsin Press. Oschrin, E. (1918). Vocational tests for retail saleswomen. Journal of Applied Psychology, 2, 148–155. Ott, J. S. (Ed.) (1989). Classic readings in organizational behavior. Pacific Grove, CA: Brooks/Cole. Parsons, F. (1909). Choosing a vocation. Boston: HoughtonMifflin. Paterson, D. G. (1950). Rating. In D. H. Fryer & E. R. Henry (Eds.), Handbook of applied psychology (pp. 147–154). New York: Rinehart. Paterson, D. G., & Ludgate, K. E. (1922–23). Blond and brunette traits: A quantitative study. Journal of Personnel Research, 1, 122–127. Pear, T. H. (1947). Charles Samuel Myers: 1873–1946. American Journal of Psychology, 60, 289–296. Pear, T. H. (1948). Industrial psychology as I have seen it. Occupational Psychology, 22, 107–117. Perloff, R., & Naman, J. L. (1996). Lillian Gilbreth: Tireless advocate for a general psychology. In G. A. Kimble, C. A. Boneau, & M. Wertheimer (Eds.), Portraits of pioneers in psychology (Vol. 2, pp. 106– 116). Washington, DC: American Psychological Association. 34

Perrow, C. (1986). The short and glorious history of organizational theory. In F. J. Landy (Ed.), Readings in industrial and organizational psychology (pp. 321– 331). Chicago: Dorsey. Pickren, W. E., & Fowler, R. D. (2003). Professional organizations. In Freedheim, D. K. (Vol. Ed.) & I. B. Weiner (Ed.-in-Chief), Handbook of psychology: Vol. 1. History of psychology. Hoboken, NJ: John Wiley & Sons. Pond, M. (1926–27). Selective placement of metal workers I. Preliminary studies. II. Development of scales for placement. III. Selection of toolmaking apprentices. Journal of Personnel Research, 5, 345–368, 405–417, 452–466. Prien, E. P. (1991). The Division of Applied Psychology at the Carnegie Institute of Technology. IndustrialOrganizational Psychologist, 29, 41–45. Roback, A. A. (1917). The moral issues involved in applied psychology. Journal of Applied Psychology, 1, 232–243. Roethlisberger, F. J. (1941). Management and morale. Cambridge, MA: Harvard University Press. Roethlisberger, F. J., & Dickson, W. J. (1939). Management and the worker: An account of a research program conducted by the Western Electric Company, Hawthorne Works, Chicago. Cambridge, MA: Harvard University Press. Rogers, H. W. (1946). Biographical information. Lafayette College Archives, Easton, PA. Rogers, T. B. (1995). The psychological testing enterprise: An introduction. Pacific Grove, CA: Brooks/Cole. Rozin, P. (2006). Domain denigration and process preference in academic psychology. Perspectives on Psychological Science, 1, 365–376. Salgado, J. F. (2001). Some landmarks of 100 years of scientific personnel selection at the beginning of the new century. International Journal of Selection and Assessment, 9, 3–8. Samuelson, F. (1977). World War I intelligence testing and the development of psychology. Journal of the History of the Behavioral Sciences, 13, 274–282. Savickas, M. L., & Baker, D. B. (2005). The history of vocational psychology: Antecedents, origin, and early development. In W. B. Walsh & M. L. Savickas (Eds.), Handbook of vocational psychology: Theory, practice, research (3rd ed., pp. 15–50). Mahwah, NJ: Erlbaum. Schmidt, F. L., & Hunter, J. E. (1977). Development of a general solution to the problem of validity generalization. Journal of Applied Psychology, 62, 529–540. Schumann, D. W., & Davidson E. (2007). Early influences of applied psychologists on consumer response: 1895–1925. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 265–280). Mahwah, NJ: Erlbaum.

History of Research and Practice in I/O Psychology

Scott, W. D. (1903). Theory of advertising. Boston: Small, Maynard. Scott, W. D. (1908). The psychology of advertising. Boston: Small, Maynard. Scott, W. D. (1911). Increasing human efficiency in business. New York: Macmillan. Scott, W. D., & Hayes, M. H. S. (1921). Science and common sense in working with men. New York: Ronald.

Copyright American Psychological Association. Not for further distribution.

Scripture, E. W. (1895). Thinking, feeling, doing. New York: Flood & Vincent.

Strong, E. K., Jr. (1918). Work of the Committee of Classification of Personnel. Journal of Applied Psychology, 2, 130–139. Strong, E. K., Jr. (1927). Vocational interest test. Educational Record, 8, 107–121. Tagg, M. (1925). Industrial psychology in Russia. Journal of the Institute of Industrial Psychology, 2, 359–364. Taylor, F. W. (1911). Principles of scientific management. New York: Harper.

Selznick, P. (1949). TVA and the grass roots. Berkley: University of California Press.

Taylor, H. C., & Russell, J. T. (1939). The relationship of validity coefficients to the practical effectiveness of tests in selection. Journal of Applied Psychology, 23, 565–578.

Shafritz, J. M., & Ott, J. S. (Eds.) (1996). Classics of organization theory (4th Ed.). Belmont, CA: Wadsworth.

Tead, O. (1935). The art of leadership. New York: McGraw-Hill.

Shartle, C. L. (1949). Leadership and executive performance. Personnel, 25, 370–380.

Thompson, A. S. (1998). Morris S. Viteles. American Psychologist, 53, 1153–1154.

Simon, H. A. (1947). Administrative behavior: A study of decision-making processes in administrative organizations. New York: Free Press.

Thorndike, E. L. (1920). A constant error in psychological ratings. Journal of Applied Psychology, 4, 25–29.

Sisson, E. D. (1948). Forced choice: The new army rating. Personnel Psychology, 1, 365–381. Smith, P. C., & Kendall, L. M. (1963). Retranslation of expectations: An approach to the construction of unambiguous anchors for rating scales. Journal of Applied Psychology, 47, 149–155. Sokal, M. M. (1981). The origins of the Psychological Corporation. Journal of the History of the Behavioral Sciences, 17, 54–67. Sokal, M. M. (1984). James McKeen Cattell and American psychology in the 1920s. In J. Brozek (Ed.), Explorations in the history of psychology in the United States (pp. 273–323). Lewisburg, PA: Bucknell University Press. Sokal, M. M. (1992). Origins and early years of the American Psychological Association. American Psychologist, 47, 111–122. Spillmann, J., & Spillmann L. (1993). The rise and fall of Hugo Münsterberg. Journal of the History of the Behavioral Sciences, 29, 322–338. Sprung, L., & Sprung, H. (2001). History of modern psychology in Germany in 19th and 20th century thought and society. International Journal of Psychology, 36, 364–376. Stagner, R. (1981). Training and experiences of some distinguished industrial psychologists. American Psychologist, 36, 497–505. Starch, D. (1910). Principles of advertising. Madison, WI: University Cooperative Co. Stogdill, R. M. (1948). Personal factors associated with leadership: A survey of the literature. Journal of Psychology, 25, 35–71.

Thurstone, L. L. (1927). Law of comparative judgment. Psychological Review, 34, 278–286. Thurstone, L. L., & Chave, E. J. (1929). The measurement of attitude. Chicago: University of Chicago Press. Tonn, J. C. (2003). Mary P. Follett: Creating democracy, transforming management. New Haven, CT: Yale University Press. Tralle, H. E. (1925). Psychology of leadership. New York: Century. Trist, E. L., & Bamforth, K. W. (1951). Some social and technical consequences of the long-wall method of coal-getting. Human Relations, 4, 6–38. Vecchio, R. P. (1995). Organizational Behavior (3rd ed.). Fort Worth, TX: Dryden. Vinchur, A. J. (2005). Charles Samuel Myers and Otto Lipmann: Early contributors to industrial psychology. Industrial-Organizational Psychologist, 43, 31–35. Vinchur, A. J. (2007). A history of psychology applied to employee selection. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (193–218). Mahwah, NJ: Erlbaum. Vinchur, A. J., & Koppes, L. L. (2007). Early contributors to the science and practice of industrial psychology. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (37–58). Mahwah, NJ: Erlbaum. Viteles, M. S. (1923). Psychology in business—In England, France, and Germany. Annals of the Academy of Political and Social Science, 110, 207–220. Viteles, M. S. (1925). The clinical viewpoint in vocational selection. Journal of Applied Psychology, 9, 131–138. Viteles, M. S. (1925–1926). Research in the selection of motormen. Part I. Survey of the literature. II. 35

Vinchur and Koppes

Methods devised for the Milwaukee Electric Railway and Light Company. Journal of Personnel Research, 4, 100–115; 173–199. Viteles, M. S. (1932). Industrial psychology. New York: Norton. Viteles, M. S. (1953). Motivation and morale in industry. New York: Norton.

Copyright American Psychological Association. Not for further distribution.

Viteles, M. S. (1967). Morris S. Viteles. In E. G. Boring & G. Lindzey (Eds.), A history of psychology in autobiography (Vol. 5, pp. 415–449). New York: Appleton-CenturyCrofts. Viteles, M. S. (1974). Industrial psychology: Reminiscences of an academic moonlighter. In T. S. Krawiec (Ed.), The psychologists (Vol. 2, pp. 440–500). New York: Oxford University Press. von Mayrhauser, R. (1987). The manager, the medic, and the mediator: The clash of professional psychological styles and the wartime origins of group mental testing. In M. M. Sokal (Ed.), Psychological testing and American society (pp. 128–157). New Brunswick, NJ: Rutgers University Press. von Mayrhauser, von R. (1992). The mental testing community and validity: A prehistory. American Psychologist, 47, 244–253. Vroom, V. H. (1964). Work and motivation. New York: Wiley. Warr, P. (2007). Some historical developments in I-O psychology outside the United States. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 81–107). Mahwah, NJ: Erlbaum. Wherry, R. J. (1983). Appendix: Wherry’s theory of rating. In F. J. Landy & J. L. Farr (Eds.), The measurement of work performance (pp. 283–303). New York: Academic Press.

A. LaFarge (Ed.), The essential William H. Whyte (pp. 43–66). New York: Fordham University Press. Williams, W. (1925). Mainsprings of men. New York: Charles Scribner. Wilson, M. A. (2007). A history of job analysis. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 219–241). Mahwah, NJ: Erlbaum. Wissler, C. (1901). The correlation of mental and physical tests. Psychological Review, 3, 1–63. Woodward, J. (1958). Management and technology. London: Her Majesty’s Stationery Office. Woodworth, R. S. (1919). Examination of emotional fitness for warfare. Psychological Bulletin, 16, 59–60. Wright, T. A. (2006). The emergence of job satisfaction in organizational behavior. Journal of Management History, 12, 262–277. Yukl, G., & Van Fleet, D. C. (1992). Theory and research on leadership. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 3, pp. 147–197). Palo Alto, CA: Consulting Psychologists Press. Zickar, M. J. (2001). Using personality inventories to identify thugs and agitators: Applied psychology’s contribution to the war against labor. Journal of Vocational Behavior, 59, 149–164. Zickar, M. J. (2003). Remembering Arthur Kornhauser: Industrial psychology’s advocate for worker wellbeing. Journal of Applied Psychology, 88, 363–369. Zickar, M. J. (2004). An analysis of applied psychology’s indifference to labor unions in the United States. Human Relations, 57, 145–167.

Wherry, R. J., & Bartlett, C. J. (1982). The control of bias in ratings: A theory of rating. Personnel Psychology, 35, 521–551.

Zickar, M. J., & Gibby, R. E. (2007). Four persistent themes throughout the history of I-O psychology in the United States. In L. L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (61–80). Mahwah, NJ: Erlbaum.

Whyte, W. H. (1954, September). The fallacies of “personality” testing. Fortune, pp. 117–121. Reprinted in

Zusne, L. (1984). Biographical dictionary of psychology. Westport, CT: Greenwood.

36

Copyright American Psychological Association. Not for further distribution.

CHAPTER 2

RESEARCH STRATEGIES IN INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY: NONEXPERIMENTAL, QUASI-EXPERIMENTAL, AND RANDOMIZED EXPERIMENTAL RESEARCH IN SPECIAL PURPOSE AND NONSPECIAL PURPOSE SETTINGS Eugene F. Stone-Romero

With little exception, advances in both the science and practice of industrial and organizational (I/O) psychology and allied disciplines (e.g., human resource management, organizational behavior, organization theory) hinge on the existence of findings from sound empirical research (referred to hereinafter as research). The results of research are used for several purposes. One is to develop theory about phenomena (e.g., individual behavior in organizations). Another is to test predictions stemming from such theory (e.g., expectancy theory). Yet another is to aid in the design and implementation of interventions (e.g., job enrichment) aimed at changing individuals, groups, and organizations (e.g., worker performance, group effectiveness, organizational efficiency). Whatever the purpose of research, its soundness is a function of the degree to which it allows for valid conclusions about (a) the existence of cause–effect relations between variables (i.e., internal validity); (b) the correspondence between the constructs (i.e., units, treatments, observations, and settings)

referenced by a researcher and their empirical realizations (i.e., construct validity); (c) the statistical estimates derived from a study (i.e., statistical conclusion validity); and (d) the extent to which relations found in a specific study generalize across different settings, units, treatments, and observations (i.e., external validity; Shadish, Cook, & Campbell, 2002). In view of the foregoing, the overall purpose of this chapter is to consider the factors that influence the validity of inferences derived from research, especially those concerned with causal connections between variables. Thus, the chapter has sections that deal with such issues as (a) study design, (b) the purposes of research, (c) the facets of validity in research, (d) the types of settings in which research is conducted, (e) the types of experimental designs that can be used in research (i.e., nonexperimental, quasi-experimental, and randomized experimental), (f) the joint consideration of experimental designs and research settings, (g) the important distinction between experimental design and statistical methods, and (h) some conclusions about the design and

This chapter is a substantially revised version of Stone-Romero (2009). I thank the editor and the associate editors for very helpful feedback on an earlier version of this chapter.

http://dx.doi.org/10.1037/12169-002 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

37

Eugene F. Stone-Romero

Copyright American Psychological Association. Not for further distribution.

conduct of research on phenomena of interest to I/O psychologists. Note that a number of research methods-related issues are covered elsewhere in this handbook. These include qualitative research (chap. 3, this volume), cross-level data analytic strategies (chap. 4, this volume), validation strategies used in personnel selection (Vol. 2, chap. 13, this handbook), and issues associated with cross-cultural research (Vol. 3, chap. 23, this handbook). Thus, this chapter does not offer coverage of these topics. RESEARCH DESIGN A distinction is made here between research design and experimental design. Research design or study design is an overall plan for conducting a study that considers several components (Cook & Campbell, 1979; Fromkin & Streufert, 1976; Kerlinger & Lee, 2000; Rosenthal & Rosnow, 2008; Runkel & McGrath, 1972; Shadish et al., 2002; E. F. Stone, 1978). First, the researcher must specify the units (e.g., individuals, groups, organizations) that will be studied. Second, he or she must choose from among three general ways of studying relations between (among) variables that differ in terms of the degree to which they provide control over possible confounds. Here, the experimental design options are nonexperimental, quasi-experimental, and randomized experimental designs. Note that in this chapter, experimental design refers to the degree to which a study uses experimental methods, thus allowing for control over confounds. Control is high with randomized experimental designs, moderate with quasi-experimental designs, and very low with nonexperimental designs (see the section titled Degree to Which a Design Is Experimental). In research that uses randomized experimental or quasi-experimental designs, the investigator must devise strategies for manipulating independent variables and measuring dependent variables. And in studies of the nonexperimental variety, he or she must determine how assumed independent, mediator, moderator, and dependent variables will be measured. Consider, for example, a nonexperimental study in which a researcher hypothesizes that (a) X (e.g., job satisfaction) → M (organizational com38

mitment) → Y (job performance) and (b) the relation between X and Y differs across levels of Z (job level). In this study, the hypothesized roles of variables are as follows: X is the independent variable, M is the mediator variable, Y is the dependent variable, and Z is the moderator variable. The mediator variable transmits the effect of the independent variable to the dependent variable. And, contingent on the level of the moderator variable, the relation between X and Y varies in terms of its magnitude or form. Note that because the study is nonexperimental, the role of each of the variables is assumed. As is explained in the next section, the researcher has no sound basis for inferring, for example, that X causes Y. It may very well be the case that (a) Y causes X or (b) both X and Y are caused by an unmeasured confounding variable. Third, depending on the experimental design, the researcher must specify the strategies that will be used in manipulating or measuring variables. Fourth, he or she must determine how studied units, treatments, settings, and outcomes will be sampled (e.g., randomly versus nonrandomly). Fifth, he or she must decide whether the study is to be conducted in a special purpose or a nonspecial purpose setting (Stone-Romero, 2002, 2009). Sixth, and finally, the researcher must specify the methods (statistical versus nonstatistical) that will be used in analyzing the data produced by the study. As is explained in the next section, there is a very important distinction between the extent to which a study is experimental and the statistical methods that are used to analyze the data produced by it. Contrary to what many researchers appear to believe, statistical methods are virtually always a very unacceptable substitute for randomized experiments in terms of a researcher’s ability to make valid inferences about causal connections between variables (Cook & Campbell, 1976, 1979; Cook, Campbell, & Peracchio, 1990; Ling, 1982; Rogosa, 1987; Rosopa & Stone-Romero, 2008; Shadish et al., 2002; Stone-Romero, 2009; StoneRomero & Rosopa, 2004, 2008). Thus, when causal inference is important, a researcher should conduct research that uses either randomized experimental designs or quasi-experimental designs that allow for ruling out likely threats to internal validity (Cook & Campbell, 1979; Shadish et al., 2002).

Research Strategies in I/O Psychology

Copyright American Psychological Association. Not for further distribution.

GENERAL PURPOSES OF RESEARCH Research on phenomena of relevance to organizations and their members is conducted for a number of major purposes, most of which involve assessing relations between (among) unobservable constructs using manipulations or measures of variables that serve to operationally define the constructs (Cook & Campbell, 1979; Guion, 2002; Nunnally, 1978; Nunnally & Bernstein, 1994; Shadish et al., 2002; E. F. Stone, 1978). For example, an investigator may study the relation between the constructs of worker ability (operationally defined by scores on a measure of job knowledge) and job performance (operationally defined by supervisor ratings). Often, he or she may not have an immediate interest in determining if an observed relation is causal in nature. However, having found a relation in one study, he or she may be interested in testing for a causal relation in a subsequent investigation. For example, a study may deal with the effect of an intervention designed to improve job knowledge (e.g., training) on job performance. A second major purpose of research is to determine the effects of various types of manipulations (interventions, treatments) of unobservable constructs on criterion constructs. Some examples include the effects of (a) job enrichment on job satisfaction, (b) realistic job previews on employee turnover, and (c) preemployment interview structure (structured vs. unstructured) on biases (e.g., sex, race, age) in ratings of job applicants. In such studies, causality is an important concern. A third major purpose of research is to determine whether causal or noncausal relations between (among) variables that are found in a study with a given set of units, treatments, and observations generalize across other types of units, treatments, and observations. For example, is a stress-reduction intervention as effective for police officers in a SWAT unit as it is for surgeons in a trauma center? FACETS OF VALIDITY IN RESEARCH A major objective in research is to generate valid inferences (conclusions) about the issues it

addresses. Thus, a researcher must be concerned with the correctness (i.e., truth value) of inferences associated with the four previously mentioned facets of validity, that is, internal, statistical conclusion, construct, and external (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002). The better the overall design of a study, the greater the confidence he or she can have about the validity of his or her research-based inferences.

A Model of Research-Based Inferences A model of factors that influence the validity of research-based inferences is shown in Figure 2.1. Arrow 1 in the figure links the operational definitions of cause and effect constructs, and relates to internal validity inferences. Arrows 2 and 3 deal with factors that affect the construct validity of measures and manipulations, that is, relations between constructs and their operational definitions (sampling particulars; Shadish et al., 2002). Arrow 4 is concerned with the correctness of statistical inferences about the relation between the operational definitions and deals with statistical conclusion validity. Finally, Arrow 5 has to do with factors that moderate the causal relation between X and Y, thus influencing external validity. Note that in the interest of limiting the complexity of Figure 2.1, it considers only the construct validity of operational definitions of assumed causes and effects. However, as noted in the following section, in an actual study the researcher also would

Cause construct 2

Factors that influence statistical conclusion validity Operational definition of cause construct

Effect construct

Factors that influence construct validity

1

4 5

3

Operational definition of effect construct

Factors that influence external validity

FIGURE 2.1. Validity facets in empirical research. From Handbook of Organizational Research Methods (p. 305), by D. Buchanan and A. Bryman (Eds.), 2009. London: Sage. Copyright 2009 by E. Stone-Romero. Reprinted with permission. 39

Eugene F. Stone-Romero

consider the construct validity of its units, treatments, and settings.

Focal construct

Operational definition

σ c2

σ o2

Copyright American Psychological Association. Not for further distribution.

Construct Validity Construct validity has to do with the degree of correspondence between the constructs referenced by a researcher and their empirical realizations. For a study to have a high level of construct validity, there must be a high degree of correspondence between (a) the units, treatments, observations, and settings to which inferences are made and (b) the study’s empirical realizations (sampling particulars) of the units, treatments, observations, and settings (Shadish et al., 2002). Thus, for example, there would be high construct validity of samples of an assumed cause (e.g., worker ability) and an effect (e.g., worker performance) if the manipulations and measures used in a study were valid empirical realizations of the underlying constructs (Cronbach & Meehl, 1955; Guion, 2002; Nunnally, 1978; Nunnally & Bernstein, 1994; Shadish et al., 2002; Stone-Romero, 1994). Construct validity of treatments and observations. Research tests for relations between variables using operational definitions are of two basic types, that is, manipulations (treatments) and measures (observations). Construct validity is critical to both such definitions. Manipulations are actions taken by a researcher to vary the value of an independent variable (Aronson, Ellsworth, Carlsmith, & Gonzalez, 1990; Cook & Campbell, 1976, 1979; Cook et al., 1990; Fromkin & Streufert, 1976; Shadish et al., 2002). Some examples include redesigning jobs, introducing quality control procedures, installing a computerized human resource information system, and implementing absence control policies and procedures. Measures are observations of the values of assumed independent, mediating, moderating, and dependent variables. There are numerous strategies for measuring variables including questionnaires, personality inventories, and aptitude and/or ability measures. These are considered in the subsection titled Measuring Variables. Figure 2.2 is concerned with the construct validity of operational definitions of assumed causes and effects. It considers both construct validity and the 40

Bias

σ b2 Deficiency

σ d2

Validity

σ v2

Unreliability

σ e2

FIGURE 2.2. The construct validity of operational definitions of manipulations and measures. From Handbook of Organizational Research Methods (p. 304), by D. Buchanan and A. Bryman (Eds.), 2009. London: Sage. Copyright 2009 by E. Stone-Romero. Reprinted with permission.

factors that detract from it in terms of variance in an unobservable, focal construct (σ 2c) and its operational definition (σ2o). As shown in the figure, the variance shared by these is construct validity (σ2v). Several factors detract from it. One is bias, that is, systematic variance in an operational definition that is unrelated to the focal construct (σ2b). For example, a questionnaire measure of organizational citizenship behavior might be biased by participants responding to its items in a socially desirable manner. Construct validity also might be affected adversely by unreliability, that is, nonsystematic (random) variance in an operational definition (σ2e). For instance, the just-mentioned measure might not be internally consistent (Nunnally, 1978; Nunnally & Bernstein, 1994). However, even if an operational definition is free from bias and unreliability, it may still have a low level of construct validity because it is deficient; that is, it may not fully capture the essence of the focal construct (Blum & Naylor, 1968; Borman, 1991; Smith, 1976; Stone-Romero, 1994). For example, assume that a researcher studied the relation between job enrichment and job satisfaction. The job characteristics model of Hackman and Oldham (1976) views enrichment as a function of the core job characteristics of autonomy, task signif-

Copyright American Psychological Association. Not for further distribution.

Research Strategies in I/O Psychology

icance, skill variety, task identity, and feedback from the job. In view of this conceptual definition, a measure of job enrichment would be deficient if it only had items that dealt with feedback from the job. In Figure 2.2, deficiency is represented by the portion of variance in the focal construct that is unshared with the operational definition (σ2d). The construct validity of an operational definition of an assumed cause or effect can be affected adversely by such problems as (a) an inadequate preoperational definition of the construct (e.g., a conceptual definition that underrepresents the focal construct), (b) a single operational definition of the construct (e.g., a specific measure of organizational commitment), (c) a manipulation that motivates research participants to respond in a biased manner (e.g., demand characteristics-based responding, evaluation apprehension-based responding), and (d) an operational definition that fails to capture the degree of variability in the focal construct (Nunnally, 1978; Nunnally & Bernstein, 1994; Rosenthal & Rosnow, 2008; Shadish et al., 2002; Stone-Romero, 1994). Several of these problems (e.g., demand characteristics) are considered in the subsection titled Measuring Variables. Construct validity of units and settings. The units that are sampled in a study have a bearing on construct validity. For example, construct validity would be questionable if a study purported to deal with the decision-making behavior of managers but used undergraduate business school students as participants. In addition, the setting in which a study is conducted influences construct validity. For instance, a study of soldier behavior in combat that is conducted in a special purpose laboratory setting would lack construct validity if the attributes of the study’s setting failed to mirror those of the target setting (i.e., an actual war zone). It merits adding that, as is explained in the following section, there is a nontrivial distinction between the construct validity of sampling particulars and external validity (Shadish et al., 2002). The construct validity of sampling particulars has to do with the correspondence between the operational definitions used in the study and the constructs to which inferences are made (e.g., that undergraduate

business school students are equivalent to managers). In contrast, external validity deals with the degree to which a causal relation that is found with one set of sampling particulars (business students) also is observed with other sampling particulars (psychology students).

Internal Validity Internal validity has to do with the degree of correctness (i.e., truth value, legitimacy) of inferences about causal connections between focal constructs (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002), for example, the inference that X causes Y (symbolically: X → Y). It is important to note that there is an important distinction between actual and assumed causal relations between constructs (StoneRomero, 2009; Stone-Romero & Rosopa, 2004, 2008). Causal relations are considered to be (a) actual when they are supported by sound randomized experimental research, and (b) assumed when based on evidence from nonexperimental research. Research that uses randomized experimental designs provides the firmest basis for demonstrating that an independent variable (X) causes (i.e., produces changes in) a dependent variable (Y). The validity of causal inferences is somewhat weaker in research that uses quasi-experimental designs, and is weakest in research that uses nonexperimental designs. In virtually all nonexperimental research, a cause is assumed (XA) as opposed to actual (X), and an effect is assumed (YA) as opposed to actual (Y). However, an assumed cause can be shown to be an actual cause through appropriate experimental research. Inferences about cause are most justified when a researcher can show that (a) the cause preceded the effect in time (temporal precedence), (b) the cause and effect are related to one another (covariation), and (c) there are no rival explanations of the covariation between the cause and effect (i.e., there are no confounds). These conditions are most likely to be fulfilled in research that uses randomized experimental designs and least likely to be satisfied in research that uses nonexperimental designs (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Stone-Romero, 2002, 2007a, 2007b, 2007c, 2009; Stone-Romero & Rosopa, 2004, 2008). 41

Eugene F. Stone-Romero

It deserves adding that there are often multiple causes of any given dependent variable. For example, research shows that job performance is caused by worker ability, worker motivation, and the degree to which the worker has a clear understanding of the requirements of his or her role. Thus, a researcher could conduct a randomized experimental study in which each of these variables is manipulated independently to determine their main and interactive effects on job performance.

Copyright American Psychological Association. Not for further distribution.

Statistical Conclusion Validity Researchers typically use statistical methods (e.g., correlation, regression, analysis of variance) to test for the existence and strength of relations between variables (Hays, 1994; Maruyama, 1998; Pedhazur, 1982; Snedecor & Cochran, 1980). For example, using data from a nonexperimental study, they infer that there is a nonzero relation between XA and YA. Or using data from a randomized experimental study, they infer that the average score for units in a treatment condition (MT) is greater than the average for units in a control condition (MC). In both cases, statistical inferences are critical to inferring that the studied variables are related to one another (Cook & Campbell, 1976, 1979; Cook et al., 1990; Hays, 1994; Shadish et al., 2002). As such, the validity of the same inferences may be adversely affected by a number of factors, including low statistical power, failing to meet the assumptions of statistical tests, and conducting large numbers of statistical tests using a nominal Type I error rate that is lower than the actual (effective) Type I error rate (Cook & Campbell, 1976, 1979; Cook et al., 1990; Hays, 1994; Shadish et al., 2002). Some of these factors (e.g., statistical power) are considered in the subsection titled Statistical Significance Versus the Importance of Research Results. It merits adding that a large number of commonly used statistical methods (e.g., analysis of variance [ANOVA], correlation, multiple regression) are special cases of the general linear model (Searle, 1971). In addition, although data from studies that use randomized experimental designs are typically analyzed with ANOVA, they also can be analyzed with multiple regression. And, as is explained in detail in the next section, the use of sophisticated data analytic 42

strategies (e.g., structural equation modeling) does little or nothing to make up for data that are derived from a study that is poor in terms of construct validity and internal validity. Put somewhat differently, data that are not worth analyzing are not worth analyzing well.

External Validity External validity has to do with the degree of correctness (i.e., truth value, legitimacy) of inferences about the existence of a causal relation between two variables across different sampling particulars of units, settings, treatments, and outcomes. For example, will a goal-setting program that has been shown to improve the performance of salespeople in retail stores in the United States improve the performance of comparable workers in Mexico? A frequently used analytical strategy for deriving evidence on external validity is meta-analysis (Hedges & Olkin, 1985; Rosenthal, 1991). It would appear that at least one sampling particular should be added to the set considered by Shadish et al. (2002), that is, time. More specifically, a causal relation may vary across time periods. For example, the effect of job-related ability (e.g., training based) on performance may vary across time (e.g., Borman, 1991; Ghiselli, 1956). The effect may be relatively strong during the early stages of a worker’s employment. However, it may become weaker over time because of changes in the roles of motivation and ability as determinants of performance. External validity is threatened by any factors that serve to moderate (Stone-Romero & Liakhovitski, 2002; Zedeck, 1971) the relation between an assumed cause and a supposed effect. Stated somewhat differently, external validity is an issue when there are interactions between a treatment (X) and one or more of the other sampling particulars of a study (e.g., units, settings, outcomes; Shadish et al., 2002).

Types of Variables in Assumed Causal Models Research aimed at testing assumed causal models (as defined previously) typically considers more than a simple, two-variable sequence. For example, it may involve assumed exogenous (i.e., independent) and endogenous (i.e., mediator, dependent) variables.

Research Strategies in I/O Psychology

Behavioral beliefs Attitude toward behavior Evaluation of outcomes

Moral beliefs Behavioral intentions

Copyright American Psychological Association. Not for further distribution.

Normative beliefs

Motivation to comply

Behavior

Subjective norms

Perceived behavioral control

FIGURE 2.3. The theory of reasoned action. From Handbook of Organizational Research Methods (p. 306), by D. Buchanan and A. Bryman (Eds.), 2009. London: Sage. Copyright 2009 by E. Stone-Romero. Reprinted with permission.

Figure 2.3, based on the theory of reasoned action (Ajzen, 1988), is an illustration of such a model. It considers five exogenous variables (i.e., behavioral beliefs, evaluation of outcomes, normative beliefs, moral beliefs, and motivation to comply), four mediator variables (attitude toward behavior, subjective norms, perceived behavioral control, and behavioral intentions), and one dependent variable (i.e., behavior).

data, behaviors (e.g., job performance), and physical attributes (e.g., weight).

Variables that are measured in research can be categorized in terms of such criteria as (a) type of variable measured, (b) data collection method, (c) data source, and (d) potential for the measurement process to evoke measurement-related artifacts.

Method for collecting data. A number of methods can be used to collect and/or record data. Among them are paper and pencil measures (e.g., questionnaires), interviews (e.g., employment related), observations (e.g., work behavior), mechanical or electronic recorders (e.g., polygraphs), physical sampling (e.g., blood, urine, DNA), content analysis (e.g., coding of responses to open-ended interview questions), and searches of archives (e.g., organizational, government). It deserves adding that Internet-based data collection methods are being used with increasing frequency in research in I/O psychology and related fields (Stanton & Rogelberg, 2002).

Type of variable measured. Measurement may focus on such variables as attitudes (e.g., job satisfaction), values (e.g., individualism), beliefs (e.g., stereotypes), perceptions (e.g., of job characteristics), personality (e.g., need for achievement), mental abilities (e.g., numeric), aptitudes (e.g., mechanical), physiological states (e.g., arousal), physical abilities (e.g., grip strength), preferences (e.g., for benefits), contents of archives (e.g., annual reports of firms), demographic information (e.g., age, sex), biographical

Data sources. Sources of data include data subjects (e.g., job incumbents, teams, organizations), observers of data subjects (e.g., supervisors, peers, subordinates, and clients of job incumbents), and organizational archives (e.g., production records, financial statements, Equal Employment Opportunities Commission reports). Whatever the variable measured in a study, it is generally wise to collect data from multiple sources. Doing so can avert problems stemming from source-based biases. For example, self-appraisals of performance tend to

Measuring Variables

43

Eugene F. Stone-Romero

Copyright American Psychological Association. Not for further distribution.

overstate actual levels of performance (e.g., Harris & Schaubroeck, 1988). Potential to evoke response artifacts. Methods for measuring variables differ in terms of the degree to which they may evoke a number of response artifacts, that is, systematic or nonsystematic biases in responses to items in a measure. They include carelessness, socially desirable responding, yea-saying, nay-saying, agreement (i.e., the tendency to agree with a statement as opposed to disagreeing with its opposite, irrespective of the content of the statement), evaluation apprehension, random responding, self-generated validity, and impression management (for more on response artifacts, see Nunnally & Bernstein, 1994; Webb, Campbell, Schwartz, Sechrest, & Grove, 1981; Weber & Cook, 1972). In this regard, a considerable amount of research has been devoted to the reactivity of measures, that is, the tendency for the measurement process to affect the variables being measured (e.g., see Rosenthal & Rosnow, 1969; Webb et al., 1981). All else constant, the greater the reactivity of a measure, the greater the potential for the obtained data to be biased by such artifacts as social desirability, evaluation apprehension, and impression management. Research in I/O psychology and allied fields also has focused on the extent to which such problems as common methods variance, response–response bias, and self-generated validity influence responses to measures (Conway, 2002; Doty & Glick, 1998; Feldman & Lynch, 1988; Salancik & Pfeffer, 1978). Method variance may result in both contaminated measures of focal variables and spuriously high relations between them and measures of other variables (e.g., job characteristics, job attitudes). Note, however, that there is controversy over the degree to which correlations between measured variables are biased by method variance (cf., for example, Salancik & Pfeffer, 1978; E. F. Stone, 1992). In addition, it is not easy to determine the degree to which observed correlations between variables are attributable to (a) true covariation between them versus (b) covariation caused by method-related bias (Nunnally, 1978; Nunnally & Bernstein, 1994; Spector, 2006). 44

Method variance has frequently been assessed with the multitrait–multimethod (MTMM) matrix approach advocated by D. T. Campbell and Fiske (1959). However, as Kalleberg and Kluegel (1975) demonstrated, the validity of inferences derived from MTMM analyses is contingent on the degree to which its underlying assumptions are satisfied. For example, MTMM analyses assume that traits and methods are uncorrelated. However, research shows that there are often trait-method correlations. Thus, they recommended that the MTMM strategy be supplanted by confirmatory factor analysis (CFA). This recommendation has also been made by Lance, Hoffman, Gentry, and Baranik (2008). They argued, for instance, that in the analysis of multisource performance rating data, rater source effects may represent valid perspectives on ratee performance, as opposed to biases attributable to halo error. Method biases tend to reduce the construct validity of measures (Shadish et al., 2002) by lowering the proportion of valid systematic variance (see Figure 2.2). Note, however, that some artifacts (e.g., random responding) may reduce statistical conclusion validity by increasing the proportion of random (nonsystematic) variance in a measure (see Figure 2.2). Assuming that at least some of the covariation between measures of two or more variables is attributable to common methods variance, researchers should work toward its reduction. Conway (2002) and Podsakoff, MacKenzie, Lee, and Podsakoff (2003) offered a number of recommendations for doing this. Among them are (a) obtaining predictor and criterion data from different sources, (b) separating the time at which predictor and criterion variables are measured, and (c) counterbalancing the order in which variables are measured. It is interesting, however, that Spector (2006) viewed common method variance as an “urban legend.” As he noted, the “urban legend that there is universally shared [common method] variance in our methods is both an exaggeration and oversimplification of the true state of affairs” (p. 230). Thus, instead of focusing on common methods variance, he argued that researchers should isolate the specific hypothesized cause(s) of spurious correlation between variables and control it (them) statistically. For example, the spurious correlation may be a function of

Research Strategies in I/O Psychology

responses that are contaminated with social desirability. To control this threat, it could be measured and its “effects” controlled statistically (e.g., by partial correlation). Unfortunately, unless the variables being controlled have high levels of reliability and validity, partialling procedures may result in upwardly biased estimates of the relation of interest (Cook & Campbell, 1979; Shadish et al., 2002; Stone-Romero, 2007a; Stone-Romero & Rosopa, 2008).

Copyright American Psychological Association. Not for further distribution.

Other Construct Validity Artifacts The results of research also may be biased by two other general types of artifacts that detract from construct validity, that is, demand characteristics and experimenter expectancy effects (Rosenthal & Rosnow, 1969; Shadish et al., 2002). Demand characteristics are biases caused by participants’ beliefs about what is expected of them in a study (Orne, 1962, 1969; Weber & Cook, 1972). Consider, for example, a randomized, between-subjects experimental study in a special purpose setting (e.g., a university-based laboratory) in which participants are randomly assigned to conditions in which they (a) perform one of two tasks that differ in terms of such characteristics as variety and autonomy and (b) provide ratings of these characteristics. However, prior to performing the task, a confederate of the experimenter provides the participants with his or her (a) views about it (e.g., telling subjects that the task is boring, repetitive, and mindless) and (b) affective reactions to it (e.g., telling subjects that they won’t like working on the task). In this case, their ratings would be a function of not only the actual characteristics of the tasks but also the demand created by the information that the confederate provided to the subjects. As a result, the task characteristics manipulation would be confounded with the demand (bias) produced by the confederate-supplied information. Related to demand characteristics are experimenter expectancy effects. These are biases that result from unintentional behavior on the part of the researcher (e.g., experimenter) that increases the likelihood of a study providing support for one or more of its hypotheses. For instance, consider a randomized, between-subjects experimental study of the effects of training on learning in which the experi-

menter posits that one type of training will result in lesser learning than another. This expectancy may bias the way in which the researcher delivers training to trainees and, thus, the amount they actually learn. The experimenter expectancies in this situation would reduce the construct validity of the training manipulation. Note that demand characteristics and experimenter expectancy effects are only two of many artifacts that may threaten the construct validity of a study’s sampling particulars. Other artifacts include using only one operational definition of a construct, measuring all variables with a single method (e.g., questionnaires), research procedures that interrupt the routines of subjects (e.g., sleep patterns), manipulations that lead subjects to believe that they have experienced treatments (e.g., job-related training) that are not as desirable as those received by others, and subjects in one experimental condition (e.g., a no-training control group) being exposed to treatments that were intended only for subjects in another condition (e.g., a training condition). Information on other artifacts and strategies for dealing with them is offered in such works as Aronson et al. (1990), Rosenthal and Rosnow (1969, 2008), and Shadish et al. (2002).

Timing of Measurement in Research In studies of assumed causal models (e.g., the model shown in Figure 2.3), it is critical that assumed mediators and dependent variables be measured at appropriate times (Mathieu, DeShon, & Berg, 2008; Mathieu & Taylor, 2006; Shadish et al., 2002; StoneRomero & Rosopa, 2008). The reason for this is that, typically, the influence of causes on mediators and effects is not instantaneous. For example, in a study aimed at testing the effects of supervisor-supplied feedback on worker behavior, it may take several weeks for feedback (the cause) to influence the worker’s behavioral intentions, and for these to affect the worker’s behavior (i.e., job performance). Thus, it is critical that there be appropriate lags between the times at which (a) causes are varied (either naturally or experimentally) and (b) mediators and effects are measured. One way of determining the appropriate lags is to base them on published reports of previous research on the phenomena of interest 45

Eugene F. Stone-Romero

that provide information about time lags. Another is to conduct a pilot study to determine the time needed for a manipulation to affect mediators and dependent variables (see, e.g., Mathieu & Taylor, 2006; Mathieu et al., 2008).

Copyright American Psychological Association. Not for further distribution.

Statistical Significance Versus the Importance of Research Results Research results are almost always evaluated in terms of the criterion of statistical significance; that is, the odds that a study’s findings are a function of Type I statistical error (i.e., falsely rejecting a true null hypothesis). Statistical conclusion validity hinges on the Type I error rate used in testing a study’s hypotheses. In general, the lower the probability that a study’s results may be a function of Type I error, the more credible are its findings. The capacity for a study to yield statistically significant results is a function of two major factors (Cohen, 1988; Hays, 1994; Rosenthal & Rosnow, 2008). One is the size of the effect in the population (e.g., the population correlation between two variables, ρ). The other is the size of the sample (N) used in testing a statistical hypothesis. All else constant, the larger the N (a) the smaller the standard error of a test statistic (e.g., the standard error of the correlation coefficient) and (b) the greater the odds of rejecting the null hypothesis and concluding that there is a statistically significant relation. However, it is important to recognize that statistical significance is not equivalent to practical importance. Typically, the latter is evaluated in terms of an explained variance or effect size criterion (e.g., r2, R2, ω´ 2). For example, in personnel selection, all else constant, the greater the correlation between a predictor and a criterion, the greater the proportion of variance the predictor explains in a criterion ( J. P. Campbell, 1976; Guion, 1976, 1991, 2002). However, in some situations, relatively low validity coefficients may signal considerable practical importance. For example, as utility formulas indicate, utility (i.e., the expected payoff of a decision-based action, expressed in dollar terms or some other metric) can be high when a validity coefficient is low, but the standard deviation of performance is high (Boudreau, 1991; J. P. Campbell, 1976; Cronbach & Gleser, 1965). 46

Sampling of Study Particulars As noted previously, construct validity inferences in a study are a function of the degree to which its sampling particulars (units, treatments, outcomes, and settings) are representative of the constructs to which inferences are made (Shadish et al., 2002). A number of sampling strategies can be used in a study. They are illustrated here by considering strategies for sampling individual workers (units) in an organization. Basic sampling strategies that can be used for this purpose are of two types, that is, probability sampling and nonprobability sampling. In probability sampling, workers are selected from a population (with N members) in such a way as to ensure that the probability of selecting a sample of a given size (e.g., n = 5) is equal to the probability of selecting any other sample of the same size. The simplest way of doing so is to select workers randomly. This can be done, for example, by (a) assigning each worker in a population with N members an identifying number (w; 1, 2, . . . n) and (b) using a computer program (e.g., Excel) to identify n values of w for inclusion in the sample. It deserves adding that probability sampling is rarely used in research in I/O psychology and related fields. More complex probability sampling strategies include multistage cluster sampling and stratified random sampling (e.g., see Cochran, 1977; Kish, 1965; Snedecor & Cochran, 1980). For example, in stratified random sampling, a researcher would specify strata of employees within an organization (e.g., managerial, nonmanagerial) and then randomly select a specific number of workers from each stratum. And in multistage cluster sampling, for instance, an investigator would (a) select a random sample of units in one or more clusters (e.g., private- vs. public-sector organizations in the United States) and then (b) randomly select clusters of organizations within each of the larger clusters (e.g., health care organizations, educational institutions). Nonprobability sampling uses nonrandom strategies for selecting sample members (e.g., workers) from a target population. Among the many nonprobability sampling strategies are convenience (also referred to as accidental and haphazard) sampling, purposive sampling of heterogeneous instances, systematic sampling, and quota sampling (e.g., see

Copyright American Psychological Association. Not for further distribution.

Research Strategies in I/O Psychology

Cochran, 1977; Kish, 1965; Shadish et al., 2002; Snedecor & Cochran, 1980). The convenience sampling strategy involves selecting a nonrandom sample based on their availability to participate in a study (e.g., samples of students in university classes). The purposive sampling of heterogeneous instances strategy entails selecting a nonrandom sample of members on the basis of the researcher’s belief that they are diverse in terms of characteristics that might influence a causal relation between variables (Shadish et al, 2002). For example, a researcher interested in the relation between job enrichment and job satisfaction in an appliance manufacturing firm might select nonrandom samples of workers in various jobs (e.g., machinists, electricians, janitors). In systematic sampling, the researcher selects sample members in a methodical manner from lists of assumed population members. For instance, a researcher might select every 10th person on a list of the full-time employees of a firm. Finally, in quota sampling, the researcher selects specific numbers of sample members of different types so as to produce a sample that is roughly representative of a target population. For example, in an organization, the researcher might select specific numbers of male and female employees in managerial versus nonmanagerial jobs. However, when there are too few employees of a particular type to allow for powerful statistical analyses, employees in specific categories can be oversampled. For example, if the managerial category included 20 women and 80 men, the researcher might sample all women and 20 men. Having unequal numbers of men and women in the sample would yield underestimates of actual relations between sex and, for example, monthly salary. The reason for this is that the point-biserial correlation coefficient is greatest when there are equal proportions of cases in two groups (Ghiselli, Campbell, & Zedeck, 1981). EPISTEMOLOGICAL ISSUES The focus of this chapter is on research methods that are generally consistent with the method of science (e.g., Braithwaite, 1996; Kaplan, 1964). This method assumes that general laws about phenomena can be established using evidence from well-designed and

conducted empirical studies (Braithwaite, 1996; Cook & Campbell, 1979; Shadish et al, 2002). Typically, such research is based on a model that involves (a) observing phenomena of interest (e.g., differences in worker performance), (b) using induction to develop an explanation of the phenomenon (e.g., performance is a function of worker ability), (c) using deduction to make a prediction about the phenomenon of interest (e.g., increasing ability will result in improved performance), (d) testing the prediction through empirical research (e.g., an empirical study to assess the effects of training on performance), and (e) deriving conclusions about the validity of the hypothesis. The same model typically relies on the findings of existing research on the phenomenon of interest. As Kuhn (1970) noted, normal science involves research that is “firmly based upon one or more past scientific achievements, achievements that some particular scientific community acknowledges for a time as supplying the foundation for its further practice” (p. 10). Note that in focusing on the method of science, this chapter does not provide coverage of such epistemological perspectives as conventionalism, logical positivism, essentialism, and falsificationism. These and other perspectives are treated in other works (e.g., Cook & Campbell, 1979; Rosenthal & Rosnow, 2008; Shadish et al., 2002). It deserves adding that a number of criticisms have been lodged against what have been referred to as established, traditional, rigorous, or normal science (Braithwaite, 1996; Kaplan, 1964; Kuhn, 1970) research methods (e.g., Argyris, 1968, 1980; Aronson et al, 1990; Lawler et al., 1985; Thomas & Tymon, 1982). In general, these methods rely on the previously described model of science. Among the often overlapping critiques of rigorous scientific research are that it (a) has little or no practical utility; (b) does not inspire confidence in the work of scientists; (c) distorts important aspects of reality in organizations; (d) fails to consider phenomena that are not readily measurable; (e) has low relevance to the criteria that are important to practitioners; (f) deals with phenomena that are common knowledge; (g) is artificial; (h) is trivial; (i) is irrelevant; (j) has low operational validity, that is, it deals with variables that cannot be controlled by practitioners; (k) focuses on 47

Copyright American Psychological Association. Not for further distribution.

Eugene F. Stone-Romero

“proving” the truth of hypotheses; (l) is untimely, that is, it deals with issues that change more rapidly than the knowledge produced by scientists; and (m) involves participants who are not representative of the populations to which generalizations are to be made. As a consequence, for example, Thomas and Tymon (1982) recommended that research be designed with the practitioner in mind and that it have high levels of five general properties or characteristics: (a) descriptive relevance (i.e., it deals with phenomena that are encountered by practitioners in organizations); (b) goal relevance (i.e., it deals with outcomes that practitioners are interested in and have the capacity to influence); (c) operational validity (i.e., it focuses on independent variables that are capable of being manipulated by practitioners); (d) nonobviousness (i.e., it deals with phenomena that extend beyond the realm of common sense); and (e) timeliness (i.e., its results are available rapidly enough to be of use to practitioners who are faced with the need to solve problems on a timely basis). All else constant, there is nothing whatsoever wrong with doing research that is of value to both scientists and practitioners. However, there is no reason why such research cannot be conducted using traditional, rigorous research methods (Cook & Campbell, 1979; Fromkin & Streufert, 1976; E. F. Stone, 1981, 1982, 1987; Shadish et al., 2002; Weick, 1965). In addition, there is considerable evidence (e.g., see Locke, 1986) showing that (a) the findings of experimental research conducted in special purpose settings (e.g., university laboratories) generalizes to nonspecial purpose settings (e.g., work organizations) and (b) the arrangements found in “artificial” organizations (i.e., special purpose settings) are “strikingly similar to actual organizations” (Weick, 1965, p. 204). Moreover, it deserves stressing that unless research has high levels of internal, construct, and statistical conclusion validity, it is of little importance that it has external validity. Unfortunately, doing the sort of research that is strongly advocated by various critics (e.g., Argyris, 1968, 1980; Lawler et al., 1985; Thomas & Tymon, 1982) requires the cooperation of top-level managers in organizations. They must (a) allow researchers to design and implement interventions aimed at changing outcomes that are of interest to practitioners and 48

(b) agree to the collection of data on relevant outcomes. However, it is typically the case that managers are very reluctant to allow experimental research to be conducted in their organizations. In addition, a host of other factors militate against the conduct of randomized experiments in organizations (see, e.g., Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002). These include, for example, faulty randomization procedures, not having a sufficient number of units (e.g., individuals) to assign randomly to conditions, refusals of units to participate in experiments, treatment-related attrition from experimental conditions, and spillover of the treatment to individuals in control conditions (Cook & Campbell, 1979; Cook et al., 1990). For the most part, researchers have little or no influence over these factors. It deserves adding, however, that quasiexperimental studies are sometimes possible in organizational contexts. They are especially useful in instances in which organizations conduct interventions that can be studied by researchers. However, as Grant and Wall (in press) noted, they are quite rare. More specifically, they reported that a search of six journals over a 25-year period revealed that less than 1% of the studies published in the same journals used this design type. This is unfortunate because, as is noted in the next section, welldesigned and executed quasi-experimental studies have far greater levels of internal validity than nonexperimental studies. EXPERIMENTAL DESIGNS VERSUS RESEARCH SETTINGS Prior to describing nonexperimental, quasiexperimental, and randomized experimental designs, it is important to distinguish between experimental designs and research settings (Stone-Romero, 2002, 2007c, 2009; Stone-Romero & Rosopa, 2008). Thus, this section focuses on three setting-related issues, that is, the “laboratory” versus “field” distinction, the influence of setting type on internal validity, and the types of realism that are important in research. Taken together, these issues are very important in research aimed at testing for causal relations between variables. As such, they have nontrivial implications for both internal validity and construct validity.

Research Strategies in I/O Psychology

Copyright American Psychological Association. Not for further distribution.

The Often Inappropriate Laboratory Versus Field Distinction Frequently, a distinction is made between laboratory and field research settings (e.g., Bouchard, 1976; Cook & Campbell, 1976, 1979; Cook et al., 1990; Fromkin & Streufert, 1976; Kerlinger & Lee, 2000; Locke, 1986; E. F. Stone, 1978). Regrettably, this distinction is not very informative (J. P. Campbell, 1986; Stone-Romero, 2002, 2007c, 2009; StoneRomero & Rosopa, 2008). One of the reasons for this is that laboratories can be set up in what are typically referred to as field settings. Thus, as Stone-Romero (2007c, 2009) noted, a more appropriate distinction is between special purpose (SP) and nonspecial purpose (NSP) research settings. Settings of the former variety are created for the specific purpose of doing research and include a laboratory room at a university or a simulated work setting in an industrial park. In contrast, settings of the latter type are created for purposes other than research and include organizations created for the purpose of producing products or providing services. These include such organizations as Microsoft, Mercedes Benz, Matsushita, Sony, and the U.S. Army. SP settings have two major attributes. First, they are created for the purpose of conducting research and cease to exist when it has been completed. Second, they are designed to allow for the effective (unconfounded) manipulation of one or more independent variables. In general, SP settings have only a subset of the features or elements that are found in NSP settings (Aronson et al., 1990; Berkowitz & Donnerstein, 1982; Fromkin & Streufert, 1976; Runkel & McGrath, 1972; Weick, 1965).

Setting Type and the Validity of Causal Inferences Inferences about cause (e.g., X → Y) vary as a function of the setting in which research is conducted. Because SP settings are created explicitly for the purpose of conducting a study, research in such settings typically provides for a much greater degree of control over confounding (nuisance) variables than research in NSP settings. The important implication of this is that, to the degree that confounding variables can be controlled, a researcher can be more confident about the internal validity of a study

(Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; E. F. Stone, 1978; StoneRomero, 2007c, 2009). As such, it is generally the case that inferences about cause are more justified when research is conducted in SP than in NSP settings (Stone-Romero, 2002, 2009; Stone-Romero & Rosopa, 2004, 2008).

Realism in Research It is important that SP settings be designed with two properties in mind. More specifically, it is essential that they have experimental realism. In addition, it is often desirable for them to have mundane realism (Aronson et al., 1990; Fromkin & Streufert, 1976; Weick, 1965). For example, a researcher interested in studying the effects of variations in job design on critical psychological states and job satisfaction could create a special purpose setting in a university-based facility. For this study, it would be important to have job design manipulations (e.g., autonomy, feedback) that had desired degrees of impact on research participants, thus ensuring the experimental realism of the study (Aronson et al., 1990). However, it would not be very important for the study to have mundane realism, that is, a special purpose setting that had all of the elements common to actual work organizations. For example, the SP setting would not have to have such features as health care benefits, retirement plans, and union–management agreements. Nevertheless, the more the study had these and other elements the greater would be its mundane realism (Aronson et al., 1990; Weick, 1965) and its construct validity. In addition, the greater would be its potential to assess the capacity of the study’s assumed causal model to explain the phenomenon of interest in a nonspecial purpose (e.g., organizational) setting. It merits adding that the just-described job design study also could be conducted in a nonspecial purpose setting (e.g., an actual work organization). However, for several reasons, it would typically be much more difficult to conduct it in such a setting (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002). One reason for this is that in most NSP settings, it is very difficult to bring about changes in existing organizational arrangements (e.g., physical layout of facilities, assignment of workers to jobs, pay and fringe benefits) that are needed to 49

Copyright American Psychological Association. Not for further distribution.

Eugene F. Stone-Romero

provide for the effective manipulation of independent variables (Cook & Campbell, 1976, 1979; Cook et al., 1990). As a consequence, research in NSP settings that uses randomized experimental or quasi-experimental designs typically results in lower levels of control over extraneous or confounding variables than studies using the same types of designs in SP settings. Therefore, inferences about cause are typically more problematic for research in NSP settings than for research in SP settings. In general, studies conducted in NSP settings have higher levels of mundane realism than those conducted in SP settings. This often serves as a basis for the claim that external validity inferences are more appropriate for studies in NSP settings than in SP settings. However, the legitimacy of this argument is suspect (Dipboye & Flanagan, 1979; Fromkin & Streufert, 1976; Locke, 1986; Stone-Romero, 2002; Weick, 1965). There are several reasons for this. One is that even when studies are conducted in NSP settings, they typically involve nonrepresentative samples of subjects, settings, and operational definitions of manipulations and/or measures. Thus, the external validity of such studies is suspect. The fact that they were conducted in NSP settings often does nothing to strengthen external validity inferences. Another reason is that a major purpose of research is to show that there is a relation between two or more constructs. In this regard, numerous chapters in a book on generalizing from laboratory to field settings (Locke, 1986) provide clear evidence that, typically, relations that are found in SP settings are also found in NSP settings. The findings apply to relations involving numerous variables of interest to organizational researchers, including attributions, goal setting, participation in decision making, financial incentives, reinforcement schedules, job satisfaction, job characteristics, and job performance. It deserves adding that experimental realism is often crucial for studies that test assumed causal models. The very important reason for this is that tests of such models may fail to show support for causal connections between variables if independent variables are operationally defined in such a way as to have insufficient impact on study participants (Aronson et al., 1990; Fromkin & Streufert, 1976; Shadish et al., 2002). Thus, even though there actu50

ally may be a causal relation between X and Y, a study that used an operational definition of X that lacks experimental realism may fail to provide evidence of it. In addition, to the extent that a study lacks mundane realism, an effect found in a special purpose setting may not be found in a nonspecial purpose setting. For example, the effect of X may be present when a special purpose setting isolates it from all of the other variables that are likely to be found in NSP settings. However, in a nonspecial purpose setting the influence of X may be too weak to allow for an adequate test of the causal model. Cook and Campbell (1979) commented on this in discussing the unobtrusive treatment implementation problem. DEGREE TO WHICH A DESIGN IS EXPERIMENTAL Experimental designs used in research are of three basic types: randomized experimental, quasiexperimental, and nonexperimental. These differ from one another in terms of their potential to control confounds and thus their value in supporting causal inferences.

Randomized Experimental Designs Studies that use randomized experimental designs have three major attributes (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Stone-Romero, 2002, 2007c, 2009). Taken together, the results of studies using such designs tend to have high levels of internal validity. Manipulation of independent variables. In research that uses randomized experimental designs, the researcher manipulates the levels of one or more independent variables. For example, in a study concerned with the effects of temperature and humidity on worker performance, a researcher could vary each of them experimentally. He or she would have to have at least two levels of each independent variable that differed enough from one another to produce changes in performance. Note that the findings of previous research (including pilot studies) could be used to determine the strength of the manipulations used in a randomized experimental study. In studies that involve two or more independent variables, the researcher can test for not only the main effects of each variable but also their interac-

Copyright American Psychological Association. Not for further distribution.

Research Strategies in I/O Psychology

tion (Hays, 1994). In addition, if there are enough subjects to allow for adequate statistical power (Cohen, 1988), the researcher can have more than two levels of each variable, increasing his or her ability to accurately model the form (e.g., linear, quadratic) of the causal relations between the independent variables and the dependent variable. The capacity to manipulate levels of one or more independent variables serves to strengthen inferences about cause; that is, it leads to relatively high levels of internal validity. One reason for this is that in a study that uses a randomized experimental design, the researcher can be confident about the cause(s) preceding the effect(s) in time, that is, temporal precedence. In addition, assuming a properly conducted randomized experiment, he or she can be confident that changes in one or more dependent variables are the result of the manipulations, as opposed to confounds. Random assignment of units to study conditions. A second attribute of a randomized experimental study is that units (e.g., individuals, teams, organizations) are randomly assigned to study conditions. Assuming a sufficient number of units and effective randomization procedures, the researcher can be highly confident of the equivalence of the units in each of the conditions (in terms of mean levels of any and all measured and unmeasured potentially confounding variables) prior to the time that subjects are exposed to the study’s manipulations. As a result, he or she can be confident that posttreatment levels of the dependent variables for each of the conditions resulted from the study’s manipulations, as opposed to confounds (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Stone-Romero, 2002; 2007c, 2009). Note, moreover, that in a study in which the unit of analysis is the group, if a large number of groups are available for random assignment to experimental conditions, it is not necessary to first randomly assign individuals to the groups. Measurement of dependent variables. A third characteristic of studies using randomized experimental designs is that levels of dependent variables (outcomes) are measured. For example, in the justdescribed study, the researcher could measure the performance of workers in each of the conditions.

He or she also could measure subjects’ beliefs about the levels of manipulations they experienced. Such manipulation checks are especially important in terms of inferences about the construct validity of the manipulations (Aronson et al., 1990; Shadish et al., 2002). As is noted previously in the subsection dealing with the construct validity of treatments operational definitions, measures should be both reliable and valid. Unless outcome measures have a sufficiently high level of reliability, statistical analyses will fail to provide evidence on the effects of manipulations. This is a threat to statistical conclusion validity (Shadish et al., 2002). When tests of assumed causal models are based on data from well-designed and properly conducted randomized experimental research, (a) causal paths are well-known, (b) confounds are not an issue, and (c) model misspecification is typically not a concern. Thus, inferences about cause rest on a very firm foundation. This argument is predicated on the assumption that the randomized experiments are properly designed and executed. However, there may be instances when they “break down.” This is more likely to be a problem in NSP than SP settings (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002). Among the many causes of this are such threats to internal validity as differential attrition across treatment conditions, history, testing, resentful demoralization, and the interaction of two or more threats. These threats are defined in the paragraphs that follow. To illustrate two of these threats, consider a simple randomized experiment having the following design: R

O1A

X

O2 A

R

O1B

—

O2 B.

Here it is assumed that subjects were randomly assigned (R) to Groups A and B and complete measures of job satisfaction at Times 1 and 2. Those in Group A were exposed to a job enrichment manipulation (X), and those in Group B served as notreatment controls. Assume that the treatment had no effect whatsoever. However, the means for the measured variables had the following pattern: 51

Copyright American Psychological Association. Not for further distribution.

Eugene F. Stone-Romero

O1A ≈ O1B < O2A ≈ O2B. Had the researcher considered only the data of Group A, he or she may have falsely concluded that the O1A − O2A difference was caused by X. However, the same difference could have been caused by various threats to internal validity. For example, testing would be a credible explanation of the O1A − O2A difference if simply completing the measure at Time 1 led to a change in its value at Time 2 that was unrelated to X. And history would be a plausible explanation of the O1A − O2A difference if factors outside of the experiment (e.g., changes in hourly pay rates) that were unrelated to X caused the difference. Differential attrition would be a plausible rival explanation if the workers in the treatment condition who had the lowest levels of job satisfaction quit their jobs prior to the measurement of this variable at Time 2. To the extent that such threats are present in a randomized experiment, the validity of inferences about causal connections between variables would suffer. To illustrate another threat to internal validity, consider the following, primitive randomized experimental design: R

X

O2 A

R

—

O2 B .

The treatment (e.g., job design change) is intended to affect O (e.g., job satisfaction) positively. Because O was not measured before X was implemented, there is no way of knowing if it had the intended effect. For example, assume that X actually was not effective, but statistical analyses showed that O2A > O2B. This might lead a naive researcher to conclude that X was responsible for the difference. However, it could have been an artifact of scores on O remaining unchanged in Group A and decreasing in Group B. A possible cause of this is that individuals in Group B could have experienced resentful demoralization as a result of not experiencing X, causing their job satisfaction to decrease. Some limitations on the use of research using randomized experimental designs. In spite of the value of randomized experiments for internal validity inferences, it may be difficult or impossible to use this type of design in many circumstances (see, e.g., Cook & Campbell, 1976, 1979; Cook et al., 1990; 52

Rosenthal & Rosnow, 2008; Shadish et al., 2002). First, some variables are not subject to manipulation (e.g., the actual age and sex of study participants). Second, although it may be possible to manipulate some variables (e.g., the personality, physical health, and mental health of study participants), manipulations of these variables would be unacceptable on ethical grounds if they resulted in physical or psychological harm to participants. Third, in research that takes place in NSP settings (e.g., work organizations), a researcher may not have the ability to randomly assign units (e.g., workers, teams, plants) to treatment conditions. Fourth, in research in NSP settings, it may not be possible to isolate units from one another (spatially), leading to the breakdown of randomized experiments aimed at testing assumed causal models. Fifth, research participants may refuse to be assigned to experimental conditions on a random basis. Sixth, it may be unethical to withhold beneficial treatments from units (e.g., those in a no-treatment control group). As a result of these and other factors, researchers who are interested in examining relations between and/or among variables that are assumed to be causes, mediators (Stone-Romero & Rosopa, 2004, 2007), and effects may have to use either quasiexperimental or nonexperimental designs (Cook & Campbell, 1976, 1979; Cook et al., 1990; Rosenthal & Rosnow, 2008; Shadish et al., 2002). Computer-based simulations. Randomized experiments also can be conducted using simulationbased manipulations of variables. When such experiments are conducted on existing computer systems (e.g., a university-based computer), they take place in NSP settings. However, if a computer system were created specifically for the purpose of conducting a simulation, the setting would be of the SP variety. One type of computer-based simulation that does not involve human participants is a statistical simulation. An example of this is a study by Stone-Romero and Rosopa (2004) that was conducted in a nonspecial purpose setting. They assessed the effects of manipulated levels of several independent variables (e.g., sample size, effect size, reliability of measures, variance of measures) on the capacity (i.e., statistical power) of hierarchical multiple regression to produce evidence of direct and indirect mediation. This study

Research Strategies in I/O Psychology

Copyright American Psychological Association. Not for further distribution.

was experimental in that the values of the independent variables were manipulated in the simulation. A second type of computer-based simulation is one in which independent variables are manipulated by a computer to assess their effects on human participants (e.g., attitudes, beliefs, emotions, behaviors; Zickar & Slaughter, 2002). For example, research in an SP setting can be used to study the effects of computer-based manipulations of such variables as weather conditions (e.g., wind speed, visibility) on simulated pilot performance.

Quasi-Experimental Designs There are five major types of quasi-experimental designs: single group designs without a control condition, multiple group designs that lack one or more pretest measures, multiple group designs that have a control condition and use pretest measures, time series designs, and regression discontinuity designs. An example of each of these designs is offered in a subsection that follows. Detailed explanations of the five types of quasi-experimental designs are available in other works (e.g., Cook & Campbell, 1976, 1979; Cook et al., 1990; Rosenthal & Rosnow, 2008; Shadish et al., 2002). Whatever their specific nature, however, quasi-experimental designs have several attributes. Taken together, they typically lead to internal validity inferences that are weaker than those stemming from research that uses randomized experimental designs but stronger than those derived from studies that use nonexperimental designs. Manipulation of independent variables. As is true of research using randomized experimental designs, in research using quasi-experimental designs, the researcher manipulates one or more independent variables. Note that in organizational research it is often true that the manipulations are designed and introduced by both the researcher and representatives of the organization. Regrettably, this may threaten both the construct validity of the manipulations and the statistical conclusion validity of a study’s findings. For example, construct validity would be threatened if the design of a study called for all individuals in a treatment condition to receive a specified treatment (e.g., performance-contingent pay), but supervisors responsible for the delivery of it

delivered not only the planned treatment but also an unplanned treatment (e.g., performance-contingent praise). In addition, statistical conclusion validity would be reduced if supervisors did not deliver the planned treatment in a systematic manner (e.g., failed to give performance-contingent pay to all workers). Typically, quasi-experimental designs tend to have a very small number of manipulated variables. In addition, in quasi-experimental designs of the time series variety, an independent variable may be introduced and then removed several times, with the expectation that measured levels of the dependent variables will covary with these changes. However, whatever the specific type of quasi-experimental design, the fact that independent variables are manipulated strengthens internal validity inferences (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Stone-Romero, 2002, 2007b, 2009; Stone-Romero & Rosopa, 2004, 2008). Nonrandom assignment of units to conditions. Unlike what is true of randomized experiments, in quasi-experiments units are not randomly assigned to study conditions. For example, in a quasi-experiment conducted in a nonspecial purpose setting, workers in one operating unit of a firm may be exposed to a goalsetting manipulation, whereas workers in a geographically remote unit may serve as no-treatment controls. Because intact units are used in the study, they may differ from one another on a host of variables prior to the time the treatment is introduced in the first unit. Such differences weaken internal validity inferences (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Stone-Romero, 2002, 2007b, 2009). Measurement of assumed dependent variables. In quasi-experimental research, the researcher measures the values of one or more assumed dependent variables. The assumed qualifier is used because units are not randomly assigned to study conditions. As a result, in research using most types of quasiexperimental designs, there is no assurance that across-condition differences in the measures of assumed dependent variables were the actual effects of the treatments. For example, they may have been a product of one or more unmeasured confounding variables (e.g., history). 53

Eugene F. Stone-Romero

Examples of quasi-experimental designs. One type of quasi-experiment is a single group design without a control condition. An example of this is the one-group pretest–posttest design:

Copyright American Psychological Association. Not for further distribution.

O1A

X

O2 A .

Because of its simplicity, this design may seem attractive to a researcher. However, in most cases, the finding of a hypothesis consistent difference between O1A and O2A would not serve as a valid basis for arguing that it was caused by X. The principal reason for this is that the design does not allow the researcher to rule out most of the previously described threats to internal validity. As such, from an internal validity perspective this design has virtually no value. A second type of quasi-experiment is the multiple group design that lacks one or more pretest measures. An example of this type of design is the posttest-only design with a nonequivalent control group: ∼R

X

O2 A

∼R

—

O2 B .

Here, units are assigned to conditions on a nonrandom (∼R) basis. This design allows the researcher to compare the measured values of the assumed outcome for Groups A and B. However, even if they are hypothesis consistent, the researcher has almost no ability to argue that the difference was caused by X. One reason for this is that there is no legitimate basis for assuming that the groups were equivalent to one another prior to the time that the treatment (X) was delivered to members of Group A. A third type of quasi-experimental design is a multiple group design that has a control condition and uses one or more pretest measures. An illustration of this is the untreated control group design with dependent pretest and posttest samples and a double pretest: ∼R ∼R

O1A O1B

O2 A O2 B

X —

O3 A O3B .

Note that although units were assigned to Groups A and B on a nonrandom basis, the existence of multiple pretests allows for an assessment of several 54

threats to internal validity. For example, history can be controlled by comparing pretest and posttest measures in Groups A and B. If, for example, O1B = O2B = O3B, such threats as testing, history, and maturation can be ruled out. This strengthens internal validity inferences. A fourth type of quasi-experimental design is the time series design. An example of this is the simple interrupted time series design: O1A O26 A

O2 A O27 A

O3 A . . . O25 A

X

O28 A . . . O50 A .

In this design, members of a single group are measured at multiple pretest (e.g., 25) periods, exposed to a treatment (X), and measured at multiple posttest (e.g., 25) periods. These measures allow the researcher to rule out such threats to interval validity as testing, maturation, and instrumentation. As a result, causal inferences are far more justified with this design than the three previously described quasi-experimental designs. The fifth type of quasi-experimental design is the regression discontinuity design. It has the following structure: ∼R

O1A

X

O2 A

∼R

O1B

—

O2 B .

Units (e.g., individuals) are measured at Time 1, and their scores are used to assign them to either the treatment (A) or control (B) condition. For example, units who scored above the mean on O1 are given the treatment, whereas the remainder serve as notreatment controls. Subsequent to treatment, the researcher regresses O2 scores on O1 scores separately for units in Groups A and B. A treatment effect, for example, would be signaled by a difference in the regression coefficients for the groups. Although a researcher might infer that X was responsible for this difference, the same inference may not be valid. One reason for this is that there may have been differential levels of mortality (attrition) in Groups A and B. Some limitations on the use of quasi-experimental designs. A number of conditions limit the capacity of quasi-experimental research to provide sound tests of causal models. They include most of the factors noted previously that pertain to randomized experimental designs. In the interest of brevity, the

Copyright American Psychological Association. Not for further distribution.

Research Strategies in I/O Psychology

same factors are not considered in this subsection. Note, however, that it is often possible to do quasiexperimental research that has reasonably high levels of internal validity, especially when a researcher can demonstrate that the pattern of findings stemming from a study are consistent with an a priori set of expectations (see the discussion of the coherent pattern matching strategy in Shadish et al., 2002). In addition, as is detailed by Grant and Wall (in press), quasi-experimental studies have a number of benefits that should motivate their use when randomized experiments are not possible, including minimizing ethical dilemmas and facilitating cooperative research with practitioners.

Nonexperimental Designs Of the three general types of designs considered in this chapter, nonexperimental designs are the most frequently used in organizational research (Austin, Scherbaum, & Mahlman, 2002; Scandura & Williams, 2000; Stone-Romero, Weaver, & Glenar, 1995). Research that uses such designs has several defining attributes. Taken together, they greatly reduce the validity of causal inferences (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Stone-Romero, 2002, 2007a, 2009; Stone-Romero & Rosopa, 2004, 2008). There are two major types of nonexperimental designs: quantitative and qualitative. In research of the quantitative variety, various variables (e.g., assumed independent, moderating, dependent) are measured and quantitative estimates of population parameters (e.g., mean, variance, covariance, correlation) are estimated for individuals in one or more groups. Nonexperimental studies that consider relations between (among) variables are often erroneously referred to as “correlational studies” (e.g., Aronson et al., 1990; Shadish et al., 2002). As indicated by Stone-Romero (2002), however, the same label is inappropriate because correlation is a statistical technique that is used to estimate the strength and direction of relations between variables. It is not an experimental design type. A second type of nonexperimental design is qualitative research. This type of design is considered in depth in chapter 3 of this volume. Note, however, that there are several subtypes of qualitative studies,

including case studies, ethnographies, and grounded theory (Bott, 2007; Locke & Golden-Biddle, 2002; Shadish et al., 2002; Yin, 1989). In the case study, for example, the researcher conducts an in-depth, typically subjective, examination of one or more units (individual, group, organization) for the purpose of describing phenomena of interest (e.g., communication processes, maturation of units). In general, the data produced by such research are not quantitative, rendering them unsuitable for statistical analyses. Measurement of assumed independent variables. In nonexperimental research, variables that are assumed to be causes are measured, as opposed to manipulated. For example, a researcher interested in studying the relation between job enrichment and job satisfaction could measure the former variable, operating on the assumption that differences in observed levels of job enrichment reflect actual differences on this variable. However, because job enrichment is measured, as opposed to manipulated, inferences about a causal connection between it and job satisfaction would rest on a very shaky empirical foundation. Stated somewhat differently, the same study would not have a high level of internal validity. In addition, as is detailed in a subsection that follows, inferences about cause would not be strengthened whatsoever by the application of so-called causal modeling procedures to data derived from the study. This point is echoed by methodologists in numerous academic disciplines (Brannick, 1995; Cliff, 1987; Freedman, 1987; Holland, 1986; Ling, 1982; Mathieu et al., 2008; Millsap, 2002; Rogosa, 1987; Rosopa & StoneRomero, 2008; Stone-Romero, 2009; Stone-Romero & Rosopa, 2004, 2007, 2008). Nonrandom assignment of units to assumed conditions. In nonexperimental studies, units (e.g., workers) are not randomly assigned to study conditions. Rather, the researcher collects data from units that have levels of assumed independent variables that may have resulted from various unknown causes. For example, in the just-described job enrichment study, the researcher would have to assume that the self-reported levels of measured job enrichment corresponded to the actual levels of enrichment of workers’ jobs. This assumption would, for example, be erroneous if the self-reports varied systematically 55

Eugene F. Stone-Romero

as a function of factors other than objective levels of enrichment (e.g., the cognitive ability levels of subjects; see E. F. Stone, Stone, & Gueutal, 1990).

Copyright American Psychological Association. Not for further distribution.

Measurement of assumed dependent variables. As is true of research using quasi-experimental designs, studies using nonexperimental designs deal with measured levels of what are assumed to be dependent (outcome) variables. As a result, studies using nonexperimental designs have very low levels of internal validity. Some limitations on the use of nonexperimental designs. In general, research that uses nonexperimental designs is easier to conduct than research that uses either randomized experimental or quasi-experimental designs (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002). This is especially true of studies conducted in NSP settings (e.g., work organizations). One important reason for this is that organizations (and other types of social systems) are far more willing to allow for studies in which assumed causes are measured as opposed to manipulated. However, because of their questionable internal validity, nonexperimental designs should be avoided by researchers who are interested in testing causal models. Value of longitudinal, nonexperimental designs. In longitudinal, nonexperimental research, assumed causes, mediators, and effects are measured. Because of this, it is virtually always the case that two of the three requirements for making causal inferences (i.e., temporal precedence and ruling out rival explanations) cannot be satisfied (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Rosopa & Stone-Romero, 2008; StoneRomero, 2002, 2007c, 2009; Stone-Romero & Rosopa, 2004, 2007, 2008). For example, McDonald (1999) noted that longitudinal studies are open to the objection that an apparently causal relation between an earlier and a later measure may just represent a relation of temporally stable traits of the person, or perhaps aspects of an unfolding developmental sequence that does not allow conceptual manipulation of the earlier measured attribute. (p. 370) 56

As a result, the internal validity of longitudinal studies of the nonexperimental variety is typically quite low. Thus, frequently expressed calls for longitudinal research as a means of testing causal propositions are ill-advised. Nevertheless, some researchers (e.g., Frese, Garst, & Fay, 2007) believe that causal inferences are aided by the analysis of longitudinal data using time-lagged structural equation modeling (SEM). Regrettably, as is noted in the following paragraphs, unless causes are manipulated, the internal validity of such research is suspect; that is, the use of so-called causal modeling methods does virtually nothing to enhance the internal validity of findings stemming from nonexperimental studies. Moreover, as Frese et al. (2007) noted, their “longitudinal study cannot rule out the existence of unknown and changing third [confounding] variables” (p. 1099). Note, moreover, that if an unknown variable (e.g., X3) is responsible for covariation between two other variables (e.g., X1 and X2) that are measured longitudinally (at Times 1, 2, 3, . . . g), the spurious correlation between X1 and X2 (i.e., r12) will be observed at all such time periods. For example, if exposure to stressful environmental conditions in a war zone is responsible for a spurious observed correlation between depression (X1) and anxiety (X2), the spurious r12 correlation will be seen at all time periods in which X1 and X2 are measured. Collecting data on X1 and X2 at three or more periods will do nothing whatsoever to reduce the spuriousness. In addition, the use of sophisticated data analytic techniques (e.g., SEM) will be of no value whatsoever in terms of ruling out the operation of one or more confounding variables. In short, longitudinal research of the nonexperimental variety does nothing to aid in valid causal inference. However, because of design considerations (i.e., the manipulation of independent variables), longitudinal data are of considerable value in making causal inferences in well-designed quasi-experimental studies. INTERSECTION OF RESEARCH SETTING AND EXPERIMENTAL DESIGN As noted previously, in any given study, a researcher has to make decisions about a number of important issues, including its experimental design and setting.

Research Strategies in I/O Psychology

This section considers the combinations of these two design features. The strengths and weaknesses of such combinations are considered by Stone-Romero (2009).

Copyright American Psychological Association. Not for further distribution.

Randomized Experiments in NSP Settings It is possible to conduct randomized experiments in NSP. However, for reasons considered previously, studies of this type are relatively rare. Nevertheless, several factors may serve to enhance the likelihood of doing experimental research in NSP settings (Cook and Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002). The odds of doing it increase when (a) the demand for a given treatment (e.g., a laborsaving device) is greater than its supply; (b) an innovation (e.g., new computers) is to be introduced, but it cannot be delivered to all units at once; (c) units are isolated from one another in time (e.g., in basic training units in the military); (d) units are separated geographically and the level of inter-unit communication is low (e.g., fast food restaurants in various regions of the country); (e) there is a need for change and multiple treatments of unknown efficacy can be introduced; (f ) units can be assigned to conditions randomly as opposed to on the basis of need or merit; (g) units have no preferences for the type of treatment they are to receive; (h) it is possible to create an organization for the sole purpose of conducting an experimental study; (i) the organization gives the researcher control over units for the purpose of a study; and (j) units expect to be assigned treatments on a random basis. Note that in some of the just-noted cases, what was originally a nonspecial purpose setting is temporarily converted to an SP setting for the duration of a study. Research by E. F. Stone, Gueutal, Gardner, and McClure (1983) illustrates the use of a randomized experimental design in a nonspecial purpose setting. They were interested in the degree to which the type of organizations with which data subjects had dealings (the independent variable) influenced their privacyrelated values, beliefs, attitudes, and several other outcome variables (the dependent variables). Study participants were randomly selected from several geographical regions within the state of Indiana and randomly assigned to conditions in which they were asked to consider their dealings with one of six types of organizations (e.g., their employer, the

Internal Revenue Service, law enforcement agencies). They then responded to structured interview items dealing with the dependent variables in terms of a specific type of organization. Results of univariate and multivariate analyses revealed that organization type affected reports of privacy-related values, beliefs, and attitudes.

Randomized Experiments in SP Settings Randomized experiments are frequently conducted in SP settings. One important reason for this is that they allow the researcher to manipulate independent variables while controlling for possible confounds through strict control over the study’s setting. Although the label is poor (Stone-Romero, 2007c; Stone-Romero & Rosopa, 2008), the term laboratory experiment is commonly used to describe research that uses this type of design (e.g., Fromkin & Streufert, 1976; Locke, 1986). A study by D. L. Stone and Stone (1985) provides an example of a randomized experiment in a special purpose setting. They were interested in factors that affected individuals’ beliefs about the accuracy of performance feedback and their self-perceived task competence (i.e., task-based esteem). On the basis of relevant theory and research, they hypothesized that self-esteem would be influenced by both the favorability of feedback and its consistency. To test these hypotheses, they conducted a randomized experimental study in a special purpose setting (i.e., a university laboratory facility). The study involved role-playbased manipulations of (a) feedback favorability (acceptable vs. superior) and (b) the consistency of feedback (consistent vs. inconsistent) from two feedback agents. After working on an in-basket task, subjects were randomly assigned to one of the four study conditions. They then received performance feedback and completed measures of the dependent variables. Results of analyses of variance showed support for virtually all of the hypothesized relations, including those concerned with interaction effects.

Quasi-Experiments in SP Settings Although it is possible to conduct quasi-experimental research in SP settings, studies of this type appear nonexistent in I/O psychology and allied fields. A search of the social science literature using the 57

Copyright American Psychological Association. Not for further distribution.

Eugene F. Stone-Romero

PsycINFO database from 1872 to June 2007 failed to reveal a single published study of this nature. However, quasi-experimental research in SP settings is not at all uncommon in other disciplines. For example, it is found in clinical psychology (e.g., Blanchard et al., 1997). Research by Blanchard et al. (1997) provided an example of a study that used a quasi-experimental design in an SP setting. They were interested in determining the effects of thermal biofeedback interventions for the treatment of vascular headaches. The research took place in an SP setting, that is, a laboratory created for the study of stress and anxiety disorders and associated treatments. There were four experimental conditions: (a) thermal biofeedback for hand warming (TBFHW), (b) thermal biofeedback for hand cooling (TBFHC), thermal biofeedback for temperature stability (TBFTS), and (c) suppression of alpha brain waves (SABW). To assess the effects of these treatments, the researchers conducted a randomized experiment in which 70 patients were randomly assigned to one of four conditions. Results of repeated measures t tests showed reductions in headache intensity for all but the subjects in the TBFTS condition. Although the overall experimental design was randomized experimental, within each of the treatment conditions the design was quasi-experimental because subjects who completed premeasures of headache intensity were exposed to the treatment and then completed postmeasures of headache intensity. In addition, Blanchard et al. conducted within-condition internal analyses (Aronson et al., 1990) with subsets of the subjects in the thermal biofeedback conditions who they referred to as learners (i.e., those who demonstrated a criterion level of proficiency at controlling their hand temperature). The use of the internal analyses converted what was initially a randomized experiment to a quasi-experiment. Results of the same analyses showed that subjects in the TBFHC and TBFTS conditions showed reductions in headache intensity, whereas those in the TBFHW condition did not.

Quasi-Experiments in NSP Settings There are numerous instances of studies using quasi-experimental designs in NSP settings in I/O 58

psychology and related fields. Research of this type is far more common than randomized experiments in such settings. One important reason for this is that quasi-experiments are far less disruptive of ongoing organizational structures, processes, and practices than randomized experiments. A study by Hackman, Pearce, and Wolfe (1978) illustrated the use of a quasi-experimental design in an NSP setting. They were interested in the effects of changes in job design (the independent variable) on measures of several job characteristics, job performance, and absenteeism (the assumed dependent variables). To determine the effects of such changes, they conducted a quasi-experimental study using 94 employees of a bank. Data on the assumed dependent variables were collected before and after changes in the job characteristics. Results showed that the job design manipulations resulted in expected changes in measures of several job characteristics and absenteeism. However, changes in performance between the pre- and postintervention periods only were found for employees who were relatively high in terms of growth need strength. The researchers noted that the results of their study needed to be replicated using studies with randomized experimental designs.

Nonexperiments in NSP Settings Nonexperimental studies in NSP settings are very common in organizational research. In fact, StoneRomero et al. (1995) showed that for 1,929 articles published in the Journal of Applied Psychology for the period of 1975 to 1993, the percentage of studies that used nonexperimental designs was much greater than that of studies using either randomized experimental or quasi-experimental designs. Similar findings have been reported by both Austin et al. (2002) and Scandura and Williams (2000). Research by Stone-Romero, Stone, and Hyatt (2003) provided an example of the use of a nonexperimental design in an NSP setting. They were interested in determining the degree to which individuals viewed 12 personnel selection procedures (e.g., application blank, interview, mental ability test, work sample) to be invasive of their privacy. To obtain data on this issue, they conducted two nonexperimental studies in NSP settings. In Study 1,

Copyright American Psychological Association. Not for further distribution.

Research Strategies in I/O Psychology

Thurstone-type scaling procedures were used to obtain invasiveness scale scores for each of the 12 procedures, and in Study 2, participants provided direct ranks of their invasiveness. In addition, in Study 2, participants ranked the 12 procedures in terms of a number of factors that were viewed as possible antecedents of invasiveness (e.g., procedure reveals negative information, procedure invades the body, procedure erroneously discredits applicants). Results of Study 1 revealed considerable differences in the relative invasiveness of the procedures. In addition, the results of Study 2 were highly consistent with those of Study 1. Moreover, Study 2 showed a number of very strong correlations between invasiveness and its assumed antecedents.

Nonexperiments in SP Settings Nonexperimental research is often conducted in SP settings. A common example of this is questionnairebased research conducted in SP facilities that are dedicated to research (e.g., a university-based behavioral laboratory). Independent of the experimental design used in a study and the type of setting (SP vs. NSP) in which it takes place, a researcher must determine the degree to which its data provide support for one or more hypotheses. In quantitative (as opposed to qualitative) research, this determination is typically based on the results of statistical tests. Thus, the next section considers the distinction between statistical methods and experimental design. Of particular interest and importance is the use of the findings of statistical tests as a basis for supporting inferences about causal relations between (among) variables. A study by Stone (1979) illustrates the use of a nonexperimental design in a special purpose setting. A considerable body of research shows that perceptions of various types of stimuli are related to field independence, that is, an individual difference dimension that deals with the ability of individuals to perceive stimuli independent of the context in which they are embedded. On the basis of a literature review, Stone hypothesized that field independence (i.e., an assumed cause) would be related to individuals’ perceptions of task characteristics (i.e., assumed effects). Thus, he performed two nonexperimental studies in which subjects (a) completed a measure of field inde-

pendence, (b) performed a task in a special purpose setting (i.e., a university-based laboratory room), and provided perceptions of the characteristics of the task. Note that all subjects performed the same task. Thus, Stone reasoned that variations in perceptions of task characteristics would be a function of differences in field independence. Among the various findings were that field independence correlated positively with perceptions of task variety, task identity, and feedback in both studies. In addition, several other correlation coefficients were of modest magnitude (e.g., r = −.24) but were not statistically significant. STATISTICAL METHODS VERSUS EXPERIMENTAL DESIGN The joint consideration of experimental design and statistical methods gives rise to an extremely important point: The type of experimental design used in a study is independent of the statistical methods that can be used to analyze data from its use. Thus, for example, data from a simple randomized experiment involving a control group and a treatment group can be analyzed with a number of statistical methods, including (a) an independent groups t test; (b) analysis of variance; (c) bivariate correlation, using dummy codes for the independent variable; (d) ordinary least-squares (OLS) regression; (e) path analysis (PA); and (f ) SEM. Note, in addition, that the F stemming from the ANOVA will equal the squared t from a test of mean group differences. In addition, the estimate of the proportion of variance explained by the ANOVA and the t test will equal the r2 derived from the bivariate correlation analysis. What’s more, the F from the ANOVA will equal the F for the test of the significance of R2 derived from an OLS analysis. Moreover, in PA using a correlation matrix as the input, the path coefficient from the independent variable to the dependent variable will equal the bivariate correlation coefficient and will result in the same t value as that obtained from the t test. Note that statistical methods and experimental design are separate issues in terms of inferences derived from a study. Consider, for example, a statistical test of the difference between the mean level of a variable for two groups. It only provides information on the degree to which a test statistic (e.g., t) 59

Copyright American Psychological Association. Not for further distribution.

Eugene F. Stone-Romero

differs from a value that would be expected by chance, given the sampling distribution associated with the test (e.g., difference between independent means, assuming equal population variances). The findings of the same statistical test (t) say nothing whatsoever about the design that was used in the study of two groups. It may have been a nonexperiment, a quasi-experiment, or a randomized experiment. Thus, inferences about statistical significance would be independent of inferences about causes of the difference between the means (i.e., the internal validity of the study). One very important implication of the foregoing is that a test of an assumed causal model that uses data from a nonexperimental study and statistical methods that purport to yield information about causality does nothing whatsoever to change the nature of the study’s experimental design. Stated somewhat differently, causal modeling methods are incapable of converting data from a nonexperimental study into data that provide a firm and legitimate basis for inferences about cause. Experimental design serves as the basis for inferences about cause, whereas statistical methods allow for inferences about the existence of relations between variables and the probability that an observed estimate of the strength of a relation resulted from sampling error. Note, moreover, that statistical methods are “blind” to the experimental design of the study that produced the data being analyzed.

Causal Modeling Although data from virtually all nonexperimental studies provide a very weak basis for making causal inferences, a number of so-called causal modeling techniques are currently being used with such data to test assumed causal models, including PA, hierarchical multiple regression (HMR), hierarchical linear modeling (HLM), and SEM (Baron & Kenny, 1986; Blalock, 1964, 1971; Bollen, 1989; Bryk & Raudenbush, 1992; Cohen, Cohen, West, & Aiken, 2003; Kenny, 1979; Maruyama, 1998; Millsap, 2002). A large number of examples of such analyses can be found in articles published in the major journals in I/O psychology and allied disciplines (e.g., Academy of Management Journal, Personnel Psychology, Journal of Applied Psychology, and Organizational Behavior 60

and Human Decision Processes) in the last 4 decades. Indeed, several studies have shown that the use of SEM for testing structural and/or measurement models has increased markedly since software for performing such tests has become available (Aguinis, Pierce, Bosco, & Muslin, 2009; Austin et al., 2002; Stone-Romero et al., 1995). In view of this, it is important to consider the conditions that are vital to valid inferences about causal relations between variables.

Conditions Vital to Inferences About Cause All else constant, internal validity is a function of the experimental design of a study. The reason for this is that the conditions vital to causal inference (i.e., temporal precedence, correlation, and absence of confounds) are most likely to be satisfied in research that uses randomized experimental designs, less likely to be satisfied in studies that use quasiexperimental designs, and highly unlikely to be satisfied in studies that use nonexperimental designs. Even if one relaxed the temporal precedence requirement, the validity of causal inferences using data from nonexperimental studies would hinge on the assumed causal model being properly specified. Among the relevant specification considerations are the proper ordering of variables in the model being tested, the inclusion of all important causes (exogenous variables) and mediator variables in the model, the correct specification of the form (e.g., linear, nonlinear) of the relations considered by the model, the correct specification of the causal paths in the model (including those associated with reciprocal causation), and the use of measures of variables that have high levels of construct validity. In this regard, it is critical to understand that it is virtually impossible to satisfy these requirements in any nonexperimental study (Bollen, 1989; Ling, 1982; Rogosa, 1987; Rosopa & Stone-Romero, 2008; StoneRomero, 2009; Stone-Romero & Rosopa, 2004, 2008). As a consequence, inferences about the validity of assumed causal models that are based on such research rest on a very weak empirical foundation. As such, they are almost never justified. In this regard, Cliff (1987) noted that when a researcher is analyzing data from nonexperimental research, “it is

Research Strategies in I/O Psychology

Copyright American Psychological Association. Not for further distribution.

not possible to isolate the empirical system sufficiently so that the nature of the relations among variables can be unambiguously ascertained” (p. 119). Consistent with this view, Rogosa (1987) wrote that “attempts to answer experimental . . . research questions with nonexperimental data seem fundamentally askew” (pp. 193–194). And, in accordance with these arguments, Ling (1982) argued that “causal inference[s] from correlational [i.e., nonexperimental] data (in the absence of controlled experiments) . . . are at best a form of statistical fantasy” (p. 490).

Model-Data Consistency Versus Model-Reality Consistency It is quite common for researchers who use causal modeling procedures to argue that they have shown support for an assumed causal model as a result of model-data consistency, that is, showing consistency between it and the covariances among variables considered by a nonexperimental study. The modeldata consistency issue is considered in relation to Figures 2.4a–2.4c, which show several possible causal models that are consistent with an observed correlation between variables X and Y. Researchers who test assumed causal models using data from nonexperimental research typically advance three claims: 1. The premise: If my assumed causal model (e.g., that shown in Figure 2.4a) is correct, then the findings of my nonexperimental study will be consistent with it. 2. The research findings: The findings of my study are consistent with my assumed causal model; that is, there is a nonzero relation between X and Y. 3. The conclusion: Therefore, my assumed causal model is correct. Unfortunately, this conclusion is not based on sound reasoning. More specifically, it is an instance of the logical fallacy of affirming the consequent (Kalish & Montague, 1964; Stone-Romero, 2009; Stone-Romero & Rosopa, 2004, 2008). For instance, arguing that a theoretical model is correct because research results are consistent with the model. It is easy to see why the just-noted conclusion is unwarranted. It is that the findings also may be consistent with a number of other possible causal models,

(a) X

Y

Y

X

(b)

X

(c) Z

Y FIGURE 2.4. Alternative causal models that are consistent with an X–Y correlation. From Handbook of Organizational Research Methods (p. 319), by D. Buchanan and A. Bryman (Eds.), 2009. London: Sage. Copyright 2009 by E. Stone-Romero. Reprinted with permission.

including (a) the model shown in Figure 2.4b, which posits that Y → X, and (b) the model shown in Figure 2.4c, which specifies that Z → X and Z → Y, implying that the correlation between X and Y is noncausal. Assuming that the other models (see Figures 2.4b and 2.4c) fit the data as well as the researcher’s assumed causal model (see Figure 2.4a), the researcher would have no firm basis for concluding that his or her preferred model is superior to the others. However, to buttress the argument that his or her assumed causal model is correct, the researcher may argue that it is theory consistent. Unfortunately, this is of little or no value in terms of supporting claims about cause. The reason for this is that a number of different theories may be used to explain any observed covariance. Stone-Romero and Rosopa (2004, 2008) and Stone-Romero (2009) provided several illustrations of this. One has to do with the relation between job satisfaction 61

Copyright American Psychological Association. Not for further distribution.

Eugene F. Stone-Romero

and job performance. Researchers associated with the human relations movement in I/O psychology have argued that job satisfaction causes job performance (e.g., see Pinder, 1998). However, expectancy theorists view job satisfaction as a consequence of job performance (e.g., Porter & Lawler, 1968). Still others have posited that job satisfaction and job performance are reciprocally related to one another. Given these competing views, the researcher is in no position to argue that one theory is more supported by the data than another. A second example of competing theoretical explanations concerns the relation between job satisfaction and organizational commitment. One group of researchers has argued that organizational commitment causes job satisfaction (e.g., Bateman & Strasser, 1984; Koslowsky, 1991; Lance, 1991; Vandenberg & Lance, 1992; Wiener & Vardi, 1980). Another set has asserted that job satisfaction causes organizational commitment (e.g., Williams & Hazer, 1986). Yet another group has contended that job satisfaction and organizational commitment are reciprocally related to one another. Finally, some researchers believe that the relation between these two variables is spurious (e.g., Currivan, 1999). Thus, a researcher who found a positive relation between these two variables would be hard-pressed to claim that his or her theoretical stance was superior to numerous alternatives. Overall, it should be clear that the invocation of a theory is of little or no value in supporting the inference that the findings of a nonexperimental study have internal validity. Thus, theories are of little or no importance in terms of buttressing claims about the correctness of an assumed causal model for such a study. Clearly, the best strategy for showing support for such a model is to perform one or more randomized experiments that test model-based predictions. Stone-Romero and Rosopa (2008) explained how such experiments can be used to support causal inferences for a simple causal model in which the relation between X and Y is mediated by M; that is, X → M → Y. More specifically, they argued that one randomized experiment can be used to support the conclusion that X causes both M and Y, and a second can be used to show that M causes Y. Consider, for instance, research aimed at showing 62

that motivation mediates the relation between incentives and task performance. In Experiment 1 incentives can be varied experimentally to assess their impact on both motivation and task performance, and in Experiment 2 motivation can be manipulated to determine its effect on task performance. Taken together, the results of these two experiments, along with reasoning from symbolic logic (see Theorem 26 in Kalish & Montague, 1964), would provide strong support for the inference that M mediates the effect of X on Y.

Prediction Versus Causation Tests of assumed causal models often lead to results that can be used for prediction purposes. Here, the term prediction is used in a statistical sense (Guion, 1976, 1991, 2002; Hays, 1994; Pedhazur, 1982). That is, information on a set of predictor variables can be used to predict the value of a criterion of interest. The fact that such prediction is possible does not imply that the predictor is a cause of the predicted variable (Bollen, 1989; Brannick, 1995; Cliff, 1987; Freedman, 1987; Holland, 1986; Kelloway, 1998; Ling, 1982; Mathieu et al., 2008; Millsap, 2002; Rogosa, 1987; Rosopa & StoneRomero, 2008; Stone-Romero, 2009; Stone-Romero & Rosopa, 2004, 2008). For example, (a) the zip codes of individuals can be used to predict their annual income levels, (b) the weights of individuals can be used to predict their heights, and (c) the heights of individuals at age 10 years can be used to predict their heights at age 30 years. In each of these instances, the fact that the first variable can be used to predict the second does not serve as a valid basis for inferences about cause. Indeed, one can use statistical methods for postdictive purposes; that is, a researcher can use data from a currently measured variable to postdict events that took place in the past (Blum & Naylor, 1968; Stone-Romero, 2007a). For example, current assessments of neurological damage in Viet Nam veterans can be used to postdict their previous level of exposure to neurotoxins (e.g., Agent Orange). The very important implication is that the ability to predict, in a statistical sense, has no necessary implications for the understanding of causal processes. Regrettably, there is considerable evi-

Research Strategies in I/O Psychology

dence of the fact that many researchers don’t understand this fact. This issue is considered in the subsection titled Unwarranted Causal Inference in Organizational Research.

Copyright American Psychological Association. Not for further distribution.

Causal Inferences Associated With Nonmanipulable Variables In randomized experimental studies, researchers actually can manipulate such characteristics of stimulus people as their age, race, biological sex, and ethnicity (e.g., via resume-based manipulations). As such, they can test for the effects of such variables on measured outcomes (e.g., hiring decisions). As a result, they can support inferences about the impact of such variables on the outcomes. However, in nonexperimental studies, the same variables are measured, as opposed to manipulated, and are often assumed to have a number of roles. One of these is that of a moderator (Stone-Romero & Liakhovitski, 2002; Zedeck, 1971). For instance, in criterionrelated validation studies, race has been studied as a moderator of the validity of predictors of job performance (e.g., Boehm, 1972; Schmidt & Hunter, 1974). A second role is that of an assumed independent variable. For example, numerous studies have examined the relation between the stigmas of targets and their treatment by others (e.g., Stone-Romero & Stone, 2007). Third, and finally, such variables are often used as controls in statistical analyses. It is interesting, however, that that researchers typically fail to provide a sound rationale for this use of nonmanipulable variables. Even if they do, statistical controls are seldom effective (Stone-Romero, 2007a). There are numerous reasons for this, one of which is that when control variables are measured in an unreliable manner, their “effects” are not fully removed by partialling procedures (e.g., calculation of partial regression coefficients in multiple regression analyses). An important issue in nonexperimental studies is the degree to which researchers can make valid inferences about the effects of nonmanipulable variables. In view of internal validity considerations, it is inappropriate to argue about the effects of various nonmanipulable variables. It deserves adding that the vast majority of nonmanipulable variables that are studied in nonexperimental studies represent lit-

tle more than surrogates for other nonmanipulable variables. For example, biological sex is a “stand-in” for variables that differ between men and women, including values, motives, attitudes, aptitudes, and abilities. Thus, in such research, it would be far better to measure such variables directly, as opposed to measuring the surrogate of sex.

Unwarranted Causal Inference in Organizational Research As noted previously, the appropriateness of causal inferences varies as a function of a study’s experimental design. Unfortunately, however, inappropriate inferences about causal connections between variables (i.e., inferences that are not justified by the type of experimental design of a study) are quite common in reports of the findings of nonexperimental studies. Unwarranted causal language (e.g., the use of such terms as causes, effects, and influences) can be found in very large percentages of the articles published in the major journals of I/O psychology and allied disciplines. Evidence of this comes from a study by Stone-Romero and Gallaher (2006). They content analyzed 161 randomly sampled articles that were published in the 1988, 1993, 1998, and 2003 volumes of Personnel Psychology, Organizational Behavior and Human Decision Processes, the Academy of Management Journal, and the Journal of Applied Psychology. The articles reported the findings of studies that used randomized experimental, quasi-experimental, and nonexperimental designs. Stone-Romero and Gallaher searched the same publications for instances of the inappropriate use of causal language in their title, abstract, and results and/or discussion sections. The analysis revealed that unwarranted causal language appeared one or more times in 58 of the 73 articles that reported the results of research that used nonexperimental designs (79% of such articles) and 14 of the 18 articles that reported the findings of research that used quasi-experimental designs (78% of such articles). For examples of articles that contain one or more instances of unwarranted causal inference, see Stone-Romero and Rosopa (2008) or Stone-Romero (2009). Overall, what the studies considered by StoneRomero and Gallaher (2006) show is that researchers 63

Eugene F. Stone-Romero

Copyright American Psychological Association. Not for further distribution.

often make unwarranted claims about cause on the basis of studies that do not provide a basis for such claims. In addition, they often assume, quite incorrectly, that data analysis, as opposed to sound experimental design, affords an adequate foundation for advancing causal inferences. Clearly, if researchers are interested in making such inferences, they must conduct research using randomized experimental or sound quasi-experimental designs. TRENDS IN RESEARCH DESIGN AND METHODS There are a number of notable trends in the designs used in research in I/O psychology and related fields that are detailed in reviews by Austin et al. (2002), Scandura and Williams (2000), and Stone-Romero et al. (1995). In addition, there is evidence on the types of articles that have been published in a journal that focuses on methods-related issues (Aguinis et al., 2009). These reviews are considered in the following section, and their major findings are summarized in Table 2.1.

Design Trends The Austin et al. (2002) study considered 609 articles sampled from nine volumes of the Journal of Applied Psychology in 10-year intervals starting at 1920 and ending at 2000. The Stone-Romero et al. (1995) review was based on 1,929 articles published in the same journal for each of 19 years (1975–1993). Finally, the Scandura and Williams (2000) article dealt with all articles (N = 774) published in three management-oriented journals (i.e., Academy of Management Journal, Administrative Science Quarterly, and Journal of Management) for the 1985–1987 and 1995–1997 periods. A brief summary of their findings follows. Experimental design. Research shows that the use of randomized experimental, quasi-experimental, and nonexperimental designs has fluctuated somewhat over time. Nonexperimental designs were used quite frequently in the periods considered by the reviews. In fact, they are the most frequently used type of design. Note, moreover, that the percentage of studies (POS) that used nonexperimental designs 64

was much higher in management journals (reviewed by Scandura & Williams, 2000) than in psychology journals (reviewed by Austin et al., 2002, and StoneRomero et al., 1995). The reasons for the variability in the types of experimental designs used by researchers across time are unclear. However, two possibilities merit consideration. One is that they were at least in part a function of the types of designs that were favored by the editors and editorial board members of the journals considered by the reviews. For instance, the journals considered by the Scandura and Williams (2000) review are less likely to publish the results of experimental studies than is the Journal of Applied Psychology. Another possibility is that the issues addressed by researchers may have influenced the experimental designs that they used. For example, some topics (e.g., goal setting, training) are more amenable to study by experimental means than are others (e.g., organizational commitment, job satisfaction). Whatever the reason(s) for the predominant use of nonexperimental designs in organizational research, it seems clear that greater use must be made of both quasi-experimental and randomized experimental designs. Consistent with this view, Highhouse (2009) argued that “despite the benefits of randomized experimentation . . . for making causal inferences, organizational scholarship has historically been characterized by an unhealthy over-reliance on [research based on] passive observation [i.e., nonexperimental designs]” (p. 554). Research settings. Neither Scandura and Williams (2000) nor Stone-Romero et al. (1995) provided information on the settings of the studies considered by their review. However, Austin et al. (2002) reported that the settings in which research has been conducted have varied from one decade to another. In contrast to the SP versus NSP setting distinction explained previously, Austin et al. (2002) distinguished between laboratory and field settings. Thus, the same distinction is used here. They reported that the POS conducted in (a) laboratory settings varied from 19.3 (1940) to 41.5 (1980) and was 22.4 in 2000, (b) field settings ranged from 53.7 (1980) to 80.7 (1940) and was 65.7 in 2000, and

6.60 (1995–1997) — 6.90 (1995–1997)

— — 42.40 (1995–1997) 5.10 (1995–1997)d

— — 30.70 (1985–1987) 8.3 (1985–1987) 4.0 (1985–1987) — 3.6 (1985–1987)

—

— — —

10.0 (1995–1997) 90.0 (1995–1997)

High (year)

—

— — —

15.5 (1985–1987) 61.6 (1985–1987)

Low (year)

Scandura & Williams (2000)

0.0 (1975) 0.0 (1975) 0.0 (1975)

4.00 (1975) 2.34 (1977) 0.67 (1975) 0.0 (1975) 10.16 (1977) 2.35 (1979)

— — —

32.5 (1990) 36.73 (1993)

Low (year)

8.16 (1993) 6.82 (1990) 10.20 (1993)

13.64 (1990) 12.24 (1993) 6.59 (1992) 2.56 (1983) 32.98 (1988) 16.04 (1982)

— — —

43.88 (1993) 58.67 (1975)

High (year)

Stone-Romero et al. (1995)

Note. ANOVA = analysis of variance; MANOVA = multivariate analysis of variance; ANCOVA = analysis of covariance; MANCOVA = multivariate analysis of covariance; SEM = structural equation modeling. Dashes indicate that no information on this variable was reported in the review. aIncludes between-subjects, within-subjects, and mixed designs. bIncludes ordinary least squares, weighted least squares, and logistic. cIncludes principal components analysis and common factor analysis. dFor Scandura and Williams review, category includes factor analysis and clustering techniques. eFor Scandura and Williams review, category includes SEM and path analysis.

16.4 (2000) 3.0 (2000) 13.5 (1990)

0.0 (1920) 0.0 (1920) 0.0 (1920)

41.5 (1980) 80.7 (1940) 6.8 (1990) 51.2 (1980) 16.2 (1990) 4.5 (2000) 2.7 (1990) 46.3 (2000) 9.8 (1980)

19.3 (1940) 53.7 (1980) 0.0 (1920)

Research setting Special purpose Nonspecial purpose Simulation

52.9 (1970) 78.0 (1930)

High (year)

0.0 (1920) 0.0 (1920) 0.0 (1920) 0.0 (1920) 1.8 (1940) 0.0 (1920)

14.6 (1930) 43.9 (1980)

Experimental design Experimental Nonexperimental

Data analysis technique ANOVAa MANOVA ANCOVA MANCOVA Multiple regression and variantsb Factor analysis (exploratory)c SEM Confirmatory factor analysis Path analysis/structural model Latent variable SEMe

Low (year)

Study attribute

Austin et al. (2002)

Findings from review of

Trends in Research: Percentage of Studies With Various Design and Analysis-Related Attributes

TABLE 2.1

Copyright American Psychological Association. Not for further distribution.

Research Strategies in I/O Psychology

65

Eugene F. Stone-Romero

(c) simulation settings extended from 0.0 (1920) to 6.8 and was 3.0 in 2000. Clearly, field settings are the most commonly used in I/O research.

Copyright American Psychological Association. Not for further distribution.

Data analytic strategies. A review of the articles by Austin et al. (2002), Scandura and Williams (2000), and Stone-Romero et al. (1995) shows that there are notable trends in the data analytic strategies used by researchers. Although a thorough consideration of all of them is beyond the scope of this chapter, the major ones are as follows. l. ANOVA variants. First, the use of multivariate data analytic strategies that are variants of ANOVA (e.g., multivariate analysis of variance [MANOVA], analysis of covariance, and multivariate analysis of covariance) has increased considerably. Consider, for instance, the use of MANOVA: The reviews by both Stone-Romero et al. (1995) and Austin et al. (2002) reported substantial increases in the POS using this technique. Scandura and Williams (2000) did not provide information on its use. 2. Multiple regression-based techniques. All three reviews indicated that there were marked increases in the use of multiple regression-based data analytic strategies. The largest such increase was reported by Austin et al. (2002). It appears to be a result of the rather long time period considered by their study (i.e., 1920 to 2000). 3. Structural equation modeling. The use of SEM techniques has increased greatly in the periods covered by the three reviews. This is true for analyses aimed at testing (a) measurement models (i.e., CFA), (b) SEM or path models, and (c) latent variable structural equation models (LVSEM). These increases appear to be a result of several factors. First, training in many SEM methods (covariance structure analysis, path analysis, and LVSEM) procedures was not widely available to researchers in I/O psychology and related disciplines until the 1970s. For instance, the first article on CFA appeared in 1969. Second, the software needed to conduct SEM analyses (i.e., LISREL 3) did not appear until 1976 and was both expensive and difficult to use. Today, however, SEM software is readily available and quite user friendly. For example, the AMOS software 66

for testing SEM models is now part of the widely available SPSS software package.

Methods-Related Articles Aguinis et al. (2009) reported the results of a contentanalysis-based review of 193 articles that were published in Organizational Research Methods from 1998 to 2007. Among the findings were that the percentage of articles (POAs) focusing on quantitative methods were much greater than those for qualitative methods. In addition, for quantitative methods, there were (a) increases in the POAs focusing on electronic/Web research, survey research, and multilevel data analytic methods, and (b) decreases in the POAs dealing with SEM. Moreover, for qualitative methods there were (a) increases in the POAs devoted to action research and interpretive research and (b) increases (1998–2002) followed by decreases (2002–2006) in the POAs concerned with policy capturing. These trends may suggest methods-related education and training needs for basic and researchers in I/O psychology and allied fields. However, care must be exercised in interpreting them. More specifically, articles on a specific method may focus on explanations of its proper use, controversies surrounding its use, or critical views on its use. Consider three examples of the latter possibility: Rogosa (1987) described the erroneous inferences that can result from the use of crosslagged panel correlation strategy (e.g., Kenny, 1975, 1979) for deriving evidence about causal relations using data from nonexperimental research. Kalleberg and Kleugel (1975) pointed out the problems that can result from the use of the MTMM strategy for assessing the construct validity of measures. StoneRomero and Anderson (1994) showed why moderated multiple regression is a better strategy for detecting moderating effects than testing for differences among subgroup-based correlation coefficients. In short, publications on a given topic are not always a signal that it is of value to researchers. CONCLUSIONS Several conclusions appear warranted on the basis of the foregoing. They have to do with the implications of research design for validity inferences, the distinction between experimental design and data ana-

Research Strategies in I/O Psychology

lytic strategies, and the value of methodological pluralism in empirical research.

Copyright American Psychological Association. Not for further distribution.

Research Design and Validity Inferences A researcher has many options when designing a study, including those that pertain to its experimental design, which is the most important determinant of its internal validity. As noted previously, the experimental design used in a study should be governed by its major purpose(s). If the sole interest of a researcher is to determine if two or more variables are related to one another, he or she can conduct a study that uses a nonexperimental design. However, if the researcher is concerned with determining causal relations between or among variables, he or she should opt for a randomized experimental study (Cook & Campbell, 1976, 1979; Cook et al., 1990; Shadish et al., 2002; Stone-Romero, 2002, 2007c, 2009; Stone-Romero & Rosopa, 2004, 2008). And, in instances in which experimentation is possible but units cannot be randomly assigned to treatment conditions, a quasiexperimental study should be conducted (Shadish et al., 2002; Stone-Romero, 2002, 2007b, 2009; Stone-Romero & Rosopa, 2008). Irrespective of the experimental design used in a study, a researcher must devote attention to the other facets of validity. First, to bolster inferences about construct validity, he or she must take care to ensure that the sampling particulars of a study (e.g., units, treatments, settings, and outcomes) are faithful representations of associated constructs. It is of virtually no value to show that a manipulation of X leads to changes in Y if a researcher has little or no knowledge about the constructs associated with these variables. Thus, for example, the construct validity of a study would be questionable if it dealt with the decision-making processes of neurosurgeons but used undergraduate business students as subjects. Second, a researcher has to be concerned about the validity of inferences derived from statistical analyses, that is, statistical conclusion validity. It is of almost no value for a researcher to claim that a study shows support for a hypothesized relation between two variables if this finding is based on a study that suffers, for example, from the problem of

inflated study-wise Type I error. In addition, the study would be of no value if there truly is a relation between the variables, but the study lacks the statistical power to detect it. Third, and finally, a study might be very strong in terms of internal validity, construct validity, and statistical conclusion validity, but weak with regard to external validity. Thus, if it is important to make inferences about a causal relation to sampling particulars other than those considered by a study (e.g., to other sets of units, treatments, outcomes, and settings), the researcher must conduct additional studies to support claims about external validity. It is important to add, however, that external validity is not always an important concern in research (Mook, 1983).

Experimental Design and Data Analytic Strategies From the foregoing it should be clear that there is a very important distinction between experimental design and the methods used to analyze data from a study. Regrettably, large numbers of publications in I/O psychology and allied disciplines provide clear evidence of a very prevalent problem: Researchers who are interested in showing causal relations between (among) variables seem to think that statistical methods are an appropriate substitute for sound experimental design; that is, they base conclusions about causal connections between (among) variables on the results of applying so-called causal modeling procedures (e.g., HMR, PA, SEM) to data from nonexperimental research (see Stone-Romero, 2009; Stone-Romero & Gallaher, 2006; StoneRomero & Rosopa, 2004, 2008). However, valid inferences about cause stem from sound experimental design, not the use of so-called causal modeling procedures. Consequently, when nonexperimental research is reported, researchers must be careful to avoid causal interpretations of their findings. Recall, for example, that an observed relation between two variables (e.g., X and Y) can be consistent with multiple assumed causal models (see Figure 2.4), and a researcher typically has no basis for legitimately arguing that one such model (e.g., Figure 2.4a) is superior to the alternatives (e.g., Figures 2.4b and 2.4c). 67

Eugene F. Stone-Romero

Copyright American Psychological Association. Not for further distribution.

Methodological Pluralism In studies of virtually all phenomena in I/O psychology and related disciplines, methodological pluralism is highly desirable (Rosenthal & Rosnow, 2008; Shadish et al., 2002; Webb et al., 1981). The call for methodological pluralism is consistent with the fiat mistura (i.e., let there be a mixture) prescription of Fromkin and Streufert (1976). In early stages of research on a phenomenon, nonexperimental studies may offer valuable clues about the correlates of a variable that is of interest to an investigator. In addition, the results of nonexperimental research may provide a basis for the formulation of hypotheses about causal connections between the same variable and its assumed antecedents. However, such hypotheses are best tested using either randomized experimental or appropriate quasi-experimental designs. Another highly important dimension of methodological pluralism relates to the empirical realizations of the constructs considered by a study. In any given study, it is vital to demonstrate that there is a clear connection between any given construct and its sampling particulars (e.g., manipulations and/or measures of variables). However, in subsequent studies it is critical to demonstrate that the observed relation between studied variables is not an artifact of sampling particulars. Thus, for example, if a researcher used a specific measure of an assumed effect construct in one study, he or she should conduct follow-up research that uses alternative measures of it. In addition, if the researcher manipulated an independent variable in a particular way in one study, he or she should use alternate manipulations in subsequent studies.

A Final Word Sound research is vital to developing an understanding of causal relations between focal variables and their antecedents and consequences. The findings of such research can (a) contribute greatly to the development and testing of theories and (b) provide a firm basis for the design of interventions aimed at changing individuals, groups, and organizations. Thus, in terms of multiple criteria that are highly important to both scientists and practitioners in I/O psychology and related fields, it is critical that research on relations between variables have high levels of internal, 68

construct, external, and statistical conclusion validity. However, of these types of validity, internal is the sine qua non ( J. C. Campbell & Stanley, 1963).

References Aguinis, H., Pierce, C. A., Bosco, F. A., & Muslin, I. S. (2009). First decade of Organizational Research Methods: Trends in design, measurement, and data analysis topics. Organizational Research Methods, 12, 69–112. Ajzen, I. (1988). Attitudes, personality, and behavior. Chicago: Dorsey Press. Argyris, C. (1968). Some unintended consequences of rigorous research. Psychological Bulletin, 70, 185–197. Argyris, C. (1980). Inner contradictions of rigorous research. New York: Academic Press. Aronson, E., Ellsworth, P., Carlsmith, J. M., & Gonzales, M. H. (1990). Methods of research in social psychology (2nd ed.). New York: McGraw-Hill. Austin, J. T., Scherbaum, C. A., & Mahlman, R. A. (2002). History of research methods in industrial and organizational psychology: Measurement, design, analysis. In S. G. Rogelberg (Ed.), Handbook of research methods in industrial and organizational psychology (pp. 1–33). Malden, MA. Blackwell. Baron, R. M., & Kenny, D. A. (1986). The moderatormediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Bateman, T. S., & Strasser, S. (1984). A longitudinal analysis of the antecedents of organizational commitment. Academy of Management Journal, 27, 95–112. Berkowitz, L., & Donnerstein, E. (1982). External validity is more than skin deep: Some answers to criticisms of laboratory experiments. American Psychologist, 37, 245–257. Blalock, H. M. (1964). Causal inferences in nonexperimental research. New York: W. W. Norton. Blalock, H. M. (1971). Causal models in the social sciences. Chicago: Aldine. Blanchard, E. B., Peters, M. L., Hermann, C., Turner, S. M., Buckley, T. C., Barton, K., & Dentinger, M. P. (1997). Direction of temperature control in the thermal biofeedback treatment of vascular headache. Applied Psychophysiology and Biofeedback, 22, 227–245. Blum, M., & Naylor, J. C. (1968). Industrial psychology: Its theoretical and social foundations. New York: Harper and Row. Boehm, V. R. (1972). Negro-white differences in validity of employment and training selection procedures: Summary of research evidence. Journal of Applied Psychology, 56, 33–39.

Research Strategies in I/O Psychology

Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley. Borman, W. C. (1991). Job behavior, performance, and effectiveness. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 2, pp. 271–326). Palo Alto, CA: Consulting Psychologists Press.

Copyright American Psychological Association. Not for further distribution.

Bott, J. P. (2007). Case study method. In S. G. Rogelberg (Ed.), Encyclopedia of industrial and organizational psychology (Vol. 1, pp. 68–69). Thousand Oaks, CA: Sage. Bouchard, T. (1976). Field research methods: Interviewing, questionnaires, participant observation, systematic observation, and unobtrusive measures. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 363–413). Chicago: Rand McNally. Boudreau, J. W. (1991). Utility analysis for decisions in human resource management. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 2, pp. 621–717). Palo Alto, CA: Consulting Psychologists Press.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum. Conway, J. M. (2002). Method variance and method bias in industrial and organizational psychology. In S. G. Rogelberg (Ed.), Handbook of research methods in industrial and organizational psychology (pp. 344–365). Malden, MA: Blackwell. Cook, T. D., & Campbell, D. T. (1976). The design and conduct of quasi-experiments and true experiments in field settings. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 223–326). Chicago: Rand McNally. Cook, T. D., & Campbell, D. T. (1979). Quasiexperimentation: Design and analysis issues for field settings. Boston: Houghton Mifflin. Cook, T. D., Campbell, D. T., & Peracchio, L. (1990). Quasi-experimentation. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 1, pp. 491–576). Palo Alto, CA: Consulting Psychologists Press.

Braithwaite, R. (1996). Scientific explanation. Herndon, VA: Books International.

Cronbach, L. J., & Gleser, G. C. (1965). Psychological tests and personnel decisions. Urbana: University of Illinois Press.

Brannick, M. T. (1995). Critical comments on applying covariance structure modeling. Journal of Organizational Behavior, 16, 201–213.

Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281–302.

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage. Buchanan, D., & Bryman, A. (Eds.). (2009). Handbook of organizational research methods. London: Sage.

Currivan, D. B. (1999). The causal order of job satisfaction and organizational commitment in models of employee turnover. Human Resource Management Review, 9, 495–524.

Campbell, D. T., & Fiske, D. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81–105.

Dipboye, R. L., & Flanagan, M. F. (1979). Research settings in industrial and organizational psychology: Are findings in the field more generalizable than in the laboratory? American Psychologist, 34, 141–150.

Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental designs for research. Chicago: Rand McNally.

Doty, D. H., & Glick, W. H. (1998). Common method bias: Does common methods variance really bias results? Organizational Research Methods, 1, 374–406.

Campbell, J. P. (1976). Psychometric theory. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 185–222). Chicago: Rand McNally.

Feldman, J. M., & Lynch, J. G. (1988). Self-generated validity and other effects of measurement on belief, attitude, intention, and behavior. Journal of Applied Psychology, 73, 421–435.

Campbell, J. P. (1986). Labs, fields, and straw issues. In E. A. Locke (Ed.), Generalizing from laboratory to field settings: Research findings from industrial-organizational psychology, organizational behavior, and human resource management (pp. 269–279). Lexington, MA; Lexington Books.

Freedman, D. A. (1987). As others see us: A case study in path analysis. Journal of Educational Statistics, 12, 101–128.

Cliff, N. (1987). Comments on Professor Freedman’s paper. Journal of Educational Statistics, 12, 158–160. Cochran, W. G. (1977). Sampling techniques (3rd ed.). New York: Wiley. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Frese, M., Garst, H., & Fay, D. (2007). Making things happen: Reciprocal relationships between work characteristics and personal initiative in a four-way longitudinal structural equation model. Journal of Applied Psychology, 92, 1084–1102. Fromkin, H. L., & Streufert, S. (1976). Laboratory experimentation. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 415–465). Chicago: Rand McNally. 69

Eugene F. Stone-Romero

Ghiselli, E. E. (1956). Dimensional problems of criteria. Journal of Applied Psychology, 40, 1–4.

Kenny, D. A. (1975). Cross-lagged panel correlation: A test for spuriousness. Psychological Bulletin, 82, 887–903.

Ghiselli, E. E., Campbell, J. P., & Zedeck, S. (1981). Measurement theory for the behavioral sciences. New York: W. H. Freeman.

Kenny, D. A. (1979). Correlation and causality. New York: Wiley.

Copyright American Psychological Association. Not for further distribution.

Grant, A. M., & Wall, T. D. (in press). The neglected science and art of quasi-experimentation: Why-to, when-to, and how-to advice for organizational researchers. Organizational Research Methods.

Kerlinger, F. N., & Lee, H. B. (2000). Foundations of behavioral research (4th ed.). Belmont, CA: Wadsworth. Kish, L. (1965). Survey sampling. New York: Wiley.

Guion, R. M. (1976). Recruiting, selection, and job placement. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 777–828). Chicago: Rand McNally.

Koslowsky, M. (1991). A longitudinal analysis of job satisfaction, commitment, and intention to leave. Applied Psychology: An International Review, 40, 405–415.

Guion, R. M. (1991). Personnel assessment, selection, and placement. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 2, pp. 327–397). Palo Alto, CA: Consulting Psychologists Press.

Kuhn, T. S. (1970). The structure of scientific revolutions (2nd ed., enlarged). Chicago: University of Chicago Press.

Guion, R. M. (2002). Validity and reliability. In S. G. Rogelberg (Ed.), Handbook of research methods in industrial and organizational psychology (pp. 57–76). Malden, MA: Blackwell. Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16, 250–279. Hackman, J. R., Pearce, J. L., & Wolfe, J. C. (1978). Effects of changes in job characteristics on work attitudes and behaviors: A naturally occurring quasi-experiment. Organizational Behavior and Human Performance, 21, 289–304. Harris, M., & Schaubroeck, J. (1988). A meta-analysis of self-supervisory, self-peer, and peer-supervisory ratings. Personnel Psychology, 41, 43–62. Hays, W. L. (1994). Statistics (5th ed.). Fort Worth, TX: Harcourt Brace. Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. New York: Academic Press. Highhouse, S. (2009). Designing experiments that generalize. Organizational Research Methods, 12, 554–566. Holland, P. W. (1986). Statistics and causal inference. Journal of the American Statistical Association, 81, 945–960. Kalish, D., & Montague, R. (1964). Logic: Techniques of formal reasoning. New York: Harcourt, Brace, & World. Kalleberg, A. L., & Kluegel, J. R. (1975). Analysis of the multitrait-multimethod matrix: Some limitations and an alternative. Journal of Applied Psychology, 60, 1–9. Kaplan, A. (1964). The conduct of inquiry. New York: Intext. Kelloway, E. K. (1998). Using LISREL for structural equation modeling: A researcher’s guide. Thousand Oaks, CA: Sage. 70

Lance, C. E. (1991). Evaluation of a structural model relating job satisfaction, organizational commitment, and precursors to voluntary turnover. Multivariate Behavioral Research, 26, 137–162. Lance, C. E., Hoffman, B. J., Gentry, W. A., & Baranik, L. E. (2008). Rater source factors represent important subcomponents of the criterion construct space, not rater bias. Human Resource Management Review, 18, 223–232. Lawler, E. E., Mohrman, A. M., Mohrman, S. A., Ledford, G. E., Cummings, T. G., & Associates. (1985). Doing research that is useful for theory and practice. San Francisco: Jossey-Bass. Ling, R. F. (1982). Review of Correlation and causality by David A. Kenny. Journal of the American Statistical Association, 77, 489–491. Locke, E. A. (1986). Generalizing from laboratory to field settings: Research findings from industrialorganizational psychology, organizational behavior, and human resource management. Lexington, MA: Lexington Books. Locke, K., & Golden-Biddle, K. (2002). An introduction to qualitative research: Its potential for industrial and organizational psychology. In S. G. Rogelberg (Ed.), Handbook of research methods in industrial and organizational psychology (pp. 99–118). Malden, MA: Blackwell. Maruyama, G. M. (1998). Basics of structural equation modeling. Thousand Oaks, CA: Sage. Mathieu, J. E., DeShon, R. P., & Bergh, D. D. (2008). Mediational inferences in organizational research: Then, now, and beyond. Organizational Research Methods, 11, 203–223. Mathieu, J. E., & Taylor, S. R. (2006). Clarifying conditions and choice points for mediational type inferences in organizational behavior. Journal of Organizational Behavior, 27, 1031–1056.

Research Strategies in I/O Psychology

McDonald, R. P. (1999). Test theory: A unified treatment. Mahwah, NJ: Erlbaum. Millsap, R. E. (2002). Structural equation modeling: A user’s guide. In F. Drasgow & N. Schmitt (Eds.), Measuring and analyzing behavior in organizations (pp. 257–301). San Francisco: Jossey-Bass. Mook, D. G. (1983). In defense of external invalidity. American Psychologist, 38, 379–387. Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill.

Copyright American Psychological Association. Not for further distribution.

Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill. Orne, M. T. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American Psychologist, 17, 776–783. Orne, M. T. (1969). Demand characteristics and the concept of quasi-controls. In R. Rosenthal & R. L. Rosnow (Eds.), Artifact in behavioral research (pp. 143–179). New York: Academic Press. Pedhazur, E. J. (1982). Multiple regression in behavioral research: Explanation and prediction (2nd ed.). New York: Holt, Rinehart and Winston. Pinder, C. C. (1998). Work motivation in organizational behavior. Upper Saddle River, NJ: Prentice-Hall. Podsakoff, P. M., MacKenzie, S. B., Lee, J-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879–903. Porter, L. W., & Lawler, E. E. (1968). Managerial attitudes and performance. Homewood, IL: Irwin-Dorsey. Rogosa, D. (1987). Causal models do not support scientific conclusions: A comment in support of Freedman. Journal of Educational Statistics, 12, 185–195. Rosenthal, R. (1991). Meta-analytic procedures for social research (Rev. ed.). Newbury Park, CA: Sage. Rosenthal, R., & Rosnow, R. L. (1969). Artifact in behavioral research. New York: Academic Press. Rosenthal, R., & Rosnow, R. L. (2008). Essentials of behavioral research (3rd ed.). New York: McGraw-Hill. Rosopa, P. J., & Stone-Romero, E. F. (2008). Problems with detecting assumed mediation using the hierarchical multiple regression strategy. Human Resource Management Review, 18, 294–310.

Scandura, T. A., & Williams, E. A. (2000). Research methodology in management: Current practices, trends, and implications for future research. Academy of Management Journal, 43, 1248–1264. Schmidt, F. L., & Hunter, J. E. (1974). Racial and ethnic bias in psychological tests: Divergent implications of two definitions of test bias. American Psychologist, 29, 1–8. Searle, S. R. (1971). Linear models. New York: Wiley. Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Boston: Houghton Mifflin. Smith, P. C. (1976). Behaviors, results, and organizational effectiveness: The problem of criteria. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 745–775). Chicago: Rand McNally. Snedecor, G. W., & Cochran, W. G. (1980). Statistical methods (7th ed.). Ames: Iowa State University. Spector, P. (2006). Method variance in organizational research: Truth or urban legend? Organizational Research Methods, 9, 221–232. Stanton, J. M., & Rogelberg, S. G. (2002). Beyond online surveys: Internet research opportunities for industrial and organizational psychology. In S. Rogelberg (Ed.), The encyclopedia of industrial and organizational psychology (pp. 275–294). Beverly Hills, CA: Sage. Stone, D. L., & Stone, E. F. (1985). The effects of feedback favorability and feedback consistency on selfperceived task competence and perceived feedback accuracy. Organizational Behavior and Human Decision Processes, 36, 167–185. Stone, E. F. (1978). Research methods in organizational behavior. Santa Monica, CA: Goodyear. Stone, E. F. (1979). Field independence and perceptions of task characteristics: A laboratory investigation. Journal of Applied Psychology, 64, 305–310. Stone, E. F. (1981). Action science: Pulling the trigger on research rigor. Contemporary Psychology, 26, 629–630. Stone, E. F. (1982). In defense of rigorous research. Contemporary Psychology, 27, 581. Stone, E. F. (1987). Review of Doing research that is useful for theory and practice, by E. E. Lawler et al. Journal of Occupational Behavior, 8, 87–91.

Runkel, P. J., & McGrath, J. E. (1972). Research on human behavior: A systematic guide to method. New York: Holt, Rinehart, & Winston.

Stone, E. F. (1992). A critical analysis of social information processing models of job perceptions and job attitudes. In C. J. Cranny, P. C. Smith, & E. F. Stone (Eds.), Job satisfaction: How people feel about their jobs and how it affects their performance (pp. 21–44). Lexington, MA: Lexington-Heath.

Salancik, G. R., & Pfeffer, J. (1978). A social information processing approach to job attitudes and task design. Administrative Science Quarterly, 23, 224–253.

Stone, E. F., Gueutal, H. G., Gardner, D. G., & McClure, S. (1983). A field experiment comparing informationprivacy values, beliefs, and attitudes across several 71

Eugene F. Stone-Romero

Copyright American Psychological Association. Not for further distribution.

types of organizations. Journal of Applied Psychology, 68, 459–468.

pedia of measurement and statistics (pp. 995–999). Beverly Hills, CA: Sage.

Stone, E. F., Stone, D. L., & Gueutal, H. G. (1990). The influence of cognitive ability on responses to questionnaire measures. Journal of Applied Psychology, 75, 418–427.

Stone-Romero, E. F., & Rosopa, P. J. (2008). The relative validity of inferences about mediation as a function of research design characteristics. Organizational Research Methods, 11, 326–352.

Stone-Romero, E. F. (1994). Construct validity issues in organizational behavior research. In J. Greenberg (Ed.), Organizational behavior: The state of the science (pp. 155–179). Hillsdale, NJ: Erlbaum.

Stone-Romero, E. F., & Stone, D. L. (2007). Cognitive, affective, and cultural influences on stigmatization and its impact on human resource management processes and practices. Research in Personnel and Human Resource Management, 26, 117–167.

Stone-Romero, E. F. (2002). The relative validity and usefulness of various empirical research designs. In S. G. Rogelberg (Ed.), Handbook of research methods in industrial and organizational psychology (pp. 77–98). Malden, MA: Blackwell. Stone-Romero, E. F. (2007a). Non-experimental designs. In S. Rogelberg (Ed.), The encyclopedia of industrial and organizational psychology (pp. 519–521). Beverly Hills, CA: Sage. Stone-Romero, E. F. (2007b). Quasi-experimental designs. In S. Rogelberg (Ed.), The encyclopedia of industrial and organizational psychology (pp. 655–657). Beverly Hills, CA: Sage. Stone-Romero, E. F. (2007c). Experimental designs. In S. Rogelberg (Ed.), The encyclopedia of industrial and organizational psychology (pp. 239–241) Beverly Hills, CA: Sage. Stone-Romero, E. F. (2009). Implications of research design options for the validity of inferences derived from organizational research. In D. Buchanan & A. Bryman (Eds.), Handbook of organizational research methods (pp. 302–327). London: Sage. Stone-Romero, E. F., & Anderson, L. (1994). Relative power of moderated multiple regression and the comparison of subgroup correlation coefficients for detecting moderating effects. Journal of Applied Psychology, 79, 354–359. Stone-Romero, E. F., & Gallaher, L. (2006, May). Inappropriate use of causal language in reports of nonexperimental research. Paper presented at the meeting of the Society for Industrial and Organizational Psychology, Dallas, TX. Stone-Romero, E. F., & Liakhovitski, D. (2002). Strategies for detecting moderator variables: A review of theory and research. Research in Personnel and Human Resources Management, 21, 333–372. Stone-Romero, E. F., & Rosopa, P. J. (2004). Inference problems with hierarchical multiple regression-based tests of mediating effects. Research in Personnel and Human Resources Management, 23, 249–290. Stone-Romero, E. F., & Rosopa, P. J. (2007). Tests of mediating effects. In N. J. Salkind (Ed.), The encyclo-

72

Stone-Romero, E. F., Stone, D. L., & Hyatt, D. (2003). Personnel selection procedures and invasion of privacy. Journal of Social Issues, 59, 343–368. Stone-Romero, E. F., Weaver, A. E., & Glenar, J. L. (1995). Trends in research design and data analytic strategies in organizational research. Journal of Management, 21, 141–157. Thomas, K. W., & Tymon, W. G. (1982). Necessary properties of relevant research: Lessons from criticisms of the organizational sciences. Academy of Management Review, 7, 345–352. Vandenberg, R. J., & Lance, C. E. (1992). Examining the causal order of job satisfaction and organizational commitment. Journal of Management, 18, 153–167. Webb, E. J., Campbell, D. T., Schwartz, R. D., Sechrest, L., & Grove, J. B. (1981). Nonreactive measures in the social sciences (2nd ed.). Boston: Houghton Mifflin. Weber, S. J., & Cook, T. D. (1972). Subject effects in laboratory research: An examination of subject roles, demand characteristics, and valid inference. Psychological Bulletin, 77, 273–295. Weick, K. E. (1965). Laboratory experimentation with organizations. In J. G. March (Ed.), Handbook of organizations (pp. 194–260). Chicago: Rand McNally. Wiener, Y., & Vardi, Y. (1980). Relationships between job, organization, and career commitments and work outcomes: An integrative approach. Organizational Behavior and Human Performance, 26, 81–96. Williams, L. J., & Hazer, J. T. (1986). Antecedents and consequences of satisfaction and commitment in turnover models: A reanalysis using latent variable structural equation methods. Journal of Applied Psychology, 71, 219–231. Yin, R. K. (1989). Case study research: Design and method. Newbury Park, CA: Sage. Zedeck, S. (1971). Problems with the use of “moderator” variables. Psychological Bulletin, 76, 295–310. Zickar, M. J., & Slaughter, J. E. (2002). Computational modeling. In S. Rogelberg (Ed.), The encyclopedia of industrial and organizational psychology (pp. 184–197). Beverly Hills, CA: Sage.

CHAPTER 3

QUALITATIVE RESEARCH STRATEGIES IN INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY

Copyright American Psychological Association. Not for further distribution.

Thomas W. Lee, Terence R. Mitchell, and Wendy S. Harman

In the overlapping areas of industrial and organizational (I/O) psychology, human resource management, and organizational behavior, interest in qualitative research waxes and wanes. About 30 years ago, for example, Van Maanen (1979) argued in a very compelling fashion for an unrealized value in applying qualitative research methods. Further, he issued what may well be the earliest call to organizational researchers to adopt these methods more commonly. Twenty years later, Lee, Mitchell, and Sablynski (1999) assessed how qualitative methods had come to be defined and how broadly these methods had been used in vocational and organizational psychology. Ten years later, Harman, Lee, and Mitchell (2009) updated a portion of the earlier Lee, Mitchell, and Sablynski review and focused on how qualitative methods had been applied to substantive organizational issues comparing the United States and Europe. They did not directly ask, or update, how specific and major kinds of qualitative methods have been applied. In other words, Harman et al. addressed specific content issues. In contrast, we focus here less on theoretical content and more on methodological topics in qualitative research. As such, we rely on the original structure by Lee, Mitchell, and Sablynski, and we update and review methods currently in use. Further, it should be noted that Van Maanen likely addressed a somewhat broader audience, namely, management researchers, whereas Lee, Mitchell, and Sablynski and Harman et al. likely addressed perhaps narrower audiences (namely, vocational psychologists, by Lee, Mitchell, and Sablynski, 1999, and European management researchers, by Harman et al., 2009). In this chapter,

we seek to reintroduce qualitative methods to a very broad audience of applied psychologists. We believe that the topics studied by I/O psychologists, and psychologists in other areas, are becoming increasingly complicated. In I/O psychology, for example, context has long been recognized as critical to understanding people in organizations (Hackman & Oldham, 1976; Johns, 2006). Perhaps as a result of more complicated topics and the general recognition for a role of context in our theories and research, many researchers apply cross-level analytical techniques such as hierarchical linear modeling to better understand how higher level contextual variables directly or indirectly affect behavior. (See also chap. 4, this volume.) I/O psychologists may see, for example, context more as a substantive or control variable as they separate the effect of employees embedded within work groups, of work groups embedded within departments, of departments embedded within firms, or of firms embedded within industries. To better understand these new and evolving topics, researchers need to pursue new or different methodological tools. We submit that knowledge of qualitative methods offers additional tools to all psychologists with which to understand not only context as a substantive or control variable but many other phenomena as well. Qualitative research will not, and should not, replace quantitative methods, but it has an important role in contributing to, and supplementing researchers’ understanding of, behavior in organizations. As a disclaimer, however, we do not cover all types of qualitative research, nor do we consider all major or relevant issues or controversies in this

http://dx.doi.org/10.1037/12169-003 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

73

Copyright American Psychological Association. Not for further distribution.

Lee, Mitchell, and Harman

domain. Instead, we offer more of a primer with the explicit intent of generating interest for a particular set of research methods. If at least some people opt to pursue these methods in more detail, we believe that this chapter accomplishes its purpose. (For additional information, please see the July 2008 and January 2002 issues of Organizational Research Methods.) In the sections to follow, we explicitly address what qualitative research is and when to use it, purposes of qualitative research, common research designs, common analytical techniques, and hybrid studies. Although we imply clear distinctiveness, the separation and identification of specific purposes for qualitative methods, research designs, and analytic techniques are somewhat arbitrary on our part. More specifically, these labels are used for purposes of discussion. In the actual practice of qualitative research, the lines of separation among these research labels are most often fuzzy. Throughout this chapter, we offer examples of what we believe to be qualitative exemplars in I/O psychology, human resource management, and organizational behavior. We intentionally offer a broad set of examples, in part to illustrate the range of possible applications of qualitative methods by I/O psychologists. WHAT IS QUALITATIVE RESEARCH, AND WHEN DO YOU USE IT? Although scholars have argued over what is and is not qualitative research, four general attributes are generally agreed on. First, it occurs in natural settings. Generally speaking, it is not conducted in a laboratory, although Gersick’s (1989) laboratory study is a wonderful exception. (Gersick, 1989, e.g., studied the punctuated equilibrium model of changes in work groups with simulated teams composed of master of business administration students in a laboratory. The groups’ meetings were videotaped, and the transcriptions were later analyzed with qualitative methods.) Second, qualitative data derive from the participants’ perspective. Typically, the researcher should not impose immediate interpretations. After substantial analysis, however, theoretical inductions (e.g., grounded theory) and/or particular interpretations (e.g., critical theory) become quite legitimate. Third, qualitative research should be reflexive (i.e., flexible). Qualitative designs 74

can be readily changed to fit the fluid or dynamic demands of the research setting. In our view, this attribute most easily differentiates qualitative from traditional quantitative research, which might be characterized as more rule driven or algorithmic (e.g., experiments, survey research). (See also chap. 2, this volume.) Fourth, qualitative instrumentation, observation methods, and modes of analysis are not standardized. Instead, the individual psychologist serves as the main research instrument. (Although many recent and impressive software packages greatly facilitate data analysis, these packages do not substitute—yet—for the qualitative researcher’s insight and inductive reasoning.) In contrast to the four general attributes of qualitative methods (above), more traditional quantitative studies can occur in the laboratory, field settings, or both (as opposed to “only” field settings). Typically, quantitative studies impose a particular theoretical perspective through which the researcher can best understand the data (as opposed to adopting the participant’s perspective). Also, traditional quantitative studies adopt stricter standards on consistency through experimental controls or randomization to eliminate alternative explanations (as opposed to a reflexive standard). Finally, traditional quantitative studies require more rigorous standards for the assessment of reliability and validity of measurement (as opposed to the researcher himself or herself serving as the research instrument). Further, two major themes characterize much of the published qualitative research. First, it is often a process of data reduction that simultaneously enhances and leads to inductions about theory. In other words, a vast amount of subjectively gathered data (e.g., context-specific observations) are logically but consistently reduced to enhance interpretive coherence. Second, and noted previously, qualitative research involves few, if any, standardized, wellresearched, or otherwise objectively observable instruments. For example, there are no widely accepted, off-the-Web, and commercially available tools to purchase. Elsewhere (Lee, Mitchell, & Sablynski, 1999), we have offered a somewhat rough analogy between generic qualitative research and exploratory factor analysis (EFA; e.g., an ocular EFA). In their separate ways, both families of techniques take large amounts of data and reduce them

Copyright American Psychological Association. Not for further distribution.

Qualitative Research Strategies in I/O Psychology

to a more meaningful whole. With EFA, latent traits are often inferred, whereas with qualitative methods, coherent categories are often induced. With both methods, greater understanding is achieved. In a more specific sense, qualitative research is well suited for issues of vivid description of realworld phenomena, rich interpretation of those deeply contextualized phenomena, and the development of theoretical understanding of those phenomena. For example, it addresses the questions, What is occurring? How is it occurring? and What constructs should I evoke to explain it? New constructs linked together in new ways often produce novel theoretical insights. In contrast, it is not well suited for issues of prevalence, generalization, or calibration, which may be issues better addressed by probabilistic sampling and statistical inference. In short and generally speaking, qualitative and quantitative research methods address separate but no less important research issues. VARIED PURPOSES OF QUALITATIVE RESEARCH In Lee, Mitchell, and Sablynski (1999), we characterized the purposes of most (but not all) qualitative research to address inferences about theory generation or elaboration but less so about theoretical testing. Although separation of these purposes is not always clear-cut, these descriptors are useful, nevertheless, for purposes of discussion. Perhaps their most well-known and readily understood application, theoretical generation occurs when the study’s design produces formal and testable propositions. Even when such propositions are not offered, the testable inferences are often easily made. Theoretical elaboration typically occurs when preliminary ideas, frameworks, or earlier theory drive the study’s design. Most often, testable hypotheses are not required. Theory testing occurs when formal hypotheses or a formal theory determines the study’s hypotheses and design. The goal of theory testing is most often accomplished through quantitative research strategies but can be done with qualitative methods. An example of “pure” theory generation, that is to say the author described the article as generating theory and provided propositions, is Hardy, Phillips,

and Lawrence’s (2003) article on interorganizational collaboration. In looking into the depth and scope of interactions between collaborating organizations, for instance, they used “a qualitative, multicase comparative research design” (p. 328) focusing on building theory from systematic coding and comparison of collaboration across cases holding organizational characteristics constant. The study examined a branch of Mère et Enfant operating in the West Bank and Gaza. The purpose of this branch of the organization is to provide medical and nutritional services in clinics in the area by using an outreach program in rural communities. Their education services focus on improving child nutrition and rehabilitating malnourished children, reducing infant mortality, and raising awareness of the importance of good nutrition. Additionally, Mère et Enfant in this region trains health care professionals in the areas of breastfeeding and weaning, and diarrhea management. They research nutritional and other issues related to the health of Palestinian children, and they provide information and education about nutrition and poverty. The organization collaborates with several other nongovernment organizations (NGOs) operating in similar environments. The results of the study allowed for theory generation and testable propositions as to the depth and nature of interactions between the focal NGO, Mère et Enfant, and various collaborating organizations. Collaborations were found, for example, to produce strategic, knowledge creation, and influence effects, yet these were found to have differential relevance across collaborations. Rather, the type of interaction between collaborating organizations determines what influence effects are realized. In an exemplar of theoretical elaboration, Maitlis (2005) extended the knowledge on organizational sensemaking in three British orchestras by using qualitative methods to study a traditional organization operating in a changing environment. Shrinking government subsidies, greater competition for corporate donors, increasing and varying forms of entertainment, and the type of industry contribute to dynamism in the environment. Qualitative methods are particularly well suited for organizations experiencing any or all of these ongoing influences. In particular, Maitlis investigated whether discernable patterns of sensemaking interactions affect different 75

Copyright American Psychological Association. Not for further distribution.

Lee, Mitchell, and Harman

social processes within these orchestras. Over 2 years, she conducted 120 semistructured interviews that included multiple repeat interviews with key informants, informal conversations, and observations of 107 meetings. Analytically, Maitlis sought to ascertain patterns of sense giving and sense taking in her data. First, she wrote 27 narratives that initially summarized her data. Second, she identified the key stakeholders (i.e., prominent players) in the sensemaking process (e.g., formal and informal leaders) described in these narratives. Third, she identified specific sense-giving behaviors and activities (e.g., calling meetings, offering plausible descriptions and explanations of environmental cues, issuing warnings, expressing opinions, writing reports). Fourth, Maitlis counted the number of stakeholders and judged the intensity and frequency of their sensegiving behaviors and activities. Fifth, she reduced the 27 narratives to 4 by creating a 2 × 2 matrix with high and low leader sense giving crossed by high and low other stakeholder sense giving. Sixth, she examined the interview quotes and related field notes for each of her four cells (e.g., each cell might have different descriptors of duplicity, deal making, and inclusion within decision making). Finally, she returned to the complete data set, and in an iterative fashion, Maitlis elaborated on and/or increased the descriptive richness of the underlying patterns (or themes) and how these patterns of sensemaking might affect organizational outcomes. Another exemplar is the research by McCleese, Eby, Scharlau, and Hoffman (2007). These researchers conducted a qualitative study on career plateaus to fill in multiple gaps in the existing literature. They looked at the untested assertion that plateauing is stressful, the specific and immediate coping responses to plateauing, and the differential effects of different types of plateauing. The qualitative study empirically tested hypotheses regarding stress, depression, and coping responses for employees who had reached a hierarchical plateau (i.e., could not move up in the organization); a job content plateau (i.e., mastered the job); and the combination of the two, or someone who is doubly plateaued (i.e., mastered the job and cannot move up in the organization) from a variety of industries. Using content analysis on both open- and closed-ended questions 76

administered through a Web site, they were able to discern that plateaued (hierarchically or job content) employees report more depression than the general population, double plateaued (hierarchically and job content) employees report higher levels of depression than just hierarchically plateaued employees, and hierarchically plateaued employees reported a higher percentage of mental coping strategies than the other types of plateaued employees. Perhaps most intriguing, they uncovered 27 distinct coping strategies used by those interviewed that represent seven coping metathemes of discuss problem, job withdrawal, job involvement, nonwork activities, mental coping, nothing, and side work. In an example (although not necessarily an exemplar) of hypothesis testing, Lee, Mitchell, Wise, and Fireman (1996) tested specific hypotheses deduced from the unfolding model of voluntary turnover (Lee & Mitchell, 1994) on a sample of nurses. In particular, the unfolding model holds that turnover occurs through four distinct prototypical paths. Each path has a specific and unique sequence of variables (e.g., shocks, deliberations, image, image violations, job satisfaction, search, alternatives, eventual quitting). Lee et al. (1996) first asked the nurses to describe their most recent job quit. Second, very specific follow-up questions (often requiring a yes–no response) were asked that described the presence, absence, and sequencing of variables. The nurses’ description of the presence, absence, and sequencing of variables was then compared to the four theorized paths. Falsification occurs if the variables do not confirm the predictions on the basis of the unfolding model’s paths, whereas corroboration occurs if the variables do conform to the model’s paths. COMMON RESEARCH DESIGNS In this section, we present three common qualitative research designs, namely, case study research, ethnography, and in-depth interview studies. It is important to note, however, that “common” refers to the research in I/O psychology. These designs may well be less common to other areas of psychology.

Case Study Research In the qualitative research by organizational psychologists and management researchers, case study

Copyright American Psychological Association. Not for further distribution.

Qualitative Research Strategies in I/O Psychology

research may well be the most common design. First, case study research necessarily seeks to generate, elaborate, or test scientific theories and models. Second, case study research can occur within an in-depth investigation of a single-case situation (e.g., the common “N = 1 situation”) or across indepth investigations of multiple cases (e.g., Lee et al., 1996, adopted an N = 44 design with individual nurses serving as cases). Third, case study research most often seeks to generate testable research propositions. In what has become the classic case study, which many organizational scholars often seek to emulate, Eisenhardt (1989) asked, “How are fast strategic business decisions made?” and “How does decision speed link to performance?” (p. 544). She answered her questions with eight cases studies, with each case representing a microcomputer firm. In the logic of case study research, each case should be seen as an experiment with seven replications. Most important, the case data serve to confirm or disconfirm testable inferences (or hypotheses) drawn from the other cases. Her specific data collection methods included (a) semistructured interviews with each CEO, (b) follow-up interviews with other members of the top management team (TMT), (c) questionnaires completed by TMT members, and (d) secondary published sources (e.g., industry reports, internal documents, information observations of daily behavior, observations of strategy and staff meetings). From these multiple data sources, preliminary analysis resulted in decision stories (i.e., timelines describing how the decisions unfolded). Next, the decision stories were examined for similarities and differences across pairs of cases. Third, similarities and differences across multiple cases were identified and general (and cross-case or replicated) propositions tentatively induced. Fourth, the tentative propositions were then tested on each case. Specifically, the cases were revisited to learn whether the data confirmed the proposed relationships. If not, the proposition was revised. After much iteration, a final list of propositions was reported. In another recent and exemplary case study design, Plowman et al. (2007) studied how church leaders’ decision to offer breakfast to homeless people led to radical change in the church and its environ-

ment. From a dinner conversation regarding alternatives to the traditional church school program on Sunday morning came the idea of serving hot breakfast to homeless people that quickly grew to serving 200 people. A few months after the first breakfast, a physician volunteer opted out of the food serving line and began seeing anyone who wanted to discuss a medical problem. Within a short time, full-scale medical, dental, and eye clinics emerged as part of the Sunday morning program, and within a few years, a 501(c)(3) spin-off of the church was receiving city grants, providing a day center for several thousand homeless people, and serving over 20,000 meals a year. Soon, programs for legal assistance, job training, laundry services, and shower facilities followed. In turn, homeless people began joining the church, singing in the choir, and ushering at the major worship service. Existing theory did not easily explain these changes. As a result, this study followed the Eisenhardt (1989) design and induced six specific extensions to the existing theory on the nature of small, continuous, and radical change. Further, Plowman et al. (2007) triangulated across data from interviews with church leaders, members, employees, and volunteers. In addition, they examined newspaper stories and church documents and reports, and conducted informal observations. The outcome of this study, for example, was a set of six testable research propositions, such as “organizational tension, created by the number and intensity of contextual conditions, encourages the emergence of small change and amplification into radical change” and “the use of symbols accelerates a small change into radical change, given a high level of organizational tension” (p. 538).

Ethnography Ethnography may be the best-known form of qualitative research. It requires that extensive time and effort be expended observing and gathering in-depth information about the research site and its participants. Lee (1999, pp. 89–99) identified four kinds of observational involvement. On the one hand, a researcher can be the “complete observer”: She or he passively remains in the background, observes the actions of others, and records daily field notes of what is 77

Copyright American Psychological Association. Not for further distribution.

Lee, Mitchell, and Harman

observed within its natural context. In other words, the researcher seeks to “blend into the organizational context.” On the other hand, the researcher can be the “complete participant”: He or she becomes a complete but covert organizational member. Specifically, the researcher hides one’s scientific intentions, role, and field note taking and may, in some cases, actually assume a position within the organization. In between these two extremes, the researcher might be a “participant observer,” in which he or she becomes a full organizational member and overtly conducts the study’s data gathering. Alternatively, one researcher might be an “observer participant,” in which she or he participates as a nonmember in the organization and overtly conducts data gathering. In what might be the exemplar of observer participant ethnography, Barley (1990) greatly clarified the process through which new technology affects organizational structure. Specifically, he investigated how new technology changed nonrelational roles among organizational members, which in turn changed relational roles among these people. Next, the changed relational roles rearranged social networks among people, which then served to sustain or modify institutional structures. Methodologically, Barley studied two hospital sites as they adopted new radiology techniques. Each day for 1 year, he spent 6 to 8 hours collecting observer participant data on nonrelational and relational roles. At the end of this observational phase, sociometric surveys assessed social networks to corroborate the observations, whether network structures existed in their anticipated form (i.e., on the basis of the observations) and whether the networks affected the work of the radiology departments. From this substantial amount of data, a remarkably rich and deep understanding of process was reported. A more recent exemplar is Ashcraft’s (2001) article on feminist bureaucracy and what form a hybrid of the feminist and bureaucratic structures takes in an organization that assists survivors of domestic violence. Ashcraft used ethnographic methods to conduct the research, spending over 230 participant observation hours (as a volunteer at the shelter, a trainee and trainer in the volunteer training program, and observing staff meetings) and conducting 60 interviews (approximately 90 minutes each at varying levels of the organization, including staff 78

and volunteers) with the organization. The findings show, for example, rich detail regarding how a feminist bureaucracy embraces strategic incongruity and how it (i.e., the feminist bureaucracy) is determined to maintain interaction between egalitarian and hierarchical processes, with leaders expected to enact egalitarian practices.

In-Depth Interview Studies In-depth interview studies are often conducted because the underlying theory is too complicated to quantify with traditional methods (e.g., Lee et al., 1996), too insufficiently developed (e.g., Loscocco, 1997), or too narrowly interpreted (e.g., Rynes, Bretz, & Gerhart, 1991). In an exemplar of an interviewbased inquiry, Rynes et al. (1991) extended the substantial body of knowledge on the job search process. In particular, they noted that most early interview studies reported substantial effects of recruiters on job choice, whereas more recent studies that adopted cross-sectional survey designs reported less compelling results on the effects of recruiters on job choice outcomes. As such, they suggested the need for designs that have greater sensitivity to dynamic effects, more complicated explanations, and fewer demand characteristics than typical cross-sectional survey research. Thus, they asked college students to describe their job search processes and strategies in their own words and over time. Specifically, Rynes et al. applied a critical incident interview technique to 41 graduating seniors at two periods in time, late January through early February and late March through early May, which allowed for substantial variation in job search behaviors and job offers received. The first set of interviews focused on initial impressions of organizational fit, whereas the second set of interviews focused on later recruitment impressions. From these data, rich, dynamic, and naturally occurring processes were reported. For example, the interview data indicated long delays between recruitment phases (although common) can result in negative inferences by the job seekers (e.g., something is wrong with the company) and in job searchers accepting positions elsewhere. Most notable, the findings of this 1991 in-depth interview study contradicted the survey-based recruitment studies reported in the 1980s.

Copyright American Psychological Association. Not for further distribution.

Qualitative Research Strategies in I/O Psychology

In a more recent, exemplary in-depth interview study, Pratt, Rockmann, and Kaufmann (2006) built theory on how professional identity is created over time. In particular, they noted that the process through which professional identity develops over time is not well understood. As such, they interviewed medical residents and attending physicians in internal medicine, radiology, and surgery over a 6-year period, as well as surveyed and observed these physicians periodically. Following a groundedtheory–like process (described below), the authors created provisional first-order codes, integrated (or reduced) these codes into a more coherent set of categories, and then aggregated the categories into larger theoretical dimensions. In turn, these dimensions were then descriptively applied to the residents in general internal medicine, radiology, and surgery. The outcome of this data collection and analysis, for example, was an inductive model of how work and identity evolve over time to result in a customized professional identity, in other words, how work– identity integrity violations—that is, an incongruence between the identity of the person as a professional and the demands of the job—influence change in the individual. The medical professionals’ identities changed to fit the job; the job did not change to fit them. When violations were minor, the residents experienced identity enhancement. When violations were major, the residences would either “patch” or “splint” their professional identities. Those with a set idea of what their professional identity is were more likely to patch, or add to, their identity. Those without a specific lens through which to interpret their professional identity were more likely to splint, that is, they would adapt a previously held identity in response to a major violation. COMMON ANALYTIC TECHNIQUES In addition to purpose and common research designs, a third way to describe qualitative research is through modes of analytic technique. Typically, however, qualitative researchers prefer not to separate issues of design from those of data analysis (Denzin & Lincoln, 1998a, 1998b, 1998c). Instead, the preference is to describe design and analysis together because a hallmark of qualitative research is

its flexibility to respond to the demands of field settings as issues, contexts, and organizational members change over time. Nonetheless, we opt to separate design and analysis for purposes of discussion. Although space constrains us to the three techniques described below, it should be noted that there are other analytical techniques as well (e.g., analysis of documents, focus groups, hermeneutics, protocol analysis, tracer studies; for excellent general references on qualitative design and analyses, see Cassell & Symon, 1995; Denzin & Lincoln, 1998a; Miles & Huberman, 1994).

Grounded Theory In organizational psychology and management research, Glaser and Strauss’s (1967) general method of grounded theory may well be the most commonly applied technique, although Strauss and Corbin (1998, pp. 159–162) offered updated modifications to the original idea. Regardless of the original or revised version, the goal of grounded theory is to move toward a data-based and inductive new theory and a deep understanding of (organizational) phenomena. On a regular and iterative basis, grounded theory explicitly requires (a) generation of hypotheses from data, (b) testing of these hypotheses on new data that may generate revisions to the original hypotheses, and (c) testing these revised hypotheses on still more new data. Further, this data-driven and inductive process may occur from a few weeks to several years, and learning and theory creation is a constant and ongoing activity. In a strict sense, the process should end when theoretical saturation is reached or when no new learning occurs through further hypothesis generation, testing, and revision. Needless to say, grounded theory, as originally described, presented a demanding analytic procedure. In actual application, however, Lee (1999) suggested caution in a strict interpretation of the label. In particular, he noted that many authors purport to apply grounded theory, but few studies offer sufficient explanation for a reader to be certain that the ideal form of grounded theory is indeed achieved or nearly achieved. Lee, Mitchell, and Sablynski (1999) suggested that this ambiguity is due to (a) overly terse descriptions, (b) overemphasis on one aspect of grounded theory to the exclusion or near exclusion of 79

Copyright American Psychological Association. Not for further distribution.

Lee, Mitchell, and Harman

other aspects, and (c) authors’ insufficient familiarity with the technique. Further, Locke (1996) offered a harsher evaluation and interpreted these ambiguous descriptions as suggesting grounded theory is not correctly applied. In rereading much of the qualitative research in organizational psychology and management research, it has been, and remains, our view that very few of the identified studies meet the complete or pure spirit of grounded theory. With that said, we judge that many studies are instead grounded-theory–like and offer very meaningful contributions to theory and research in organizations. Two studies in particular are exemplars. Sutton (1991) investigated the effects of conflicts between the expressed emotions of bill collectors (which derive from professional norms), their true feelings, and the expressions from debtors. In applying grounded theory, Sutton (a) had extensive interactions with key informants, (b) trained as a bill collector, (c) worked as a bill collector for 20 hours, (d) held focus groups with bill collectors, (e) interviewed supervisors, (f) observed other bill collectors, and (g) examined printed materials. Further, Sutton stated, “I developed hunches about these norms, compared these ideas to new data from the site, and then used the new data to help decide whether to retain, revise, or discard these inferences” (p. 250). In other words, he followed the logic of grounded theory and produced, for example, a very rich and deep description and explanation of debt collectors’ behaviors and inner emotions. In a setting in which hundreds of thousands, if not millions, of dollars can result from a 20-minute pitch of a screenplay to a producer, Elsbach and Kramer (2003) studied Hollywood pitch meetings, More important, they extended researchers’ knowledge on creativity and developed a “dual-process model of creativity judgments.” In applying grounded theory, Elsbach and Kramer (a) interviewed a total of 36 writers, agents, and producers; (b) observed 28 actual pitches; (c) attended screenwriting classes; and (d) examined archival materials. Further, they applied a four-stage design. In Stage 1, they iteratively examined their text data for strong inferences on creative pitches and identified 15 cues for creativity and 4 cues for uncreativity. In Stage 2, the 19 cues were 80

converted to seven more general prototypes. In Stage 3, the data were revised with a focus on the relationship between the “pitcher and catcher”; the specific outcome of this third analysis was to identify the dynamic back-and-forth nature of the process. In Stage 4, another set of 14 informants (i.e., writers, agents, and producers) were interviewed to allow a cross-check on the emerging inferences (or another view of the data). As a final validity check, 4 independent informants (a former studio head, an agent, a writer, and a writer-director) offered global judgments on the explanatory adequacy of the induced model. The outcome of these data analyses—involving multiple stages and sources of input—was, for example, a theoretical model that explains the creative process during high-intensity meetings.

Pattern Matching Although grounded theory may be the most common analytic technique in general and with case study research in particular, pattern matching is not an uncommon analytical strategy as well. It can be applied to tests of formal hypotheses, an explicit theory, or a less formal conceptual structure. In its essence, the researcher anticipates or predicts a particular pattern of variables, phenomena, or outcomes. This pattern can be deduced from a formal theory or induced from preliminary observations; it can be static, dynamic, or contingent; and it can be simple or complicated. Within one or multiple cases, however, this pattern is assessed or measured against actual data with the intent of serving as a benchmark for interpreting case data or for falsifying or corroborating formal theories. In an example of pattern matching (which was mentioned earlier), Lee et al. (1996) tested their unfolding model of voluntary turnover. According to the theory, employees quit by means of four prototypical processes. Most important, each process is defined by a unique set of variables that occur at specific times. As operationalized, these variables are defined to occur as a sequence of binary events (i.e., it must happen or it must not happen at a particular point in a sequence of other binary variables). Theoretically, each quit event should be classifiable into one and only one of these processes. From semistructured interviews designed to elicit how this

Copyright American Psychological Association. Not for further distribution.

Qualitative Research Strategies in I/O Psychology

process unfolds, the quittings of 44 nurses were classed into these prototypes. One indication of corroboration or falsification of the unfolding model is the simple number of correct and incorrect classifications. The outcome of this study was the first empirical corroboration of their theory. Three years later, Lee, Mitchell, Holtom, McDaniel, and Hill (1999) reported a survey-based test of their model. In an exemplar of pattern matching, Maguire, Hardy, and Lawrence (2004) adopted a modified form of Eisenhardt’s (1989) case research design. Specifically, they studied how the concept of “institutional entrepreneurship” helped explain the evolution of HIV/AIDS treatments in Canada. From their extensive interview data, they conducted a four-stage study. First, they developed a basic timeline based on narrative accounts of the treatment’s evolution. Second, Maguire et al. juxtaposed changes in medical practices and information exchanges onto the time line. Third, they identified the key entrepreneurs who were judged responsible for these changes. Finally, Maguire et al. applied pattern matching to determine and understand the processes through which these innovative changes occurred. The outcome of their study was, for example, a very rich explanation of a timely social issue and expanded understanding of institutional entrepreneurship.

Hybrid Studies In our experience, much of the tension between qualitative and quantitative research derives from insufficient information about how qualitative researchers go about their work. Fortunately, increased researcher comfort, acceptance, and application of qualitative methods are beginning to be seen. In particular, Lee, Mitchell, and Sablynski (1999) recommended that researchers consider a “blended design,” in which advantages from inductive qualitative studies and deductive quantitative studies are combined into one publication. In an exemplary investigation on the effects of affect on creativity using a blended design, Amabile, Barsade, Mueller, and Staw (2005) combined qualitative narratives with quantitative survey data. More specifically, 222 individuals from 26 project teams completed daily diaries. In addition to specific closedended questions, two open-ended items resulted in

narrative descriptions of creative processes. Through content analysis, specific measures of creativity and affect were derived. (Participants also completed monthly peer-rating questionnaires.) In turn, the qualitative and quantitative data were analyzed with regression methods that included lagged variables. The outcome of the data and its analysis was delineation of the organizational affect–creativity cycle. As another example, Felps et al. (2009) proposed, tested, and corroborated the effect of coworkers’ job embeddedness on an individual’s voluntary turnover by means of a three-step process. First, a cross-level effect of coworkers’ job embeddedness on individual turnover was established in Sample 1. (Coworker and individual job satisfaction, organizational commitment, and perceived alternatives were controlled.) Second, qualitative focus group data that were content analyzed with text-based software provided preliminary evidence that this cross-level effect was mediated through others’ job search behavior. Third and in an independent sample, coworkers’ job search (as measured with a questionnaire) fully mediated the effect of coworkers’ job embeddedness on individual turnover. (Similar to the first step, coworker and individual job satisfaction, organizational commitment and perceived alternatives, and individual job search were controlled.) The outcome of this study is a new area of inquiry for turnover researchers and those scholars interested in contagion processes across levels. In our judgment, such hybrid designs hold substantial promise for enhanced cumulative research through closer examination of dynamic relationships and process variables. Looking to the future, we expect that such blended designs will become more common in organizational psychology and management research. (For additional applications on hybrid designs, see Currall, Hammer, Baggett, & Doniger, 1999, on corporate boards of directors; Yauch & Seudel, 2003, on the work climate and organizational effectiveness.) CONCLUDING NOTE We hope to have provided some insight into the reasons behind engaging in qualitative research, how it is currently being undertaken, and what the benefits are 81

Copyright American Psychological Association. Not for further distribution.

Lee, Mitchell, and Harman

for understanding the phenomena under examination using these methods. In particular, Aguinis, Pierce, Bosco, and Muslin (2009) recommended that qualitative methods such as case studies, natural field observations (e.g., participant observer methods), and interviews should be included more routinely in doctoral student education. At this point, we mention some of the challenges that the qualitative researcher faces. First, it is very time intensive. As scholars, we have multiple demands on our time, and qualitative research may simply require too much time, especially in the early years of one’s career. Second (and related to the first point) it may be professionally risky. The time commitment most likely will slow any other research progress resulting in lower research quantity. Because qualitative methods remain less well known in I/O psychology in the United States, getting qualitative research published can be quite challenging (Zedeck, 2003). Third, some organizational scholars claim we have too much theory (see, e.g., the editorial forum in December 2007 issue of the Academy of Management Journal on this topic). As mentioned earlier, the vast majority of qualitative research generates or elaborates theory. More theory testing through qualitative means, however, might help counter this issue. Finally, although we have tried to use examples directly relevant for I/O psychologists (e.g. job search, turnover, careers, diversity), it is clear to us that many topics for I/O psychology, such as training or performance appraisal, have been examined mostly with quantitative procedures. Thus, although the above challenges are not insurmountable, they are nevertheless present and require researchers’ attention. Qualitative research provides a different and enriching window for observing behavioral phenomenon and can be invaluable for providing a different perspective on topics that are in need of some renovation and creative new thinking.

References

Ashcraft, K. L. (2001). Organized dissonance: Feminist bureaucracy as hybrid form. Academy of Management Journal, 44, 1301–1322. Barley, S. R. (1990). The alignment of technology and structure through roles and networks. Administrative Science Quarterly, 35, 61–103. Cassell, C., & Symon, G. (1995). Qualitative methods in organizational research. Thousand Oaks, CA: Sage. Currall, S. C., Hammer, T. H., Baggett, L. S., & Doniger, G. M. (1999). Combining qualitative and quantitative methodologies to study group processes: An illustrious study. Organizational Research Methods, 2, 5–36. Denzin, N. K., & Lincoln, Y. S. (1998a). Collecting and interpreting qualitative materials. Thousand Oaks, CA: Sage. Denzin, N. K., & Lincoln, Y. S. (1998b). The landscape of qualitative research: Theories and issues. Thousand Oaks, CA: Sage. Denzin, N. K., & Lincoln, Y. S. (1998c). Strategies of qualitative inquiry. Thousand Oaks, CA: Sage. Eisenhardt, K. M. (1989). Making fast strategic decisions in high-velocity environments. Academy of Management Journal, 32, 543–576. Elsbach, K. D., & Kramer, R. M. (2003). Assessing creativity in Hollywood pitch meetings: Evidence for a dual-process model of creativity judgments. Academy of Management Journal, 46, 283–301. Felps, W., Hekman, D. R., Mitchell, T. R., Lee, T. W., Harman, W. S., & Holtom, B. C. (2009). Turnover contagion: How coworkers’ job embeddedness and coworkers’ job search behaviors influence quitting. Academy of Management Journal Gersick, C. J. G. (1989). Marking time: Predictable transitions in task groups. Academy of Management Journal, 32, 274–309. Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory: Strategies for qualitative research. Chicago: Aldine. Hackman, J. R., & Oldham, G. R. (1976). Motivation through design of work—Test of a theory. Organizational Behavior and Human Performance, 16, 250–279. Hardy, C., Phillips, N., & Lawrence, T. B. (2003). Resources, knowledge, and influence: The organizational effects of interorganizational collaboration. Journal of Management Studies, 40, 321–347.

Aguinis, H., Pierce, C. A, Bosco, F. A., & Muslin, I. S. (2009). First decade of Organizational Research Methods trends in design, measurement, and dataanalysis topics. Organizational Research Methods, 12, 69–112.

Harman, W. S., Lee, T. W., & Mitchell, T. R. (2009). Comparing U.S. and European application of qualitative methods. Manuscript in preparation, University of Washington.

Amabile, T. M., Barsade, S. G., Mueller, J. S., & Staw, B. M. (2005). Affect and creativity at work. Administrative Science Quarterly, 50, 367–403.

Johns, G. (2006). The essential impact of context on organizational behavior. Academy of Management Review, 31, 386–408.

82

Qualitative Research Strategies in I/O Psychology

Lee, T. W. (1999). Using qualitative methods in organizational research. Thousand Oaks, CA: Sage.

sion, and coping responses. Journal of Vocational Behavior, 71, 282–299.

Lee, T. W., & Mitchell, T. R. (1994). An alternative approach: The unfolding model of voluntary employee turnover. Academy of Management Review, 19, 51–89.

Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis (2nd ed.). Thousand Oaks, CA: Sage.

Copyright American Psychological Association. Not for further distribution.

Lee, T. W., Mitchell, T. R., Holtom, B. C., McDaniel, L. S., & Hill, J. W. (1999). The unfolding model of voluntary turnover: A replication and extension. Academy of Management Journal, 42, 450–462.

Plowman, D. A., Baker, L. T., Beck, T. E., Kulkarni, M., Solansky, S. T., & Travis, D. V. (2007). Radical change accidentally: The emergence and amplification of small change. Academy of Management Journal, 50, 515–543.

Lee, T. W., Mitchell, T. R., & Sablynski, C. J. (1999). Qualitative research in organizational and vocational psychology, 1979–1999. Journal of Vocational Behavior, 55, 161–187.

Pratt, M. G., Rockmann, K. W., & Kaufmann, J. B. (2006). Constructing professional identity: The role of work and identity learning cycles in the customization of identity among medical residents. Academy of Management Journal, 49, 235–262.

Lee, T. W., Mitchell, T. R., Wise, L., & Fireman, S. (1996). An unfolding model of voluntary employee turnover. Academy of Management Journal, 39, 5–36.

Rynes, S. L., Bretz, R. D., Jr., & Gerhart, B. (1991). The importance of recruitment in job choice: A different way of looking. Personnel Psychology, 44, 487–521.

Locke, K. (1996). Rewriting The Discovery of Grounded Theory after 25 years. Journal of Management Inquiry, 5, 239–245.

Strauss, A., & Corbin, J. (1998). Grounded theory methodology: An overview. In N. K. Denzin & Y. S. Lincoln (Eds.), Strategies of qualitative inquiry (pp. 158–183). Thousand Oaks, CA: Sage.

Loscocco, K. A. (1997). Work–family linkages among self-employed women and men. Journal of Vocational Behavior, 50, 204–226. Maguire, S., Hardy, C., & Lawrence, T. B. (2004). Institutional entrepreneurship in emerging fields: HIV/AIDS treatment advocacy in Canada. Academy of Management Journal, 47, 657–679. Maitlis, S. (2005). The social processes of organizational sensemaking. Academy of Management Journal, 48, 21–49. McCleese, C. S., Eby, L. T., Scharlau, E. A., & Hoffman, B. H. (2007). Hierarchical, job content, and double plateaus: A mixed-method study of stress, depres-

Sutton, R. I. (1991). Maintaining norms about expressed emotions: The case of bill collectors. Administrative Science Quarterly, 36, 245–268. Van Maanen, J. (1979). Reclaiming qualitative methods for organizational research: A preface. Administrative Science Quarterly, 24, 520–526. Yauch, C. A., & Steudel, H. J. (2003). Complementary use of qualitative and quantitative cultural assessment methods. Organizational Research Methods, 6, 465–481. Zedeck, S. (2003). Editorial. Journal of Applied Psychology, 88, 3–5.

83

CHAPTER 4

ADVANCES IN ANALYTICAL STRATEGIES

Copyright American Psychological Association. Not for further distribution.

David Chan

In industrial and organizational (I/O) psychology, we continuously make inferences about work behaviors and recommendations for organizational interventions based on findings from empirical studies. One of the major components in the interpretation of findings is a sense-making process in which the researcher selects and applies an inferential analytical technique to the data for the purpose of testing hypotheses or making some substantive inferences. Clearly, in addition to good theory, design, and measurement, the adequacy of the hypothesis testing and substantive inferential process is directly affected by the appropriateness of the choice and application of the analytical technique for the specific data set in question. In recent years, important theoretical advances have been made in I/O psychology. Examples include the specification of mediating mechanisms that connect predictors and criteria (e.g., Barrick, Stewart, & Piotrowski, 2002), expansion of the job criterion space (e.g., Borman & Motowidlo, 1993; Campbell, McCloy, Oppler, & Sager, 1993; Sackett, 2002), conceptualizations of dynamic criteria and various facets of changes over time (e.g., Chan & Schmitt, 2000; Hofmann, Jacobs, & Baratta, 1993), person– environment fit (e.g., Edwards, 1994; Kristof, 1996), and multilevel phenomena in organizations (e.g., Chan, 1998a; Kozlowski & Klein, 2000; Morgeson & Hofmann, 1999). These theoretical advances raise important empirical questions that need to be answered by adequate data analyses. As explained in the following sections, traditional analytical tech-

niques are often limited in their adequacy in addressing these empirical questions. The purpose of this chapter is to summarize and discuss recent advances in analytical strategies which, when appropriately applied, will provide more adequate approaches to ensure adequate linkages that connect substantive inferences from data with theory, construct, measurement, and design. Note, however, that in some situations, using more sophisticated techniques is not necessarily the best approach, especially when statistical assumptions are violated, psychometric properties are poor, or adequate theory is lacking. In any empirical research, the relevant theory, design, measurement properties, nature of data, and types of intended inferences should drive the selection of analytical strategy. In addition, poor data obtained through inadequate theory, design, or measurement will not become worthier of analysis by increasing the sophistication level of the analytical strategies. This chapter is organized into four sections. The first section discusses advances in analytical strategies for modeling relationships between constructs, specifically on relationships that go beyond the bivariate prediction paradigm. Strategies that will be discussed include mediation analyses, interaction analyses, combination of interactions and mediations, and structural equation modeling (SEM). The second section discusses issues in modeling multilevel phenomena in which the data set has a nested hierarchical structure. First, the problems of singlelevel analytical approaches to multilevel data are

http://dx.doi.org/10.1037/12169-004 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

85

Copyright American Psychological Association. Not for further distribution.

David Chan

explicated. Next, the critical conceptual issues that underlie multilevel modeling are discussed in terms of composition models. Finally, multilevel analyses, including hierarchical linear modeling and multilevel SEM, are described. The third section discusses advances in strategies for modeling changes over time. Fundamental questions on changes over time are explicated, and limitations of traditional techniques for analyzing changes over time are illustrated. Latent variable approaches to modeling changes over time, as well as the multilevel nature of longitudinal data, are discussed. The fourth section discusses strategies for modeling measurement invariance across groups. The fundamental issue of measurement invariance and differential item functioning is explained first, followed by a discussion of itemresponse theory techniques and multiple-group means and covariance structure analysis. The chapter ends with an appendix on data analysis software and internet resources on selected analytical strategies. MODELING RELATIONSHIPS BETWEEN CONSTRUCTS In the early years of I/O psychology, especially in the area of personnel selection research, some empirical studies proceeded on a bivariate prediction paradigm that is focused on describing and comparing different predictors (e.g., selection procedures). The primary objective was searching for predictors that maximize prediction of the criterion (e.g., job performance). Examples of typical research questions were “Can personality tests predict job performance?” and “Do cognitive ability tests or personality tests predict job performance better?” Such questions are descriptive and directly comparative in nature, and traditionally they have been addressed by using basic analytical strategies such as correlation and regression techniques. A correlation coefficient indexes the direction and magnitude of the bivariate relationship between a personality test and job performance. A multiple regression analysis may be conducted to compare the relative validity of an ability test with that of a personality test in predicting the job performance criterion. When researchers examined only bivariate relationships in their studies, they probably had constructs and theoretical relationships in mind when 86

they included the predictors and criteria. However, the absence of a clear theoretical model that linked predictors and criteria plus the emphasis to empirically establish the predictor-criterion bivariate relationship resulted in the impression that such studies were typically lacking in theoretical bases and the efforts were largely driven by the aim to search for the optimal set of predictors that offered the highest validity in predicting the criterion. Such descriptive and comparative research may be consistent with the applied goal of maximizing prediction, but they have low-explanatory value and therefore impede scientific progress in I/O psychology. Theoretical advances in the past few decades saw significant shifts from the bivariate prediction paradigm to include three or more study variables focusing on the nature of the constructs represented by the variables and the relationships between constructs. This shift is characterized by an approach that emphasizes theoretical understanding of the phenomena being studied. Accordingly, descriptive research questions have been replaced by more construct-oriented and explanatory ones such as “If there is a causal relationship between the predictor and the criterion, what is the variable that transmits the effect of the predictor on to the criterion?” and “Is the nature of the predictor-criterion relationship, in terms of direction and/or magnitude, dependent on the values of a second predictor?” The first of these two example theory-driven questions may be addressed by mediation analyses, and the second may be addressed by interaction/moderation analyses. The two types of analyses may also be combined into an integrated analytical strategy to address more complex questions pertaining to whether the direction and/or strength of a mediation is dependent on a variable not in the mediation chain (i.e., moderated mediation) and what variable is transmitting the effect of an interaction on to the criterion (i.e., mediated moderation/interaction). Even more complex relationships between constructs represented by the relationships among the observed variables may be modeled using SEM. The following subsections describe the advances made in these analytical strategies and how they may be used to address theoretical questions on interconstructs relationships, many of which have important practical implications.

Advances in Analytical Strategies

Copyright American Psychological Association. Not for further distribution.

Mediation Analyses This section discusses mediation analysis as a statistical technique to test a hypothesized causal chain. Throughout the discussion, it is important to remind ourselves that causality is a function of theory and design and the mediation analysis or any other analytical strategies are statistical techniques that help us empirically “infer” and support the hypothesized causal relationship. Analytical strategies per se do not conclusively demonstrate causality. In a mediation relationship, the mediating variable M (e.g., work motivation) is an intervening variable that transmits the effect of an antecedent variable, X (e.g., job satisfaction), on to a dependent variable, Y (e.g., job performance). Establishing a mediation in the form of the X → M → Y causal chain (e.g., satisfaction → motivation → performance) is important because it provides a better understanding of the relationships among the constructs represented by the variables X, M, and Y. In addition to the scientific value of gaining a better understanding of the fundamental process by which X affects Y (i.e., indirectly through M), a mediation analysis offers applied value in terms of potentially enabling us to develop more effective interventions to affect Y by focusing on the mediating variable M, which is a more proximal cause of Y than the more distal cause X. This is especially relevant when X is not readily malleable. In the above example, if X represents conscientiousness instead of satisfaction, then the practical value of the mediation relationship is clear because conscientiousness, unlike satisfaction, is not readily malleable, but motivation is and it is possible to devise interventions to increase motivation, which in turn should increase performance. Despite the apparent conceptual clarity of the mediation relationship, there has been at least 2 decades of debate in regard to statistical issues in the analytical approaches used to test a mediation. Much of the debate is focused on the causal-steps approach as described in Baron and Kenny (1986) and Judd and Kenny (1981), which is the most popular approach used to test a mediation. This approach, which uses either partial correlation or multiple regression to test the mediation, involves performing separate significant tests on the overall relation between X and Y, the relation between X and M, and

the relation between M and Y controlled for X, and comparing these relations. According to this approach, if one or more of the three zero-order (i.e., bivariate) relations, that is, between X and Y, between X and M, and between M and Y, are nonsignificant, then one should conclude that the hypothesized mediation X → M → Y is not supported. If all three zero-order relations are significant, then one should proceed to test the significance of (a) the relation between M and Y controlling for X and (b) the relation between X and Y controlling for M. The mediation X → M → Y is supported if (a) is significant and (b) is nonsignificant. There are three major problems with this causal-steps approach. First, simulation studies (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002) have demonstrated that the approach has very low ability to detect mediated effects. Second, it does not provide a numerical estimate of the strength of the mediation, that is, the indirect effect from X to Y through M. Without an estimate of the strength of the indirect effect, the approach does not appear to meet the statistical aim of performing a mediation analysis. That is, by focusing on statistical significance testing only without estimating effect size (i.e., magnitude) of the indirect effect, the approach indicates the presence or absence of a mediation relationship without providing an estimate of the strength of the mediation relationship. Third, the approach requires that the overall relation between X and Y is significant before proceeding with subsequent steps in the test. In other words, if the overall X–Y relation is not significant, then the approach cannot proceed to the next step. Intuitively, this may seem to make good sense because one is forced to conclude that a mediation does not exist or is meaningless because there is no effect of X to be transmitted on to Y in the first place. However, this requirement is in fact flawed because a mediation in the sense of an X → M → Y causal chain can surely exist in the absence of a significant overall X–Y relation. Note that the test of the overall X–Y relation tends to have less statistical power than the test of the X → M relation and the test of the M → Y relation. This is in fact a common occurrence because the indirect effect is almost always smaller (and rarely equal but never higher) than each of the two direct effects (X → M, M → Y). The requirement of 87

Copyright American Psychological Association. Not for further distribution.

David Chan

a significant overall X–Y relation has failed to make the distinction between the sample estimate and the population parameter. To put it simply, there can be a true X → M → Y causal chain in the population, but the sample estimate of the overall X–Y relation (i.e., the indirect effect of X on Y through M) is nonsignificant because of low statistical power or a small effect size. Note that when a true causal chain exists, the indirect effect, which is computed by taking the product of the X → M direct effect and the M → Y direct effect, has to be nonzero in the population, because neither of the two direct effects is nonzero. Path analysis (or SEM if measurement errors of the observed variables are also estimated), addresses the problems of the causal-steps approach by fitting the causal-chain path model to the data and estimating the magnitude of the direct effects from X to M and from M to Y, as well as the indirect effect from X to Y through M (see MacKinnon, 2008). Path analysis provides a numerical estimate of the indirect effect, which is simply the product of the two direct effects in the causal-chain path model. Although the statistical significance of this indirect effect may be computed, the indirect effect clearly exists insofar as it is a nonzero parameter and there is no requirement for the overall relation between X and Y to be significant to proceed with testing the mediation. A mediation relation can occur only if there are in fact causal effects to be transmitted from the antecedent on to the dependent variable. Therefore, in all statistical mediation analysis, there is an assumption that the relations that link the variables are causal in nature (hence a causal chain) and also an assumption that the causal ordering of the variables is correctly specified in the mediation analysis. Note that the viability of these causal assumptions is dependent on the adequacy of the theory and nature of the study design, and strong inferences of causality require manipulation of variables in experimental design studies. Methodological and statistical issues on mediation analysis continue to be debated in publications and conferences. For a recent overview and debate of the unresolved issues, see the series of articles in the feature topic of a special issue of Organizational Research Methods (Mathieu, DeShon, & Bergh, 2008). An interesting overview of the truths and 88

myths associated with the tests of mediations is provided by LeBreton, Wu, and Bing (2009).

Interaction Analyses A statistical interaction is said to occur when the direction or magnitude of the relation between X (e.g., cognitive ability) and Y (e.g., task performance) is dependent on a third variable Z (e.g., conscientiousness). In other words, the relationship between X and Y varies according to Z. When an X × Z interaction effect on Y exists, the failure to take into account the interaction leads to an incomplete interpretation or even a misinterpretation of the X–Y relationship. When X and Z are categorical variables with two or more levels and Y is a continuous variable, the well-known statistical model called analysis of variance (ANOVA) is the most widely used analytical strategy to test the X × Z interaction effect. When either X or Z is continuous, the multiple regression model is the most widely used strategy. This is because multiple regression allows for X and Z to be either categorical or continuous. Although ANOVA and multiple regression analysis differ in various statistical features and they were historically developed from different study design traditions, both techniques are in fact specific instances (where there is only one dependent variable in the model) of the more general statistical model called the general linear model, which allows for linear transformations and linear combinations of multiple dependent variables. The general linear model is the statistical foundation of most of the multivariate analytical strategies, such as canonical correlation, factor analysis, cluster analysis, and multidimensional scaling. The general linear model is not to be confused with generalized liner model, which was formulated to generalize the ordinary regression model to include other statistical models such as logistic regression and Poisson regression. Unlike experimental design studies in social or cognitive psychology in which the X and Z are categorical variables that are independently manipulated into discrete levels (conditions) to test their effects on the dependent variable Y, the X and Z variables in I/O psychology could be either categorical (e.g., sex) or continuous (e.g., cognitive ability) variables.

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

Hence, interaction effects in I/O psychology are most commonly tested by using hierarchical (moderated) multiple regression because, unlike ANOVA, which requires the interaction to be a joint effect of categorical variables, multiple regression allows the interaction to be a joint effect of predictors in which any of the predictors may be either categorical or continuous. When the criterion (Y) is a categorical variable (e.g., promoted vs. not promoted), a hierarchical logistic regression can be performed to test the interaction effect. Interaction analyses allow us to go beyond the simple bivariate prediction paradigm to test interesting hypotheses about the X–Y relation by assessing how the strength of the X–Y relation is moderated by Z (e.g., understanding that the magnitude of a positive relation between X and Y increases as Z increases) or how a direct interpretation of the zeroorder X–Y relation is meaningless or misleading when Z is not considered (e.g., a disordinal interaction showing that X is positively related to Y when Z is low but negatively related to Y when Z is high). Disordinal interactions are especially important because when they exist the interpretation of the main effects is often misleading or meaningless. For example, in a study demonstrating that proactive personality (X) may be either adaptive or maladaptive, Chan (2006) performed separate interaction analyses, which showed that employees’ proactive personality positively predicted several desirable work-relevant criteria (Ys, e.g., satisfaction, job performance) when their situational judgment ability (Z) is high but negatively predicted the same outcomes when Z is low. The findings on the disordinal interaction effects of proactive personality and situational judgment ability on work-relevant criteria show that a highly proactive personality may be either adaptive or maladaptive depending on the individual’s level of situational judgment ability and hence caution against direct interpretations of bivariate associations between proactive personality and work-relevant criteria. The standard two-step procedure for performing an interaction analysis by using hierarchical regression involves first entering the X and Z predictor variables as a single block in Step 1 of the regression of Y (the criterion) and then next entering the X × Z product term as a single predictor variable in Step 2. The

aim is to examine whether entering this interaction term will result in a significant and substantial increase in proportion of criterion variance accounted for, reflecting the presence and magnitude of the X × Z interaction effect on Y. If significant, the interaction can be plotted to explicate the nature of the interaction pattern. Details of performing interaction analysis by using hierarchical regression, as well as the standard issues of statistical assumptions, multicollinearity, shrinkage of the R 2, and regression diagnostics, are readily available in statistical textbooks or chapters (e.g., Aiken & West, 1991) and beyond the scope of this chapter.

Combining Mediations and Interactions Mediations and interactions may be combined into one integrated analysis to better understand or elaborate a mediation effect or an interaction effect. Baron and Kenny (1986) distinguished between moderated mediation analysis and mediated moderation analysis. The former is to elaborate on a mediation effect and the latter is to elaborate on a moderation (i.e., interaction) effect. A moderated mediation analysis can be performed to examine whether the magnitude of a mediation effect, in terms of the indirect effect from X to Y through M in the causal chain X → M → Y, is dependent on the values or level of a moderator (i.e., interacting) variable Z. In moderated mediation analysis, the moderator Z is often a categorical variable such as group membership in which the goal is to examine how the indirect effect of X through M to Y in the same causal chain may differ in strength or even the sign (positive or negative effect) of the X → M effect or M → Y effect across groups. In these situations, the purpose of performing the moderated mediation analysis is either to test the generalizability of the mediation effect across different groups or test a specific hypothesis of group difference in the strength or direction (sign) of the mediation effect according to some theory of substantive group differences. For example, we could test the moderated mediation hypothesis that for a given task, the strength of the mediation in the causal chain conscientiousness (X) → motivation (M) → task performance (Y) is moderated by sex (Z) such that the mediation effect is stronger among males than among 89

Copyright American Psychological Association. Not for further distribution.

David Chan

females. We could further test specific hypotheses that the stronger mediation effect is due to a larger magnitude of the X → M and/or the M → Y relation among males than among females. Note that the grouping variable Z may be a naturally occurring group membership variable (e.g., sex, country) or a manipulated independent variable (i.e., the groups are conditions in an experimental design). In addition, Z may be a continuous variable that has been categorized to create two or more groups. For example, intelligence, a continuous variable, may be categorized to create three groups corresponding to low, moderate, and high intelligence. In these grouping situations, the cutoff scores used to create the groups are critical in that they should produce meaningfully different group memberships. If individuals from truly different groups were wrongly categorized into the same group or individuals from truly the same group were wrongly categorized into different groups, then the results obtained may mask true moderated mediation effects or produce spurious moderated mediation effects. A mediated moderation (or synonymously, a mediated interaction) analysis can be performed to test whether an interaction effect (X × Z interaction) on a criterion/dependent variable (Y) is transmitted through a mediator M such that M is both the proximate effect of the interaction between X and Z and the proximate cause of the dependent variable Y. It is interesting to note that although a mediated interaction/moderation is an interaction in the sense that we are explaining how the X × Z interaction effect on Y occurs (i.e., through M), the explanation is given in terms of ordering the effects into a causal chain represented a mediation relation, in which the causal effect of the antecedent variable represented by the statistical interaction term (X × Z interaction) is transmitted on to the dependent variable Y through the mediator M. Given the potential theoretical and practical value of analyses that combine mediations and moderations into an integrated analysis, it is somewhat surprising that very few studies in I/O psychology have used moderated mediation or mediated moderation analyses to elaborate mediation or moderation/interaction effects, respectively. Future researchers should consider how the substantive processes that underlie the 90

relationships between constructs derived from theory may be more adequately tested by combining mediation and moderation into an integrated analysis. A study of race differences in situational judgment test performance published more than a decade ago (Chan and Schmitt, 1997) incidentally provides an illustration of how a mediated moderation analysis may be used to test substantive hypotheses that advanced selection research in important ways through conceptual, methodological, and practical contributions. In this study, Chan and Schmitt (1997) started by making the distinction between the format of testing used by a given selection test (i.e., the specific test method) and the content of the test (i.e., the intended test constructs). Using a theory of cultural differences between ethnic groups associated with experiences of and reactions to reading requirements and social interactions, the authors argued that the test method can have differential effects on test scores for Black versus White Americans and can therefore affect the construct validity of the test scores. They then compared the test scores of Black and White Americans on a video-based situational judgment test with a written paper-and-pencil version (i.e., two different methods of testing) of the same test. Consistent with the theory, they found that Black–White group differences on the video-based version of this test were substantially less than differences on the written version (i.e., difference of one fifth of an SD unit vs. one SD unit, favoring the White group). In other words, they established an Ethnicity × Test Method interaction effect on test performance. To provide a more rigorous test of the theory, Chan and Schmitt (1997) also measured the perceptions of the face validity of these instruments and the reading comprehension levels of the participants. Face validity and reading comprehension were hypothesized as the mediators that transmit the effect of the Ethnicity × Test Method interaction on to the test performance variable. The authors conducted a mediated moderation analysis by using hierarchical moderated regression and showed that the magnitude of the Ethnicity × Test Method interaction effect (X × Z) on test performance (Y) that was observed in the study was reduced substantively when the two mediators (Ms: face validity, reading comprehension) were statistically controlled for in the moderated

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

regression. In other words, they found that part of the Black–White differences in test scores across the two different methods of test presentation were due to Black–White differences in reading comprehension and reactions to the tests. Now, if the intent is to measure constructs other than reading comprehension, for example, the correlations with reading comprehension would be evidence that the test is contaminated. Even if reading comprehension were an important part of job performance, it would be best to measure it directly with another test and keep the situational judgment test, which is usually intended to measure other constructs, as free of the influence of reading comprehension as possible. This would allow for better interpretation of predictorcriterion relationships and the proper application of any weighting schemes used in combining various elements of a selection test battery. The theory-driven study design in Chan and Schmitt (1997) was a relatively complex quasiexperiment that manipulated test method and coded ethnicity as factors of an experiment, measured facevalidity perceptions and reading comprehension as mediators, and measured test performance as the dependent variable. The hypotheses derived from the authors’ theory and tested in their quasi-experimental study design were able to be directly addressed by conducting an appropriate mediated moderation analysis. For a summary of the technical statistical issues involved in testing moderated mediation and mediated moderation, including recent advances in performing these analyses, it is recommended that the reader see Edwards and Lambert (2007) and MacKinnon (2008). Given the interest in moderating effects of demographic and other categorical variables in I/O psychology, the reader may also refer to Aguinis (2004), which provides a comprehensive summary of the issues involved in regression analysis for categorical moderators including fully worked-out examples of how to conduct and interpret the moderator analyses.

Structural Equation Modeling In many research situations in I/O psychology, the nature of the study variables (e.g., personality traits) and practical constraints such as administrative,

social, political, legal, and ethical issues prevent us from conducting true experiments or even quasiexperiments to empirically establish causal relations from the study data. Very often, our studies are designed, either by constraint or choice, to collect and analyze data on measured variables rather than those that result from manipulation of independent variables. Given the interest in causal relations among the many constructs represented by variables obtained under such study conditions, researchers were ready to embrace multivariate analytical strategies such as SEM, which provided high flexibility in modeling causal relations based on variances and covariances of a large set of measured variables. In addition to the flexibility of SEM, the surge of interest and sustained use of SEM analyses in I/O psychology was associated with the increased availability and ease of use of SEM software such as LISREL (Jöreskog & Sörbom, 1996) and EQS (Bentler, 2004), among others. In SEM, a hypothesized causal model that specifies the directional paths linking all of the study variables is posited based on theoretical arguments by the researcher and fitted to the observed data. The SEM model consists of the measurement aspects, which specify the relations linking the observed variables (indicators) to the latent variables (constructs) to which the observed variables were intended to reflect (measure), and the structural aspects, which specify the causal relations linking the latent variables representing the substantive constructs in the study. An example of an SEM model is shown in Figure 4.1, which specified the causal relations linking four latent variables (Constructs A, B, C, D), with each latent variable being measured by three observed variables (Indicators V1 to V12). The SEM model hypothesized that altruism (Construct A) and loyalty (Construct B) have independent direct causal effects on likeability (Construct C), which in turn has a direct causal effect on adjustment (Construct D). In SEM analysis, the hypothesized causal model is fitted to the study data by using the variance– covariance matrix of the observed variables as data input to test the specific hypothesized causal model. Various practical model fit indices (e.g., adjusted goodness-of-fit, nonnormed fit index) are computed 91

David Chan

Construct A (e.g., altruism)

V1

V2

Copyright American Psychological Association. Not for further distribution.

Construct B (e.g., loyalty)

V4

V5

Construct C (e.g., likeability)

V3

V7

V8

V9

Construct D (e.g., adjustment)

V10

V11

V12

V6

FIGURE 4.1. A diagrammatic representation of an SEM model. The model consists of the measurement aspects that specify the relations that link the observed variables (indicators), V1 to V12, to the latent variables (constructs), Construct A to Construct D, to which the observed variables were intended to reflect (measure) and the structural aspects that specify the causal relations that link the latent variables, which represent the substantive constructs in the study.

by the SEM software. These indices, which provide information on the extent to which the model provides a good fit to the observed data, are used as evidence to conclude that the hypothesized causal model is a good or poor representation of the data and therefore is empirically supported or not supported. Because the various indices have different statistical assumptions, the evidence is stronger when the various indices lead to the same substantive conclusion on the extent of model fit. SEM also allows the researcher to test and compare other alternative causal models based on competing theories or conceptual arguments with the hypothesized model for the purpose of identifying the most adequate model based on relative fit, parsimony, and other conceptual or empirical considerations. Many comprehensive and excellent books and articles on SEM, both addressing basic and advanced issues, are readily available (e.g., Bollen, 1989; MacCallum, Browne, & Sugawara, 1996; Maruyama, 1998). In addition, examples of applications of SEM analyses to real data in substantive research are commonly found in studies published in I/O psychology journals (e.g., Journal of Applied Psychology). However, it is useful to reiterate several critical points on 92

the use of SEM, which often did not receive sufficient emphasis in technical expositions of the analytical strategy. SEM is intended to be a confirmatory analytical strategy insofar as the hypothesized model to be fitted to the data requires a theorization of the causal relations linking the variables and the purpose of the SEM is to assess whether the model fits the data. However, in practice it is probably not uncommon that some researchers use SEM in exploratory ways, such as constructing theories to fit a model that happens to fit the data well but is in fact one of many atheoretically generated models fitted to the data in a trial-and-error fashion. When a model provides a poor fit to the data, the statistical source of the poor fit may be readily found by looking at the causal links that were nonsignificant and the modification indices associated with potential causal links that were not included in a model. By simply removing the hypothesized links that turned out to be nonsignificant and including new links to pairs of constructs that exhibited large modification indices, it is possible to easily “trim” the original model until one gets a good-fitting model. Such atheoretical and data-driven modeltrimming approaches will guarantee the researcher to derive a model that will show a good fit to the data

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

but one that is likely atheoretical and capitalizes on the idiosyncrasies of the particular sample producing the data set which have no causal relevance. Hence, model assessment and model selection have to be driven by good theory underlying model specification. A hypothesized model can have a good fit to the data, explain much of the variances and covariances among the observed variables, have significant coefficients representing the relations between variables in the model, and still be a wrong causal model. This is because multiple alternative models with different specifications of causal relations among the variables may fit the same data set similarly well. Hence, it is important to specify, fit and compare different competing models that are premised on different theoretical conceptualizations of the causal relations among the variables thereby leading to different configurations of the causal relations (i.e., different causal models). The soundness of the theory associated with the causal model and the adequacy of the design and measurement, which affect the quality of the data obtained, are important considerations that directly affect the validity of an SEM model. For example, consider the case of an SEM model that hypothesized that organizational citizenship behavior causes newcomer adaptation. If this model provided a good fit to the data, but the study design or measurement was such that data on organizational citizenship behavior were collected after data on newcomer adaptation were collected, then there is a mismatch in temporal sequence between the hypothesized causal relation (A before B) and the measurement (B before A). That is, the good model fit to the data would not provide evidence for any strong causal inference given that the temporal sequence of A and B did not fulfill the necessary condition that the cause temporally precedes the effect. Finally, a good-fitting model should be cross-validated in a replication study to ensure that the good fit was not because of capitalization on chance or a unique data set in the original study. In summary, appropriately applied with good theory, design, and measurement, SEM can provide useful tests of competing theoretical models and help provide partial evidence for hypothesized causal flows, linking various constructs based on modeling covariation among observed variables.

MODELING MULTILEVEL PHENOMENA Traditionally, the majority of the areas of research in I/O psychology (e.g., personnel selection, performance appraisal, job satisfaction, work motivation) has approached the focal constructs or phenomenon under investigation from a microperspective that focuses almost exclusively on individual-level variables such as traits and the individual’s job performance. However, many constructs and phenomena of interest examined by I/O psychologists are in fact multilevel in nature, involving multiple levels of analysis such as the individual, group, department, organization, and even higher levels such as state and country. More specifically, even if the unit of theory or focus of study is at one level of analysis, say the level of the individual, the data that the researcher obtained from the individual responses are in fact multilevel in nature in the sense that the data are hierarchically structured or nested; that is, the units of observation at one level of analysis are nested (meaning grouped) within units at a higher level of analysis. For example, a study may consist of employees nested within work teams, resulting in a two-level research with the individual employees at the lower level (Level 1) and teams at the higher level (Level 2). When the phenomenon is multilevel and the data set is hierarchical with lower level units (e.g., individual employees) nested within higher units (e.g., work teams), restricting the conceptualization and analytical strategy to the individual level analysis at best limits our understanding and at worst leads to misleading and wrong substantive inferences. The adequate examination of multilevel constructs and data sets involve addressing complex conceptual, measurement, and data-analysis issues. Fortunately, researchers have developed useful organizing frameworks that help to clarify conceptualizations and decide on measurements or operationalizations of similar constructs at multiple levels, as well as identify the types of relevant evidence to support the multilevel hypotheses. Specifically, in the past 2 to 3 decades, many I/O psychologists have contributed to multilevel research by developing conceptual organizing frameworks (e.g., Chan, 1998a; Kozlowski & 93

David Chan

Copyright American Psychological Association. Not for further distribution.

Klein, 2000; Morgeson & Hofmann, 1999; Rousseau, 1985) and many scholars have helped advance multilevel research by providing relatively nontechnical summaries and applications of existing multilevel analytical strategies (Bliese, 2000; Bryk & Raudenbush, 1992; Ostroff, 1993), which helped users in I/O psychology to apply appropriate techniques to their multilevel data sets. The following sections summarize these conceptual bases and analysis issues that concern multilevel analytical strategies (for more details, see Chan, 2005).

Problems of Single-Level Analytical Approaches to Multilevel Data Single-level analytical approaches to multilevel data are problematic because these approaches fail to take into consideration the hierarchical nested structure of the multilevel data, thereby resulting in numerous statistical and conceptual fallacies. This failure is best illustrated when the researcher uses the traditional single-level regression model to analyze multilevel data. When applied to analysis of multilevel data, a major assumption of the traditional single-level regression model is violated. This is the assumption that the observations are randomly sampled from a homogeneous population and therefore are statistically independent. The violation of this independence assumption in multilevel data leads to wrong sample sizes being used in statistical computations and wrong standard errors obtained, and these in turn lead the researcher to wrongly estimate the precision of parameters and draw inaccurate substantive inferences from the data. Chan (2005) provided a simple example of a multilevel data set of students nested within classes to illustrate the problems of ignoring the dependency problem in multilevel data and violating the independency assumption of single-level regression models. Students in the same class are more similar than students from different classes because of various factors such as classroom learning environment and assignment of students to classes based on prior academic achievement. These grouping effects (i.e., class effects) associated with the nested structure of the data are ignored in the single-level regression model. 94

Consider a data set with a total sample size of 2,000 students made up of 50 classes of varying class size. The researcher is interested in predicting student examination scores from student intelligence and class size. To apply the single-level regression analysis, the researcher will first disaggregate class size to the student level (i.e., ignore the fact that there are only 50 different class sizes and assign each of the 2,000 students a class size as if there were 2,000 observations on the class-size variable) and then run a regression analysis at the individual level of analysis by regressing student examination scores on student intelligence and class size. However, there are only 50 classes, and therefore the true sample size (independent observations) for the class size variable is in fact 50. By disaggregating the contextual variable (i.e., class size), the sample size for the variable has been inflated from 50 to 2,000, and the regression analysis treats the 2,000 disaggregated values on the contextual variable as 2,000 independent observations when they are not in fact independent. The regression analysis uses this grossly inflated sample size to compute statistical significance tests, and therefore the results are likely to lead to spurious inferences because the large sample size increases the probability of a Type I error. The inflated sample size also artificially reduces the standard errors (because standard error decreases as sample size increases), leading the researcher to overestimate the precision of the respective parameters. In the single-level regression analysis, the standard error estimates are also inaccurate because they fail to take into account the variance and covariance components associated with the effects of grouping (i.e., the nested-structure effects) of the data (see, e.g., Aitkin, Anderson, & Hinde, 1981). When the single-level regression analysis is applied to aggregated data, the converse problem will occur. Suppose the researcher is now interested in predicting examination scores from class size. In attempting to apply a single-level analysis, the researcher might compose or aggregate the student examination scores data to the class level so that the two variables are on a single and same level of analysis. The researcher could, for example, compute the mean of the examination scores of all the students within a class to form “class mean exami-

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

nation scores,” which is a variable aggregated from the lower level units. The researcher then performs the single-level regression by regressing class mean examination scores on class size. This analysis is problematic because in aggregating student examination scores within each class, information on the lower level units (i.e., the different values of student intelligence within each class) is lost, and the sample size is reduced. The consequences are increased standard errors and reduced statistical power (i.e., a higher probability of committing a Type II error). The single-level model approach to multilevel data often leads to substantive inferences that are inadequate or misleading. These erroneous inferences are inferential fallacies of the wrong level in the sense that they refer to a mismatch or disconnect between the level of data analysis and the level of substantive conclusion. Put simply, data are analyzed at one level but the conclusion is made at another (wrong) level. The two most well-known fallacies of the wrong level are the ecological fallacy and the atomistic fallacy. The ecological fallacy occurs when the researcher mistakenly makes substantive conclusions at the lower level based on the results of aggregated data analyzed at the higher level. The fallacy can be very serious because the correlation or regression coefficients from the data at the aggregated level can be very different in magnitude from the corresponding correlation or regression coefficients from the data at the lower level. The ecological fallacy is so named because of the mistaken reliance on the ecological correlations, a term made popular by Robinson’s (1950) study, which refers to the correlations at the aggregated level. Robinson found that the correlation between percentage of Black people and illiteracy level in several geographic regions (r = .95), computed based on the two aggregated variables that were at the region level (i.e., the ecological correlation), is very much higher than the corresponding correlation computed based on the two variables at the individual level (r = .20). This disparity in the magnitude of correlation clearly shows the serious extent of the error in making substantive conclusions about the variables at the individual level based on the results of analysis of aggregated variables. To conclude on the basis of the ecological cor-

relation of .95 that there is a very strong relationship between being Black and being illiterate at the individual level is to commit an ecological fallacy (also known as the Robinson effect) because the adequate conclusion should be based on the individual-level correlation of .20, which is likely to be a conclusion that only a small or trivial relationship exists. Whereas the ecological fallacy refers to the failure to take into account the aggregation problem, the atomistic fallacy refers to the failure to consider the disaggregation problem. The atomistic fallacy occurs when the researcher mistakenly makes substantive conclusions at the higher level on the basis of the results from data analyzed at the lower level. In the earlier example, the researcher may disaggregate class size and class mean examination scores (both are variables at the class level) to the individual student level by assigning all students within the same class the same mean examination score and same class-size value. The individual-level correlation between the two disaggregated variables will be different in magnitude (probably much larger) than the corresponding higher level correlation. Fallacious inferences occur when the researcher uses the lower level correlation as the parameter estimate to represent the relationship between class size and class mean examination scores and make substantive conclusions about the relationship. The technique of hierarchical liner modeling provides an appropriate analytical strategy that will address the problems of statistical dependency in multilevel data as illustrated in the “class effects” example described previously and avoid committing the associated inferential fallacies. This technique is explicated in the section after the next. Although there are statistical explanations for the disparity of correlations at different levels of analysis (see Robinson, 1950; Kreft & de Leeuw, 1998), the major reasons for any observed large disparities are likely to be conceptual in nature pertaining to the lack of construct isomorphism across levels (isomorphism exists when the conceptual content of the construct is essentially the same and remain unchanged from one level to another level). Specifically, when a lower level variable is aggregated to form a variable at the higher level, one cannot assume that the two 95

David Chan

variables at different levels of analysis are representing the same construct. This leads us to the next section, on the need to specify adequate composition models. Composition models are fundamental because they provide the conceptual basis for selecting the appropriate multilevel analysis.

Copyright American Psychological Association. Not for further distribution.

Composition Models On one hand, we cannot assume that the conceptual content will remain identical (or isomorphic) across levels as the researcher composes the higher level variable from the lower level counterparts. On the other hand, we need to avoid atheoretical situations in which the researcher freely composes “new” constructs at the higher level by aggregating the lower level units, representing an established construct in different ways (e.g., computing the mean or the variance to represent different higher level constructs). These two problems are addressed by formulating and applying appropriate composition models (Chan, 1998a). Composition models specify functional relationships between constructs at different levels thereby providing a systematic framework for mapping the transformation across levels. For example, a composition model may be formulated to specify how the individual level construct of self-efficacy may also be conceptualized at the group level and how self-efficacy may be transformed upward to form the higher level construct of group efficacy, including how the similar constructs at the two different levels are related conceptually and operationally through measurement and statistical analyses. The explicit transformation relationships provide conceptual precision in the target construct, which in turn aids in the derivation of test implications for hypothesis testing. Hence, composition models help test the assumption of construct isomorphism across levels. Composition models also help to guide the development and validation of newly proposed constructs in multilevel research, and hence avoid ending up with a multitude of construct labels, all of which purportedly refer to scientific constructs but in reality have no incremental explanatory value. Chan (1998a) developed a typology of composition models that specify functional relationships between constructs at different levels. Specifically, 96

Chan’s typology described five different basic forms that composition models can take. These are (a) additive, (b) direct consensus, (c) referent-shift consensus, (d) dispersion, and (e) process composition. A theory of the focal construct in a multilevel study may contain one or more of the five composition forms. In Chan’s typology, each composition model is defined by a particular form of functional relationship specified between constructs at different levels. Corresponding to each form of functional relationship is a typical operational process by which the lower level construct is combined to form a higher level construct. The operational combination process is the typical form, as opposed to a necessary consequence of the functional relationship specified. Chan also gave several suggestions on what constitutes the evidence needed to support the relevant functional relationships and to establish that appropriate combination rules are applied (see Chan, 1998a, Table 1). A description of all of the five composition models is beyond the scope of this chapter (for details, see Chan, 1998a). The central point to note here is that the researcher needs to explicate an adequate composition model and not uncritically assume construct validity for a “new” higher level construct aggregated from the lower level units. For example, a composition model of workgroup climate for justice would specify how the group-level construct of justice climate is derived from the established individuallevel construct of justice perceptions and how this new group-level construct can be empirically validated. The nature of the aggregated construct can be very different, depending on the specific composition used. Different compositions models, when applied to the same individual-level construct of justice perceptions, would result in different grouplevel climate constructs with different measurement implications. To illustrate how composition models drive conceptualizations and measurement of constructs, lets consider the aggregation of justice perceptions at the individual level to a group-level construct by using a direct consensus composition versus using a dispersion composition. In a direct consensus composition, which is the most familiar and popular form of composition among multilevel researchers, “within-group

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

consensus of the lower level units” is used as the functional relationship for specifying how the construct conceptualized and measured at the lower level is functionally isomorphic to another form of the construct at the higher level (i.e., the analytical construct). On the basis of this composition model, the researcher may construe psychological climate for justice as an individual’s perception of fairness in the work environment in terms of the psychological meaning and significance to the individual ( James, 1982). The researcher may then derive a conceptual definition for a group-level construct called group climate for justice, which refers to the shared assignment of meanings among individuals within the workgroup. In this conceptualization, within-group agreement among individual climate perceptions indicates shared assignment of psychological meaning. It is this “sharedness” that constitutes functional equivalence between the climate constructs at the two levels. Hence, the definition of group climate for justice is essentially the same as psychological climate for justice, except that the former refers to the shared perceptions among the individuals. The conceptual relationship between the two forms of the construct at different levels then drives the manner in which the lower level construct composes to the higher level construct. In terms of the operational combination process, the conceptual relationship would specify the manner of and preconditions for combining the individual lower level measurements to represent the higher level measurement. For example, withingroup agreement indexes, such as the rwg index (James, Demaree, & Wolf, 1984), may be calculated, and some cutoff level of within-group agreement of psychological climate responses is used to justify aggregation of the individual-level responses to form the group-level construct of group climate for justice. High within-group agreement indicates consensus and justifies aggregation of individual climate responses to represent scores on the group climate variable. Typically, in direct consensus composition, the mean of the individual-level responses within a group is used to represent the group’s value on the group-level construct after passing the selected cutoff for within-group agreement. In this example, the aggregation procedure and preconditions, together with the conceptual definition of group climate for

justice (i.e., the group-level construct) determine the meaningfulness and validity of the operationalization of the group-level construct. If we apply a dispersion composition model to the aggregation of individual-level justice climate perceptions to the group level, then we could derive a group-level construct called justice climate strength, which is distinct from the group-level construct called justice climate level derived from a direct consensus composition. The following paragraphs describe how a dispersion construct at the group level could be composed. Recall that in direct consensus composition, within-group agreement of scores from the lower level units or attributes is used to index consensus. The researcher hopes to achieve a high agreement at the lower level to justify aggregation to represent variables at the higher level. In this composition model, consensus is a necessary condition for construct validity at the higher level, and high within-group agreement constitutes an empirical or statistical precondition to be fulfilled for the operational combination process to be legitimate. In contrast, dispersion composition would, based on some theory, conceptualize the degree of within-group agreement of scores from the lower level units or attributes as a focal construct as opposed to a statistical prerequisite for aggregation. That is, instead of treating within-group variance as error variance (which is what the direct consensus composition models does), withingroup variance (i.e., the within-group dispersion of scores) could serve as an operationalization of a focal construct. In essence, the dispersion composition model focuses on the use of within-group dispersion (i.e., variance or agreement) to specify the functional relationship in composition of a dispersion construct. Continuing with our climate example, the researcher may propose the construct of climate strength conceptualized as the degree of withingroup consensus of individual climate perceptions and index the construct using within-group variance or some dispersion measure of individual climate responses. The dispersion measure then is correlated with some criterion variable such as a measure of group cohesion to test the researcher’s hypothesis 97

Copyright American Psychological Association. Not for further distribution.

David Chan

(e.g., that group climate strength is associated positively with group cohesion). To summarize, the above examples show that the two aggregated constructs, namely, group justice climate level and group justice climate strength, although aggregated from the same individual-level climate construct, are distinct constructs with different conceptual definitions and measurement implications driven accordingly by their respective composition model. For example, in the aggregation to form values for the group-level construct of justice climate level, the direct consensus composition model would treat within-group agreement of individual-level climate responses as a statistical hurdle to be cleared to justify an aggregation method that would use the group’s mean to represent the group’s value on the aggregated construct. On the other hand, in the aggregation to form values for the group-level construct of justice climate strength, the dispersion composition model would use each group’s within-group agreement of individual-level climate responses to represent the group’s value on the aggregated construct. Because group climate level and group climate strength are distinct constructs, they are likely to have different antecedents, correlates, and outcomes. They may also interact to affect an outcome variable. For example, the magnitude of the relationship between group climate level and a criterion variable (e.g., group performance) may be dependent on the values of group climate strength. Applying the appropriate composition models (multiple models can be applied in the same study) would clarify the conceptualization and measurement of the focal variables in a multilevel study. Recently, several researchers, particularly those in climate research, have used various composition models to formulate and test interesting multilevel and cross-levels hypotheses, and they have produced results that have significantly advanced the field (e.g., Klein, Conn, Smith, & Sorra, 2001; Schneider, Salvaggio, & Subirats, 2002). Chan’s typology of composition models provides an organizing framework for existing focal constructs, facilitating scientific communication in multilevel research. Researchers can be more confident that they are referring to the same construct when it is explicated according to the same form of composition. 98

Meaningful replications and extensions of current findings then are possible. Apparent contradictory findings may be reconciled, and debates may be clarified. Organizing existing constructs also aids cumulation of research findings by providing a framework for performing meaningful meta-analytic studies in multilevel research. The typology also provides a conceptual framework for developing and validating new focal constructs and multilevel theories. It could help compose new explanatory constructs from established ones. In addition, being cognizant of different models allows the researcher to consider alternative designs, measurements, and data analyses for testing competing hypotheses, modifying existing theories or developing new ones, or performing a more rigorous test of the original hypothesis. In short, an adequate typology of composition models is important because it helps to clarify conceptualizations and measurement of similar constructs at different levels of analysis.

Hierarchical Linear Modeling Having formulated and applied one or more adequate composition models to obtain reliable and valid measures of the relevant aggregated grouplevel constructs, the next step is to apply an appropriate analytical strategy so that the hierarchical nested structure of the multilevel data is taken into account. The two major types of established techniques are (a) multilevel regression models or more generally known as hierarchical linear models (HLM) and (b) multilevel latent variable models. HLM are discussed in this section, and multilevel latent variable models are described in the next section. We use the term HLM to refer to the multilevel regression analyses that probably constitute the most common class of multilevel techniques used in multilevel research. Conceptually, HLM may be viewed as traditional regression models (as in singlelevel regression) with additional variance terms to represent variables specifically associated with the hierarchical nature of the multilevel data. To understand the importance of these additional terms, we need to first examine the single-level regression model that does not contain these terms. Consider

Advances in Analytical Strategies

the traditional ordinary least squares (OLS) simple regression model for predicting y from x,

Copyright American Psychological Association. Not for further distribution.

yi = b0 + b1 x i + e i ,

(1)

in which b1 is the slope or regression coefficient for x, b0 is the intercept, and ei is the error (or residual) term. As long as the prediction of y from x is imperfect (i.e., the value on the actual y is not the same as the value on the predicted y), which is almost always the case in psychological research, ei is nonzero. In OLS regression, we make three assumptions about ei. First, ei is assumed to be normally distributed with a mean of zero (this is called the normality assumption). Second, it is assumed that the variances of ei are constant across different values of x (this is called the homoscedasticity assumption). Finally, it is assumed that the ei from each observation is not correlated with (i.e., independent of) the eis from the other observations in the sample (this is called the independence assumption). Although some scholars have shown that violations of the normality and homoscedasticity assumptions in OLS regression analyses tend to have little impact (e.g., Berry & Feldman, 1985), it is well established that the violation of the independence assumption can have serious impact and could lead to inaccurate or incorrect inferences. Recall that in multilevel data, the lower level observations within the group (i.e., within the higher level unit) are typically not independent insofar as they tend to share some similar characteristics or are exposed to the same effects by virtue of being in the same group. When the within-group observations are not independent, the errors from these withingroup observations are more similar to one another than would be expected by chance. In other words, the errors from the observations in the sample will be correlated with group membership. That is, the eis are correlated (dependent), and the independence assumption of OLS regression is violated. When the observations are not independent (and therefore the eis are correlated), the line of best fit defined by the regression equation no longer represents the sources of variance in y adequately. Specifically, two sources of variance are not represented. First, the source of variance from group membership (i.e., there are between-group differences in mean y values) is not

represented by the regression equation. In other words, the group differences in the b0 parameter estimate are not represented. Second, the source of variance from between-group differences in x–y relationship (i.e., the different groups have different x–y relationships) is not represented by the regression equation, which specifies only a single x–y relationship. In other words, the group differences in the b1 parameter estimate are not represented. Therefore, the traditional single-level regression model is inadequate for analyzing multilevel data because there are no additional variance terms in the regression equation to represent these two between-group sources of variance. HLM analyses solve these problems by including these additional (multilevel) variance terms. To illustrate an HLM analysis, consider a twolevel data set in which j refers to the group membership (e.g., class) that the lower level units, i (e.g., students), belong. We are interested in predicting or explaining the lower level variable, y (e.g., math performance), from a lower level variable, x (e.g., gender of student), and a higher level variable, z (e.g., teacher experience). Considering first only the lower level variables, we can now rewrite Equation 1 as follows so that each observation’s (i) group membership is represented: yij = b0 j + b1j x ij + e ij .

(2 )

The first-order multilevel questions of interest here are whether the intercepts (b0) will differ across the j groups and whether the slopes (b1) will differ across the j groups. In other words, do classes differ in mean math performance (i.e., differences in intercepts), and does gender predict math performance differently across classes (i.e., differences in slopes)? The secondorder multilevel questions of interest are, given that they exist, whether we can predict these group differences in parameter estimates. For example, can we predict class differences in mean math performance and class differences in the gender-math performance relationship using teacher experience? To examine the first-order questions, the multilevel model will allow the intercepts as well as the slopes to vary randomly across groups. This is why HLM models are often referred to as random coefficient 99

Copyright American Psychological Association. Not for further distribution.

David Chan

models. Equation 2 is known as the Level 1 regression, representing the prediction of the lower level y variable (i.e., student math performance) from another lower level predictor variable x (i.e., student gender). In this model, across all j classes, the intercepts have a distribution with some mean and variance. Similarly, across all j classes, the slopes have a distribution with some mean and variance. As mentioned above, the first questions of interest are establishing that intercept variance, as well as the slope variance, is significantly different from zero. The next step in the multilevel modeling procedure is to address second-order questions, namely, to predict or explain these coefficient variances by introducing some variable at the higher level. In the math performance example, this refers to introducing teacher experience (z), a class-level variable, to predict or explain the class differences in mean math performance (i.e., variance in intercepts), and the class differences in gender-math performance relationship (i.e., variance in slopes). This is done by estimating Level 2 regressions as follows: b0 j = γ 00 + γ 01z j + μ 0 j , and b1j = γ 10 + γ 11z j + μ1j .

yij = γ 00 + γ 10 x ij + γ 01z j + γ 11 x ij z j + μ1j x ij + μ 0 j + e ij .

(5)

(3) (4)

Equation 3 predicts class differences in mean math performance (b0j) from teacher experience (z). A significant and positive γ01 indicates that classes with more experienced teachers tend to have higher mean math performance. Conversely, a significant and negative γ01 indicates that classes with more experienced teachers tend to have lower mean math performance. Equation 4 states that the relationship between student gender and student math performance (i.e., the gender effect on math performance) is dependent on teacher experience. A significant and positive γ11 indicates that the gender effect on math performance is larger in classes with more experienced teachers. Conversely, a significant and negative γ11 indicates that the gender effect on math performance is smaller in classes with more experienced teachers. That is, γ11 indexes the moderator relationship in which teacher experience moderates the relationship between student gender and student math performance. 100

Unlike b0j and b1j in Equation 2, which are random coefficients that vary across groups (classes), the regression coefficients γ00, γ01, γ10, and γ11 in Equations 3 and 4 do not have the subscript, j, to indicate group membership. These coefficients are fixed coefficients (as opposed to random coefficients), which are not assumed to vary across groups (classes). The remaining between-group variance in the b coefficients after accounting for the effect of the group-level variable, z, is the residual error variance at the group level. As shown in Equations 3 and 4, these group-level residual errors are represented by the μ terms. These errors are assumed to have a mean of zero and to be independent of the errors (eij) at the individual (student) level; however, the covariance between the two higher level error terms is typically not assumed to be zero. Substituting Equations 3 and 4 into Equation 2 and rearranging terms, we obtain the following regression equation:

Equation 5 represents a basic HLM model involving two levels (student level and class level). To better understand this equation, we can replace the abstract variable notations with the variable labels as follows: Math performance ij = γ 00 + γ 10 genderij + γ 01teacher experience j + γ 11genderijteacher experiencej + μ1j genderij + μ 0 j + e ij .

(6 )

There are three noteworthy points about the multilevel regression Equation 5. First, the component [γ00 + γ10xij + γ01zj + γ11 xij zj] of the equation has all the fixed coefficients and this constitutes the fixed part of the regression model. The other component, [μ1j xij + µ0j + eij], which has the error terms, is the random part of the regression model. Typically, the substantive interest is to maximize the fixed part and minimize the random part. Second, note that the γ01zj represents the effect of teacher experience, which is a higher level variable, on math performance, which is

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

a lower level variable, and it is therefore referred to as a cross-levels effect. In addition, the interaction term, γ11 xij zj, represents the interaction effect of a lower level variable (i.e., student gender) and a higher level variable (i.e., teacher experience) on a lower level variable (i.e., student math performance). Therefore, this interaction effect is referred to as a cross-levels interaction effect. Third, note that the error term, μ1j, is associated with xij. This indicates that the errors will be different for different values of xij. In other words, there is heteroscedasticity of errors, and this heteroscedasticity characteristic of multilevel data is explicitly taken into account in the multilevel regression model. In contrast, recall that the traditional single-level regression model assumes homoscedasticity of errors, an assumption that is typically violated in multilevel data. At this point, the reader who is familiar with OLS regression may wonder why it is worth going through all the above complicated equations when we can simply analyze multilevel data using a “contextual analysis” through an OLS regression. In other words, why not simply disaggregate the group-level variable, z (e.g., teacher experience), to the lower level units (students) and perform an OLS regression of student math performance on student gender and the disaggregated contextual variable teacher experience? That is, why not simply run the following OLS regression equation? yij = b0 + b1j x ij + b2 z j + e ij

(7 )

The answer is that this contextual analysis through OLS regression is problematic because, unlike multilevel regression models, it fails to take into account the critical characteristics of the hierarchical nature of multilevel data. First, the errors, eij, in the OLS regression Equation 7 are assumed to be independent. As explained earlier, this independence assumption is almost always violated in multilevel data. Second, the sample size for determining the statistical significance of b2 in Equation 7 is the number of lower level observations (i.e., students), which is incorrect, instead of the number of classes, which is the correct sample size for the group-level variable, z (i.e., teacher experience). Hence, by using the wrong (inflated) sample size, the standard errors obtained

are wrong (misestimated precision), and it is easier to obtain statistical significance for the effect of z (increasing Type I error). In contrast, multilevel regression models evaluate the effect of z by using the correct sample size (i.e., number of classes), resulting in correct standard errors. For more details on HLM, refer to Bryk and Raudenbush (1992) and Hofmann, Griffin, and Gavin (2000).

Multilevel Latent Variable Modeling Multilevel latent variable modeling (sometimes called multilevel SEM) is essentially multilevel analysis implemented within an SEM framework, and therefore it has all the advantages of SEM, such as the ability to incorporate measurement errors in the model, flexibility in specifying various causal models and comparing competing models, and ability to test for equivalence or hypothesized differences of corresponding parameters across groups. Multilevel latent variable modeling proceeds through a multistage process of fitting SEM models hierarchically. The analysis begins with specifying and establishing a good-fitting model at the lower level of analysis, such as the level of the individual worker in a study of workers nested within work teams. After establishing the good-fitting model at the lower level (i.e., Level 1 at the level of the individual worker), the higher level unit (i.e., Level 2 at the level of the work team) is added to the model to form a multiplegroup model in which the groups refer to the higher level units. The multiple-group model incorporates the statistical dependencies among the lower level units, which are nested within a same higher level unit. To proceed with the multilevel analysis in this multiple-group model, the Level 1 (individual worker) model is tested for each unit (work team) at Level 2. In this multilevel analysis, the Level 2 variable (work team) is the predictor, and the criterion or dependent variables are the various types of parameters from the individual Level 1 models. By implementing the multilevel analysis in the multiplegroup SEM framework, the analysis allows nested multiple-group model comparisons by following the same statistical logic of standard multiple-group SEM, thereby enabling the researcher to test competing theories of cross-levels effects in which the higher level variables may affect the lower level variables. 101

Copyright American Psychological Association. Not for further distribution.

David Chan

Technical details of how to conduct multilevel latent variable modeling is beyond the scope of this chapter. Excellent technical introduction to this analytical strategy is available in Chou, Bentler, and Pentz (1998) and Muthén (1994, 1997). Conceptual and statistical issues in multilevel analysis continue to be of interest to I/O researchers, given the multilevel nature of organizational phenomena. Examples of current issues are available in a series of articles in a special issue of Organizational Research Methods (Bliese, Chan, & Ployhart, 2005), which addresses issues and future research directions on multilevel methods. The topics discussed included the need to continue moving beyond notions of isomorphism in developing and testing aggregate-level constructs, the potential value of using discontinuous growth models to understand transitions in longitudinal studies within a multilevel framework, and issues on the ability to decompose variance in multilevel modeling of dichotomous and other nonnormal outcome data. MODELING CHANGES OVER TIME For many decades in I/O psychology, predictorcriterion relationships have been described in terms of static models of the criterion without any attention paid to the temporal aspects of the criterion constructs, including what and how changes may occur over time. Consider the example of job performance models. An individual’s job performance may change over time in various ways (e.g., increase or decrease in level, changes in the number or nature of underlying dimensions) and these intraindividual changes are important for understanding and predicting job performance. For example, when performance changes over time either in terms of level or dimensionality, using a sample of job incumbents with varying levels of job tenure in a validation study could affect and confound estimates of validity and the interpretation of predictor-criterion relationships. Advances in longitudinal analytical strategies, especially those that involve latent variable modeling, provide us with both the conceptual basis and statistical method to hypothesize, test, and interpret criterion (e.g., performance) changes over time, which in turn allows us to draw practical implications (e.g., person102

nel selection issues) such as changes in test validities, changes in mean levels of the criterion, changes in rank order of individuals’ criterion scores, and changes in criterion dimensionality (i.e., number and/or nature of dimensions). The analysis of change over time has to be guided by the conceptualization of change over time. By first specifying the specific facet of change over time, appropriate longitudinal designs and data analytic techniques can be applied to implement research that answers important questions relating to criterion changes over time. In the example of job performance changes, these questions may include the nature of new performance dimensions associated with changes in job demands or different points in time over the individual’s job tenure; describing, predicting, and explaining the form of the intraindividual change trajectory (e.g., linear vs. quadratic, increasing vs. decreasing) and individual differences in the rate of intraindividual change; and modeling associations among performance dimensions and the trajectories by which they change over time.

Fundamental Questions on Changes Over Time The various specific facets of change over time are related to distinct fundamental questions that may be asked of the nature of change that may occur over time. Chan (1998b) explicated nine such questions. These questions highlight the complexities involved when considering change over time and the importance of clarifying the specific question asked of the change phenomenon and relating it to the analytical strategy and the substantive inferences made from data. These questions, addressed in detail in Chan (1998b), are briefly summarized here. 1. Is an observed change over time (and observed between-group differences in change over time) due to meaningful systematic differences or random fluctuations resulting from measurement error? If measurement error is not adequately taken into account when specifying the data analysis model and estimating the parameters, results of the analyses can be severely affected by measurement error. The classic independence of

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

errors assumption, which is common among many traditional data analysis procedures, may be violated when assessing change over time in longitudinal designs, particularly when the longitudinal data are collected on measurement occasions closely spaced together by using identical measures. 2. Is the change over time reversible? The question on the reversibility of change over time can be construed in terms of the functional form of the intraindividual growth (change) trajectory. For example, monotonically increasing or decreasing (e.g., linear) functional forms represent irreversible (within the time period studied) change in the sense that there is no returning or restoring to previous levels on the focal variable, at least during the period under investigation. On the other hand, a nonmonotonic functional form (e.g., an inverted U) would represent reversible change over time. 3. Is the change over time proceeding in one single pathway or through multiple different pathways? Two (or more) groups of individuals may follow the same or different trajectories as they proceed from one time point to another (through intervening time points measured in the study). For example, in a four-time point study (e.g., organizational newcomer adaptation study with adaptation outcomes measured at four time points equally spaced at one-month measurement interval), two groups (e.g., locals and expatriates) may have the same value on the focal variable at initial status (Time 1) and at endpoint (Time 4), but one group follows a positive linear trajectory and the other follows a positively accelerated monotonically increasing trajectory. That is, change from one value of the focal variable at Time 1 to another value at Time 4 could proceed through multiple different pathways. 4. Is the change on the quantitative variable proceeding in a gradual manner, or is it best characterized as large magnitude shifts at each time interval? Quantitative change over time may proceed gradually as characterized by a linear trajectory with a low slope or it may be characterized in terms of large magnitude changes as represented by a high slope.

5. Is the change over time (or across groups) to be considered as alpha, beta, or gamma change? Golembiewski, Billingsley, and Yeager (1976) distinguished three types of change: alpha, beta, and gamma. Alpha change refers to changes in absolute levels given a constant conceptual domain and a constant measuring instrument. For example, if job satisfaction was adequately measured both at Time 1 and Time 2 in terms of reliability and validity such that the same construct was measured at both time points and with the same precision, then the difference in the satisfaction scores between the two time points represent an alpha change in satisfaction and the change may be directly interpreted as a change in the absolute level of job satisfaction. We can meaningfully speak of alpha change only when there is measurement invariance of responses across time. Measurement invariance across time exists when the numerical values across time waves are on the same measurement scale. Measurement invariance could be construed as absence of beta and gamma changes. Beta change refers to changes in absolute level complicated by changes in the measuring instrument given a constant conceptual domain. Beta change occurs when there is a recalibration of the measurement scale. That is, in beta change, the observed change results from an alteration in the respondent’s subjective metric or evaluative scale rather than an actual change in the construct of interest. For example, because of the respondent’s increased leniency in ratings over time, a rating of 6 given at Time 2 may be defined by the respondent as was rating of 5 at Time 1. Gamma change refers to changes in the conceptual domain. Gamma change (i.e., change in the meaning or conceptualization of the construct(s) of interest) can take a variety of forms. For example, in the language of factor analysis, the number of factors (a factor representing a construct) assessed by a given set of measures may change from one time point to another. To illustrate, in a study of changes in performance over time, performance may undergo a type of gamma change represented by factorial integration of performance measurement so that performance components (factors) become 103

David Chan

Copyright American Psychological Association. Not for further distribution.

6.

7.

8.

9.

increasingly interrelated over time such that performance at early time points are best represented as multiple distinct and relatively uncorrelated factors, at mid time points are best represented as multiple highly correlated factors and at later time points are best represented as a single factor. Is the change over time occurring at the individual, group, or both levels of conceptualization? Change over time can be conceptualized and assessed at the individual level, group level (e.g., team, department), or both levels. Any analytic technique that is restricted to only one level of conceptualization and analysis is limited in an important way because the assumption of no or “irrelevant” change at the other level is not tested. In addition to detecting interindividual differences in intraindividual change, can we predict (and hence increase our understanding of) these differences? Individuals may systematically differ in the way they change over time. We can increase our understanding if the longitudinal modeling can incorporate additional variables and assess their efficacy in predicting the different aspects of these individual differences (e.g., individual differences in rate of change, individual differences in trajectory forms). Are there cross-domain relationships in change over time? Changes in one focal variable may be systematically related to changes in another focal variable. For example, during the period of newcomer adaptation, the rate of change in information seeking may be positively correlated with the rate of change in task mastery. An adequate longitudinal-modeling procedure would allow us to explicitly model these crossdomain relationships. Do the various relationships with respect to specific facets of change over time vary or remain invariant across groups? Different groups may either share or differ in the various specific facets of intraindividual changes. An adequate longitudinal-modeling procedure would allow us to explicitly model and test the various hypotheses concerning between-group differences or similarities in change over time.

104

Limitations of Traditional Techniques for Modeling Changes Over Time Chan (1998b) provided a detailed description of these nine questions and explained why traditional techniques such as difference scores analysis, repeated measures ANOVA, and time series are limited in their ability to adequately address these questions. To illustrate the limitations of traditional techniques for modeling changes over time, consider time-series models, which are probably the most commonly used longitudinal data analysis technique. Time-series models were developed to describe a relatively long series of observations typically consisting of at least 20 or 30 time points. In general, time-series models may be classified into time domain and frequency domain models. Autoregressive integrated moving average (ARIMA) models are representative of time domain models (e.g., Box & Jenkins, 1976), whereas spectral analysis models are representative of frequency domain models (Larsen, 1990). Time domain and frequency domain models differ in how they represent the same time-series information. Time domain models analyze the longitudinal data and make inferences based on the autocorrelations in the sequence of observations. Autocorrelation refers to the correlation between later items in a time series and earlier items (when the time series is completely random, the autocorrelation is zero). The time series is expressed in terms of autoregressive or some other time-based parameters. In these models, a given observation in time is characterized as a weighted function of past observations of the same underlying process. These time-series models, such as ARIMA models, are typically used for forecasting purposes. Frequency domain models, on the other hand, express and account for the time-series data in terms of trigonometric functions such as sine and cosine functions. These functions are used to represent rhythms or cycles assumed to underlie the timeseries data. Clearly, the choice between the two classes of models is dependent on the nature of the research question at hand. For example, questions that forecast time points call for time domain models whereas those that assess rhythms or cycles within the data call for frequency domain models. Although both classes of time-series models have potential applied value in substantive longitudinal

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

research in I/O psychology (e.g., time domain models can be applied to the study of predicting future job-performance rankings from past job-performance rankings; frequency domain models can be applied to the study of mood variability at the workplace), the requirement of a large number of repeated measurements in the longitudinal design limits the actual applied value of these time-series models, at least in the current state of I/O research. It is more important to note that as explained in Chan (1998b), time-series models are not well equipped to assess the various aspects of the intraindividual change discussed above. For example, time-series models cannot be readily used to model interindividual differences in intraindividual changes. It is possible to fit a time series to an individual’s repeated observations (hence compare different individuals’ function by comparing distinct time-series models) or to the summary statistics of a group of individuals (hence compare different groups’ functions by comparing distinct time-series models), but it is not possible to do both at the same time. That is, it is not possible, within a single time-series model, to examine a group’s intraindividual change function at the aggregate (group) level and, at the same time, individual differences in intraindividual change functions. One fundamental question on intraindividual change is whether the same construct is in fact being observed over time and, if so, whether it is being assessed with the same precision. This issue of measurement invariance (of repeated responses on the identical measure) over time is statistically fundamental because virtually all of the traditional techniques such as time-series models, repeated measures ANOVA, and difference scores analysis are applied in a manner that assumes, rather than directly tests, the assumption of measurement invariance of intraindividual repeated responses over time. In addition, depending on the research question, certain measurement invariance questions may be theoretically interesting in their own right (i.e., reflecting a substantive intraindividual change process), apart from the issue of reflecting a statistical hurdle to be cleared prior to assessing substantive intraindividual change. For example, a lack of measurement invariance of responses over time may reflect a substantive intraindividual

change process associated with a type of gamma change in performance dimensions over time.

Latent Variable Approaches to Modeling Changes Over Time Latent variable approaches are well suited for longitudinal modeling because they can explicitly take into account both cross-sectional and longitudinal measurement errors. Hence, the researcher is able to model a variety of error covariance structures and assess any distorting effects that cross-sectional or longitudinal measurement errors may have on the various parameter estimates of true change. In addition, latent variable approaches are highly flexible and powerful because a variety of latent variable (i.e., SEM) models can be fitted to the longitudinal data to describe, in alternative ways, the change over time. Latent growth modeling (LGM) offers a direct and comprehensive assessment of the nature of true intraindividual changes and interindividual differences in these changes. LGM also allows these differences to be related to individual predictors. An LGM model can be elaborated into a multiple-indicator latent growth model (MLGM). MLGM is essentially an LGM analysis in which the focal variable of change is modeled as a latent variable represented by multiple indicators. Technical details of LGM and MLGM are described in Chan (1998b). LGM represents the longitudinal data by modeling interindividual differences in the attributes (i.e., parameters) of intraindividual changes over time (i.e., individual growth curves). In an LGM analysis, we can estimate the means and variances of the two growth parameters (intercept and slope factors) and examine whether the two parameters are correlated with each other. The LGM analysis can also be used to examine associations between the growth parameters and individual difference predictor variables. For example, in newcomer adaptation research, we can use LGM to predict initial status and rate of change in information seeking from proactive personality (Chan & Schmitt, 2000). Different univariate latent growth models can also be combined to form a multivariate latent growth model. In a multivariate growth model, parameters from different change trajectories 105

Copyright American Psychological Association. Not for further distribution.

David Chan

can be correlated to examine cross-domain associations (i.e., relationships between two focal variables being examined for intraindividual change over time). For example, in a study of interpersonal relationships, rate of change in relationship building can be correlated with rate of change in social integration. One or more predictors can also be included in the multivariate model, thereby allowing hypotheses in regard to differential predictions (using the same individual predictor) of intraindividual change across domains can be tested. Finally, LGMs (univariate or multivariate) can be fitted simultaneously to different groups of individuals (e.g., gender, ethnic, occupational, experimental groups), and multiple-group LGM analyses can be performed to test for acrossgroups invariance of one or more of the specified relationships in the latent growth model. To incorporate measurement invariance concerns in the model specification, LGM can be extended to an MLGM in which the focal variable of change is modeled as a latent variable assessed by multiple indicators as opposed to a manifest variable typically the case in prior work on LGM. The use of multiple indicators in a latent growth model allows both random and nonrandom measurement errors to be taken into account when deriving the intercept and slope–shape factors. The use of multiple indicators to assess the focal construct allows reliable (nonrandom) variance to be partitioned into true-score common (construct) variance and true-score unique variance. True-score unique variance is nonrandom, and it is that portion of variance in a measure that is not shared with other measures of the same construct. In LGM, the same measures are repeatedly administered over time. Hence, a failure to partition nonrandom variance into true construct variance and unique variance leads to distorted (inflated) estimates of true change in the focal construct over time. Because only scale–composite level but no item-level (multiple indicator) information on the focal variable is used in the standard LGM, the procedure does not provide the isolation of nonrandom error variance from reliable variance and it takes only random errors into consideration. MLGM addresses the problem. Chan (1998b) demonstrated how the above questions on measurement invariance, functional forms 106

of intraindividual changes, and other fundamental questions on change over time may be answered in an integrative two-phase latent variable analytical procedure that combines longitudinal means and covariance structures analysis and multiple-indicator latent growth modeling. In Phase 1 of the procedure, longitudinal mean and covariance analysis, which is similar to longitudinal factor analysis except that both the indictor intercepts and factor means are also estimated, is used to examine issues of measurement invariance across time and across groups. Establishing invariance provides evidence that results of subsequent growth modeling constituting Phase 2 of the procedure are meaningful. By building invariance assessments as the first logical step to longitudinal modeling, this integrative procedure contrasts with the analytical models that left untested the assumption of measurement invariance across time or groups. In addition to invariance assessments, Phase 1 of the procedure helps in the preliminary assessment of the basic form of intraindividual change by identifying the constraints on the patterns of true score (factor) means and variances over time. In Phase 2, multipleindicator longitudinal growth modeling is used to directly assess change over time by explicitly and simultaneously modeling the group and individual growth trajectories of the focal variable as well as their relationships to other time-invariant predictors, time-varying correlates (i.e., growth trajectories in a different domain), or both. As explained in Chan (1998b), longitudinal mean and covariance analysis and multiple-indicator latent growth modeling together provide a unified framework for directly addressing the various fundamental questions on change over time. Longitudinal covariance structures analyses such as longitudinal factor analysis, longitudinal means and covariance structures analysis, and latent growth modeling are appropriate when the latent variables are continuous in nature. When the latent variables are discrete (i.e., categorical) in nature, latent class analysis is appropriate. When latent class modeling is applied to discrete longitudinal data, the analysis is known as latent transition analysis, which allows the researcher to specify and test stage-sequential development or changes over time. An excellent introduction to latent class analysis and latent transition

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

analysis is provided by Collins and Wugalter (1992). Recently, Muthén (2004) developed an inclusive framework known as general growth-mixture modeling, which combines latent growth models and latent class models. This general framework allows the researcher to identify latent classes characterized by different patterns of latent growth. These mixture models are useful because they allow us, in a single integrated analysis, to identify groups of individuals with qualitatively different growth trajectories. Not all latent variable approaches are suited for modeling changes over time. For example, autoregressive latent modeling, which is one of the simplest latent variable approaches, is not adequate for the analysis of longitudinal data representing intraindividual change over time. Autoregressive models estimate scores on a variable based on values of the same variable. Proponents of the inclusion of autoregressive models in the longitudinal modeling of intraindividual change argue that the autoregressive effect (the effect of the Time 1 measure on the Time 2 measure of the same variable) is a legitimate competing explanation for an observed effect and therefore must be included before causal inferences can be made in regard to the influence of other predictors of change over time. The inclusion of autoregressive effects in longitudinal modeling of intraindividual change is problematic because they tend to remove all potentially important predictors of change except those that predict changes in rank order of the observations over time. For example, in a monotonically stable growth process in which all individuals increase at a constant rate (i.e., linearly) while maintaining the same rank order, the important predictors of the individual slopes would be eliminated with the inclusion of autoregressive effects. The autoregressive model fails when intraindividual change is accompanied by high-rank-order stability over time (Stoolmiller & Bank, 1995). In addition, the autoregressive effect is questionable as a true causal effect, and researchers have argued that proponents of the application of autoregressive latent models in longitudinal modeling have misinterpreted the autoregressive effect as a parameter that represents true causal effect when it is in fact a stability coefficient that represents the boundary or initial values of the system. For more compre-

hensive discussions of the problems associated with including autoregressive effects in longitudinal modeling of intraindividual change, see Rogosa and Willett (1985) and Stoolmiller and Bank (1995).

The Multilevel Structure of Longitudinal Data The preceding section on modeling multilevel phenomena discusses the “traditional” type of multilevel data in which individuals are nested within groups. In modeling changes over time by using longitudinal data, we are in fact dealing with a type of multilevel data in which the multilevel structure is less obvious. Longitudinal data are obtained from measurements repeated on the same individuals over time, and hence a multilevel structure is established with the repeated observations over time (Level 1) nested within individuals (Level 2). Although the multilevel analysis of cross-sectional grouped data is concerned with interindividual differences associated with group membership, multilevel analysis of longitudinal data is concerned with modeling intraindividual change over time. Although multilevel regression models can also be used to analyze these changes over time (e.g., Bryk & Raudenbush, 1992), the issues of changes over time are often very complex and may involve facets of change over time (e.g., conceptual changes in the constructs, changes in calibration of measurement, various types of time-related error-covariance structures) that are not readily handled by multilevel regression models. In modeling change over time, we are primarily concerned with describing the nature of the trajectory of change and accounting for the interindividual differences in the functional forms or parameters of the trajectories by relating them to explanatory variables. The explanatory variables may be in the form of experimentally manipulated or naturally occurring groups, time-invariant predictors, time-varying correlates, or the trajectories of a different variable. Latent growth modeling and its extensions are well suited to address these issues. Chan (1998b) provided a detailed review of these issues and the application of latent growthmodeling techniques, as well as an overview comparison between latent variable models and multilevel regression models. 107

David Chan

Copyright American Psychological Association. Not for further distribution.

MODELING MEASUREMENT INVARIANCE ACROSS GROUPS In virtually all areas of I/O psychology, we often make direct comparisons between two or more groups of individuals (e.g., male vs. female, White vs. Black, supervisors vs. coworkers, Culture A vs. Culture B) in their responses to the same set of items or measures. On the basis of absolute differences in the scores on the measurement scale, substantive inferences are made about between-group differences in the level of the construct purportedly represented by the items or measures. The validity of these inferences is dependent on the often untested assumption that, across groups, the same items or measures are measuring the same construct and measuring it with the same precision. When this assumption of measurement invariance is in fact violated, absolute differences in scores between groups, and therefore inferences based on these differences, are likely to be misleading or not meaningful. Hence, measurement invariance is often a statistical hurdle that should be cleared before making direct between-group comparisons of scores. On the other hand, measurement invariance or lack thereof may also reflect or represent substantive between-group differences that are of theoretical interest. This section introduces the conceptual and data-analysis issues in measurement invariance and describes the various advances made in analytical strategies for modeling measurement invariance across groups. I first explicate the notion of measurement invariance and differential item functioning across groups. I then discuss the corresponding analytical strategies including item response theory (IRT) techniques and multiple-group means and covariance structures analysis.

Measurement Invariance and Differential Item Functioning Very often, I/O researchers are interested in betweengroup (e.g., gender) differences on latent trait variables such as personality traits or cognitive abilities. It is meaningful to make direct between-group comparisons only if there is measurement equivalence of item responses across groups. An item is said to be not equivalent (i.e., the item functions differentially) across groups when individuals with equal levels on 108

the test latent trait but different group membership respond differently to that item. If there is a lack of measurement equivalence on a substantial number of items on a measure, then we cannot even assume that the same construct is being assessed across groups by the same measure. To understand differential item functioning (DIF) across groups, we need to distinguish the type of item parameter that differs across groups. The two types of parameters are item difficulty and item discrimination. Conceptually, item difficulty is a location parameter that relates the latent trait (e.g., ability) and the mean item response (in the case of dichotomous items, this is the probability of a particular [correct] response). The location parameter corresponds to the value on the latent trait scale at which the mean item response is at a fixed value or alternatively, it corresponds to the mean item-response value at which the latent trait is at a fixed value. Whether value on the latent trait scale or the mean item response is fixed depends on the type of item-response model in question. In the context of IRT models, the convention is to fix the probability of correct item response (i.e., mean item response) at .50 so that the item difficulty parameter refers to the corresponding value on the latent trait scale. In the context of factor analytic item-response models, the convention is to fix the value on the latent trait scale at 0 so that the item difficulty parameter refers to the corresponding mean item-response value (Ferrando, 1996). Hence, the higher the value of the item difficulty parameter, the higher the latent trait level is required for an examinee to have a .50 probability of endorsing the particular response or, equivalently, the lower the item mean (or probability of correct response) is obtained for an examinee with a value of 0 on the latent trait scale. Conceptually, item discrimination refers to the extent to which the item is able to distinguish between individuals high and those low on the latent trait variable. The higher the discrimination parameter, the narrower the range of the latent trait scale the item is able to make this distinction. Uniform DIF exists when only the item difficulty parameter differs across groups, that is, when there is no interaction between ability level and group membership. Nonuniform DIF exists when the item discrimination parameter differs across groups,

Advances in Analytical Strategies

that is, when there is an interaction between ability level and group membership.

Copyright American Psychological Association. Not for further distribution.

Item-Response Theory Attempts to detect DIF based on classical test theory, such as transformed item difficulty and ANOVA methods, are technically flawed in that they confound item difficulty and item discrimination, and the use of these methods could create artifactual DIF and miss true DIF (for review, see Angoff, 1993). IRT methods solved these problems by distinguishing between the item difficulty and discrimination parameters. IRT models specify a nonlinear monotonic function to account for the relationship between individual level on the latent trait variable and the probability of a particular item response. IRT models were developed in the context of dichotomous items. Recently, a variety of polytomous IRT models have been proposed to account for responses on items with a polytomous ordered response format such as the Likert-type rating scale. However, as noted by Reise, Widaman, and Pugh (1993), these polytomous IRT models require very large sample sizes to obtain accurate parameter estimates. The likelihood ratio chi-square test, which is the only standard measure of model fit, is extremely sensitive to sample size so that with large sample sizes, most models will produce statistically significant chi-square values, resulting in the rejection of these models even if they are theoretically reasonable. Unlike methods such as confirmatory factor analysis, there are no practical fit indices in IRT methods to reduce the dependence on sample size when assessing model fit. In addition, there are no modification indices from a full measurement equivalent model to indicate the likely changes in model fit because of the freeing of the equality constraint (across groups) of each item parameter. However, it is possible to implement IRT models in the MULTILOG program (Thissen, 1991), which allows the user to create significance tests of the between-group difference in specific parameters. Asymptotically, these tests should lead to similar conclusions as using the modification indices except for any difference due to the nature of the underlying item-response model (i.e., IRT vs. factor analytic models).

The Mean and Covariance Structures Model The means and covariance structures analysis (MACS) model proposed by Sorbom (1974) provides a flexible and useful method for detecting DIF, both uniform and nonuniform, in polytomous items that are assumed to approximate a continuous scale. Within the MACS model, responses on items that use a polytomous-ordered response format such as the Likert-type rating scale (e.g., a job satisfaction measure that uses a 5-point Likert-type rating scale) are considered to be approximations of responses on a continuous line. Consider an item response, xij, which is the observed response of the individual i to item j, where x is a number on a continuous scale. Assume that the item responses on the test are explained by one latent trait variable or factor, ξ (i.e., the test latent trait; e.g., job satisfaction). The MACS model represents the relationship between x and ξ in a linear regression of x on ξ:

(

)

x (ijg ) = μ (jg ) + λ (jg ) ξ(i g ) + e (ijg ), e (ijg )⬃N 0, σ 2j ( g ) , (8) where the regression intercept, μj, is the mean response to item j when ξ is 0, the regression slope or coefficient, λj, is the expected change in the scale response per unit change in ξ or the “factor loading” for item j, eij is a stochastic error term, and g refers to group membership. Within the MACS item-response model, the item intercept corresponds with the item difficulty parameter and the item factor loading corresponds with the item discrimination parameter. (2) (G) Hence, the invariance of μj (i.e., μ(1) j = μj = . . . = μj ) (2) (G) and λj (i.e., λ(1) j = λ j = . . . = λ j ) across groups implies the absence of DIF, and the lack of invariance implies DIF. Specifically, a between-group difference in μj indicates uniform DIF (i.e., differences in item difficulty) and a between-group difference in λj indicates nonuniform DIF (i.e., differences in item discrimination). The advantage of the MACS model over the standard confirmatory factor analysis model is that unlike the latter, which estimates the item factor loading but not the item intercept, the former estimates both factor loading and intercept so that both nonuniform and uniform DIF can be examined. Reise et al. (1993) applied the stan-

109

Copyright American Psychological Association. Not for further distribution.

David Chan

dard confirmatory factor analysis model to the assessment of measurement equivalence (DIF) across groups and thus was able to examine nonuniform DIF (i.e., differential item discrimination) but not uniform DIF (i.e., differential item difficulty). Note that although between-group differences in unique factor variance for the jth item, σ 2j, provide information on differential precision across groups and hence are relevant to measurement equivalence, σ2j does not correspond with the two item parameters of substantive interest (item difficulty and item discrimination) in the DIF literature. Hence, in most applications of MACS analysis to DIF, the researcher is not concerned with the invariance (or lack thereof) of σ 2j across groups. Of course, by ignoring group differences in σ 2j, the researcher is testing a weaker form of invariance than would be the case if he or she had also tested for invariance in σ 2j. To detect uniform and nonuniform DIF, the MACS model is simultaneously fitted to the groups of interest (e.g., testing for DIF on a job satisfaction measure across an American sample and a Chinese sample of respondents). This is accomplished by fitting a hierarchical series of nested models to the item responses and systematically testing the statistical significance of between-group differences in item parameters. This modeling procedure can be easily implemented by using the familiar multiple-group covariance structure analysis (incorporating means as well) procedure in any of the popular SEM software. Issues of differential item functioning and measurement invariance across groups continue to be of great interest in I/O psychology. For reviews of the conceptual and statistical issues in measurement invariance, see Chan (1998b), Schmitt (1982), and Vandenberg and Lance (2000). CONCLUDING REMARKS This chapter has provided an overview of the advances in data analytical strategies, many of which are being applied in substantive research in I/O psychology and some of which are being actively debated. Undoubtedly, some of the discussions in this chapter will soon be outdated given the speed with which the technical sophistication of data analytic techniques is developing, catalyzed by powerful com110

puters and intelligent software programs. However, it is worthy to remind ourselves that no amount of sophistication in an analytical technique can turn invalid inferences resulting from inadequate design, measurement, or data into valid inferences. None of the analytical strategies discussed in this chapter, or anywhere else for that matter, has magical solutions for a study if there is poor research design, instrument development, and data collection. The substantive application of any analytical strategy has to be based on adequate theoretical foundations. APPENDIX: DATA ANALYSIS SOFTWARE AND INTERNET RESOURCES

Data Analysis Software Multilevel regression analyses may be performed by using widely available multipurpose statistical packages such as SAS, SPSS, and S-PLUS. For example, multilevel analyses can be performed by using the PROC MIXED procedure in SAS and the commands in the Advanced Models module in SPSS. Examples of programs specifically developed for multilevel regression analyses include HLM (Bryk, Raudenbush, & Congdon, 1996), MIXOR/MIXREG (Hedeker & Gibbons, 1994), and VARCL (Longford, 1987). Latent variable models can be specified and tested by using any of the widely available structural equation modeling programs such as AMOS (Arbuckle, 1999), EQS (Bentler, 2004), and LISREL (Jöreskog & Sörbom, 1996), although the procedures for multilevel latent models are somewhat difficult to implement at times because the programs were not specifically written for multilevel analyses. MPLUS (Muthén & Muthén, 2004) is a structural equation modeling program that has specifically incorporated features for estimating multilevel models and is well suited to specify and test a variety of different multilevel latent variable models, including mixture of latent class and latent growth models. Undoubtedly, the features of the above programs are likely to change as technology and knowledge change.

Internet Resources There are several useful Internet resources on multilevel and latent variable analysis. For multilevel

Copyright American Psychological Association. Not for further distribution.

Advances in Analytical Strategies

analysis, it is useful to begin by going to comprehensive websites that give a variety of information on multilevel research including publications, newsletters, workshops, multilevel datasets, software reviews, and useful links to other websites. Examples include the UCLA Multilevel Modeling Portal (http:// www.ats.ucla.edu/stat/mlm/) and the website of the Center for Multilevel Modeling (http://multilevel. ioe.ac.uk/index.html). The latter Web site provides a comprehensive list of references on multilevel modeling, an excellent set of reviews of computer software for performing multilevel analyses, and a library containing multilevel datasets that you can download for purposes of teaching and training in the application of multilevel models. There is also an active Internet discussion list where subscribers discuss conceptual and statistical problems in multilevel modeling ranging from elementary to advanced issues (http://www.jiscmail.ac.uk/lists/ multilevel.html). SEMNET is an excellent electronic mail network for anyone interested in discussing with researchers on any topics related to SEM and latent variable analyses. The website has an amazing archive of the discussions, organized by month dating back to 1993 (http://bama.ua.edu/archives/ semnet.html). The RMNET is a question-and-answer network for members of the Research Methods Division of the Academy of Management. The questions may be related to any research method issue concerning design, measurement, and data analysis. Subscribers to RMNET include a diversity of researchers, ranging from beginning graduate students to established scholars who have published on advances in analytical strategies. More information including how to join the RMNET are available on http://division. aomonline.org/rm/rmnet.html. Several researchers also maintained their personal websites on specific analytical strategies. For example, David Kenny’s website provides useful information on mediation analysis (http://davidakenny.net/cm/ mediate.htm) and Herman Aguinis’s website provides useful information on interaction analysis (http:// mypage.iu.edu/∼haguinis/mmr/iindex.html). Internet resources are updated very rapidly and the reader should stay abreast by using search engines

or contacting relevant professional organizations, such as Division 5 (Evaluation, Measurement and Statistics) of the American Psychological Association and the Research Methods Division of the Academy of Management.

References Aguinis, H. (2004). Regression analysis for categorical variables. New York: Guilford Press. Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Aitkin, M. A., Anderson, D., & Hinde, J. (1981). Statistical modeling of data on teaching styles. Journal of the Royal Statistical Society. Series A (General), 144, 419–461. doi:10.2307/2981826 Angoff, W. H. (1993). Perspectives on differential item functioning methodology. In P. W. Holland & H. Wainer (Eds.), Differential item functioning (pp. 3–23). Hillsdale, NJ: Erlbaum. Arbuckle, J. L. (1999). AMOS (Version 4.0) [Computer software]. Chicago: Smallwaters. Baron, R. M., & Kenny, D. A. (1986). The moderator– mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. doi:10.1037/00223514.51.6.1173 Barrick, M. R., Stewart, G. L., & Piotrowski, M. (2002). Personality and job performance: Tests of the mediating effects of motivation among sales representatives. Journal of Applied Psychology, 87, 43–51. doi:10.1037/0021-9010.87.1.43 Bentler, P. M. (2004). EQS structural equations program manual [Computer software manual]. Encino, CA: Multivariate Software. Berry, W. D., & Feldman, S. (1985). Multiple regression in practice. Newbury Park, CA: Sage. Bliese, P. D. (2000). Within-group agreement, nonindependence, and reliability: Implications for data aggregation and analysis. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations (pp. 349–381). San Francisco: Jossey-Bass. Bliese, P. D., Chan, D., & Ployhart, R. E. (2007). Multilevel methods: Future directions in measurement, longitudinal analyses and nonnormal outcomes. Organizational Research Methods, 10, 551–563. doi:10.1177/1094428107301102 Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley. Borman, W. C., & Motowidlo, S. J. (1993). Expanding the criterion domain to include elements of contextual performance. In N. Schmitt & W. C. Borman (Eds.),

111

David Chan

Personnel selection in organizations (pp. 71–98). San Francisco: Jossey-Bass. Box, G. E. P., & Jenkins, G. M. (1976). Time series analysis: Forecasting and control. San Francisco: Holden Day. Bryk, A., & Raudenbush, S. W. (1992). Hierarchical linear models. Thousand Oaks, CA: Sage. Bryk, A., Raudenbush, S. W., & Congdon, R. (1996). HLM: Hierarchical linear and nonlinear modeling with HLM/2L and HLM/3L programs. Chicago: Scientific Software.

Copyright American Psychological Association. Not for further distribution.

Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. (1993). A theory of performance. In N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations (pp. 35–70). San Francisco: Jossey-Bass. Chan, D. (1998a). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. doi:10.1037/00219010.83.2.234 Chan, D. (1998b). The conceptualization of change over time: An integrative approach incorporating longitudinal means and covariance structures analysis (LMACS) and multiple indicator latent growth modeling (MLGM). Organizational Research Methods, 1, 421–483. doi:10.1177/109442819814004

native. Organizational Behavior and Human Decision Processes, 58, 51–100. doi:10.1006/obhd.1994.1029 Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and mediation: A general analytical framework using moderated path analysis. Psychological Methods, 12, 1–22. doi:10.1037/ 1082-989X.12.1.1 Ferrando, P. (1996). Calibration of invariant item parameters in a continuous item response model using the extended Lisrel measurement submodel. Multivariate Behavioral Research, 31, 419–439. doi:10.1207/s15327906mbr3104_2 Golembiewski, R. T., Billingsley, K., & Yeager, S. (1976). Measuring change and persistence in human affairs: Types of change generated by OD designs. Journal of Applied Behavioral Science, 12, 133–157. Hedeker, D., & Gibbons, R. D. (1994). A random effects ordinal regression model for multilevel analysis. Biometrics, 50, 933–944. doi:10.2307/2533433 Hofmann, D. A., Griffin, M. A., & Gavin, M. B. (2000). The application of hierarchical linear modeling to organizational research. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations (pp. 467–511). San Francisco: Jossey-Bass.

Chan, D. (2005). Multilevel research. In F. Leong & J. Austin (Eds.), The psychology research handbook (2nd ed., pp. 401–418). Thousand Oaks, CA: Sage.

Hofmann, D. A., Jacobs, R., & Baratta, J. (1993). Dynamic criteria and the measurement of change. Journal of Applied Psychology, 78, 194–204. doi:10.1037/00219010.78.2.194

Chan, D. (2006). Interactive effects of situational judgment effectiveness and proactive personality on work perceptions and work outcomes. Journal of Applied Psychology, 91, 475–481. doi:10.1037/0021-9010.91.2.475

James, L. R. (1982). Aggregation bias in estimates of perceptual agreement. Journal of Applied Psychology, 67, 219–229. doi:10.1037/0021-9010.67.2.219

Chan, D., & Schmitt, N. (1997). Video-based versus paperand-pencil method of assessment in situational judgment tests: Subgroup differences in performance and face validity perceptions. Journal of Applied Psychology, 82, 143–159. doi:10.1037/0021-9010.82.1.143

James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85–98. doi:10.1037/00219010.69.1.85 Jöreskog, K., & Sörbom, D. (1996). LISREL 8 user’s reference guide [Software manual]. Chicago: Scientific Software.

Chan, D., & Schmitt, N. (2000). Interindividual differences in intraindividual changes in proactivity during organizational entry: A latent growth modeling approach to understanding newcomer adaptation. Journal of Applied Psychology, 85, 190–210. doi:10.1037/0021-9010.85.2.190

Judd, C. M., & Kenny, D. A. (1981). Process analysis: Estimating mediation in treatment evaluations. Evaluation Review, 5, 602–619. doi:10.1177/ 0193841X8100500502

Chou, C. P., Bentler, P. M., & Pentz, M. A. (1998). Comparisons of two statistical approaches to study growth curves: The multilevel model and the latent curve analysis. Structural Equation Modeling, 5, 247–266.

Klein, K. J., Conn, A. B., Smith, D. B., & Sorra, J. S. (2001). Is everyone in agreement? An exploration of within-group agreement in employee perceptions of the work environment. Journal of Applied Psychology, 86, 3–16. doi:10.1037/0021-9010.86.1.3

Collins, L. M., & Wugalter, S. E. (1992). Latent class models for stage-sequential dynamic latent variables. Multivariate Behavioral Research, 27, 131–157. doi:10.1207/s15327906mbr2701_8

Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations (pp. 3–90). San Francisco: Jossey-Bass.

Edwards, J. R. (1994). The study of congruence in organizational behavior research: Critique and a proposed alter112

Advances in Analytical Strategies

Kreft, I. G. G., & de Leeuw, J. (1998). Introducing multilevel modeling. London: Sage. Kristof, A. l. (1996). Person–organization fit: An integrative review of its conceptualizations, measurement, and implications. Personnel Psychology, 49, 1–49. doi:10.1111/j.1744-6570.1996.tb01790.x

Copyright American Psychological Association. Not for further distribution.

Larsen, R. (1990). Spectral analysis. In A. von Eye (Ed.), Statistical methods in longitudinal research (Vol. 2, pp. 319–350). Boston: Academic Press. LeBreton, J. M., Wu, J., & Bing, M. N. (2009). The truth(s) on testing for mediation in the social and organizational sciences. In C. E. Lance & R. J. Vandenberg (Eds.), Statistical and methodological myths and urban legends (pp 109–144). New York: Routledge. Longford, N. T. (1987). VARCL: Software for variance component analysis of data with hierarchically nested random effects (maximum likelihood). Princeton, NJ: Educational Testing Service. MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130–149. doi:10.1037/1082-989X.1.2.130 MacKinnon, D. P. (2008). Introduction to statistical mediation analysis. Mahwah, NJ: Erlbaum. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83–104. doi:10.1037/ 1082-989X.7.1.83 Maruyama, G. M. (1998). Basics of structural equation modeling. Thousand Oaks, CA: Sage. Mathieu, J. E., DeShon, R. P., & Bergh, D. D. (2008). Mediation inferences in organizational research. Organizational Research Methods, 11, 203–223. doi:10.1177/1094428107310089 Morgeson, F. P., & Hofmann, D. A. (1999). The structure and function of collective constructs: Implications for research and theory development. Academy of Management Review, 24, 249–265. doi:10.2307/ 259081 Muthén, B. O. (1994). Multilevel covariance structure analysis. Sociological Methods & Research, 22, 376–398. doi:10.1177/0049124194022003006 Muthén, B. O. (1997). Latent variable modeling of longitudinal and multilevel data. In A. Raftery (Ed.), Sociological methodology (pp. 453–480). Boston: Blackwell. Muthén, B. (2004). Latent variable analysis: Growth mixture modeling and related techniques for longitudi-

nal data. In D. Kaplan (Ed.), Handbook of quantitative methodology for the social sciences (pp. 345–368). Newbury Park, CA: Sage. Muthén, B. O., & Muthén, L. (2004). Mplus user’s guide [Software manual]. Los Angeles: Muthén & Muthén. Ostroff, C. (1993). Comparing correlations based on individual level and aggregated data. The Journal of Applied Psychology, 78, 569–582. doi:10.1037/ 0021-9010.78.4.569 Reise, S. P., Widaman, K. F., & Pugh, R. H. (1993). Confirmatory factor analysis and item response theory: Two approaches for exploring measurement invariance. Psychological Bulletin, 114, 552–566.doi:10.1037/0033-2909.114.3.552 Robinson, W. S. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15, 351–357. doi:10.2307/2087176 Rogosa, D. R., & Willett, J. B. (1985). Understanding correlates of change by modeling individual differences in growth. Psychometrika, 50, 203–228. doi:10.1007/BF02294247 Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level and cross-level perspectives. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (pp. 1–7). Greenwich, CT: JAI Press. Sackett, P. R. (2002). The structure of counterproductive work behaviors: Dimensionality and relationships with facets of job performance. International Journal of Selection and Assessment, 10, 5–11. doi:10.1111/ 1468-2389.00189 Schmitt, N. (1982). The use of analysis of covariance structures to assess beta and gamma change. Multivariate Behavioral Research, 17, 343–358. doi:10.1207/s15327906mbr1703_3 Schneider, B., Salvaggio, A. N., & Subirats, M. (2002). Climate strength: A new direction for climate research. The Journal of Applied Psychology, 87, 220–229. doi:10.1037/0021-9010.87.2.220 Sorbom, D. (1974). A general method for studying differences in factor means and factor structures between groups. British Journal of Mathematical and Statistical Psychology, 27, 229–239. Stoolmiller, M., & Bank, L. (1995). Autoregressive effects in structural equation models: We see some problems. In J. M. Gottman (Ed.), The analysis of change (pp. 261–278). New Jersey: LEA. Thissen, D. (1991). MULTILOG: Multiple, categorical item analysis and test scoring using item response theory. Chicago, IL: Scientific Software, Inc. Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature. Organizational Research Methods, 3, 4–69. doi:10.1177/109442810031002 113

CHAPTER 5

ORGANIZATIONS: THEORY, DESIGN, FUTURE

Copyright American Psychological Association. Not for further distribution.

George P. Huber

This chapter examines three subjects: organization theory, organization design, and organizations in the future. Organization theory is a field of study and knowledge. Organization design is a routine carried out primarily by managers but sometimes by interested parties with the motivation, authority, and resources to establish the organization, such as legislators or venture capitalists. Examination of the changing nature of organizational environments provides an opportunity to consider the contributions and limits of organization theory and organization design. With an acknowledgement of my bounded knowledge and resources and the consequences thereof and without any intention, the chapter reflects the influence primarily of literature published in North America and, to a more limited extent, in Western Europe. ORGANIZATION THEORY Organization theory, as the term is currently used, is a field of study and knowledge centrally concerned with explaining the nature and circumstances of individual organizations and collectives of organizations. (Here, by an organization’s nature, I mean the attributes of its leadership, strategy, core process, structure, employees, culture, and routines and the propensities and properties associated with combinations of these features, such as its competences. By its circumstances, I mean the current direction and speed

of change of its size, maturity, performance, and status with regard to survival.) Why is this population of organizations growing or declining? Why is this organization performing so well or poorly? Why are some organizations in this industry rigidly structured, whereas others are loosely structured? Instrumental to explaining the nature and circumstances of organizations, organization theorists are concerned with determining (a) what factors influence the founding, effectiveness, persistence, and demise of individual organizations and collectives of organizations and in what ways do these factors exert their influence, and (b) why and how do organizations and collectives of organizations change. The explanations associated with such questions are frequently referred to as theories. Some wellknown works in the field (Hage, 1965; March & Simon, 1958; Mintzberg, 1979, 1983; Williamson, 1991) contain sets “of related propositions that specify relationships among variables,” a condition that satisfies accepted definitions of social science theory (cf. Kerlinger, 1986, p. 9). Here, I often use the term theory in place of explanation because such use in the field is commonplace (cf., Walsh, Meyer, & Schoonhoven, 2006) and because its range of meanings encompasses all of the explanations to be examined (Random House Webster’s Unabridged Dictionary, 2001). Because organizational behavior is sometimes used to mean the field of study and knowledge focused on the behavior of humans and

I acknowledge the help of James Fredrickson, Pamela Haunschild, Andrew Henderson, Benjamin Herndon, David Jemison, Martin Kilduff, Jørn Flohr Nielsen, Francisco Polidoro, and members of The American Psychological Association Handbook of Industrial and Organizational Psychology’s editorial board, all of whom read all or a substantial portion of an earlier version of the chapter and made numerous useful suggestions.

http://dx.doi.org/10.1037/12169-005 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

117

Copyright American Psychological Association. Not for further distribution.

George P. Huber

human groups in organizations but is also sometimes used more broadly to include this field of study and also the field of organization theory, I use organization science to refer to the field of study encompassing both organizational behavior and organization theory. Organization studies is a term generally used interchangeably with organization theory but sometimes used interchangeably with organization science. Although philosophers and people active in governmental and other influential societal institutions have theorized about the nature of organizations for thousands of years (March, 1965, p. ix; Exodus 2:18 [King James Version]; Starbuck, 2003, pp. 146–147), much of what is regarded as organization theory has been developed during the last 5 decades (Augier, March, & Sullivan, 2005; March, 2007). Paraphrasing March (1965, p. ix), March (2008) noted that “the study of organizations could claim a chronology but not a pedigree” (p. 317), that is, the study of organizations does not reflect a series of causally connected events over time. Rather, as I demonstrate, although bodies of work have often drawn on previous work, new bodies of work have often been prompted by other stimuli, for example, by observations or reports of changes in the characteristics or effectiveness of groups of organizations or by discomfort with existing beliefs or with voids in the literature. During much of this 5-decade period, the field has been in a transition. Historically, the field was dominated by its mother science, sociology. As described by Hinings (see Hickson et al., 1988), in the 1960s began the shift of those studying organizations (at the macro level) from departments of sociology to business schools/ management centres. Initially this was a shift of people from one to the other . . . [but] it has increasingly been an institutionalized split with the management centres producing their own intellectual product, the Ph.D. in organizational analysis. The professional basis for the study of organization has changed drastically. . . . Organizational theory is no longer the sociology of 118

organizations. . . . It still draws on sociology, but to an ever decreasing extent. . . . The paradigm shift is from sociology to organization theory. (p. 2) The state of the matter, as described by Hinings in 1988 (Hickson et al., 1988), is today much more apparent than it was 2 decades ago. Writing in 2005, Augier et al. (2005) asserted that “by 1980, a field had been defined . . . and the center of gravity of organization studies had moved decisively to a business school locale” (p. 356). In an article important to those interested in the path dependency of the field, the authors also explained how this shift in institutional focus was accompanied by a geographical shift as well: the development of a dominant and parochial organization science community “in Anglophone North America” that was “involved the intertwining of history, time, place, and agency” (p. 357; see also March, 2004, 2007). It seems worth noting that in parallel with the movement of organization studies to business schools, the scope of organization theory came to include some theories based on economic concepts. Perhaps the theories found a welcoming environment. Perhaps the environment fostered the theories. The remainder of this Organization Theory section of the chapter first discusses the nature of organizations, then describes the more prominent organization theories, and concludes by examining some issues currently active in the field. This is followed by the chapter’s section on Organization Design.

Nature of Organizations This subsection, on the nature of organizations, first formally defines organizations and the concepts that organization theorists use in their discourse about organizations. It then describes the three most widely referenced perspectives on organizations: the rational system perspective, the natural system perspective, and the open system perspective. Organizational characteristics. Organizations are socially constructed, goal-directed, boundarymaintaining, hierarchically differentiated, open systems of human activity. Socially constructed indicates that living entities interact in shaping and reshaping

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

the system. Goal directed implies that although different components of the organization may have multiple and different goals, each component has at least one assigned activity that contributes to the focal goal of either continued organizational survival or a specific achievement. Boundary maintaining refers to the organization’s authorized routine of distinguishing between members and nonmembers. Hierarchically differentiated1 means that one or more hierarchical components have the authority to assign activities to one or more hierarchically subordinate components. Open systems are systems that interact with their environment. Although the open systems perspective is currently dominant in the field, the field actually originated within the closed system perspective to be described shortly. Among the organizational features that are most prominent in the organization science literature and in the business and lay presses are the organization’s leadership, strategy, core process, structure, employees, culture, and routines. I examine these features, and the various attributes of individual features, in depth when I turn later in the chapter to the subject of organization design. Specific attributes of features influence organizational outcomes. For instance, many studies have contributed to the widely held belief among organization theorists that an organizational structure (structure being an example feature) that is highly formalized (formalization is an example attribute of structure) is associated with high performance (high performance is an example outcome) in stable organizational environments. Propositions such as this could form the basis of guidelines for the practice of organization design. Traditional perspectives on organizations. Driven by the varying interests, observations, and interpretations in the organization science community, organization theorists have come to acknowledge different perspectives on organizations, each lending itself to a different emphasis when thinking about organizations (Astley & Van de Ven, 1983; Clegg, Hardy, Lawrence, & Nord, 2006; Scott & Davis, 1

2007; Ulrich & Barney, 1984). I describe in the following subsections the three that are most commonly referenced: the rational system, natural system, and open system perspectives. Each is well grounded in the traditional organization theory literature. Reviewing these perspectives provides a sense of the history of the field and a context for the organization theories to be examined subsequently. Unlike theories, perspectives need not be scientifically validated to endure. For a perspective to be accepted into the lexicon of the field, it is necessary only that many of the field’s members believe that the perspective’s defining characteristics are or were manifested in a large number of organizations. Many in the organization theory community are comfortable with accepting more than one or two of these perspectives as reasonable descriptions of what organizations are. Rational system perspective. The rational system perspective emphasizes the pursuit of specific focal goals, such as profitability, and the rationalization and formalization of processes and structures for attaining these goals: “Organizations are collectivities oriented to the pursuit of relatively specific goals and exhibiting relatively highly formalized social structures” (Scott & Davis, 2007, p. 29). The perspective calls attention to the purposefulness of many organizational endeavors and is part of the worldview of many managers. Early authorities drew on their experience and observations to identify organizational attributes that they believed contributed to performance (Fayol, 1949; Gulick & Urwick, 1937; Mooney & Reily, 1947; Taylor, 1911; Weber, 1947) and, in so doing, generally manifested this perspective. Examples of these organizational attributes included unity of command (employees reporting to only one manager) and unambiguous lines of authority. Prominent sociologists also included the idea of focal goals in their definitions of organizations (Parsons, 1956, p. 64; Perrow, 1970, p. 171). The rational system perspective has been criticized for not acknowledging issues such as the diversity of

J. G. Miller (1978) provides the most extensive analysis of the formal distinction between a group and an organization (see also J. G. Miller, 1972). Following J. G. Miller (1978), if in a collective of cooperating specialists, no subset of one or more members has formal authority over any other subset of one or more members, the collective is a group. If a subset of one or more the collective’s members has formal authority over one or more other members, even if that authority is limited in its realm—for example to matters in the realm of the subset’s specialty—the collective is an organization.

119

Copyright American Psychological Association. Not for further distribution.

George P. Huber

member’s goals, environmental forces, information deficiencies, and the bounded rationality of decision makers. The perspective does not deny these issues, central to the natural system and open system perspectives described below. It simply does not attend to them. The rational system perspective served as an important foil in the classic works by March and Simon (1958) and Cyert and March (1963), in which these authorities argued that (in contrast to the rational system perspective) members’ goals vary, organizational goals tend not to be entirely operational, and the rationality of decision makers and organizations is quite limited. Natural system perspective. In keeping with the observations of March and Simon (1958) and the varied insights of other early writers (Gouldner, 1954; McGregor, 1960; Perrow, 1961), it is clear that individual organizational members have multiple goals, goals that vary over time and situations, and goals that tend to vary from the goals of other members. Consequently, it is reasonable to expect some behaviors that are not in keeping with the goals of other members or that are not directly aligned with the focal goals of the organization. However, people generally join organizations to satisfy personal goals that they cannot satisfy without the assistance of other people. Thus, an organization is “a coalition of groups and interests, each attempting to obtain something from the collectivity by interacting with others, and each with its own preferences and objectives” (Pfeffer & Salancik, 1978, p. 52). Nevertheless, to attain and retain membership in the organization and its work groups, members must carry out some activities in pursuit of the organization’s focal goals (or of subordinate goals instrumental to achieving the focal goals). As a consequence of these considerations, the natural system perspective is that “organizations are collectivities whose participants are pursuing multiple interests, both disparate and common, but who recognize the value of perpetuating the organization as an important resource” (Scott & Davis, 2007, p. 30). Open system perspective. Neither of these first two perspectives acknowledges the importance of the organization’s external environment. This absence invites the open system perspective, famously and elaborately articulated by Katz and 120

Kahn (1966). The open system perspective emphasizes that organizations depend on their environments for their basic resources, such as employees; materials with which to produce goods or services; and monies from customers or clients or legislatures with which to pay employees, purchase materials, or carry out the functions that ensure the organization’s legitimacy (Pfeffer & Salancik, 1978). The perspective also emphasizes the interdependent, exchange-based relationship between organizations and their environments: Organizations produce goods and services that other organizations and societal components require (thereby acquiring the resources on which their survival depends). An organization–environment exchange of great importance is the flow of human resources, as people enter and leave various levels and forms of membership in the organization. This exchange of people between the organization and its environment has considerable effect on the organization’s ability to satisfy the demands of external stakeholders, as new members can bring new and necessary knowledge and skills, whereas departing members can take knowledge and skills with them. A third emphasis of the perspective is that organizations are shaped and constrained by regulative, normative, and cultural–cognitive environmental forces in their environment. Attention to these phenomena led Scott and Davis (2007) to the following definition of the open system perspective: “Organizations are congeries of interdependent flows and activities linking shifting coalitions of participants embedded in wider material-resource and institutional environments” (p. 32). I turn now to describing the more prominent organization theories. The first seven of the theories explain the nature or circumstances of organizations in terms of cause–effect relationships. In this way, they differ from the above perspectives on organization, which merely describe and define organizations. Elements of the open systems perspective appear in all of these theories, and elements of each of the other two perspectives are visible in some. The theories generally differ from one another with regard to their independent variables and in some instances with regard to their dependent variables.

Organizations: Theory, Design, Future

Copyright American Psychological Association. Not for further distribution.

Organization Theories Organization theories cannot be ordered with certainty on an ordinal time scale because their dates of origin are unclear. For the most part, the theories did not evolve from previous organization theories, but rather they evolved from theories in other fields, followed from new insights, or were prompted by observations and interpretations of changes in organizational environments or in organizational properties or processes. The theories can be roughly ordered on continuous or nominal scales associated with three types of variables. One is the unit of analysis, proceeding on a nominal or categorical scale from communities of different types of organizations downward in scope to individual organizations (which may be subunits of larger organizations). Another scale is the dependent variable (e.g., survival vs. demise, performance, or simply a categorically assessed demographic feature such as structural form). The third scale is the amount of influence that powerful actors can have on organizational features or performance. The ordering of the following descriptions approximates the orderings of the theories on these scales. After describing the individual theories, I provide a summary and a discussion of controversial issues. Population ecology theory and the evolutionary view. The population ecology and evolution approaches to the study of organizations are concerned with collectives of organizations as the unit of analysis. An example of a collective is a population of organizations. A population consists of all the organizations within a particular geographical and temporal boundary that have a common form, in which form generally refers to a distinguishing attribute or a particular collection of organizational attributes (usually, but not necessarily, structural attributes). Another example of a collective is a community of organizations. A community consists of a set of organizations having different forms and collaborating to construct “a regulated controlled social environment that mediates the effect of the natural environment” (Astley & Van de Ven, 1983, p. 251). An organizational field is defined similarly (DiMaggio & Powell, 1983, p. 143).

Population ecology portrays the organizational landscape as consisting of diverse populations (or subpopulations) in competition for scarce resources such as members, capital, and legitimacy (Hannan & Carroll, 1992; Hannan & Freeman, 1989). Some populations possess combinations of attributes well suited to obtaining particular resources within the prevailing environmental conditions and will tend to survive, at least until the conditions change. Other populations, competing for the same resources and subject to the same constraints, do not fit the environmental conditions as well and will tend not to survive. Thus, population ecology theory explains the nature and circumstances of populations in terms of the suitability of the attributes of the organizations in the population relative to the attributes of organizations in competing populations. In its early formulation (Hannan & Freeman, 1977), the approach posited that organizations are so inertial relative to environmental changes that it is hardly necessary to consider the possibility of organizational adaptation to the changed environment. But the focus of the theory has been broadening. Lest this brief discussion portray the current state of the theory too narrowly, I note Hannan and Freeman’s more recent observation that “the current diversity of organizational forms reflects the cumulative effect of a long history of variation and selection, including the consequences of founding processes, mortality processes, and merger processes” (Hannan & Freeman, 1989, p. 20), and Aldrich and Reuf’s (2006) observation—alluding to the work of Barnett and Carroll (1995) and Dobrev, Kim, and Carroll (2002)—that “ecological models now explicitly include transformation processes” (Aldrich & Reuf, 2006, p. 37; see also Singh, House, & Tucker, 1986). Like population ecology, the evolutionary view deals with variation and selection. However, this view is broader in the scope of issues included. For example, it is more explicit about how variation comes to be; it calls attention to the conditions under which newly founded organizations, versus existing organizations, are the source of variation in organizational forms. “If organizations are relatively inert after they are created, then new organizations 121

Copyright American Psychological Association. Not for further distribution.

George P. Huber

are the primary source of variety in populations. . . . If, however, organizations change significantly and frequently over their life course, then existing organizations are the major source of diversity in populations” (Aldrich & Reuf, 2006, p. 13). Possible sources of variation from existing organizations are (a) changes required by environmental events such as the availability of new technologies; (b) changes imposed by agentive and powerful actors in and near the organization, possibly reacting to intrapreneurial initiatives; (c) new entrants into the organization (March, 1991); and (d) idiosyncratic path-dependent development of organizational routines (Nelson & Winter, 1982). The evolutionary view is also more elaborate than population ecology in that it emphasizes the adaptation and persistence of populations. Populations adapt through experiencing the demise of those of its member organizations whose distinguishing attribute or particular collection of organizational attributes differed from, and were less effective than, the attributes of those of its member organizations that survive. In that some members survive, the population is said to persist even though its membership is somewhat different.2 Institutional theory. Beliefs about the appropriate values to hold and norms of behavior are present to some degree in all social systems. When the beliefs about values and norms are widely and intensely held, one speaks of the systems as possessing “strong cultures” (O’Reilly and Chatman, 1996, p. 158). Institutionalization is the process by which such beliefs come to take on “a rule-like, social fact quality” (Aldrich & Ruef, 2006, p. 39), a “taken for granted” quality (Meyer & Rowan, 1977; DiMaggio & Powell, 1991). Such beliefs often exert strong influence and are frequently thought of as institutional forces. The behaviors associated with institutional beliefs need not be interpersonal; they can be related to how work gets done or the ways things are done. Thus, institutionalization, in the case of 2

organizations, can also refer to “the processes by which an organization ‘takes on a special character’ and ‘achieves a distinctive competence or, perhaps, a trained or built-in incapacity’ (Selznik, 1996)” (Scott & Davis, 2007, p. 73). Associated with institutional theory is the belief that organizations and their leaders generally have very little discretion in choosing their organization’s nature or actions. In support of this belief, institutional theorists point to the constraints that result from the prior institutionalization of hard-to-change norms and from the legal and moral requirements that society imposes through entities such as regulatory bodies, special interest groups, and trade and professional associations. Simply put, institutional theory explains the nature of an organization or of a population as a consequence of its being shaped principally by institutional forces. From its early days (DiMaggio & Powell, 1983; Meyer & Rowan, 1977) institutional theory has been undergoing considerable revision and elaboration (Haunschild & Chandler, 2008; Scott, 1987, 2008), to the point at which institutional theorists make a stark distinction between old and new institutionalism (also known as neoinstitutionalism; DiMaggio & Powell, 1991; Scott, 2008; Selznik, 1996). However, both the old and new approaches share a skepticism toward rational-actor models of organization, and each views institutionalization as a state-dependent process that makes organizations less instrumentally rational by limiting the options they can pursue. Both emphasize the relationship between organizations and their environments, and both promise to reveal aspects of reality that are inconsistent with organizations’ formal accounts. Each approach stresses the role of culture in shaping organizational reality. (DiMaggio & Powell, 1991, p. 12)

When considering how a population can adapt, that is, change, and still be considered to persist, that is, be the same population, it may be instructive to consider that the experiences of individual organizations often result in adaptations that manifest themselves as new organizational attributes. Even though the organization is thus different, it still retains its identity; it is still the same organization. The adaptation and yet persistence of populations is analogous. Retention of lessons learned experientially by an individual organization is captured in the concepts of organizational learning and memory (Argote, 1999; Huber, 1991, 2004). Frequent reference to these concepts in the organization theory literature suggests that they are readily understood or at least accepted. The idea that populations learn and have memory is less frequently referenced and perhaps is less well understood.

122

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

Drawing on their review of recent work, Haunschild and Chandler (2008) concluded that such contemporary institutional theorists see “institutions as providers of the framework within which actors are able to define and pursue their interests” (p. 631) and that these theorists also believe “firms are not always passive recipients of institutional forces, but have the strategic potential both to select from, and also to influence and change, the institutional logics that are prevalent in their environment when it is in their best interest to do so” (p. 632). The future of this more liberal perspective within institutional theory is uncertain. With the exceptions noted previously, population ecology, the evolutionary view, and institutional theory explain the nature and circumstances of organizations primarily as consequences of environmental forces. The next five theories to be discussed explain organizational natures and circumstances partly or largely as consequences also of the intentions of powerful human actors. Resource dependence theory. Resource dependence theory leapt into prominence with the publication of The External Control of Organizations: A Resource Dependence Perspective (Pfeffer & Salancik, 1978). “The resource dependence approach focuses on strategic actions undertaken by organizations to manage interdependencies with other organizations in their environment” (Aldrich & Ruef, 2006, p. 50). This focus follows from the idea that an organization’s survival depends on its ability to acquire and maintain resources and that many of these resources are controlled by other organizations. Organizations are not autonomous, but are constrained by a network of interdependencies with other organizations. Interdependence, when coupled with uncertainty about what the actions will be of those with which the organization is interdependent, leads to a situation in which survival and continued success are uncertain, and therefore, organizations take actions to manage external interdependencies, although such actions are inevitably never completely successful and produce new patterns of

dependence and interdependence. (Pfeffer, 1997, p. 63) Thus, a central theme in resource dependence theory is that the environment, as represented by other organizations, is a dominant determinant of organizational actions or nonactions. A second central theme is the importance of power (the obverse of dependence) for understanding both intraorganizational and interorganizational behavior. With regard to intraorganizational processes, an internal entity’s power is viewed as arising from the entity’s capability for coping with critical organizational uncertainties. With regard to interorganizational relations, power is viewed principally as the control by one organization over resources needed by another organization (Aldrich, 2008). Alongside this theme is the idea that much of an organization’s activity is directed at controlling these relationships so as to satisfy its needs for resources. “The emphasis on power, and the careful analysis of repertoires for pursuing it, is the distinctive hallmark of resource dependence theory” (Scott & Davis, 2007, p. 233). From the above, one can conclude that resource dependence theory explains an organization’s circumstances as outcomes of the organization’s actions and success (or lack thereof) in attaining power over those organizations that possess resources on which the organization depends. Transaction cost theory. Transaction cost theory, sometimes called transaction cost economics, attempts to explain why organizations conduct some of their activities inside of their boundaries and have other activities conducted by external entities (Williamson, 1981, 1994). Employees and departments are costly to house and maintain and are not always fully used. Why not contract with an outside entity to do the work when needed? Or, for that matter, why not contract with an outside entity that specializes in producing the needed component product or service and can therefore do it faster, better, or cheaper? One answer to such questions is that contracts are costly to create, monitor for compliance, and enforce. The outside entity, called an agent, is assumed to have goals that do not coincide with those of the organization, so the organization is at greater risk when contracting its work to an agent than when 123

Copyright American Psychological Association. Not for further distribution.

George P. Huber

assigning its work to an employee or internal unit. This causes organizations to include in the contract incentives intended to motivate the agent to fulfill the organization’s goals. The exchanges of goods and services, and compensation for these, across the organization’s boundary are transactions, and the monitoring, enforcement, and risks associated with contracts involve costs. The explanations for when and how best to contract out work can be set forth as theoretical propositions that could be said to constitute a theory (see David & Han, 2004, pp. 41–42, for propositions based on their review of empirical studies), and thus, the term transaction cost theory. Drawing on Granovetter (1985) and Nilakant and Rao (1994), Aldrich and Ruef (2006) pointed out that transaction cost economics “may overstate the role of individually oriented economic incentives in organizations and understate the role of social exchange: reciprocity, cooperation, and trust” (p. 57). The second answer to the question of why not contract out work is that the purchased product may not achieve its full value to the firm until it is integrated into the organization’s operating system. If the fitting process is complex or ambiguous, it is often the case that the product’s features will be more integration-ready if those receiving and producing the product or service had worked closely together (e.g., worked in the same organization) during the product’s creation. Clearly, the risk involved in contracting work out is not simply a consequence of the difference in the goals of the organization and the agents but is also a consequence of the efficacy of the technology for integrating the products generated by the component units. Transaction cost theory focuses on the decision and contracts regarding where the organization’s work is done: inside or outside of the organization. It explains the boundaries of an organization in terms of past choices on the basis of the relative costs and risks associated with contacting out work versus having the work done in the organization. A third answer to the question of why not contract out work is that in some situations a network form of organization among independent entities is more desirable than either having work in house or contracting in a market context (Powell, 1990; Uzzi, 1996). For example, a network member’s fulfillment 124

of responsibilities to the network may determine the member’s continued membership in the network or the member’s future opportunities to participate in like networks, either of which may be valuable resources. Contingency theory and congruence theory. I noted earlier several organizational features that have been shown to be related to organizational effectiveness (leadership, strategy, core process, structure, employees, culture, and routines). Donaldson (2001) stated that “the essence of the contingency theory paradigm is that organizational effectiveness results from fitting features of the organization, such as its structure, to contingencies that reflect the situation of the organization (T. Burns & Stalker, 1961; Lawrence & Lorsch, 1967; Pennings, 1992; Woodward, 1965)” (p. 1), in which “a contingency is any variable that moderates the effect of an organizational feature on organizational performance” (p. 7). Donaldson (2001, p. 4) pointed out that contingency theories exist for structure, leadership, human resource management, strategic decision making, and other organizational features. The contingency theory of most interest to organization theorists is structural contingency theory, and within that theory the most studied contingencies are organization’s environment, size, and strategy. “Some combinations of the contingency and organizational structure are better than others for performance. For each level of the contingency variable, there is a level of the organizational structural variable that produces the highest performance and thereby constitutes the fit” (Donaldson, 2001, p. 185; see also Pennings, 1992, and Pfeffer, 1997). For example, environmental dynamism moderates the relationship between the rigidity of an organization’s structure and the organization’s performance; in turbulent environments, rigid structures are insufficiently adaptable and rigidly structured organizations perform poorly (T. Burns & Stalker, 1961; Mintzberg, 1979). In another example, where the feature is the organization’s human resource routines, societal norms and labor market conditions (both components of the environment) moderate the relationships between an organization’s human resource routines and its success in hiring and

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

retaining employees; when the labor market contains few qualified workers, organizations must apply recruiting programs that are more creative or energetic or provide employment opportunities that are more attractive (Ployhart, 2006). Professionals in the area of organization design are prone to communicate with their clients or readers using a congruence model, a box and arrow model with the boxes containing organizational features such as those considered here but often also including leadership and the environment and with the arrows showing the principal relationships (Burke, 2002; Burton & Obel, 2004; Burton, DeSanctis, & Obel, 2006; Mackenzie, 1986; Peters & Waterman, 1982, pp. 9–10). These models and writings generally include the concept of fit, drawing on the contingency theory concept of an attribute level and a performance level moderated by a contingency. They also include within the concept of fit what is called congruence theory. Congruence theory holds that there are interdependencies among organizational features such that for each level of an attribute of one of the features, there is a level of an attribute of an interdependent feature that results in the highest performance. For example, employee performance depends in part on the organization’s human resource routines and in part on the organization’s culture, and because these two features are interdependent, the attribute levels of the two features must be congruent to obtain the highest employee performance (Nadler & Tushman, 1997). Thus, the essence of congruence theory is that organizational effectiveness results from congruence among multiple attributes of the organization. Environmental forces select out configurations (Meyer, Tsui, & Hinings, 1993) of attributes that result in unacceptable performance levels, leaving on the organizational landscape configurations that become named and included in the management and organizational literature as parts of typologies (Gresov & Drazin, 1997). I discuss these matters in more detail in the Organizational Design section of the chapter. In view of the above, it appears that contingency theory and congruence theory explain the nature of an organization as a consequence of the organization possessing a configuration of attributes that sufficiently satisfy the contingencies and that are

sufficiently congruent with each other that the organization is surviving. The processes through which an organization obtains its attributes are not specified and could include managerial initiatives or organizational responses to exogenous and endogenous events or to environmental or institutional pressures. Network theory. Network theory has a long history of empirical studies. By the 1950s, communication networks were used by psychologists in small-group experiments to investigate effects of the distribution and use of information (Leavitt, 1951; Shaw, 1954), and graph theory was a fashionable research topic in management science (Harary, 1959; Luce, 1950). In the 1970s, survey data were used to investigate the distribution of information and influence among scientists in industrial research laboratories (Allen, 1977; Tushman, 1978). A good deal of network research has focused on the sharing of information among individuals within and across organizations or organizational subunits (Granovetter, 1973; Hansen, 1999; Nebus, 2006). With the explosion of network (or virtual) organizations in the business world, much recent work has focused on interorganizational relationships, particularly the transfer of knowledge, but also the transfer of trust, reputation, and other resources (Lavie, 2006; Stuart, 2000). With the advent of computers, the possibility of doing sophisticated analyses of the flow of information and other resources in large networks accelerated the application of the theory to organizations (Burt, 1992; Human & Provan, 1997; Reagans & McEvily, 2003; Rosenkopf & Padula, 2008), with ideas about the advantages of brokerage and closure in networks (Burt, 2007) and centrality (Bonacich, 1987) becoming dominant in explaining both the formation and the outcomes of various interorganizational arrangements. Network theory generally explains organizational performance as a function of the organization’s network structure and of the attributes of the network’s nodes (Provan & Kenis, 2008). That network analyses are used in many fields other than organization science caused some organization scientists to consider whether such analyses, even if they were used in organizational settings, justified thinking of network theory as an organization 125

George P. Huber

Copyright American Psychological Association. Not for further distribution.

theory (Salancik, 1995; Wellman, 1988). Drawing on the definition of organization theory as centrally concerned with explaining the nature and circumstances of individual organizations and collectives of organizations, organization theorists are able to lay claim to network theory as an organization theory because networks as a structural variable are now used to explain organizational change and performance (Kraatz, 1998; Parkhe, Wasserman, & Ralston, 2006). Strategic choice theory. The unit of analysis in Child’s (1972) strategic choice theory is the organization as represented by its dominant coalition, generally the organization’s upper management (Cyert & March, 1963). The theory is a direct response to the claim of organization theorists holding to population ecology, institutional, and contingency explanations that the nature and circumstances of observed organizations are determined by, for example, environmental, technological, size, and institutional constraints. The argument behind this claim of external determinism was that the organization’s dominant coalition would either adhere to these constraints, or, if it did not, the environment would select out the organization. In contrast to this argument, strategic choice theory holds that the dominant coalition can reasonably believe that at least to a modest extent it can enact (i.e., choose or create) its external environment (Weick, 1969, p. 130), its core technology, and its size, and through such manipulation of contingencies it can free itself to choose the attributes of its other features. In Child’s (1997) words, those holding the determinism explanation “stress environmental selection rather than selection of the environment” (p. 45). The coalition’s choices are still constrained by its concern with choosing attributes that allow the organization to achieve a “satisficing” (March & Simon, 1958, pp. 140–141) level of performance. If performance exceeds this “satisficing” level (and one is assuming that this level represents a degree of return that is at least sufficient to secure resources required for the fulfillment of present and future plans), then the decisionmaking group may take the view that 126

the margin of surplus permits them to adopt structural arrangements which accord better with their own preferences, even at some extra administrative cost to the organization. (Child, 1972, p. 11) Thus, in contrast to the theories described above, strategic choice theory specifically acknowledges agency (Eisenhardt, 1989a; Emirbayer & Mische, 1998). In essence, strategic choice theory explains the nature and circumstances of an organization as consequences of the intentions of the dominant coalition, subject to the coalition’s perception of constraints posed by the environment, the organization’s condition, and the performance requirements of key stakeholders, and subject to the coalition’s success in getting its intended organizational attributes emplaced and maintained. Recent essays and theory pieces (e.g., Hambrick, 2007; Hambrick & Finkelstein, 1987; Kaiser, Hogan, & Craig, 2008) emphasized that strategic choice theory makes no claim that dominant coalitions are omniscient; whereas dominant coalitions do “have considerable influence over the form and fate of their companies . . . we don’t mean that that they matter only positively” (Hambrick, 2007, p. 341). Rather, “evidence shows that leaders do indeed affect the performance of their organizations—for better or for worse” (Kaiser et al., 2008, p. 96). The seven theories discussed above conform to the organization theory paradigm (regarding scientific paradigms, see Kuhn, 1970). That is, the theories are characterized by their positivism; they are subject to empirical refutation. The two theories discussed immediately below do not conform to these characteristics of the paradigm. Critical management theory and postmodern theory. Critical management theory is sometimes referred to as critical management studies (CMS) and postmodern theory is sometimes referred to as postmodernism. Critical management studies (CMS) offers a range of alternatives to mainstream management theory with a view to radically transforming management practice. The common core is deep

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

skepticism regarding the moral defensibility and the social and ecological sustainability of prevailing conceptions and form of management and organization. CMS’s motivating concern is neither the personal failures of individual managers nor the poor management of specific firms, but the social injustice and environmental destructiveness of the broader social and economic systems that these managers and firms serve and reproduce. (Adler, Forbes, & Willmott, 2007, p. 119) Critical theorists view organizations as systems of domination in which one class of actors exploits others, and differences in interests, far from being negotiated and reconciled, are typically resolved by the more powerful suppressing the weaker. (Scott & Davis, 2007, p. 215) Important references regarding CMS are Academy of Management Review (1992), Administrative Science Quarterly (1998), and Alvesson and Deetz (2006). Although CMS is ideologically motivated, its assertions serve as propositions and its empirical works often offer supporting evidence. Thus, it satisfies the requirements of a theory (cf. Kerlinger, 1986, p. 9). In some contrast, “postmodernism is a recent intellectual development and still a work in progress” (Scott & Davis, 2007, p. 216). At the present time, postmodernism exists more as a critique of conventional understandings of organization than as an alternative model of organizing. . . . As noted, it is not clear what a “post modern” organization might look like. Presumably its culture would support diversity, pluralism, and ambiguity (see Martin, 1992). (Scott & Davis, 2007, pp. 217–218) Focusing more on the postmodern movement than on its role as an organization theory, Kilduff and Mehra (1997) called the attention of organization theorists to the commitments of postmodernists “to breaking down disciplinary boundaries, chal-

lenging conventional wisdom, and giving voice to viewpoints and perspectives hitherto silenced” (p. 476) and to the “epistemological problematics that postmodernists have surfaced” (p. 476). Research and findings in organization theory. In the preceding, I examined seven prominent organization theories that are congruent with the positivistic and empirical verification characteristics of the field’s paradigm, and also two others, critical management theory and postmodern theory, that are not congruent with these characteristics. In the present discussion of research in organization theory and in the remainder of the chapter, when I speak of organization theories, I mean only theories that are congruent with these characteristics of the field’s paradigm. In this section, I begin with a discussion of how organization theory research is done. I then comment on some differences among the theories and an important implication of these differences. Finally, I describe some evolving shifts in foci of organization theory researchers. Research in the field of organization theory proceeds in much the same general manner as does research in other behavioral and social science fields. Early on, some process leads to the generation of a hypothesis. The range of these processes in organization theory is wide and includes puzzling over contradictory findings in current or previous studies, speculations about the untested boundaries of theories, recognition of unexamined phenomena in the environment, and findings from field studies using grounded theory or from computer simulation studies. (Computer simulations are experimental manipulations of variables within computer models of field phenomena. The results are sometimes used to generate hypotheses to be tested with field data but are sometimes portrayed as validations of previously developed hypotheses. For examples, see March, 1991; Repenning, 2002; Siggelkow & Levinthal, 2003; Zhiang, Zhao, Ismail, & Carley, 2006.) The hypothesis is then empirically tested. The data sources used in tests of organization theory hypotheses include, for example, publicly available or researcher-developed archival data bases, field survey questionnaires, fieldlike laboratory simulations, field interviews, and ethnographic 127

Copyright American Psychological Association. Not for further distribution.

George P. Huber

observations. The methodology for hypothesis testing in organization theory is nearly always statistical analysis. However, when the data elements are complex and not subject to decomposition, or idiosyncratically context dependent, or not reliably scalable, as might be the case for ethnographic observations or interviews, researchers’ interpretations of carefully collected and recorded observations are accepted as confirmatory evidence (Eisenhardt, 1989b; Hargadon & Sutton, 1997; Meyer, 1982). Comparisons of the relative predictive validity of the prominent organization theories are problematic for both theoretical and empirical reasons. One obstacle to making such comparisons follows from the fact that the various theories tend to be specialized by the unit of analysis for which the theory was originally developed (e.g., communities, populations, networks, autonomous organizations in general; for-profit organizations in particular). Very few studies have attempted to test the validity of the theories using units of analysis other than that for which the theory was developed. A second obstacle is that the theories focus on different constructs. For example, population ecology theory focuses on economic (broadly defined) competition as the principal determinant of survival, whereas institutional theory focuses on legitimacy (adherence to societal norms) as the principal determinant of survival. Thus, the predictive validity of population ecology would tend to be high in a capitalistic society, and the predictive validity of institutional theory would tend to be high in a socialistic or theocratic society. The relative predictive validity of the two theories would depend on the societal context in which the comparison is to be made. A third obstacle to comparing the predictive validity of the theories results from the fact that validation studies of the different theories tend to operationalize organizational performance using different performance measures (e.g., duration of survival; survival vs. demise; or changes in sales, profitability, market share, or market value), making comparisons of relative predictive validity problematic. Finally, the theories make different assumptions about the role of managerial choice as a determinant of organizational outcomes. Some theories (network theory, contingency theory, resource 128

dependence theory, transaction cost theory, and strategic choice theory) either admit to or focus on managers’ choices of organizational attributes (such as structural forms) as an important outcome variable (in addition to organization performance). Other theories (population ecology theory, institutional theory) tend to ignore or dismiss the choicemaking role of managers. Because the influence of managerial discretion varies across research contexts, the predictive validity of theories is influenced by the mix in the levels of managerial discretion that exist in the validation studies used to establish the theory’s predictive validity. A conclusion that follows from the above conditions is that it is generally not possible to make meaningful global conclusions about the relative predictive validity of the prominent organization theories. I noted earlier that much of what is regarded as organization theory has been developed in the last 5 decades (Augier et al., 2005; March, 2007). It seems important to point out that organization theory research has been, and is, undergoing change. Davis and Marquis (2005) described the baseline condition from which much of the change is evolving: A number of paradigms for the study of organizations were elaborated during the mid-1970s, including transaction cost economics, resource dependence theory, organizational ecology, new institutional theory, and agency theory in financial economics. These approaches reflected the dominant trends of the large corporation of their time; increasing concentration, diversification, and bureaucratization. (p. 332) Relative to today, organizational environments at that time were less complex, less dynamic, and (in the case of for-profit organizations) less competitive. Subsequent to the 1970s, however, advances in information technology and increases in the several phenomena associated with globalization created great changes in organizational environments and, hence, in the nature of organizations. Increases in the speed of environmental change caused organizational learning, innovation, and restructuring to become commonplace processes. Alliances, net-

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

works, and virtual organizations became commonplace structures. Organizations became much more proactively interdependent with other organizations. As a result, organizational fields and organizations as members of communities, rather than organizations as individual actors, became the foci of research interest to many organization theorists. The processes in which organizations engaged as they encountered environmental threats and opportunities with greater frequency became the foci of many of those organization theorists whose interest was individual organizations. Certainly there are exceptions, but as a whole, organization theorists seemed to gradually lose interest in the theories associated with the “paradigms” noted by Davis and Marquis (2005, p. 332). I elaborate on these matters as the chapter proceeds and in the Concluding Observations and Commentary. Section summary and controversial issues. After reflecting on the seven organization theories described earlier, it is likely that the following question comes to mind: Is there a common theme that runs across them or an overarching organization theory that encompasses all of them? The answer is yes. There is an assertion that seems to capture the essence of each of the theories and that seems not to be contradicted by any of the theories. It could be called fit theory. In the context of organization theory, fit theory explains the nature and circumstances of organizations as outcomes determined by the goodness-of-fit between (a) the organization’s attributes and (b) the constraints, threats, and opportunities posed by the organization’s external and internal environments. (To elaborate the fit theory assertion, I note that when examining the subject of organization design, the organization’s attributes at any given time will be shown to be the outcome of the design process, the design implementation and maintenance processes, and the effects on these processes of environmental forces that have caused the current attributes to differ from the designer’s original intentions. There, it will also be apparent that the organization’s circumstances, e.g., its performance, are negatively influenced by incongruities among the organization’s attributes. See also Drazin & Van de Ven, 1985; D. Miller, 1992.)

A second question might also come to mind: Are there any significant controversial issues in the field? It is beyond the scope of this chapter to examine the controversial issues currently active within and among the individual theories just described, but it seems appropriate to call attention to three controversial issues that span the theories. One concerns the appropriate focus of organization theory research: Should organization theorists direct, or even limit, their work to that which will serve to improve the performance of organizations by, for example, conducting research that can result in guidance for organization designers, or should they focus their work exclusively on explaining, rather than improving, the nature and circumstances of organizations? A second controversial issue also concerns appropriate focus: Should organization theorists maintain a value-free approach in their studies, as is indicated by the positivistic characteristic of the field’s paradigm (Kuhn, 1970), or should they focus their work on improving the lot of less powerful organizational members and stakeholders, as claimed by critical management theorists? It seems unlikely to me that either of these value-laden issues will be resolved through research or debate among researchers. To the extent that progress is ever made on these issues, my view is that even partial resolution will come about only as a consequence of institutional forces operating through the allocation or withholding of resources by different societal components. For example, it may be that institutions such as research funding agencies would fund either heavily or minimally studies directed at improving the lot of those less powerful. Or, it may be that departmental faculty groups or academic leaders would encourage or discourage such studies. The third issue concerns the effectiveness of top managers in influencing the attributes and performance of organizations. As in the above, Child’s (1972) article was a reaction to what he saw as the incomplete understanding of many organization theorists concerning how the structural attributes of organizations come to be. In forthright contrast to the apparent understanding of these theorists, Child stated that “in shifting attention toward choice, we are led to account for organizational variation directly through reference to its sources rather than 129

Copyright American Psychological Association. Not for further distribution.

George P. Huber

indirectly through reference to its supposed consequences” (p. 14). Here, Child is suggesting that variation in organizational attributes might better be explained as a function of the source of the attributes (managerial action) rather than as a function of their outcome (organizational performance or survival). He is emphasizing that dominant coalitions, or the managers they designate, choose organizational attributes (although not without considering performance requirements and not necessarily with omniscience). Once positioned, the attributes may be in place for long periods of time (Schein, 1983; Swaminathan, 1996). Thus, during the interval between when an organization’s attributes are emplaced by the dominant coalition or its subordinates and when the attributes are changed to fit the coalition’s revised view of the organization’s environment or circumstances, the observed attributes are the consequence of strategic choice, not environmental selection. The above thoughts address the role of individuals in determining the attributes organizations possess. But what about organizational performance? Do managers typically make wise choices? Part of the difference in points of view regarding the answer to this question is explained by the time horizons of those holding to different theories. Population ecologists might argue that whatever attributes managers choose will turn out in the long run not to be efficacious because the environment in effect when the choices were made will eventually change and the attributes will be ill-fitted to it. Therefore, the environment will trump managerial choice in the long run. In rebuttal, strategic choice theorists might argue that this argument is irrelevant because organizational attributes must be chosen and emplaced to serve during the interval between the present and the long run (that is, from the present until when the attribute becomes sufficiently incongruent with the changed environment that it must be replaced). Some organization theorists doubt the ability of top managers to make wise choices. For example, The organizations that can be observed at any point in time are the survivors of a selective process that has eliminated a 130

large fraction of the underlying population. . . . If individuals [such as top managers, venture capitalists, or organization theorists without information on nonsurvivors] try to learn from this sample of organizations, they will get a misleading impression of the determinants of corporate performance. (Denrell, 2003, p. 239; parenthetical insert added) This fact may be the basis for the skepticism of some organization theorists concerning the irrelevance of strategic choice and organization design: “Investigators haven’t learned what particular combinations of internal organizational characteristics are most effective in permitting organizational survival or growth” (Aldrich, 2008, p. 57). This level of skepticism implies that (even research-based) beliefs about appropriate levels of attributes for specific situations are so inaccurate that they are useless. It seems likely that the level of inaccuracy would vary, from very small to very large, depending on the level of bias in the available sample and the level of information those drawing the conclusions possess, and can adequately process, about the nonsurvivors (see Denrell, 2003). Other researchers forthrightly assert that toplevel managers do influence organizational performance, presumably through their influence over organizational attributes, but not always to the organization’s benefit. A review of longitudinal studies that account for year, industry, and company demonstrates that “top executives (at least in the U.S.) have considerable influence over the form and fate of their specific companies. . . . So, when upper echelon researchers assert that executives matter, we don’t mean that they matter only positively. . . .” And, as we proposed in our initial 1984 article, we anticipate that TMTs [top management teams] matter even more. (Hambrick, 2007, p. 341) The authors then summarize the evidence showing that leaders do indeed

Organizations: Theory, Design, Future

affect the performance of organizations— for better or for worse. (Kaiser et al., 2008, p. 96)

Copyright American Psychological Association. Not for further distribution.

Some organization theorists appear to disparage the idea that managerial actions must, of necessity, involve perceptions of the environment. Organizations, like people, supposedly create their environments by imposing order on information about the environment. The imagery of enactment suggests that the environment is in the eye of the beholder, an organizational dream. If organizations are understood to relate to social, economic, and political changes in such a surreal way, they can hardly shape the broader processes. (Hannan & Freeman, 1989, pp. 31–32) Environments are defined in terms of members’ perceptions, resources, uncertainty, and other disparate schemes. (Aldrich, 2008, p. 57) Observations such as these last two seem not to reflect the well-supported proposition in psychology that the effects of environmental attributes or stimuli on human actions are necessarily mediated by the actor’s perceptions of the attributes or stimuli. As noted in the earlier discussions of population ecology and institutional theory, recent works of some population ecologists and institutional theorists indicate that segments of these research communities are recognizing the relevance of managerial choice as an influence on the attributes, performance, and survival of organizations (Haunschild & Chandler, 2008). Some organization theorists have proposed ways in which these opposing perspectives can be reconciled (Astley & Van de Ven, 1983; Beckert, 1999; Hrebiniak & Joyce, 1985; Zammuto, 1988). It seems appropriate at this point to note that many organization theorists contribute to process theories closely related to the seven organization theories discussed above. Prominent examples of the processes investigated are organizational learning (Argote, 1999; Huber, 1991; March, 1991), organizational decision making (Baum & Wally, 2003; Chattopadhyay, Glick,

& Huber, 2001; S. J. Miller & Wilson, 2006), and organizational change (Brown & Eisenhardt, 1997; Pettigrew, Woodman, & Cameron, 2001; Schwarz & Huber, 2008). Studies in these areas focus on the antecedents and effectiveness of the processes, where effectiveness is generally assessed in terms of intermediate-level variables that contribute to the higher level effectiveness variables that characterize the organization theories: performance and survival. Accordingly, it can be argued that work in these three areas is encompassed within organization theory. The idea of managerial choice with respect to organizational attributes introduces the subject of organization design. Do organization theories influence the practice of organization design? Because some organizations manifest configurations of attributes that are consistent with guidelines that could have been derived from one or more of the above theories, one might think so. But, rather than reflecting management’s knowledge about organization theories, might not these observed configurations result from the organization’s experiential or vicarious learning or from institutional forces or selection by the environment, as described earlier? I turn now to examining the material from the organization theory community that bears on the routine of organization design. ORGANIZATION DESIGN The practice of organization design must be nearly as old as the existence of organizations, with powerful, interested persons shaping organizations to suit their needs and preferences (subject to perceived environmental constraints). The practice has long been conducted without the benefit of findings from organization science. Today, however, as a result of increases in the number of college graduates with management-related degrees and of increases in the amount of management-related material appearing in the business press, the potential for organizational designs to be influenced by organization theory is significantly greater than ever before. Beginning in the last quarter of the 20th century, numerous management consultancies and consulting organization theorists identified groups of organizational features 131

Copyright American Psychological Association. Not for further distribution.

George P. Huber

that they believed should be used in guiding the design of client organizations. They also established guidelines for selecting organizational attributes for these features that, for various contingencies, they believed would contribute to organizational performance. The McKinsey 7-S framework is an early, well-known example (Peters & Waterman, 1982, pp. 9–10; see, for other examples, Burke, 2002; Burton, DeSanctis, & Obel, 2006; Mackenzie, 1986; Mintzberg, 1981, 1983). Most of these prescriptions or models were developed using principles consistent with contingency theory or congruence theory reasoning as derived from early works such as those included in Donaldson (1995) or modern explications of the theory (see Donaldson, 2001). At least one model has been formalized to the point at which its algorithms are encompassed in a computer program (Burton & Obel, 2004). As a verb, organization design refers to the process of choosing the attributes that the designer intends for the organization to possess, attributes of some of the organizational features mentioned earlier: strategy, core technology, structure, employees, culture, and routines. The intended attributes are those that follow from the trade-offs made between the designer’s preferences on the one hand and the constraints imposed by resource availability and by powerful and interested stakeholders on the other (see Hambrick & Finkelstein, 1987). As a noun, an organization’s design refers to the outcome of the design process, the design implementation and maintenance processes, and the effects on these processes of forces that caused the emerged design to differ from the designer’s intended design. Over time, external and internal forces cause the organization’s attributes to change, thus causing the organization’s design to change from the design originally implemented. These changes generally occur most frequently early in an organization’s life. For readers interested in designing an organization, I recommend examining the above organization design references or others like them. In the following section, I describe prominent material from the organization theory community that underlies the content of most organization design tutorials and literature. I begin with the basic material: the attributes of organizational structure and 132

the traditional structural forms that organizations take. Then, I describe organization design in terms of choosing specific combinations of organizational attributes.

Organizational Structure and Forms A feature of great interest to an organization’s employees and managers is the organization’s structure: that is, the organization’s relatively fixed set of formal authority and responsibility relationships and information sources and flows. Structure is important to employees because it influences the nature of their roles, coworkers, supervisors, status, and job security. These in turn influence employee satisfaction and identification with the organization (Ashforth & Mael, 1989). Structure is important to managers because it is easy to visualize and communicate. By changing authority and responsibility relationships and information sources and flows, managers can relatively quickly reallocate human and material resources to those subordinates whom the leader feels will give the resources more effective direction and oversight. More so than with some other organizational features, changes in structure can be achieved with directives. This is because structures tend to be more malleable than strategies, core technologies, and cultures. Further, because the tangibility of structure enables it to be readily described or observed, it is easy for leaders to communicate and demonstrate to owners, employees, and other stakeholders through structural changes that the leaders are actually taking action to solve a problem or exploit an opportunity. Related to this, its tangibility and malleability make structural change an attractive and frequently used device for signaling change. Given its importance in the contexts just mentioned, it will be useful at this point to examine organizational structure in some detail. We must keep in mind, however, that for maximum effectiveness an organization’s structure must accommodate the organization’s core technology, and the core technology must facilitate the organization’s strategy. Structural forms cannot be chosen with complete freedom. Organizational structures are described both with continuous dimensional scales and with typologies of organizational forms. I begin by

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

describing the most commonly discussed structural dimensions. (Many attributes are specific locations on these structural dimension scales.) Then, drawing on the organization theory literature and especially the well-known works of Mintzberg (1979, 1981, 1983), Porter (1980), and Miles and Snow (1978, 1994), I describe the most frequently discussed forms of traditionally structured organizations. (Recall from my previous discussion of organizational populations that an organization’s form generally refers to its possession of a distinguishing structural attribute or a particular collection of structural attributes.) Attributes of organizational structure. Organizational structure has traditionally been thought of as being concerned with how to divide and coordinate the work and how to provide an optimum balance between the capability of being efficient and the capability of being adaptable. As will be seen when I examine in the next section of the chapter the nature of future organizations, organization structure is now viewed more broadly to cope with more complex issues such as exploration versus exploitation, ambidexterity, and networks of organizations. Structure has been conceptualized as having several dimensions. Specialization of jobs or organizational units can be thought of as vertical specialization (the number and distinctiveness of hierarchical levels) or horizontal specialization (the functional scope and distinctiveness of horizontally located positions or units). The variety and intensity of the organization’s integration mechanisms (its personnel, routines, and structures involved with communication and coordination) are consequences of, and therefore positively related to, the degree of specialization. The complexity of an organization is high if it has high levels of specialization and integration. The extent of the organization’s formalization has to do with the rigidity versus flexibility of policies, routines, and job descriptions. Other scales for characterizing formalization are mechanistic versus organic and formal versus informal. Centralization is a measure of the scope of authority to make decisions or give directives. Vertical centralization is high if authority is concentrated at the top of an organization. An organization

with vertical decentralization has such authority located at lower and middle management levels. Organizations are horizontally centralized, if a small minority of multiple functions possesses all or nearly all of the authority, or are horizontally decentralized, if the authority is widely distributed among units at the same level. I turn now to examining the traditional structures present in the common discourse of both the academic and practitioner communities. Each of these structural forms can be thought of as a generic combination of particular organizational attributes. Following this examination of traditional forms, I examine nontraditional organizational forms, including some network organizations and virtual organizations. Traditional structural forms. The most commonly referenced traditional structural forms are simple structure, functional structure, matrix structure and cross-functional teams, machine bureaucracy, professional bureaucracy, adhocracy, divisionalized form, and conglomerate. Each of these also goes by other names. All of these traditional structural forms are readily observed today. Their relative frequencies vary across industries, nations, and cultures. The frequencies also vary across time, as the strengths of relevant environmental forces change. Simple structure. Simple structures are units having one or a very few top managers who direct and coordinate the organization’s activities and a group of operators who do the basic work. Within both the top management and the operator groups there is relatively low differentiation of functions, and the structure is organic. Start-up firms typically adopt a simple structure. Functional structure. Functional structures can develop through different processes. One pattern occurs when, as the organization matures, its members and groups informally begin to specialize in their activities and as a result become more expert in whatever organizational task it is that they collectively perform. To take advantage of these proficiencies, the organization structures itself into specialized functional units. Another pattern occurs when the organization grows to a size that is too large to be directed by one person, so subgroups are 133

Copyright American Psychological Association. Not for further distribution.

George P. Huber

formed and a second echelon of managers is created to direct the operations of these groups. In these circumstances, either preferences or rational choices often result in the groups specializing by function. A functional structure can also develop when competitive or institutional pressures cause the organization to develop or hire experts to carry out specialized functions more effectively or efficiently than can employees without the specialized training. In a manufacturing firm, there might be separate departments for the functions of research, development, engineering, manufacturing, marketing, sales, shipping, accounting, and finance. Some departments might be quite specialized, such as for various types of engineering work. Over time, functional units tend to focus narrowly on their specific responsibilities and become insular, forcing coordination to be managed at the upper organizational level. “Hence, the logic of the functional structure is centrally coordinated specialization” (Miles & Snow, 1994, p. 38). Matrix structures and cross-functional teams. The insularity of the units in a functional structure interferes with communication and cooperation and often leads to delays in decision making and in operations as issues are driven to upper levels and issue-resolving directives come back down to the operations level. To overcome these problems, organizations sometimes adopt a matrix structure, a structure where operating units and support units have formal, joint authority and responsibility for completing certain tasks. However, depending on the priorities associated with their circumstances, organizations often have matrices in which the level of authority between the operation and functional managers is not equal (Ford & Randolph, 1992; Larson & Gobeli, 1987). In matrix structures, the coordination is achieved between managers. Crossfunctional teams are composed of members with different expertise. Often the members are on loan from a functional department. In cross-functional teams, the coordination is achieved among the team members. 3

Machine bureaucracy. Bureaucracy was famously articulated as an ideal organizational form by Weber (1947). Mintzberg (1979, 1983) visualized two forms of bureaucracy, each achieving coordination through standardization. An organization required to be highly efficient or highly reliable by an external entity—such as its clientele, parent organization, or superordinate governmental agency— tends to standardize its core technology and to specialize and standardize its jobs. To such an organization, Mintzberg gave the term machine bureaucracy. As noted by Mintzberg (1979, 1983), who first explicated the specifics of this form and the two forms described next, the defining characteristic of the machine bureaucracy is standardization of work, an attribute that requires a variety of planning, control, and service units. These units gain a limited degree of informal power which, in turn, results in a certain amount of horizontal decentralization. Nevertheless, to attain high levels of the efficiency or reliability demanded by external entities, machine bureaucracies organize their value-adding functions into specialized operations departments which, to the extent that they must be coordinated, are coordinated at or near the top of the organization. Thus, overall, machine bureaucracies are rigidly structured and vertically centralized. These attributes of machine bureaucracies apply to almost all governmental organizations. Professional bureaucracy. As contrasted with the standardization of work that characterizes the machine bureaucracy, the defining characteristic of professional bureaucracies is standardization of skills. Examples of professional bureaucracies are accountancies, consultancies, architectural firms, and universities.3 The professional bureaucracy is a highly decentralized structure. . . . A great deal of power over the operating work rests at the bottom of the structure, with the professionals of the operating core. Often, each works with his

At the time that Mintzberg (1979) explicated the nature of the professional bureaucracy form, most hospitals manifested primarily the attributes of professional bureaucracies. However, in recent decades, as the demands for efficiency (for example, the demand for cost control by third-party payers such as insurance companies and governmental agencies) and reliability (for example, the litigious demand for error-free health care) resulted in increased standardization and hierarchical control, hospitals have become hybrids of professional bureaucracies and machine bureaucracies.

134

Organizations: Theory, Design, Future

Copyright American Psychological Association. Not for further distribution.

own clients, subject only to the collective control of his colleagues. (Mintzberg, 1983, p. 195). Some professional bureaucracies have an abundance of paraprofessional and support units to free the professionals to make the most of their expertise. Adhocracy. Toffler (1970), in his insightful work, Future Shock, foresaw the growth of an organizational form particularly well-suited to the increasing complexity and dynamism of the world of work. He viewed it as a constellation of temporary project teams, and called it “ad-hocracy” (Toffler, 1970, p. 113). Mintzberg (1979) was the first to richly describe what is now termed an adhocracy as an organizational form. One version, an operating adhocracy, draws on the multiple types of expertise possessed by its members to create and provide innovative solutions to its clients’ problems, as might an avant-garde film company. As innovation projects are completed, the cross-expertise team members often move on to become members of different teams. Because of the need for team direction to be a function of expertise and because of the dynamism of the adhocracy’s internal environment, coordination is not achieved through hierarchy or through standardization but rather through mutual adjustment by the teams as they exchange, lend, and share resources and allocate responsibilities. To these ends, the operating adhocracy’s structural attributes include high horizontal job specialization, teams that include various groupings of line managers and staff and operating experts, and a highly organic structure with widely distributed power. In contrast to an operating adhocracy, the administrative adhocracy version undertakes projects to serve itself, as might a space exploration agency. The administrative component of an administrative adhocracy is structured organically to use the expertise and creativity of its members to identify and develop plans for exploiting opportunities. The operating core that carries out the projects developed by the administrative component is distanced from the administrative component and may be a separate unit of the organization or may be composed of one or more independent contractors.

Divisionalized form. The divisionalized form was used in the governance of ancient civilizations, ancient armies, and the Roman Catholic Church. It was invented for the modern business world at DuPont and independently at General Motors in the early decades of the 20th century to solve problems of corporate governance (Chandler, 1962). The divisions of an organization structured as a divisionalized form can be specialized by product, customer type, or geographical area, and generally have little formal interaction. The organization’s headquarters, besides containing the top management team, contains those support and control groups that serve or control certain activities in all divisions. Typically, these activities would include some human resources services, legal services, some financial control groups, and occasionally a research and development group. The individual divisions do not usually duplicate these groups (although they may have units that serve the division and that work with the support and control groups located at the headquarters). To maximize the possibility that what the headquarters’ top managers support and what control groups learn or know can be used in managing multiple divisions, the organizations generally include only divisions that operate in similar markets or produce similar products. The organization’s headquarters coordinates the divisions through its allocation of resources among them, which is based primarily on their performance (such as profitability in the case of for-profit organizations) or need (such as a high ratio of demand for services to organizational resources in the case of a governmental agency). In that the divisions compete for resources with the other divisions, they tend to organize themselves as machine bureaucracies so as to achieve efficiency and to enable their top managements to exert maximum control. Conglomerate. A conglomerate is an organization having largely the same properties as the divisionalized form. The defining difference between the two forms is that in the conglomerate the multiple divisions operate in different, unrelated domains, thus benefiting minimally if at all from interdivisional knowledge transfer. Mintzberg (1983) portrayed the conglomerate as the end outcome of opportunistic, 135

George P. Huber

Copyright American Psychological Association. Not for further distribution.

rather than planned growth. Thus, it is not surprising that in 1980, the conglomerate firm . . . was perhaps the dominant corporate form in the United States. Yet, by 1990 this form had in effect become deinstitutionalized. . . . First, diversified firms were taken over at a high rate and their unwanted parts were typically sold off, and second, the less diversified firms that survived shunned the strategy of conglomerate growth. . . . By 1990, the largest industrial firms in the United State became considerably less diversified. (Davis, Diekman, & Tinsley, 1994, p. 547) In the preceding, I portrayed traditional organizational forms as combinations of structural attributes. But organization design entails more than choosing only structural attributes. It also entails choosing attributes of the organization’ strategy, core process, employees, culture, and routines that, on the basis of contingency theory and congruence theory, satisfy the contingencies and are not incongruent with each other.

Designing as Choosing Configurations of Organizational Attributes Some organizations are designed by their founding entrepreneurs, who may be influenced by early investors or other influential stakeholders. Some are designed by the founding intrapreneurs (innovators functioning within an organization), generally influenced by resource controllers such as executives of the parent organization. Other organizations are designed by external actors, such as upper managers of acquiring organizations or the legislators who guided the legislation that founded the new governmental agency. Because governmental organizations can generally do nothing that is not specified by law (for example, their mandates greatly constrain the scope of their domain or other components of their strategy), whereas for-profit organizations can generally do anything that is not prohibited by law (for example, they can choose from and enact any of a variety of strategies, subsequent changes forced by 136

competitors not withstanding), designers of forprofit organizations have on the whole more influence over more attributes of more features of the organization to be designed than do designers of government organizations. (Other differences between these two organizational types are discussed later in the chapter.) Given the greater freedom of their designers, it is more instructive to focus our discussion on the design of for-profit organizations (what I refer to as firms). To avoid distracting complications associated with redesigning existing organizations, I assume here an idealized situation in which the designer is an entrepreneur creating an organization where one does not exist. Organization design is portrayed here as the process of choosing configurations of organizational attributes (i.e., as choosing, for a particular organization, specific attributes for the features of strategy, core technology, structure, employees, culture, and routines). I later discuss the situation in which the organization’s leadership is a design feature. Immediately below are descriptions of the three organizational features most often treated in the organization theory and strategic management literatures as the features most critical to the organization’s performance. Strategy, core technology, and structure. The organizational features of strategy, core technology, and structure are the three features most commonly thought of when upper level managers or strategic management researchers think of organization design. Their centrality in organization design research and practice leads them to be viewed by many as the organization design features. Strategy. In most instances, when entrepreneurs begin thinking about creating an organization to help achieve their goals, they already have in mind a product or service, or a market that they see as likely to be responsive to a product or service that they could provide. (Henceforth, when I speak of products, I mean to include services.) Either way, the first organizational attribute chosen is likely to be a specific domain for the organization, that is, a product–market combination or a mix of products and markets. Choosing the domain and choosing the major processes and associated structures for

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

generating products and moving them to their markets are first steps in the entrepreneur’s strategy: the major intended and emergent initiatives taken by the organization’s leadership to enhance the organization’s performance in its external environment (Nag, Hambrick, & Chen, 2007, p. 944). Examples of initiatives with respect to introducing a specific product to a specific market might be acting as a first mover or an early follower or offering the highest performance or the best value. If one of these types of initiatives becomes a pattern, it would be regarded as an emergent strategy (Mintzberg, 1978). Core technology. The means for generating products and services and moving them to their markets includes the core technology (or core process), the primary input–transformation–output process by which the organization adds value to material or intellectual products or adds perceived well-being to customers or clients. For example, for a real estate firm the core technology is the process of transforming an aspiring home buyer into a person who is a satisfied homeowner. Some attributes of core technologies are standardized (vs. variable) and longlinked assembly (vs. hub-and-spoke assembly). As a consequence of legal requirements and constraints imposed by the industry’s professional associations, the real estate firm’s core process would be standardized but with subroutines for various features of the property, the buyer, and the seller. The many steps involved in marketing the property, negotiating the contract, and producing the transaction characterize a long-linked assembly but would have within this process some hub-and-spoke stages in which participants came together to coordinate their efforts. Structure. Core processes, and most of the organization’s other important processes, are managed primarily through the organization’s structure: the organization’s relatively fixed set of formal authority and responsibility relationships and information sources and flows. (Of course, informal sources of influence and communication generally supplement this formal structure in managing the organization’s processes.) Some commonly discussed attributes of structure are (a) mechanistic (vs. organic) and (b) vertically centralized (vs. verti-

cally decentralized). A real estate firm allows its salespeople flexibility and grants them discretion in many noncritical matters but has rigid rules to enforce laws and industry norms, and its top management tends to reliably enforce ethical standards among the salespeople. From what has been said earlier, it is clear that one of the designer’s tasks is to choose combinations of attributes for strategy, core technology, and structure that are efficacious for satisfying the environmental contingencies and are congruent with each other. Common wisdom in the management community is that commonly observed combinations (often articulated by academics or consultants), called configurations, work well. Common configurations of the three fundamental design features. The 1970s saw an eruption of interest in organization design among academics (Anderson & Paine, 1975; Miles & Snow, 1978; Mintzberg, 1979; Porter, 1980; Segal, 1974). Of particular interest was the identification of configurations of organizational attributes that seem to contribute to firm effectiveness. Three typologies of commonly combined attributes (Miles & Snow, 1978, 1994; Mintzberg, 1979, 1983; Porter, 1980) have been sustained in academic discourse. Mintzberg’s typology. Mintzberg (1979, 1983), focusing primarily on structure, set forth a configuration typology comprising five of the eight structural forms noted above: simple structure, machine bureaucracy, professional bureaucracy, adhocracy, and the divisionalized form. For each of these forms, Mintzberg (1983) identified the specific attribute that characterizes the form for each of thirteen organizational features, mostly aspects of structure but also including the nature of the core technology. The attributes seem congruent with the respective structural form and would logically be part of the organization’s configuration of attributes. Mintzberg described the five configurations as “pure forms” for the purposes of discussion and analysis but, with regard to the routine of organization design, he called attention to the fact that multiple simultaneous demands on an organization can force the perfectly logical use of structural hybrids of the forms (Mintzberg, 1979, p. 474). He also 137

George P. Huber

noted that in some situations a firm can choose any of multiple forms:

Copyright American Psychological Association. Not for further distribution.

The restaurant can structure itself like a Simple Structure, Machine Bureaucracy, or Professional Bureaucracy, depending on whether it wishes to remain a small, classic greasy-spoon, grow large through the mass distribution of basics, such as steak and lobster, or develop the gourmet skills of its chefs. (p. 474) Porter’s typology. Porter’s (1980) focus was generic competitive strategies, “internally consistent strategies (which can be used singly or in combination) for creating a defendable position in the long run or outperforming competitors in an industry” (p. 34). Porter described three such strategies: overall cost leadership (maintaining an industry-wide, low cost position), differentiation (maintaining something that is perceived industry-wide as unique; e.g., product features, brand image, customer service), and focus (focusing on a particular buyer group, segment of a product line, or geographic market). Although Porter’s focus was on strategy, for each strategy he briefly mentioned many organizational attributes that are congruent with that strategy. Porter also viewed his strategies as pure, arguing that the firm failing to develop its strategy in at least one of the three directions—a firm that is “stuck in the middle”—is in an extremely poor strategic situation. . . . The firm stuck in the middle . . . either loses the high-volume customers who demand low prices or must bid away its profits to get this business away from low-cost firms. Yet it also loses highmargin businesses—the cream—to the firms who are focused on high-margin targets or have achieved differentiation overall. (Porter, 1980, pp. 41–42) Miles and Snow’s typology. Drawing on their acquaintance with the organization theory literature and their own studies, Miles and Snow (1978; see also Miles, Snow, Meyer, & Coleman, 1978) derived a typology of four strategic types: defenders, prospec138

tors, analyzers, and reactors. The authors presented each of the first three types as a configuration of strategic, technological, and structural attributes that defines the type and that appears optimal in terms of contingency theory. For each strategic type, many other attributes besides those associated with the three fundamental design features are described as being congruent with the type. Defenders are organizations which have narrow product–market domains. Top managers in this type of organization are highly expert in their organization’s limited area of operation but do not tend to search outside of their domains for new opportunities. As a result of this narrow focus, these organizations seldom need to make major adjustments in their technology, structure, or methods of operation. Instead, they devote primary attention to improving the efficiency of their existing operations. (Miles & Snow, 1978, p. 29) To attain the efficiencies associated with their intended strategy of holding their market through offering low-cost products, defenders’ core technologies are relatively inflexible and highly routinized and their structures tend to be bureaucratic (rigid and highly vertically centralized). Most defenders are older, well-established firms that use long-established relations with customers, suppliers, and regulators to help them defend their domain. Thus, they seek, and contribute to maintaining, stable environments. Many defenders are also large and enjoy economies of scale that contribute to their low operating costs. Power and rewards are generally associated with production (by which I refer to both goods and services) and finance. An example would be a company’s division that was charged with generating profits for the parent company by manufacturing standard automobile tires. Another example would be a family-owned ethnic restaurant in a stable ethnic neighborhood that stuck with its traditional menus, hours of service, and so forth. In that case, organizational controls would be the consequence of traditions rather than bureaucratic formalization.

Organizations: Theory, Design, Future

Copyright American Psychological Association. Not for further distribution.

Prospectors are organizations which almost continually search for market opportunities, and they regularly experiment with potential responses to emerging environmental trends. Thus, they are often the creators of change and uncertainty to which their competitors must respond. However, because of their strong concern for product and market innovation, these organizations usually are not completely efficient. (Miles & Snow, 1978, p. 29) Prospectors seek out and contribute to complex, dynamic environments. They tend to be young firms, not yet committed to a particular product or market. With regard to exploiting new products or markets, prospectors are first movers. To maintain the flow of new products, a prospector must dedicate much of the technological core to producing prototypes. Thus, the core is flexible and much of the knowledge associated with it is embedded in people rather than in equipment and procedures. Prospectors’ structures tend to be decentralized and oriented toward coordinating their multiple endeavors rather than engaging in top-down direction. Power and rewards are generally associated with the research and development and marketing functions. An example would be a small start-up software company looking for markets for the software it was developing. Another example would be a large fastfood company that continually experimented with different menus, venues, and hours of operation. Analyzers are organizations which operate in two types of product–market domains, one relatively stable, the other changing. In their stable areas, these organizations operate routinely and efficiently through use of formalized structures and processes. In their more turbulent areas, top managers watch their competitors closely for new ideas, and then they rapidly adopt those which appear to be the most promising. (Miles and Snow, 1978, p. 29) Analyzers tend to be fast followers. Successful product innovations are further developed in ways

that result in lower costs and higher earnings that can be used to fund additional innovations. Analyzers have a dual technological core, with a flexible component for the changing domain and stable component for the stable domain. Their structures reflect attempts to balance needs for effectiveness and efficiency, including use of matrix structures and formalized conflict resolutions processes. Power and rewards are generally associated with the marketing and applied research functions. An example would be a multidivisional company that focused on developing new products and getting market share quickly so as to justify transferring the product to an existing division or to create a new division around the product. Reactors are organizations in which top managers frequently perceive change and uncertainty occurring in their organizational environments but are unable to respond effectively. Because this type of organization lacks a consistent strategy–structure relationship, it seldom makes adjustment of any sort until forced to do so by environmental pressures. (Miles and Snow, 1978, p. 29) An example would be a retail chain without the talent or processes that allow it to keep its product lines up to date and that, instead, tries to achieve an acceptable sales level by adjusting its marketing practices. It can be discerned from the descriptions of defenders, prospectors, and analyzers that each of these strategic types enacts, contributes to, and works best in a different environment: the defender in a relatively simple and stable environment, the prospector in a relatively complex and dynamic environment, and the analyzer in a mixed environment. Miles and Snow (1978) argued that in the environment for which each of the types is best suited, each will outperform the others, a line of reasoning clearly in keeping with contingency theory. They further concluded that “if management does not choose to pursue one of these ‘pure’ strategies, then the organization will be slow to respond to opportunities and is likely to be an ineffective performer in its industry” (Miles & Snow, 1978, p. 14). 139

Copyright American Psychological Association. Not for further distribution.

George P. Huber

Studies by Doty, Glick, and Huber (1993), Hambrick (1983), Jennings and Seaman (1994), and Priem (1994) offered support to these assertions. Miles and Snow’s rich descriptions of the types show that the attributes of the employees, cultures, and routines of these three strategic types also vary in ways that seem congruent with the attributes of the strategies, technologies, and structures of the respective types. From the above, it appears that the organizations observed by Miles and Snow (1978) were designed in accord with contingency theory and congruence theory. However, because contingency theory and congruence theory were not mature at the time that Miles and Snow made their observations and were just beginning to be discussed in business school classrooms or the business press, it is more likely that the configurations of attributes observed by Miles and Snow either were the outcomes of vicarious or experiential learning by the managers of the organizations observed or were the configurational survivors of ecological selection processes out of which organizations with other combinations of attributes had been selected in the environments sampled by Miles and Snow. Besides the three fundamental design features just examined, three other organization design features that are prominent in the organization science literature and in the business and lay presses have been found to be associated with organizational performance. These are the organization’s employees, culture, and routines. Employees, culture, and routines. Employees, culture, and routines are pervasive concepts in the thinking of lower to mid-level managers and organization behavior researchers (although individual managers or researchers may use terms for the concepts other than those I use here). The attention of researchers to organizational culture and routines increased considerably in the 1990s. These increases may have been stimulated in part by the increased organizational downsizing of the 1980s and the organizational learning and forgetting issues that follow from the associated employee turnover. Employees. Except in the situation of very tiny entrepreneurial firms, the number of employees and the skills of the employees are rarely independent 140

design variables. Rather, they are largely an outcome of choices concerning the organization’s primary input–transformation–output process, the quantity of goods or services to be generated, and the quality level of these goods or services. Once the production technology and production level choices are made, managers responsible for production choose the required number, skill types, and skill levels of employees necessary to cause the core technology to generate the required production level. The natures of employees needed are also an outcome of the choice of strategy. For example, an overall cost leadership strategy would require that the number and skill levels of employees be as low as possible while still meeting customer expectations, whereas a differentiation strategy would require more than the minimum number and skill types and levels so that high levels of innovation or quality could be assured. For the firm to be competitive, of course, its employees must be as productive as their skills and the organization’s resources and internal environment permit. Employee productivity is partly a function of employee motivation, which is a property of the employee and is influenced by some of the organizational routines to be discussed shortly. Organizational culture. An organization’s culture consists of the values and norms that guide communications and behaviors in the organization (O’Reilly & Chatman, 1996). Strong cultures are those in which these values and norms are intensely and widely held. Examples of scales for describing organizational culture are the following: meritocracy versus seniority (or some other demographic variable), integrity versus expediency, change resistant versus change welcoming, uniform versus variable (or varied), competitive versus cooperative, and risk taking versus risk avoiding. In work environments in which employee behaviors cannot be monitored, perhaps not even by other employees, culture is often the most, if not the only, effective mechanism for influencing employee behavior. Consider, for example, service organizations, and recall from my earlier definition that an organization’s core technology is the primary input–transformation–output process by which the organization . . . adds perceived well-being to customers or clients. Especially in service organizations (in which the technology

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

for transforming a customer to a state of greater perceived well-being or heightened satisfaction is embedded in the employee’s mind, as it is in those of attendants, waiters, and waitresses in upscale restaurants), a culture that values high-quality personal service is absolutely necessary if the organization’s strategy involves marketing to customers who value and are willing to pay for such service (Zeithaml, Bitner, & Gremler, 2005). Strong cultures that are congruent with other organizational attributes amplify the positive or negative effects of these attributes and consequently can contribute to an organization’s performance (Peters & Waterman, 1982; Collins & Porras, 1997), but strong cultures may be resistant to change and can therefore be dysfunctional in fast-changing environments. The culture–performance link is more complex than some of the early writings on organizational culture indicated (Saffold, 1988; Sorensen, 2002). Routines. Routines refer to organizational processes not part of the technological core. Examples include (a) employee recruiting, training, and compensation routines; (b) management development and succession planning routines; (c) public relations routines; (d) budgeting and financial control processes; (e) safety and security processes; and (f) management actions directed at instilling in managers and employees values and norms congruent with achievement of the organization’s focal goals. The particular employee recruiting, development, and compensation routines (and supervisory practices) that lead to appropriate skill types and skill levels and high levels of motivation are partly contingent on the skills and values that employees possess when they become employees. Thus, these routines and the employees for which the routines are designed constitute together a dyadic configuration. Similarly, the specific management actions that succeed in instilling in managers and employees values and norms congruent with achievement of the organization’s focal goals (such as exhibiting, communicating about, and rewarding behaviors consistent with these values and norms) are contingent on the values and behavioral norms that employees possess when they join the organization. Thus, these culture-influencing actions and the managers and employees at whom they are directed also constitute a dyadic configuration.

In the description of the resource-based view of the firm, it was indicated that the organization’s leadership is among the organizational features thought to influence organizational performance. Consequently, before leaving this section on the design of traditional organizational forms I examine leadership as a design variable. Leadership as an organizational design feature. Previously, I described organization design as a process undertaken by an entrepreneur. Of course, little of what was said would be different had it been undertaken by the top management team of an ongoing organization. But an organization’s leadership, which I operationalize here as its top management or top manager, is not always well suited for the organization of which it is a feature. The most common misfit occurs when the organization’s current critical contingency (e.g., low workforce productivity, poor market reaction to new products, lack of innovation, relations with powerful stakeholders) is not currently within the scope of the leadership’s abilities or inclinations to address. In such instances, the firm’s overseers (e.g., its directors or some other powerful and interested entity, such as executives of the organization’s parent) replace the organization’s leadership with a leadership possessing attributes more congruent with the current or anticipated critical contingency. Thus, this aspect of organizational redesign sometimes takes place as a result of actions by powerful entities not necessarily within the organization. The above discussion implies that an organization will perform better if its attributes satisfy contingencies and are congruent with one another, that is, if fit is achieved. Fit is more easily understood, achieved, and maintained in stable environments, but when designing for a turbulent environment, in which the organization must be capable of flexing or adapting, fit must be thought of somewhat more broadly. One matter, not a great departure from the discussions to this point, is to protect the organization’s critical internal processes by designing a protective capability at the organization–environment interface, a matter to be examined immediately below. Later, I discuss some design-based approaches for achieving flexibility and adaptability. 141

George P. Huber

Designs for Difficult or Dynamic Environments

Copyright American Psychological Association. Not for further distribution.

I mentioned earlier that many organization theory researchers have turned to studying the processes used by today’s organizations as they engage environmental threats and opportunities with greater frequency. Much of this research is directed at determining the effectiveness of various structures for attaining flexibility and, more generally, for managing the organization’s adaptation processes. Designing the organization–environment interface. An organization’s core process is where value is added, whether the process is in the form of a smoothly running production line or a highly productive research laboratory. For the core process to be maximally effective, unwanted environmental disturbances must be minimal. This condition is approached by including in the organization’s design specific structures and routines that minimize the organization’s need to modify its core process even temporarily. Thompson (1967) was the first to formally address this issue: “Under norms of rationality, organizations seek to seal off their core technologies from environmental influences . . . to buffer environmental influences by surrounding their technical cores with input and output components” (pp. 19–20). Examples of buffering against irregularly arriving or nonstandardized inputs are the routines of stockpiling materials and providing training programs for new employees. Maintaining inventories of products or employees with which to respond to customer orders are example routines for buffering against irregular output requirement. Such routines require structures, which generate costs. Thompson’s alternatives to structures for dealing with environmental instability or turbulence are routines requiring little structure for their implementation but that have their own costs: ■

■

smoothing transactions, such as offering inducements to customers to access the organization when traffic is slow or charging more for service during peak periods; forecasting arrival of fluctuations in disturbances and scheduling resources to address the fluctuations, such as hiring additional sales clerks at peak seasons; and

142

■

prioritizing disturbances according to their importance and allocating the core’s resources accordingly, such as determining the criticality of the injured in accidents and allocating key resources to those most in need.

Two other routines associated with designing the organization–environment interface are monitoring the environment to identify threats and opportunities so they can be dealt with in a timely manner and searching the environment for knowledge that can be used to create competitive advantage. I address these matters further in the chapter’s section on future organizations. Exploration, exploitation, and ambidexterity. Building on March’s (1991) distinction between an organization’s exploration (for new knowledge) and exploitation (of existing knowledge), Levinthal and March (1993) stated that “an organization that engages exclusively in exploration will ordinarily suffer from the fact that it never gains the returns of its knowledge” (p. 105) and “an organization that engages exclusively in exploitation will ordinarily suffer from obsolescence” (p. 105). Such reasoning is persuasive, and today many organization theorists would agree with March’s (1991) conclusion that “maintaining an appropriate balance between exploration and exploitation is a primary factor in system survival and prosperity” (p. 71). Congruence theory postulates that those attributes (of features such as leaders, structures, and cultures) that are appropriate for engaging in exploratory actions are different from those appropriate for exploiting the organization’s current capabilities and knowledge resources. Miles and Snow (1978) made the same general point when they described prospector and defender organizations as having core technologies and structures that differ across the two strategic types. How can organizations simultaneously pursue two strategies such as exploration and exploitation—each requiring different core technologies, structures, employees, cultures, and routines—and still be sufficiently effective that they can compete and survive? Is there a structure with which organizations can pursue the exploration and exploitation strategies simultaneously and not suffer the costs of incongruity?

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

The organizational design school’s answer to this last question is affirmative. The organizational competence required is organizational ambidexterity, first articulated by Duncan (1976) and currently defined by Raisch and Birkinshaw (2008) as “an organization’s ability to be aligned and efficient in its management of today’s business demands while simultaneously being adaptive to changes in the environment” (p. 375). O’Reilly and Tushman (2008) provided a somewhat more proactive definition: “the ability of a firm to simultaneously explore and exploit” (p. 185). But what organization design attributes provide this “ability”? The primary approach to providing organizational ambidexterity is to structurally separate subunits for exploration and exploitation. However, these separate units are held together by a common strategic intent, an overarching set of values, and targeted structural linking mechanisms to leverage shared assets. These internally inconsistent alignments and the associated strategic tradeoffs are orchestrated by a senior team with a common fate incentive system and team processes capable of managing these inconsistent alignments in a consistent fashion. (O’Reilly & Tushman, 2008, p. 193; see also Gibson & Birkinshaw, 2004; Raisch & Birkinshaw, 2008) The above addressed the situation of two conflicting contingencies: the exploration strategy and the exploitation strategy. Is there a structural approach to dealing with a wider variety of conflicting contingencies? Again, the answer is affirmative. The structural mechanism is loose coupling. Loose coupling and temporary suboptimality. When organizational subunits are structurally or processually linked, but nevertheless operate somewhat independently, the system of subunits is said to be loosely coupled. Competing contingencies, or an incongruity between the requirements of a contingency and some organizational condition (e.g., cultural inertia; the availability of specific resources; powerful, obstructive managers), can make complete congruence among certain desirable organiza-

tional attributes infeasible or less than optimal. In such cases, organizations often perform adequately if managerial or employee inventions or interventions, or well-chosen design compromises in which none of the competing contingencies is fully satisfied but all are partially satisfied, enable less than completely congruent attribute combinations to result in satisfactory performance. A simple example is using buffer inventories to deal with the situation in which production capabilities cannot be totally synchronized with demands for the product. Here, the inventory creates a loose coupling between the incongruent attributes of capacity and demand. Dissimilarity between adjacent cooperating units is also often associated with loose coupling. As an example, in a cross-industry analysis (330 U.S. manufacturing industries), Schilling and Steensma (2001) found that “in general, the heterogeneity of industries’ production processes in terms of inputs and demands is positively associated with their levels of modularity” (i.e., with their looseness of coupling; p. 1161). Loose coupling is often temporary. Consider, for instance, that in a tightly coupled system, optimization of a subunit’s activities can, by interfering with the effectiveness of other subunits, contribute to suboptimization of the organization’s overall performance. But when temporary optimization of a subunit’s activity is critical to the organization, as might be the case when the subunit is responsible for completing a critical project by a certain date, it can be optimal from the organization’s overall effectiveness if, temporarily, the subunit is loosely coupled and thus able to optimize its activities, even at the expense of optimal performance by adjacent units. This is suggested by DuBois and Gadde’s (2002) study in the construction industry in which the researchers found that tight coupling within individual projects and loose couplings across the network of organizations involved in the overall construction effort facilitated short-term productivity. For an in-depth description of instances of loose coupling in large firms, see Thompson, Hochwarter, and Mathys (1997). Loose coupling allows for variation and thus experimentation, which permits learning and adaptation (Weick, 1976), at least at the individual unit 143

George P. Huber

Copyright American Psychological Association. Not for further distribution.

level, but possibly not at the organizational level (DuBois & Gadde, 2002). When different compromise designs provide equivalent functionality, the condition is called equifinality (Gresov & Drazin, 1997). For an in-depth description of the diverse applications of the loose coupling concept, see Orton and Weick (1990). Flexibility and dynamic capabilities. Dynamic environments present high rates of often novel threats and opportunities. To survive, organizations must be flexible. In particular, they must be capable of rapidly revising or reconfiguring their attributes to attain new capabilities. The capabilities for attaining these new capabilities are called dynamic capabilities, “routinized activities directed to the development and adaptation of operating routines” (Zollo & Winter, 2002), which enable the firm “to integrate, build, and reconfigure internal and external competences to address rapidly changing conditions” (Teece, Pisano, & Shuen, 1997, p. 516). Examples given by Eisenhardt and Martin (2000) are “product development, strategic decision making, and alliancing” (p. 1105). My sense, however, is that there is a need to distinguish between, on the one hand, rather operational routines for creating new products, strategies, and structures (e.g., the Eisenhardt and Martin, 2000, examples) and, on the other hand, higher order routines, that is, metaroutines for rapidly revising or reconfiguring organizational attributes to attain new capabilities (as is implied by Teece et al., 1997; Zollo & Winter, 2002). The challenge is, of course, to develop these latter routines so that they are in place when needed. Zollo and Winter (2002) emphasized an organization’s experiential learning and active knowledge management (see also Dosi, Nelson, & Winter, 2000), whereas Volberda (1998) emphasized a more proactive, planned design approach that seems to encompass both forms of routines.

Specialized Firms and Network Organizations As demonstrated earlier and again in this section, network organizations are not new, but they are more pervasive and tend to be much larger than in earlier eras. They are also greatly characterized by 144

and dependent on modern forms of information and communication technology. Although information technology is almost universally an enabler of today’s network organizations, it is not the cause of their proliferation. A major factor leading to the growth in the number of network organizations is the increase in organizational specialization, as explained immediately below. Newly specialized firms. Traditionally structured organizations became more complex during the 20th century. This increased complexity resulted in part from the increases in size permitted by increases in information and transportation technologies and in part from the confrontation between (a) the greatly increased scope and depth of technical knowledge needed to produce marketable products and (b) the limits to the scope of expertise and number of experts that individual managers could coordinate. The organizational design consequences were increases in employee and departmental specialization and in the number and variety of interdepartmental coordination mechanisms required to integrate the specialized knowledge into products (i.e., greater organizational complexity within firms). One way this complexity can be conceptualized structurally is as a value chain (Porter, 1985; Quinn, 1992), the chain of value-adding functions needed to transform raw material or ideas and information into a product marketable to the ultimate consumer. An example of a value chain is the sequence of specialized functions that would be incorporated in departments of an integrated oil company: seismic studies, experimental drilling, developmental drilling, infrastructure development, mixed blending, transportation out, marketing, and distribution (Quinn, Anderson, & Finkelstein, 1996). In for-profit organizations, however, beginning generally in the 1980s, a second consequence of departmental specialization resulted in less organizational complexity within firms. It became apparent in many firms that the departments that carried out functions that added value to the firm’s products and that served as stages in the firm’s value chain varied in the knowledge and capabilities they possessed and in the value they added, this last in part because the departments varied in the extent to

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

which they were effective in using their knowledge and capabilities. Not only did departments vary in the value they added to the firm’s value chain, they varied in the value they added when compared with similar departments in other organizations. Thus came the idea that firms should (a) specialize in the value-adding function in which they could or hoped to achieve or approach “best in world” (Quinn, 1992, p. 37) capability (called core competence); (b) keep within the firm the core-competence function and related stages that were world class; and (c) outsource all other functions (after considering both internal and external transaction costs), that is, they should specialize and become specialists. Using this approach to develop its strategy and structure, a firm could go in either or both of two directions. It could provide its best-in-world functional service to many other firms and thus grow, or it could engage in alliances with other firms that could provide the other high-quality functional services necessary for the collective alliance to market a viable product (Quinn, 1992; Quinn et al., 1996). Of course, both of these directions are available to start-up firms as well. Engaging in alliances means becoming a node in a network (even if the network is only a dyad) and, hence, the importance of network organization. Network organizations. Network organizations are of two types: traditional hierarchical network organizations and virtual organizations. The difference between the types is that unlike traditional network organizations, virtual organizations often do not manifest one or both of two concepts typical of organizations discussed earlier: the boundarymaintaining routine of distinguishing between members and nonmembers or the property of being hierarchically differentiated. Hierarchical network organizations. One traditional network organizational form is the hierarchical contractor form in which the prime contactor central node contracts with partner (e.g., subcontractor) organizations (which compose the periphery of the network and may have their own subcontractor organizations) to produce the required product (e.g., an office building). As described by Miles and Snow (1992, p. 64), the contractor form can be stable or

temporary, and individual partners often participate in multiple networks. Partners informally coordinate with each other when necessary, but the ultimate coordinative authority rests with the prime contractor. A second traditional network organization is the franchise form, in which top managers and systems and technical experts are at the central node (the headquarters) and customer sales or services are carried out by end node franchise units. The headquarters might choose to use any of the strategies described by Porter (1980) but generally requires that all franchises use the same strategy and maintain certain common attributes so that the firm and its franchisees can benefit from a brand name associated with the strategy and so that customers can know what to expect from any franchise. In this organizational form, franchises typically do not share knowledge or coordinate with each other but, to the contrary, are in competition for resources from the central node. A traditional network form finding more application is what Quinn et al. (1996) called the starburst network. Examples of this form are mutual fund groups and movie studios. A parent firm with a central intellectual competency spins off other organizations with which it maintains continuing relationships. “These spinoffs remain partially or wholly owned by the parent, usually can raise external resources independently, and are controlled primarily by market mechanisms” (Quinn et al., 1996, p. 19). In that a network organization has a defined membership of independent organizations, it conforms to the boundary-maintaining aspect of the definition of an organization, presented early in the chapter. This helps distinguish it from a marketplace of organizations exchanging goods and services. Virtual organizations. Virtual organizations are a network of traditional organizations and “are typified by member organizations contributing competencies to an entity that need not have legal existence or location properties (e.g., physical locations) that are easily identified” (Cooper & Muench, 2000, p. 191). It is important that “the boundaries of more traditional forms of organization are likely to be sharp and fixed over long intervals of time, whereas the (virtual organization) boundaries . . . 145

Copyright American Psychological Association. Not for further distribution.

George P. Huber

are likely to be vague and fluid” (p. 190). Thus, whether the term virtual organization is meant to accept the boundary-maintaining requirement of network organizations is unclear. Many virtual organizations are composed of differently specialized organizations contributing their specialized expertise to other member organizations or directly to the attainment of the network’s focal goal. In some situations, network members may fulfill their responsibilities to other members or the network in accord with contracts, as in agency theory. In other networks, members may fulfill their responsibilities because doing so determines the member’s continued membership in the network or the member’s future opportunities to participate in like networks, either of which may be important to the member’s future survival. Uses of the term virtual organization vary and probably encompass the spider’s web network form (Quinn et al., 1996). Spider’s web networks materialize when independent organizations recognize an opportunity or problem that is relevant to their interests and that requires the expertise and participation of other organizations. The intention of the participating organizations can be to establish either a permanent or project network organization. Examples include political action groups, securities exchanges, and library and research consortia. The independent organizations, generally not having a permanent communication system, tend to coalesce and coordinate through personal contacts, only delegating authority to a project leader organization when tighter coordination becomes necessary. The participating organizations are generally self-motivated to work toward whatever is the common end, but motivation is enhanced by other organizations stimulating “a sense of interdependency and identity with the problem at hand” (Quinn et al., 1996, p. 22). The component organizations of virtual organizations are often geographically distant from one another and, other than for spider’s web networks, tend to differ from each other in the products or services they offer. These two factors, differentiation and distance, generate difficulties in communication and thus coordination. The increases in communication capabilities made available through the 146

advances in information technology of the past few decades have enabled all organizations, but especially virtual organizations, to attenuate their communication difficulties and thus be much more effective than they otherwise would be. This phenomenon has contributed enormously in the past 2 decades to growth in the variety, number, and size of virtual organizations and suggests the need to consider the nature and circumstances of future organizations. ORGANIZATIONS IN THE FUTURE Will organizations in the future be different from those of their predecessor organizations that operated in the same domain? If so, how and why? From the earlier review of prominent organization theories, it is apparent that the answers to these questions must account for the fact that the nature of an organization is determined in part by the organization’s environment. I begin this section by considering the nature of future organizational environments. Because, as I explain shortly, the environments of for-profit organizations will change to a greater extent than will those for public sector organizations, it is more fruitful to examine the changes in these environments and then to highlight what the differences will be between the environments of for-profit and public sector organizations. In an earlier work (Huber, 2004), I analyzed in some depth forthcoming changes in the environments of business organizations and the effects of these changes on the nature of business firms. There, I concluded, using arguments I review shortly, that a strong case can be made that future business environments will manifest each of the following properties: ■

■

■ ■ ■

The amount of available scientific knowledge will be significantly greater, and its rate of growth will be increasing. The effectiveness of information, transportation, and manufacturing technologies will be greater. Environmental complexity will be greater. Environmental dynamism will be greater. Environmental competitiveness will be greater.

How these properties might influence the nature of future business organizations can best be under-

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

stood if we examine the root causes of the properties. I turn shortly to this task, but first I must address the differences between the future environments of for-profit and public sector organizations. The task environments of public sector and forprofit organizations will, of course, contain the same kinds of features, such as increasing complexity, dynamism, and competitiveness. The task environments will differ, however, in the extent to which they influence the nature and circumstances of organizations of the two types. For example, increases in the effectiveness of information, manufacturing, and transportation technologies affect the competitive environment of for-profit organizations greatly through facilitating competition from other firms for funding from customers, but public sector organizations compete for funding from the legislative and executive branches of government, and this competition is influenced hardly at all by advances in technologies. Similarly, technological advances contribute greatly to the ability of firms to produce new products but contribute much less to the ability of pubic sector organizations to produce new services. Further, the usefulness of new services in the eyes of the funding institutions of public sector organizations is quite subject to the influence of precedents and traditions, whereas the usefulness of new products in the eyes of the customer is much less subject to these influences. In general, whereas the kinds of changes in the environments of the public sector and for-profit organizations will be similar, their magnitudes and therefore their effects will be different. Further, the effects of environmental changes on pubic sector organizations will be more constrained by institutional forces than are the effects on for-profit organizations.

Forthcoming Changes in Business Environments Growth in scientific knowledge. Consider, as a component of the business environment, scientific knowledge. Why is growth in scientific knowledge relevant to changes in organizations? The answer is that more so than in earlier eras, advances in scientific knowledge contribute to improvements in the effectiveness of technology (Heilbroner, 1995; Mokyr, 1990, 2002), and improvements in technol-

ogy, as will become clear, have momentous effects on the nature of organizations. The amount of available scientific knowledge will be significantly greater and its rate of growth will be increasing, for four reasons. The first is straightforward: Knowledge leads to more knowledge. One discovery leads to another. Knowledge is its own generative raw material: the more you have the more you get, and the more you get the more you have in a geometric progression. Another reason is that there will be more scientists generating knowledge. This growth in the number of scientists follows from the economic development cycle, in which emerging economies import new-to-the-nation technologies, use low labor costs and supportive governmental policies to become competitive economies, and use the increased national wealth to support their own scientific community. A third reason new knowledge will be created at a greater rate is that increases in the capability and application of communication technologies greatly increase the availability of whatever knowledge exists, thus enabling a larger proportion of existing knowledge to serve as material with which to generate still more knowledge. The fourth reason why scientific knowledge will grow at an increasing rate is that scientific findings contribute to the development of new technologies, some of which contribute to advances in science, creating a synergistic spiral of advances in both communities. Data supporting this description of this synergistic coevolution of science and technology are included in Huber (2004). Increases in the variety and effectiveness of technologies. Consider, as a component of the business environment, technology. The effectiveness of technologies in the future will be increasing at an increasing rate for three of the reasons that scientific knowledge will increase at an increasing rate. That is, (a) on a worldwide basis there will be more engineers and other technologists creating new technologies and improving existing technologies, (b) increases in the capability and application of communication technologies will greatly increase the availability of whatever technology-enhancing knowledge exists, and (c) advances in science will contribute to the development of new technologies, 147

George P. Huber

Copyright American Psychological Association. Not for further distribution.

some of which will in turn contribute to the development of other new technologies.4 Increases in environmental complexity. New technologies lead to new products. They sometimes also replace existing products, but the effect is generally a net increase in the variety and number of products. An example is the technologies for highspeed transmission of information over long distances. Beyond the technologies of flags and blinking lights (which were, with a few exceptions, replaced), there came copper cable, short and long radio waves, fiber-optic cable, and lasers. Each new technology spawned new manufacturers and thus a greater variety of firms. But many of the existing firms did not disappear and thus there are a greater number of firms. More generally, as new products become available, there evolves a greater variety and number of users and consumers, firms, competitors, suppliers, regulatory agencies, and trade associations, all of which interact to create a more complex environment. Business environments in the future will be more complex.5 Increases in environmental opportunities, threats, dynamism, and competitiveness. Increases in the number and variety of (a) technologies available for improving performance, (b) suppliers available from which to obtain superior product components, and (c) firms available with which to partner advantageously will increase the rate at which opportunities to improve performance become available. Similarly, increases in the number and variety of (a) technological changes that threaten a firm’s current product lines, (b) competitors that threaten a firm’s market share, and (c) regulatory agencies that constrain a firm’s options will increase the rate at which threats to performance arrive. Thus, business environments in the future will be more dynamic, with increases in both opportunities and threats. In addition, as a consequence of improvements in informa-

tion and transportation technologies, knowledge and knowledge workers are both more mobile, thus increasing the capability of competitors. Further, as a result of advances in information, manufacturing, and transportation technologies, each of the more numerous competitors will be able to attack a firm’s markets more rapidly. That is, competitors will be able to produce their new product more quickly, market it more rapidly and intensively, and deliver it more quickly than ever before. Business environments in the future will be more competitive. What will be the effects of these changes in business environments on firms? I turn now to answering this question.

Changes in the Nature of Business Organizations Recall that by the nature of an organization I mean the attributes of its leadership, strategy, core process, structure, employees, culture, and routines, and the propensities and properties associated with combinations of these features, such as its competences. Immediately below, I describe changes in the nature of organizations that occurred in the 20th century. Afterward, I describe forthcoming changes in the nature of business organizations, changes that follow from the changes in the properties of future business environments. Historical changes. Certainly the environments of business organizations have become more complex, dynamic, and competitive in recent decades. Thus, if there had not been much environmentally induced change in the nature or circumstances of organizations in the recent past, it would be difficult to believe that there would be much in the near future. But the 20th century saw significant changes in organizational structures: ■

The multidivisional corporate form was invented and became commonplace.

4

It would seem that this increase in the effectiveness of technologies will soon slow as upper limits are reached. The dominant design concept (Tushman & Murmann, 1998) helps clarify why, even though proportional improvements within any one dominant technological design tend to decline across time and to become even incremental, the overall rate at which a given technology advances can nevertheless accelerate due to the arrival of new dominant designs. As an example, the overall performance of wire-connected telephones (a once-dominant design) improved incrementally for decades, but a qualitative improvement in performance took place when a new dominant design, the cellular phone, arrived to dominate its predecessor. Similarly, nonvoice data transmission technologies such as the telegraph and television (each a once dominant design) improved, but more significant technological improvements in nonvoice data transmission came through new dominant designs, such as e-mail and the internet.

5

Data supporting this conclusion are noted in Huber (2004, pp. 27–28, 42–43). It is important to recognize, however, that on occasion negative feedback forces arise and curtail or even reverse the growth in the variety or number of particular entities (e.g., specific products or specific types of businesses; see Huber, 2004, p. 43).

148

Organizations: Theory, Design, Future

■

■

■

Copyright American Psychological Association. Not for further distribution.

■

■ ■ ■

The assembly line and the administrative machinery for its control were invented and became commonplace. Matrix and project team structures evolved and became pervasive. Firms’ divestiture of components not directly related to their core competence became commonplace. Alliances, virtual organizations, and other network arrangements of autonomous firms increased greatly in number and in the range of their application. Firms became flatter. Outsourcing of tasks to clients became pervasive. The use of the conglomerate firm declined greatly.

Each of the changes can be traced to the increased effectiveness of newly available technologies, more dynamic environments, or more competitive environments. Five of the eight occurred in the second half of the century. Of course there were also changes in core technologies (as noted above), organizational routines (e.g., less employment security), employee qualifications (e.g., more highly educated), and organizational culture (e.g., less vertical power distance). Given this recent history, it appears likely that the nature of firms will change. Future changes in business organizations. The following covers types of changes that may be expected in organizations in the future. Changes resulting from increases in the variety and effectiveness of technologies. New or greatly improved technologies will enable the development of novel core technologies (e.g., core business processes) such as the information technologies used by Amazon and Google. They will also enable the development of new products and services such as those marketed by IBM, Google, and SAP. Changes resulting from increased complexity. We noted above forthcoming increases in the number and variety of relevant entities in the firm’s environment (new technologies, and more alliance partners, competitors, suppliers, regulators), each of which can generate threats or opportunities. Firms that adapt to these changes will be favored by the environment and firms that do not will be selected out.

It seems reasonable to expect that surviving firms will be those possessing the abilities to (a) detect and interpret a greater variety of environmental thrusts, (b) determine what, if any, action is required for a greater variety of thrusts, and (c) effectuate the chosen action for a greater variety of thrusts. Two factors that will retard the outsourcing of these tasks are the need for speed and the need for the interpreting and deciding units to be very well informed about the strengths, weaknesses, and circumstances of the organization. For these reasons, one can expect that in the future, the proportion of a firm’s structure and employees that are devoted to scanning, interpreting, deciding about, and responding to the environment will be greater. Changes resulting from increased dynamism. We also noted that forthcoming increases in the effectiveness of information, transportation, and manufacturing technologies will enable each member of each of the entity types noted just above to generate threats or opportunities more rapidly. Thus, there will be a multiplier effect: more generators of threats and opportunities, each generating its events more rapidly. Firms that respond to these changes appropriately will be favored by the environment, and those that do not will be selected out. Thus, as before, it seems reasonable to expect that surviving firms will be those possessing the abilities to (a) detect a greater number of thrusts per interval of time, (b) determine more rapidly what, if any, action is required, and (c) effectuate any chosen action more quickly. We can expect, therefore, that in the future, the firm’s units and employees that are devoted to scanning, interpreting, and deciding will be capable of acting more rapidly. This capability for more rapid action may be achieved through enhanced information infrastructure, enhanced communication routines, and enhanced employee and managerial selection and training. Changes resulting from increased competitiveness. Increases in the effectiveness of information and transportation technologies have shrunk the world physically and culturally. Although these changes have made new customers available to many firms, they have also made the customers of many firms subject to theft by other firms. Customers, to an extent never seen before, are 149

Copyright American Psychological Association. Not for further distribution.

George P. Huber

no longer buffered from information about their options. Thus, more firms are more likely to be in a state of hypercompetition (D’Aveni, 1995), with many increasingly agile competitors and no cultural, spatial, or temporal barriers to give them advantage over these competitors. It seems that one pattern of response will be to put into place the enhanced information infrastructure, communication routines, and selection and training routines such as those noted above. But in a rapidly changing competitive environment, firms cannot afford to take a purely defensive posture. Such an environment will favor firms that engage in exploration for new products (including services) and bring them to market quickly. Thus, in the future, firms will provide the structures, processes, and employees to engage in more, and more effective, search, learning, and innovation. Among these processes will be processes for development and deployment of dynamic capabilities. (In Huber, 2004, I provided descriptions of organizational structures and processes for engaging in more effective search, learning, and innovation.) These changes will be prompted by the changes in business environments (e.g., more fickle customers and more, and faster-moving, competitors). The changes will move all firms toward strategic positions more like the pure prospector firms described earlier. That is, defender firms will be forced to engage in more exploration. Analyzers will shift their distribution attention and resources toward that of the prospector form. A much greater proportion of firms will become ambidextrous. Changes in the capabilities of top managers. The top managers of future firms must be able to function in environments that are more complex and more dynamic than environments were in the past. This means either that these managers must be more cognitively complex and more capable under stress, or they must be able to develop and manage a team of subordinates with these properties, either as individuals or as a team. In this regard, it is worth noting the research demonstrating that most managers lose their ability or inclination to actively monitor their environments after a relatively short tenure (Hambrick, 2007; Henderson, Miller, & Hambrick, 2006; D. Miller, 1991). This is thought to be, at least in part, a consequence of their coming to rely on a 150

familiar set of human information sources or channels. The faster changing environments of the future will favor firms that avoid this situation. In addition, these environments will favor firms whose top managers are able to shape their organizations to compete in environments that are more complex, dynamic, and competitive than were environments in the past and are able to visualize and establish the core technologies, structures, employees, routines, and cultures that will enable their organization to function well in these environments. Future environments will be more challenging. All this suggests that in the future, the owners and boards of directors of firms will select, incentivize, and monitor top managers so as to ensure that they are active monitors of the firm’s environment and are active in adjusting the firm’s features to be aligned with changes in this environment.

Forthcoming Changes in Circumstances of Business Organizations By its circumstances, I mean an organization’s size, maturity, performance, status with regard to survival, and the current direction and speed of change of these variables. An organization’s circumstances are in part a consequence of the goodness of fit between its attributes and its environment relative to the goodness of fit of its competitors in the same environment. Conflicting effects on organizational circumstances of changes in technology, and in environmental complexity, dynamism, and competitiveness, do not permit assessment of their net effect on the circumstances of firms in the future. Heightening this predicament is the fact that other factors, such as the state of the economy and the level of governmental intervention, directly affect the circumstances of firms and also interact with and influence changes in technology and competitiveness. As a result of this combination of conditions, the following analyses address specific changes in firms’ circumstances brought about by individual environmental changes rather than addressing net changes in firms’ circumstances brought about by multiple environmental changes. Changes in size. Advances in information technology influence size in two ways. One is that the improved technologies increase the effectiveness of

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

coordination and control processes, thereby enabling some firms to become much larger than they would be otherwise. But increases in the effectiveness of coordination also facilitate the functioning of network organizations, thereby enabling component firms in the network to focus on their core competence (in which case they will become smaller unless they are able to create a large market for this competence). Advances in information and manufacturing technology also lead to automation, the use of equipment to do what employees once did. In this way, automation reduces the number of jobs in some of a firm’s occupations and therefore, all else equal, reduces the size of the firm’s workforce. Automation is, however, generally associated with changes in the firm’s strategic scope, core technology, or routines. Such changes often generate concomitant increases in the number of employees in other of the firm’s occupations, so the net effect of automation on firm size varies with the nature of other simultaneous changes. Increases in environmental competitiveness will result in the demise of some firms and may result in the growth of the surviving firms. That is, if the associated difficulties in coordination and control can be sufficiently attenuated, perhaps with new information technologies or new organizational structures, increases in environmental competitiveness can lead to increases in size for surviving firms in those industries where economies of scale retard the replacement of the failed firms. Changes in founding rates. The growth in the number and variety of new technologies will open new product and process domains and thus more opportunities for the founding of new firms. So will the lowered barriers to entry for new firms that will be associated with future business environments (e.g., increased mobility of employees, improved information technologies for reaching customers, availability of specialist firms with which to form alliances). However, without predecessor firms from which to learn about how to function in their new domain, top managers of new firms in new domains are more likely to make mistaken judgments. With minimal resources with which to buffer itself from the effects of misjudgments, a newly founded firm’s

survival is in question. On the other hand, the environmental dynamism that causes new technologies to arrive suddenly and unpredictably may give new firms some short-term advantage, as the inertial forces of existing firms may prevent them from rapidly entering the new domain. Changes in the frequency of unexpected and disastrous performance declines. Advances in technology can contribute to the design of organizational systems in which the technological components are so complex, so tightly coupled, and with features so deeply embedded that the human components of the organizational system are unable to cope in a timely manner with the crises that occur when the technological component malfunctions. Examples are malfunctions in the atomic power plants at Three Mile Island (Perrow, 1984) and at Chernobyl (“Chernobyl accident,” 2008). Distinct from the consequences of highly complex technologies deeply embedded in organizational systems are the consequences of advances in information technology that facilitate the development of organizational and societal systems so complex that the humans responsible for managing the systems are unable to anticipate their failure or quickly identify the most effective interventions. The failure of the U.S. financial system to avoid or quickly remedy the credit crisis of 2008–2009 is an example of such a system. Because organizational environments will, in the future, be more complex, many of the societal and organizational systems that link to these environments will also be more complex. Thus, two separate technology-related arguments indicate that it is reasonable to expect organizations in the future to more frequently experience unexpected and disastrous performance. See also Weick and Sutcliffe (2001) for an analysis of the dangers of tightly coupled organizations and of structural attributes and routines that may reduce the possibility of unexpected and disastrous poor organizational performance. Changes in performance, rate of demise, and variation in life span. Increases in environmental dynamism and competitiveness will contribute to increases in the arrival rate of shocks with which firms must contend. Although the changes in firms’ 151

Copyright American Psychological Association. Not for further distribution.

George P. Huber

structure, routines, and employees discussed earlier will somewhat attenuate the effects of more, and more rapidly arriving, shocks, the net effect seems likely to be a decrease in performance and a decrease in the life span of firms. Further, changes in technology, and in the products and processes of competitors, will cause many of these shocks to be novel to the firm, thus causing the firm’s responses to have a lower success rate and amplifying the decreases in performance and life span. This same reasoning and outcome applies to populations of firms as well as individual firms. At the same time, however, a possible result of the high death rate associated with severe shocks or a series of lesser shocks is reduced competition in the domain and the resultant development of oligopolies or monopolies. An example is the consequence of the dot-com bust at the turn of the 20th century; many firms failed and a few survived to dominate their respective niches. Monopolistic and oligopolistic firms tend to become large and are often able to exert considerable control over their environments. As a result, they tend to survive for a long time. Putting this observation together with the above argument concerning expected decreases in organizational life spans, it seems that an outcome of forthcoming increases in environmental dynamism and competitiveness is likely to be an increase in the variation of firms’ life spans. CONCLUDING OBSERVATIONS AND COMMENTARY This chapter has examined three subjects: organization theory, organization design, and organizations in the future. Within the Organization Theory section, I examined seven prominent theories. At this point, after reviewing the fields of organization theory and organization design and considering some ideas on the nature and circumstances of future organizations, several questions arise. What do organization theorists know about organizations? One answer to this question follows from the fact that each of the seven theories discussed earlier has been substantiated many times in many contexts. Thus, a summarizing statement from any of the theories would be something that organization theorists would claim was something that they know to be 152

valid, insofar as anything can be proved empirically. For example, in the earlier description of population ecology theory, I noted that “population ecology theory explains the nature and circumstances of populations in terms of the suitability of the attributes of the organizations in the population relative to the attributes of organizations in competing populations.” This could easily be rephrased as “the suitability of the attributes of the organizations in a population, relative to the attributes of organizations in competing populations, determines the nature and circumstances of the population.” This theoretical assertion is something that organization theorists know to be true. Similar theoretical statements from the other six theories are known to be true. Further, from Section Summary and Controversial Issues in the Organization Theories section follows an overarching theoretical assertion that seems not to be contradicted by any of the theories and that organization theorists would hold to be true: The nature and circumstances of organizations are determined by the goodness of fit between (a) the organization’s attributes and (b) the constraints, threats, and opportunities posed by the organization’s external and internal environments. Are organization theory researchers currently testing the prominent organization theories, or extensions of the theories? I know of no research on this matter, but evidence from a large convenience sample suggests that they are not. Walsh et al. (2006) reported that of the 429 submissions to the Organization and Management Theory Division for presentation at the Academy of Management’s 2005 annual meeting, the percentage of papers submitted in each of the established theoretical categories (as reported by the authors of the papers) were as follows: institutional theory, 25.4%; network theory, 16.8%; population ecology, 6.7%; agency theory, 4.5%; resource dependence theory, 3.9%; transaction cost theory, 3.4%; contingency theory, 2.5%; and stakeholder theory, 2.5%. “None of the above” accounted for 56% of the papers submitted. (p. 658) Why might it be that more than half of these active organization theory researchers were not

Copyright American Psychological Association. Not for further distribution.

Organizations: Theory, Design, Future

examining these established theories? I offer three possible explanations. One is that most of the prominent theories are regarded by many organization theorists as relatively mature (the most conspicuous exceptions being network theory and institutional theory). Therefore, these researchers might have believed that working with established theories would result in a lower likelihood of being able to make a significant contribution to the field or even a contribution sufficiently significant that the submitted paper would be accepted for presentation at the meeting. Because it is more likely that one can make a significant contribution in a new area, the researchers surveyed may have been thinking in these terms when they chose the focus of the research they submitted. A second possible explanation is that the researchers were studying current phenomena in existing organizations and environments and were thus prone not to see as relevant theories that were developed decades ago (when environments were less complex, dynamic, and competitive, and when fewer organizations were as global, changeful, or threatened as are today’s). That is, possibly with little conscious thought, researchers perceived the established theories as unlikely to provide explanations pertaining to the questions in which they were interested or to the environments in which they were testing explanations generated from sources other than the established theories. A third possible explanation is that the researchers actually were drawing on existing theories to help themselves understand new organizational phenomena but viewed their contribution as the elucidation of the phenomena they examined, not as an extension of the theories on which they drew for insights. In view of the changing nature of organizational environments, must the prominent organization theories be modified to maintain their relevance? As indicated by the data just discussed just, and as I noted in the earlier discussion of change in the foci of organization theory researchers, some of the most prominent and mature organization theories seem to be perceived by members of the organization theory community as insufficiently relevant for understanding organizations in current organizational environments. Few would argue that this is an acceptable

state of affairs. Thus, the answer to the question is yes: To maintain their relevance in view of the changing nature of organizational environments, at least some of the theories must be modified. As I suggest shortly, possible approaches to modifying the theories include updating or extension. Are organization designers currently drawing on the prominent organization theories? From my field research interviews with, and teaching of, organization design-level managers, I conclude that such managers are unfamiliar with the theories and therefore do not draw on them per se. Probably more pertinent is that in the organization design literature (cited earlier), rarely does one find reference to the individual theories or to the specific explanations associated with the individual theories. Instead, one finds frequent mention, using the terminology of managers rather than that of management theorists, of the basic idea of fit theory, that the nature and circumstances of organizations are determined by the goodness of fit between (a) the organization’s attributes and (b) the constraints, threats, and opportunities posed by the organization’s external and internal environments. These observations suggest that one possible answer to the above question is that those executives or other persons entitled to design organizations are unfamiliar with the intricacies of the theories, but that they do sense and rely on the basic idea of fit theory, described earlier as encompassing each of the seven prominent organization theories. Executives’ sense of fit theory could have been obtained from exposure to the business press, conversations with peers or consultants, or personal observations. A second possible answer to this question is that the authors of the organization design books and articles (on which executives might draw for guidance) seek to portray their works as useful for preparing organizations for evolving environments, and even though the authors might be familiar with the theories, they are prone not to see, or attempt to sell, theories that were developed decades ago as relevant to their clients’ situations. If organization theorists see the usage level of their theories as unsatisfactory, what might they do? One response would be to update or increase the scope of the prominent theories by drawing on related, or more recent, work done outside of the theory. For 153

Copyright American Psychological Association. Not for further distribution.

George P. Huber

example, strategic choice theory could be updated and extended with work on upper echelons (Hambrick, 2007; Hambrick & Finkelstein, 1987; Kaiser et al., 2008) and with work on executive motivations and personality variables (Malmendier & Tate, 2005; Sanders, 2001; Wiseman & GomezMejia, 1998). Another response would be to update the theories, or the methodologies for studying the theories, by considering what must be done to increase their validity in more complex, dynamic, or competitive organizational environments. For example, if population ecologists wanted their work to be more relevant to executives or other organization designers, they might address the fact that organization designers are making decisions with respect to future environments, not to past or even current environments. As historical data become less relevant to current and future conditions, what methodological or conceptual adjustments might population ecologists make to cause their findings to be more applicable to practitioners facing new and faster changing environments? What is the state of the field of organization theory? Organization theory at present encompasses several theories that are well substantiated, explanatory, and predictive. In this sense, organization theory is an established field. However, organization theory is also an evolving field. Further, beyond the fact that it is both established and evolving, it is highly likely that the study of organizations will become a larger proportion of the research conducted in the behavioral and social sciences in the future than it has been in the past. That organization theory is evolving is indicated in two ways. One is that organization theorists are actively contributing to and drawing on the very active research areas of organizational learning, organizational change, and strategic management. The result is that current organization theories are being extended by findings from these areas. For example, it was noted earlier that population ecologists are studying organizational learning and adaptation in spite of their earlier inclination to view adaptation as improbable and that institutional theorists are studying the influence of organizations on institutions as contrasted with an earlier view of institutional forces as essentially irresistible. Indeed, the definition of 154

organization theory as a field of study and knowledge centrally concerned with explaining the nature and circumstances of organizations suggests that work in organizational learning, organizational information processing, organizational decision making, and organizational change ought to be regarded as within the field of organization theory. More generally, examination of the content and authorship of work in the organization theory literature and of the research presented under the auspices of the Organization and Management Theory Division at the annual meetings of the Academy of Management indicates that much of the current work of researchers who see themselves in the organization theory field deals with new issues and new explanations rather than focusing on currently prominent organization theories. Much of this work focuses on processes and structures that enable organizations to be more flexible and adaptive, that is, to be more effective in evolving organizational environments. Organizations now employ a larger proportion of the world’s population than ever before and are directly involved in producing the world’s goods and services to a greater extent than ever before. Organizations are, more than ever before, the mechanisms through which economies operate, and organizational actions and circumstances have greater effects on economies and societies than ever before, as became especially clear in 2008 and 2009. That the environments of organizations will become more complex, more dynamic, and (for business organizations) more competitive will force organizations to change and thus will serve as a strong external force prompting research in organization theory. A good argument can be made that in the foreseeable future, no other behavioral or social science, with the possible exception of macroeconomics, will be more blessed or cursed with a stronger externally driven demand for new understandings than will organization science and, within organization science, organization theory.

References Academy of Management Review. (1992). 17(3), 404–611. Adler, P. S., Forbes, L. C., & Willmott, H. (2007). Critical management studies. The Academy of Management Annals, 1, 119–179.

Organizations: Theory, Design, Future

Administrative Science Quarterly. (1998). 443(2), 235–497. Aldrich, H. E. (2008). Organizations and environments. Palo Alto, CA: Stanford Business Books. Aldrich, H. E., & Ruef, M. (2006). Organizations evolving. London: Sage.

Copyright American Psychological Association. Not for further distribution.

Allen, T. (1977). Managing the flow of technology: Technology transfer and the dissemination of technological information within the R&D organization. Cambridge, MA: Massachusetts Institute of Technology Press. Alvesson, M., & Deetz, S. A. (2006). Critical theory and postmodernism approaches to organizational studies. In S. R. Clegg, C. Hardy, T. B. Lawrence, & W. R. Nord (Eds.), The SAGE handbook of organizational studies (pp. 191–217). London: Sage.

Burt, R. S. (2007). Brokerage and closure: An introduction to social capital. New York: Oxford University Press. Burton, R. M., DeSanctis, G., & Obel, B. (2006). Organizational design: A step-by-step approach. New York: Cambridge University Press. Burton, R. M., & Obel, B. (2004). Strategic organizational diagnosis and design: The dynamics of fit with OrgCon software (3rd ed.). Norwell, MA: Kluwer Academic Publishers. Chandler, A. (1962). Strategy and structure: Chapters in the history of American enterprise. Cambridge, MA: Massachusetts Institute of Technology Press. Chattopadhyay, P., Glick, W. H., & Huber, G. P. (2001). Academy of Management Journal, 44, 937–955.

Anderson, C. R., & Paine, F. T. (1975). Managerial perceptions and strategic behavior. Academy of Management Journal, 18, 811–824.

Chernobyl accident. (2008). In Encyclopedia Britannica. Retrieved December 3, 2008, from http://www. britannica.com/EBchecked/topic/109428/Chernobylaccident

Argote, L. (1999). Organizational learning: Creating, retaining, and transferring knowledge. Norwell, MA: Kluwer.

Child, J. (1972). Organizational structure, environment and performance: The role of strategic choice. Sociology, 6, 1–15.

Astley, W. G., & Van de Ven, A. H. (1983). Central perspectives and debates in organization theory. Administrative Science Quarterly, 28, 245–273.

Child, J. (1997). Strategic choice in the analysis of action, structure, organizations and environment: Retrospect and prospect. Organization Studies, 18, 43–77.

Augier, M., March, J. G., & Sullivan, B. N. (2005). Notes on the evolution of a research community: Organization studies in anglophone North America, 1945–2000. Organization Science, 16, 85–95.

Clegg, S., Hardy, C., Lawrence, T., & Nord, W. R. (Eds.). (2006). The SAGE handbook of organizational studies. London: Sage.

Barnett, W. P., & Carroll, G. R. (1995). Modeling internal organizational change. Annual Review of Sociology, 21, 217–236. Baum, J. R., & Wally, S. (2003). Strategic decision speed and firm performance. Strategic Management Journal, 24, 1107–1129. Beckert, J. (1999). Agency, entrepreneurs, and institutional change. The role of strategic choice and institutional practices in organizations. Organization Studies, 20, 777–799. Bonacich, P. B. (1987). Power and centrality: A family of measures. American Journal of Sociology, 92, 1170–1182. Brown, S. L., & Eisenhardt, K. M. (1997). The art of continuous change: Linking complexity theory and timepaced evolution in relentlessly shifting organizations. Administrative Science Quarterly, 42, 1–34. Burke, W. W. (2002). Organization change: Theory and routine. Thousand Oaks, CA: Sage.

Collins, J. C., & Porras, J. I. (1997). Built to last: Successful habits of visionary companies. New York: HarperCollins. Cooper, W. A., & Muench, M. L. (2000). Virtual organizations: Routine and the literature. Journal of Organizational Computing and Electronic Commerce, 10, 189–208. Cyert, R. M., & March, J. G. (1963). The behavioral theory of the firm. Englewood Cliffs, NJ: Prentice-Hall. D’Aveni, R. A. (1995). Hypercompetition: Managing the dynamics of strategic maneuvering. New York: The Free Press. David, R. J., & Han, S.-K. (2004). A systemic assessment of the empirical support for transaction cost economics. Strategic Management Journal, 25, 39–58. Davis, G. F., Diekman, K. A., & Tinsley, C. H. (1994). The deinstitutionalization of conglomerate firm in the 1980s. American Sociological Review, 59, 547–570.

Burns, T., & Stalker, G. (1961). The management of innovation. London: Tavistock.

Davis, G. F., & Marquis, C. (2005). Prospects for organization theory in the early twenty-first century: Institutional fields and their mechanisms. Organization Science, 16, 332–343.

Burt, R. S. (1992). Structural holes: The social structure of competition. Cambridge, MA: Harvard University Press.

Denrell, J. (2003). Vicarious learning, undersampling of failure, and the myths of management. Organization Science, 14, 227–243. 155

Copyright American Psychological Association. Not for further distribution.

George P. Huber

DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48, 147–160.

Gibson, C. B., & Birkinshaw, J. (2004). The antecedents, consequences and mediating role of organizational ambidexterity. Academy of Management Journal, 47, 209–226.

DiMaggio, P. J., & Powell, W. W. (1991). Introduction. In W. W. Powell & P. J. DiMaggio (Eds.), The new institutionalism in organizational analysis. Chicago: University of Chicago Press.

Gouldner, A. W. (1954). Patterns of industrial democracy. Glencoe, IL: Free Press. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1380.

Dobrev, S. D., Kim, T.-Y., & Carroll, G. R. (2002). The evolution of organizational niches: U.S. automobile manufacturers, 1885–1981. Administrative Science Quarterly, 47, 233–264.

Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. America Journal of Sociology, 91, 481–510.

Donaldson, L. (1995). Contingency theory. Aldershot, Hants, England: Dartmouth.

Gresov, C. & Drazin, R. (1997). Equifinality functional equivalence in organization design. The Academy of Management Review, 22, 403–428.

Donaldson, L. (2001). The contingency theory of organizations. Thousand Oaks, CA: Sage. Dosi, G., Nelson, R. R., & Winter, S. (Eds.). (2000). The nature and dynamics of organizational capabilities. New York: Oxford University Press. Doty, D. H., Glick, W. H., & Huber, G. P. (1993). Fit, equifinality and organizational effectiveness: A test of two configurational theories. Academy of Management Journal, 36, 1196–1250. Drazin, R., & Van de Ven, A. H. (1985). Alternative forms of fit in contingency theory. Administrative Science Quarterly, 30, 514–539. DuBois, A., & Gadde, L.-E. (2002). The construction industry as a loosely coupled system: Implications for productivity and innovation. Construction Management and Economics, 20, 621–631. Duncan, R. (1976). The ambidextrous organization: Designing dual structures for innovation. In R. H. Kilmann, L. R. Pondy, & D. P. Selvin (Eds.), The management of organization design (pp. 167–188). New York: North-Holland. Eisenhardt, K. M. (1989a). Agency theory: An assessment and review. Academy of Management Review, 14, 57–74.

Gulick, L. H., & Urwick, L. (Eds.). (1937). Papers in the science of administration. New York: Columbia University Press. Hage, J. (1965). An axiomatic theory of organizations. Administrative Science Quarterly, 10, 289–320. Hambrick, D. C. (1983). Some tests of the effectiveness and functional attributes of Miles and Snow’s strategic types. Academy of Management Journal, 26, 5–26. Hambrick, D. (2007). Upper echelons theory: An update. Academy of Management Review, 32, 334–343. Hambrick, D. C., & Finkelstein, S. (1987). Managerial discretion: A bridge between polar views of organizational outcomes. Research in Organizational Behavior, 9, 369–406. Hannan, M. T., & Carroll, G. R. (1992). Dynamics of organizational populations: Density, legitimation, and competition. New York: Oxford University Press. Hannan, M. T., & Freeman, J. (1977). The population ecology of organizations. American Journal of Sociology, 82, 929–934. Hannan, M. T., & Freeman, J. (1989). Organizational ecology. Cambridge, MA: Harvard University Press.

Eisenhardt, K. M. (1989b). Making fast strategic decisions in high-velocity environments. Academy of Management Journal, 32, 543–576.

Hansen, M. (1999). The search-transfer problem: The role of weak ties in sharing knowledge across organization subunits. Administrative Science Quarterly, 44, 82–111.

Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities: What are they? Strategic Management Journal, 21, 1105–1106.

Harary, F. (1959). Graph theoretical methods in the management sciences. Management Science, 5, 387–403.

Emirbayer, M., & Mische, A. (1998). What is agency? American Journal of Sociology, 103, 962–1023. Fayol, H. (1949). General and industrial management. London: Pitman. Ford, R., & Randolph, W. (1992). Cross-functional structures: A review and integration of matrix organization and project management. Journal of Management, 18, 267–289. 156

Hargadon, A., & Sutton, R. I. (1997). Technology brokering and innovation in a product development firm. Administrative Science Quarterly, 42, 716–749. Haunschild, P., & Chandler, D. (2008). Institutionallevel learning: Learning as a source of institutional change. In R. Greenwood, C. Oliver, K. SahlinAndersson, & R. Suddaby (Eds.), The Sage handbook of organizational institutionalism (pp. 627–647). Thousand Oaks, CA: Sage.

Organizations: Theory, Design, Future

Heilbroner, R. L. (1995). Visions of the future: The distant past, yesterday, today, tomorrow. New York: Oxford University Press.

Lawrence, P. R., & Lorsch, J. W. (1967). Organization and environment. Cambridge, MA: Harvard University Press.

Henderson, A. D., Miller, D., & Hambrick, D. C. (2006). How quickly do CEOs become obsolete? Industry dynamism, CEO tenure, and company performance. Strategic Management Journal, 27, 447–460.

Leavitt, H. J. (1951). Some effects of certain communication patterns on group performance. Journal of Abnormal Social Psychology, 46, 38–50.

Copyright American Psychological Association. Not for further distribution.

Hickson, D., Hinings, C. R., Clegg, S. R., Child, J., Aldrich, L., Karlpik, L. & Donaldson, L. (1988). Offence and defence: A symposium. Organization Studies, 9, 1–32. Hrebiniak, L. G., & Joyce, W. F. (1985). Organizational adaptation: Strategic choice and environmental determinism. Administrative Science Quarterly, 30, 336–349. Huber, G. P. (1991). Organizational learning: The contributing processes and the literatures. Organization Science, 2, 88–115. Huber, G. P. (2004). The necessary nature of future firms: Attributes of survivors in a changing world. Thousand Oaks, CA: Sage. Human, S. E., & Provan, K. G. (1997). An emergent theory of structure and outcomes in small-firm strategic manufacturing networks. Academy of Management Journal, 40, 368–403. Jennings, D. F., & Seaman, S. L. (1994). High and low levels of organizational adaptation: An empirical analysis of strategy, structure and performance. Strategic Management Journal, 15, 459–475. Kaiser, R. B., Hogan, R., & Craig, S. B. (2008). Leadership and the fate of organizations. American Psychologist, 63, 96–110.

Levinthal, D. A., & March, J. G. (1993). The myopia of learning. Strategic Management Journal, 14, 95–112. Luce, R. D. (1950). Connectivity and generalized cliques in sociometric group structure. Psychometrika, 15, 169–190. Mackenzie, K. D. (1986). Organizational design: The organizational audit and analysis technology. Norwood, NJ: Ablex. Malmendier, U., & Tate, G. (2005). CEO overconfidence and corporate investment. Journal of Finance, 60, 2661–2700. March, J. G. (1965). Handbook of organizations. Chicago: Rand McNally. March, J. G. (1991). Exploration and exploitation in organizational learning. Organization Science, 2, 71–87. March, J. G. (2004). Parochialism in the evolution of a research community: The case of organization studies. Management and Organization Review, 1, 5–22. March, J. G. (2007). The study of organizations and organizing since 1945. Organization Studies, 28, 9–21. March, J. G. (2008). Explorations in organizations. Palo Alto, CA: Stanford University Press. March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley.

Katz, D., & Kahn, R. (1966). The social psychology of organizations. New York: Wiley.

Martin, J. (1992). Cultures in organizations: Three perspectives. New York: Oxford University Press.

Kerlinger, F. N. (1986). Foundations of behavioral research. New York: Holt, Rinehart & Winston.

McGregor, D. (1960). The human side of enterprise. New York: McGraw-Hill.

Kilduff, M., & Mehra, A. (1997). Postmodernism and organizational research. Academy of Management Review, 22, 453–481.

Meyer, A. D. (1982). Adapting to environmental jolts. Administrative Science Quarterly, 27, 515–537.

Kraatz, M. S. (1998). Learning by association? Interorganizational networks and adaptation to environmental change. Academy of Management Journal, 41, 621–643. Kuhn, T. S. (1970). The structure of scientific revolutions. Chicago: University of Chicago Press. Larson, E. W., & Gobeli, D. H. (1987). Matrix management: Contradictions and insights. California Management Review, 29, 126–138. Lavie, D. (2006). The competitive advantage of interconnected firms: An extension of the resource-based view. Academy of Management Review, 31, 638–658.

Meyer, A. D., Tsui, A. S., & Hinings, C. R. (1993). Fit, equifinality, and organizational effectiveness: Configurational approaches to organizational effectiveness. Academy of Management Journal, 36, 1175–1195. Meyer, J. W., & Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83, 340–363. Miles, R. E., & Snow, C. C. (1978). Organizational strategy, structure and process. New York: McGraw Hill. Miles, R. E., & Snow, C. C. (1992). Causes of failure in network organizations. California Management Review, 34, 53–72. 157

George P. Huber

Miles, R. E., & Snow, C. C. (1994). Fit, failure and the hall of fame: How companies succeed or fail. New York: The Free Press. Miles, R. E., Snow, C. C., Meyer, A. D., & Coleman, H. J., Jr. (1978). Organizational strategy, structure, and process. Academy of Management Review, 3, 546–562. Miller, D. (1991). Stale in the saddle: CEO tenure and the match between organization and environment. Management Science, 37, 34–52. Miller, D. (1992). Environmental fit versus internal fit. Organization Science, 3, 159–178.

Copyright American Psychological Association. Not for further distribution.

Miller, J. G. (1972). Living systems: The organization. Behavioral Science, 17, 1–182. Miller, J. G. (1978). Living systems. New York: McGraw-Hill. Miller, S. J., & Wilson, D. C. (2006). Perspectives on organizational decision making. In S. R. Clegg, C. Hardy, T. B. Lawrence, & W. W. Nord (Eds.), The Sage handbook of organizational design (pp. 469–484). London: Sage. Mintzberg, H. (1978). Patterns in strategy formation. Management Science, 24, 934–948. Mintzberg, H. (1979). The structuring of organizations. Englewood Cliffs, NJ: Prentice Hall. Mintzberg, H. (1981). Organization design: Fashion or fit? Harvard Business Review, 59, 103–116. Mintzberg, H. (1983). Structure in fives: Designing effective organizations. Englewood Cliffs, NJ: Prentice Hall. Mokyr, J. (1990). The levers of riches: Technological creativity and economic progress. New York: Oxford University Press. Mokyr, J. (2002). The gifts of Athena: Historical origins of the knowledge economy. Princeton, NJ: Princeton University Press. Mooney, J. D., & Reily, A. C. (1947). The principles of organization. New York: Harper.

O’Reilly, C. A., & Tushman, M. L. (2008). Ambidexterity as a dynamic capability: Resolving the innovator’s dilemma. Research in Organizational Behavior, 28, 185–206. Orton, J. D., & Weick, K. E. (1990). Loosely coupled systems: A reconceptualization. Academy of Management Review, 15, 202–223. Parkhe, A., Wasserman, S., & Ralston, D. A. (2006). New frontiers in network theory development. Academy of Management Review, 31, 560–568. Parsons, T. (1956). Suggestions for a sociological approach to the theory of organizations. Administration Science Quarterly, 1, 63–85, 225–239. Pennings, J. M. (1992). Structural contingency theory: A reappraisal. Research in Organizational Behavior, 14, 267–309. Perrow, C. (1961). The analysis of goals in complex organizations. American Sociological Review, 26, 854–866. Perrow, C. (1970). Organizational analysis: A sociological view. Belmont, CA: Wadsworth. Perrow, C. (1984). Normal accidents: Living with high-risk technologies. New York: Basic Books. Peters, T. J., & Waterman, R. H., Jr. (1982). In search of excellence: Lessons from America’s best-run companies. New York: Warner Books. Pettigrew, A. M., Woodman, R. W., & Cameron, K. S. (2001). Studying organizational change and development: Challenges for future research. Academy of Management Review, 44, 697–713. Pfeffer, J. (1997). New directions for organization theory: Problems and prospects. New York: Oxford University Press. Pfeffer, J., & Salancik, G. (1978). The external control of organizations: A resource dependence approach. New York: Harper & Row.

Nadler, D. A., & Tushman, M. L. (1997). Competing by design. New York: Oxford University Press.

Ployhart, R. (2006). Staffing in the 21st century: New challenges and strategic opportunities. Journal of Management, 32, 868–897.

Nag, R., Hambrick, D. C., & Chen, M.-J. (2007). What is strategic management, really? Inductive derivation of a consensus definition of the field. Strategic Management Journal, 28, 935–955.

Porter, M. E. (1980). Competitive strategy: Techniques for analyzing industries and competitors. New York: The Free Press.

Nebus, J. (2006). Building collegial information networks: A theory of advice network generation. Academy of Management Review, 31, 615–637. Nelson, R. R., & Winter, S. (1982). An evolutionary theory of economic change. Cambridge, MA: Belknap. Nilakant, V., & Rao, H. (1994). Agency theory and uncertainty in organizations: An evaluation. Organization Studies, 15, 649–672. O’Reilly, C. A., & Chatman, J. A. (1996). Culture as social control: Corporations, cults, and commitment. Research in Organizational Behavior, 18, 157–200. 158

Porter, M. E. (1985). The competitive advantage of nations. New York: The Free Press. Powell, W. W. (1990). Neither market nor hierarchy: Network forms of organization research. Research in Organizational Behavior, 12, 295–336. Priem, R. L. (1994). Executive judgment, organizational congruence, and firm performance. Organization Science, 5, 421–437. Provan, K. G., & Kenis, P. (2008). Modes of network governance: Structure, management, and effectiveness. Journal of Public Administration Research and Theory, 18, 229–252.

Organizations: Theory, Design, Future

Quinn, J. B. (1992). Intelligent enterprise. New York: The Free Press.

Selznik, P. (1996). Institutionalism “old and new.” Administrative Science Quarterly, 41, 270–277.

Quinn, J. B., Anderson, P., & Finkelstein, S. (1996). Leveraging intellect. Academy of Management Executive, 10, 7–27.

Shaw, M. E. (1954). Some effects of problem complexity upon problem solution efficiency in different communication nets. Journal of Experimental Psychology, 48, 211–217.

Copyright American Psychological Association. Not for further distribution.

Raisch, S. & Birkinshaw, J. (2008). Organizational ambidexterity: Antecedents, outcomes, and moderators. Journal of Management, 34, 375–409. Random House Webster’s unabridged dictionary. (2001). New York: Random House.

Siggelkow, N., & Levinthal, D. A. (2003). Temporarily divide to conquer: Centralized, decentralized, and reintegrated approaches to exploration and adaptation. Organization Science, 14, 650–669.

Reagans, R., & McEvily, B. (2003). Network structure and knowledge transfer: The effects of cohesion and range. Administrative Science Quarterly, 48, 240–267.

Singh, J. V., House, R. J., & Tucker, D. J. (1986). Organizational change and organizational mortality. Administrative Science Quarterly, 31, 587–611.

Repenning, N. P. (2002). A simulation-based approach to understanding the dynamics of innovation implementation. Organization Science, 13, 109–127.

Sorensen, J. B. (2002). The strength of corporate culture and the reliability of firm performance. Administrative Science Quarterly, 47, 70–91.

Rosenkopf, L., & Padula, G. (2008). Investigating the microstructure of network evolution: Alliance formation in the mobile communications industry. Organization Science, 19, 669–687.

Starbuck, W. (2003). The origins of organization theory. In H. Tsoukas & C. Knudsen (Eds.), The Oxford handbook of organization theory (pp. 143–182). Oxford: Oxford University Press.

Saffold, G. S., III. (1988). Culture traits, strength, and organizational performance: Moving beyond “strong” culture. Academy of Management Review, 13, 546–558.

Stuart, T. E. (2000). Interorganizational alliances and the performance of firms: A study of growth and innovation rates in a high-technology industry. Strategic Management Journal, 21, 791–811.

Salancik, G. R. (1995). WANTED: A good network theory of organization. Administrative Science Quarterly, 40, 345–349.

Swaminathan, A. (1996). Environmental conditions at founding and organizational mortality: A trial-by-fire model. Academy of Management Journal, 39, 1350–1377.

Sanders, W. G. (2001). Behavioral responses of CEOs to stock ownership and stock option pay. Academy of Management Journal, 39, 891–919.

Taylor, F. W. (1911). The principles of scientific management. New York: Dover.

Schein, E. H. (1983). The role of the founder in creating organizational culture. Organizational Dynamics, 12, 13–28.

Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18, 509–533.

Schilling, M. A., & Steensma, H. K. (2001). The use of modular organizational forms: An industry-level analysis. Academy of Management Journal, 44, 1149–1168.

Thompson, J. D. (1967). Organizations in action. New York: McGraw-Hill.

Schwarz, G. M., & Huber, G. P. (2008). Challenging organizational change research [Special issue]. British Journal of Management, 19, S1–S7. Scott, W. R. (1987). The adolescence of institutional theory. Administrative Science Quarterly, 32, 493–511. Scott, W. R. (2008). Approaching adulthood: The maturing of institutional theory. Theory and Society, 37, 427–442. Scott, W. R., & Davis, G. F. (2007). Organizations and organizing: Rational, natural, and open system perspectives. Upper Saddle River, NJ: Pearson Prentice Hall. Segal, M. (1974). Organization and environment: A typology of adaptability and structure. Public Administration Review, 35, 212–220.

Thompson, K. R., Hochwarter, W. A., & Mathys, N. J. (1997). Stretch targets: What makes them effective? Academy of Management Executive, 11(3), 48–60. Toffler, A. (1970) Future shock. New York: Bantam Books. Tsoukas, H. (2003). New times, fresh challenges. In H. Tsoukas & C. Knudsen (Eds.), The Oxford handbook of organization theory (pp. 607–622). Oxford: Oxford University Press. Tushman, M. L. (1978). Technical communication in R & D laboratories: The impact of project work characteristics. Academy of Management Journal, 21, 624–645. Tushman, M. L., & Murmann, J. P. (1998). Dominant designs, technology cycles, and organizational outcomes. Research in Organizational Behavior, 20, 231–266. 159

George P. Huber

Ulrich, D., & Barney, J. B. (1984). Perspectives in organizations: Resource dependence, efficiency, and population. Academy of Management Review, 9, 471–481.

Williamson, O. E. (1981). The economics of organization: The transaction cost approach. American Journal of Sociology, 87, 548–577.

Uzzi, B. (1996). The sources and consequences of embeddedness for the economic performance of organizations: The network effect. American Sociological Review, 61, 674–698.

Williamson, O. E. (1991). Comparative economic organization: The analysis of discrete structural alternatives. Administrative Science Quarterly, 36, 269–296.

Volberda, H. K. (1998). Building the flexible firm: How to remain competitive. New York: Oxford University Press.

Copyright American Psychological Association. Not for further distribution.

Walsh, J. P., Meyer, A. D., & Schoonhoven, C. B. (2006). A future for organization theory: Living in and living with changing organizations. Organization Science, 17, 657–671. Weber, M. (1947). The theory of social and economic organization. Oxford: Henderson and Parsons.

Williamson, O. E. (1994). Transaction cost economics and organizational theory. In N. J. Smelser & R. Swedberg (Eds.), The handbook of economic sociology (pp. 77–107). Oxford: Princeton University Press. Wiseman, R. M., & Gomez-Mejia, L. R. (1998). A behavioral model of managerial risk taking. Academy of Management Review, 23, 133–153. Woodward, J. (1965). Industrial organization: Theory and practice. New York: Oxford University Press.

Weick, K. E. (1969). The social psychology of organizing. Reading, MA: Addison-Wesley.

Zammuto, R. F. (1988). Organizational adaptation: Some implications of organizational ecology for strategic choice. Journal of Management Studies, 25, 105–120.

Weick, K. E. (1976). Educational organizations as loosely coupled systems. Administrative Science Quarterly, 21, 1–19.

Zeithaml, V. A., Bitner, M. J., & Gremler, D. D. (2005). Service marketing: Integrating customer focus across the firm (4th ed.). New York: McGraw-Hill.

Weick, K. E., & Sutcliffe, K. M. (2001). Managing the unexpected: Assuring high performance in an age of complexity. San Francisco, CA: Jossey-Bass.

Zhiang, L., Zhao, X., Ismail, K. M., & Carley, K. M. (2006). Organizational redesign and restructuring in response to crises: Lessons from computational modeling and real-world cases. Organization Science, 17, 598–618.

Wellman, B. (1988). Structural analysis: From method and metaphor to theory and substance. B. Wellman & S. D. Berkowitz (Eds.), Social structures: A network approach (pp. 19–61). New York: Cambridge University Press.

160

Zollo, M., & Winter, S. G. (2002). Deliberate learning and the evolution of dynamic capabilities. Organization Science, 13, 339–351.

CHAPTER 6

STRATEGIC DECISION MAKING

Copyright American Psychological Association. Not for further distribution.

Philip Bromiley and Devaki Rau

The performance and survival of a firm depend on the strategic choices made by senior managers. Several organizational scholars have attempted to identify and classify the different types of strategic decisions made by managers in organizations (Butler, Davies, Pike, & Sharp, 1993; Cray, Mallory, Butler, Hickson, & Wilson, 1988, 1991; Hickson, Butler, Cray, Mallory, & Wilson, 1986; Mintzberg, 1990; Mintzberg, Raisinghani, & Theoret, 1976; Mintzberg & Waters, 1985; Nutt, 1984; Quinn, 1978, 1980; Schoemaker, 1993; Simon, 1960). Others have examined the processes by which managers make strategic decisions in attempts to explain the choices managers make and to improve managerial decision making. Researchers have examined decision-making processes at different levels (including individual, group, and organization); how attributes of the decision influence the decision-making process (e.g., the risk associated with the decision, whether the decision is familiar or routine, the information and time constraints on the decision); and other factors that might influence the decision-making process, such as the decision maker’s incentives. Scholarly interest in strategic decision making comes from both its importance and the characteristics of the decisions. Strategic decisions often involve high levels of complexity, have limited and questionable data, address dynamic conditions, and involve multiple actors with varied goals. This chapter describes five influential decisionmaking approaches—the behavioral theory of the firm (BTOF), behavioral decision theory (BDT), agency theory, top management teams, and cognition—that

organizational scholars have used to examine strategic decision making in firms. Unlike the BTOF and agency theory, BDT, top management teams, and cognition are not distinct theories but rather describe a variety of phenomena studied with a particular emphasis. Although we divide this chapter on the basis of approaches, these approaches represent different emphases rather than directly competing theories attempting to explain the same phenomena. With the exception of agency theory, all of the approaches take a bounded rationality view of the organization, that is, assume individuals seek to achieve their goals but with extremely stringent constraints on their abilities to obtain and process information. The cognitive literature relies on findings in cognitive psychology and often attempts to map belief structures of managers. The top management team literature adds an emphasis on the interaction of the members of the top management team and cognitive biases that they may have as a result of prior experiences. BDT offers a set of propositions about particular biases or decision patterns that top management teams and organizations may evidence, but these biases generally depend on the framing of choices (i.e., not organizational effects). Finally, the BTOF emphasizes how organizational processes influence decisions. However, although some specific propositions may allow us to differentiate among the theories for particular applications, for strategic decision making, the theories all take a somewhat consistent view where boundedly rational individuals interact in an organizational setting. The exception to this

http://dx.doi.org/10.1037/12169-006 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

161

Copyright American Psychological Association. Not for further distribution.

Bromiley and Rau

commonality, agency theory, assumes rationality in the economic sense. That is, it assumes managers can identify the very best of all possible actions in terms of returns either to the owners or to themselves. This has been relaxed somewhat by work with what has been termed behavioral agency theory, but this theory has a very different theoretical foundation than the original agency theory. These commonalities make dividing papers into one approach or another difficult. Papers often reference more than one of these approaches, for example, looking at top management teams while also considering BDT biases or agency theory incentive issues. Where authors perceive multiple theories making the same prediction, they may reference multiple relevant theories. Furthermore, researchers in strategic decision making often use the same theory at different levels of analysis. For example, work drawing on cognition may map the cognitive structures of individual managers or CEOs and relate those structures to decisions, whereas other work may ascribe a cognitive structure to the organization as a whole and tie those structures to organizational actions and outcomes. Studies often relate CEOs’ or senior managers’ incentives directly to corporate behaviors and outcomes. These studies assume that the choices of senior managers substantially influence corporate behaviors and outcomes. Because most strategic decisions involve risk, the strategic decision making literature has emphasized risk. Whereas the BTOF offers explanations of decisions based on a variety of factors and can apply to risk, interesting decisions in BDT and agency theory almost all involve risk. Consequently, after describing these approaches, this chapter discusses how they differ in their treatment of risk. We use the term risk here but discuss the construct in more detail later. We conclude the chapter with a brief look at some recent attempts to integrate the different approaches as well as attempts to extend them to other strategic decision making contexts. THE BEHAVIORAL THEORY OF THE FIRM March and Simon’s (1958) Organizations and Cyert and March’s (1963) A Behavioral Theory of the Firm developed a specific theory of organizations that has 162

strongly influenced management scholarship. The theory assumes “boundedly rational” decision makers—they attempt to achieve their goals, but serious limits in information and computation constrain their efforts. The BTOF presented itself as a contribution in economics—an attempt to offer more realistic assumptions for an economic theory. At the time, economic theorists assumed firms maximized profit and operated with almost perfect knowledge. Deviations from perfect knowledge came in very manageable forms, for example, in which the decision maker does not know the value of a variable but does know its expected value and distribution around that expected value (and where the distributions fit well-behaved statistical functions). The BTOF argues that, faced with limits on information-processing ability, managers and individuals adopt routines that drastically reduce the amount of information they need to process. Although firms may systematically set some routines, many resemble habits that develop over time without a systematic assessment of their desirability. Whereas economics assumed a firm had a single goal, Cyert and March (1963) described firms as coalitions involving a variety of stakeholders. Instead of a single goal for the organization, or even a welldefined set of goals, organizations have aspiration levels (desired levels of achievement) on various dimensions (e.g., aspirations on sales, inventories, profits). The aspiration levels depend on prior aspiration levels, prior performance, and the performance of comparable organizations. When performance exceeds an aspiration level, the firm operates according to its established routines. If performance falls below an aspiration level, the firm searches for ways to raise performance above the aspiration level. Trying to keep outcomes above aspirations results in organizations “satisficing,” or producing good-enough rather than optimal results (Simon, 1957). In the BTOF, the firm adapts on the basis of differences between aspiration and expected performance rather than substantial analysis and long-term planning. The dimensions on which a firm has aspirations depend partially on the set of stakeholders who control the firm (the dominant coalition). For example, the backgrounds and positions of the members of

Copyright American Psychological Association. Not for further distribution.

Strategic Decision Making

the top management team should influence the dimensions the team attends to and their importance. Firm behavior depends on which aspiration level triggers search. The same underlying problem might evoke search to improve marketing if the problem that appears to management as sales falling below sales aspirations or evoke search for quality improvement if the problem appears to management as high product returns due to defects. The apparent goals of the organization thus depend on the aspiration level violated. The firm uses routines (also called standardized decision rules or heuristics). Routines provide predictability (and so allow coordination) and reduce the need for individuals to process information. Some routines constitute formal standard operating procedures (Cyert & March, 1963). Gigerenzer and colleagues (Gigerenzer, 2000; Gigerenzer & Goldstein, 1996), as well as related empirical work (see, e.g., Newell, Weston, & Shanks, 2003), extended this idea of using heuristics to reduce uncertainty to propose that individual decision makers under conditions of uncertainty not only satisfice by using heuristics but also economize on the cognitive effort required to make a decision by using specific fastand-frugal heuristics. The BTOF offers four major relational concepts. First, firms do not resolve their conflicts directly but rather engage in quasi-resolution of conflict. They treat goals as independent constraints, solve things locally (with respect to the goals that are invoked locally), accept outcomes that are good enough, and pay sequential attention to goals (represented as aspiration levels). Instead of forecasting an uncertain future, firms rely on feedback based on differences between aspiration and performance to adapt and may attempt to influence their environment to reduce uncertainty. When the firm faces a problem, it searches in the area of the problem rather than looking for some general solution. Finally, organizations adapt largely by modifying their routines in simple, incremental ways. This brief outline does not begin to explain the full theory. Among the other parts of the theory that have been influential is the argument that firms generate slack and use this slack to buffer themselves from an uncertain world. Having excess cash frees

the firm from continually setting optimal cash levels. Having excess inventory or labor available lets the firm weather small changes in demand. Concepts from the BTOF have influenced many organizational theories, including institutional theory, population ecology, and organizational learning (see Argote & Greve, 2007, for a discussion of the influence of BTOF on organizational theories). Recent empirical research based on the BTOF has explored a variety of strategic decisions, including organizational change (Wezel & Saka-Helmhout, 2006), strategic positioning (Park, 2007), research and development (R&D) investment patterns and intensity (Chen & Miller, 2007; Greve, 2003), financial misrepresentation (Harris & Bromiley, 2007), and technological and organizational innovations (Massini, Lewin, & Greve, 2005). Many of these studies examined how the firm’s context influences its decisions. Chen and Miller (2007), for instance, examined how variations in firm R&D intensity depend on the specific situation facing the firm, including its performance relative to aspirations, proximity to bankruptcy, and slack. Park (2007) examined how a firm’s performance relative to its aspiration level makes it move its strategic position closer to or further from competing firms. Wezel and Saka-Helmhout (2006) focused more generally on the firm’s external environment and examined how instability in the institutional environment influences changes in an aspiration-driven firm. Empirical studies of the BTOF typically use accounting data to measure key constructs, such as performance and slack (Bromiley, 1991; Chen & Miller, 2007). These studies seldom measure aspirations directly; instead, they model aspirations as a function of the past performance of the focal firm and other referent firms (Bromiley, 1991; Chen & Miller, 2007; Greve, 2003; Wiseman & Bromiley, 1996). The literature on the BTOF includes several different models for the setting of aspiration levels by firms. March and Shapira (1992) proposed that firms shift among different aspiration levels depending on firm performance. Whereas most firms aspire to reasonable performance relative to their peers, survival may become the aspiration level if the firm is close to bankruptcy. Bromiley (1991) proposed that firms performing below their industry’s mean 163

Copyright American Psychological Association. Not for further distribution.

Bromiley and Rau

aspire to that level of performance, whereas firms performing above it aspire to improve their position. Following Bromiley (1991), Park (2007) measured aspiration as a function of social aspiration level and historical aspiration level. These studies typically used some form of regression analysis to estimate the relations between performance versus aspiration and an outcome variable. Parts of the theory have strong empirical support. Many studies find that the difference between organizational aspirations and performance influences firm behavior. Usually, such studies find that firms with low performance relative to their aspirations change more and take riskier actions than do firms with performance above their aspirations. McNamara and Bromiley (1997) found that the difference between aspirations and performance influences the riskiness of banks’ commercial loans. Bromiley (1991) found that the difference between aspirations and performance influences the uncertainty of a firm’s income stream (a proxy for risk), whereas Park (2007) found that firms change strategy more when their performance falls below versus above their aspiration levels. Harris and Bromiley (2007) found that the difference between a firm’s aspiration and performance influences misrepresentation of financial statements. The BTOF takes a reactive view of the firm. The theory allows for some forms of innovation that would come from, for example, routines to develop new processes or products. The theory allows for some routine foresight, for example, the standardized forecasting inherent in planning systems. However, it puts little emphasis on management initiative or creativity. Apart from suggesting that firms with extremely high performance can take risks without danger of falling below their aspiration level, in the theory most change comes from firm attempts to solve specific problems. The BTOF provides some direct insights on strategic decision making by organizations. First, and not discussed a great deal in strategy literature, much of what organizations do comes from their routines, and even for nonroutine decisions, routines structure the information available, the framing of the problem, where firms look for solutions, the feasible solutions, and so forth. Second, individuals and organizations 164

have aspiration levels, that is, outcomes they would like to see on various dimensions. When firms or individuals appear likely not to exceed the aspiration level, they search for ways to improve such that they will exceed the outcome. This results in firms making more changes when they fall below their aspirations than when they are above them (see March, 1994, for an excellent overview of the decision-making implications of the BTOF). BEHAVIORAL DECISION THEORY BDT describes an area of psychological research that examines individual decision making, largely emphasizing the influence of context on decision making in well-defined situations. Of the many streams in BDT, the heuristics noted in Tversky and Kahneman’s (1974) and Kahneman and Tversky’s (1979) prospect theory have influenced research on strategic decision making most. Tversky and Kahneman (1974) identified three heuristics decision makers use under uncertainty: representativeness (i.e., assuming that a small sample accurately represents a population), availability (i.e., relying on readily remembered data, which results in estimates that depend on the ease with which different kinds of events can be remembered), and adjustment and anchoring (i.e., many decisions start with a base and then adjust from there, and the decision maker does not adjust sufficiently). Kahneman and Tversky’s (1979) prospect theory has influenced research across the social sciences, including research on strategic decisions. Prospect theory attempts to explain individual choice under uncertainty. In prospect theory, individuals judge outcomes relative to a reference point. The theory assigns values to the difference between the outcome and the reference point. The function for values differs for outcomes above and below the reference point. The value function is concave for outcomes above the reference point (risk averse in normal terminology) and convex for outcomes below the reference point (risk seeking) and has a substantially steeper slope for negative versus positive outcomes. Choice depends on a sum of these values weighted by a function that depends on the probabilities of the outcomes. The weighting function generally

Copyright American Psychological Association. Not for further distribution.

Strategic Decision Making

underweights outcomes with mid-range probabilities but overweights extremely low probabilities and may assign zero weight to very low probabilities. The overall value of a gamble equals the sum of the values times the weights. The vast majority of discussions based on prospect theory in the strategic decision making literature emphasize the curvature of the value function. As with much of BDT, prospect theory attempts to explain how individuals differ from the predictions of expected utility maximization. Arkes, Hirshleifer, Jiang, and Lim (2008) identified three major differences between prospect and expected utility theories. First, prospect theory proposes that people derive utility from gains and losses relative to a reference point, whereas traditional utility theory assumes that people derive utility from total wealth or consumption. Second, prospect theory’s value function differs in the domain of gains versus the domain of losses, but, because utility functions only consider final outcomes, utility functions do not differ with reference point. Third, near the reference point, in prospect theory, a unit change for outcomes framed as losses influences value much more than a unit change for outcomes framed as gains. A multitude of studies have examined and elaborated prospect theory. In the psychological literature, recent theoretical developments in this area include characterizing the conditions related to strong risk aversion (Schmidt & Zank, 2008) and extending prospect theory to group decision making (Tamura, 2005). Recent experimental studies focus on identifying individual differences related to the reduction of uncertainty in decision making (Lauriola, Levin, & Hart, 2007) and the adaptation of reference points to gains and losses (Arkes et al., 2008). Other researchers have attempted to advance prospect theory methodologically by developing better measures of loss aversion (Abdellaoui, Bleichrodt, & Paraschiv, 2007). Although prospect theory has been extremely influential in BDT, researchers in psychology continue to develop alternative, and potentially superior, models of individual decision making under risk. Because prospect theory came from experimental results, the empirical foundation consisted of situations where the experimental circumstances imposed a particular reference point. Consequently, the theory

does not offer a sophisticated explanation for the determination of the reference point. Tversky and Kahneman (1981), however, commented that a diversity of factors determines the reference outcome in everyday life. The reference outcome is usually a state to which one has adapted; it is sometimes set by social norms and expectations; it sometimes corresponds to a level of aspiration, which may or may not be realistic. (p. 456) Although scholars in psychology continue to advance BDT, most applications in organizations and strategy rely heavily on interpretations of the original statement of prospect theory (Kahneman & Tversky, 1979) or the heuristics noted by Tversky and Kahneman (1974). Almost all strategy interpretations of prospect theory consider only the value function without addressing the other components of the theory. Bromiley (in press) demonstrated that the predictions of prospect theory depend heavily on assumptions concerning the parameters for the probability weighting function, the breadth of the distribution of outcomes, the probabilities of outcomes, and the parameters in the value function (particularly the loss aversion parameter); strategy researchers have largely ignored these complications. Some research on strategic decision making uses heuristics or prospect theory to examine specific strategic decisions. For example, at the individual level, Mullins, Forlani, and Walker (1999) used prospect theory to explain individual managers’ new product investment decisions, and Bamberger and Fiegenbaum (1996) used it to explain the impact of individuals’ reference points on strategic decisions related to human resources. Other studies have used prospect theory to examine escalation of commitment (McNamara, Moon, & Bromiley, 2002; Moon et al., 2003) and decision fiascoes by decision-making groups in organizations (Whyte, 1998). Researchers have used heuristics and prospect theory to explain strategic decisions such as organizational responses to threats and opportunities (Chattopadhyay, Glick, & Huber, 2001) and business unit divestiture (Shimizu, 2007). 165

Copyright American Psychological Association. Not for further distribution.

Bromiley and Rau

In contrast to work that only uses prospect theory explanations, Bromiley (1987) and McNamara and Bromiley (1997) included both organizational and psychological hypotheses in empirical work using archival data. Bromiley (1987) found that contrary to the conservatism or anchoring-and-adjustment arguments (Tversky & Kahneman 1974), organizational forecasts did not lie between the anchor and the actual but rather reflected clear organizationally justified biases. For example, a corporation consistently overestimated capital expenditures (because underestimation could result in serious problems from cash shortages). Similarly, a Navy unit overestimated service orders it would receive and spending because underestimation could result in the unit running out of funds but overestimation would simply result in funds left at the end of the year. Likewise, examining risk assessment on commercial loans, McNamara and Bromiley (1997) found that organizational effects dominate psychological ones. Their study of risk assessments of borrowers by commercial lenders found that organizational pressures for profitability appeared to influence borrowers’ risk ratings, with newer borrowers and larger loans receiving more favorable ratings than borrowers with longer relations with the bank and smaller loans. Recent strategy research has begun to examine how prospect theory, in combination with other theories such as the BTOF, can explain specific organizational decisions, such as the decision to divest units (Shimizu, 2007) or the decision to remain independent versus belonging to a cooperative (Gomez-Mejia, Haynes, Nunez-Nickel, Jacobson, & Moyano-Fuentes, 2007). Paralleling the research on the BTOF, some recent studies in this area have attempted to place the decisions predicted by prospect theory in the broader context of the external environment facing the organization (Lehner, 2000; Slattery & Ganster, 2002; Voss, Sirdeshmukh, & Voss, 2008). Voss et al. (2008), for instance, used prospect theory to explain where organizations exploit their current markets and abilities (exploitation) or develop new markets and abilities (exploration). Considering firms in threatening environments, they found that organizational responses (e.g., increased risk taking through increased exploration) depended on the extent to which the environment threatens current and long166

term performance, as well as the firm’s level of slack (resources not needed for immediate operation). Note the confluence of prospect theory arguments with concepts from the BTOF (exploration–exploitation and slack). Other strategy research attempts to explain aggregate firm characteristics on the basis of prospect theory. Much of this research examines the associations of risk and return across firms. Instead of a full derivation of hypotheses from prospect theory, studies in this area interpret prospect theory as predicting risk aversion for high-performance organizations and risk seeking for low-performance organizations (Fiegenbaum, 1990; Fiegenbaum & Thomas, 1986, 1988, 1990; Jegers, 1991; Wiseman & Bromiley, 1991). Essentially, these scholars have argued that high firm performance indicates the firm exceeds its target level of performance (usually assumed to be industry average performance). Such firms will value potential outcomes on the basis of the positive portion of the value function, which is risk averse. Firms below their target level of performance will judge choices on the basis of the negative portion of the value function, which is risk seeking. Fiegenbaum and Thomas (1988), for instance, used prospect theory to explain Bowman’s (1980) risk–return paradox, a finding that accounting measures of firm risk and return correlate negatively across firms. Consistent with the prospect theory argument, Fiegenbaum and Thomas (1988) found negative risk–return associations for firms below target levels and positive associations for firms above target levels. Fiegenbaum and Thomas (1995) extended prospect theory’s idea of a reference point to explain industry structure. Their results suggest that strategic groups act as reference points for firm strategies. Lehner (2000) extended this work using different risk measures. Strategy research that uses prospect theory to explain organizational decisions or firm characteristics suffers from several difficulties, both empirical and theoretical. Slattery and Ganster (2002) argued for research that examines risk taking in environments that closely resemble actual decision-making environments encountered by managers, with uncertain outcomes and meaningful consequences. Some research that directly compares the predictions from

Copyright American Psychological Association. Not for further distribution.

Strategic Decision Making

BDT with the predictions from other organizational theories finds that organizational effects appear much stronger than BDT effects. For many years, some strategy scholars have questioned the extent to which experiments generally done on undergraduate students in sterile conditions generalize to top management in complex organizational environments. Although experiments on executives constitute an improvement, they still omit the environmental complexity. From a theoretical perspective, Bromiley (in press) demonstrated that the predictions of prospect theory depend strongly on a variety of assumptions that the management literature has ignored. Furthermore, the literature seldom distinguishes between the predictions of prospect theory and the predictions of the BTOF. Many interpretations of the BTOF propose that it predicts increased risk taking for firms below their aspiration point and reduces risk taking for firms above their aspiration point. This coincides with the interpretations others have drawn from prospect theory. Thus, empirical results finding these patterns do not differentiate between the theories. The two theories, however, do differ in their predictions, particularly in relation to choices near the reference point. Whereas prospect theory was developed in situations involving all positive or all negative outcomes, most strategic choices have both positive and negative outcomes (i.e., mixed gambles). Considering only the value function, prospect theory implies that decision makers should exhibit extreme risk aversion when faced with mixed gambles (Bromiley, in press). Specifically, risk aversion in the prospect theory value function can be measured by a comparison of the value of risky outcomes with the value of a sure outcome equal to the mean of the risky outcomes. Thus, it depends on the curvature of the value function. The join between positive and negative value functions at zero creates a drastic change in curvature, which results in substantial risk aversion for mixed gambles. The strategy literature has ignored this implication of prospect theory. Firms with performance near the reference point should face even more mixed gambles than firms further from the reference point. In contrast, the BTOF predicts relatively little risk taking for firms near the reference point. Firms just above the refer-

ence point should make few changes because their current behavior is sufficient to exceed the reference point, and firms just below the reference point should make some changes, though modest, compared with firms further away from the reference point. In short, research on BDT finds corporate behavior patterns that researchers consider consistent with various heuristics–biases and prospect theory. However, whether the patterns actually reflect organizational rather than psychological effects remains an open issue. From a practical standpoint, several lessons appear. First, without formal analysis for both the probabilities themselves and the choices dependent on probabilities, management judgments under uncertainty will contain substantial errors. Second, reference points will substantially influence management behavior, with more risk taking when management perceives outcomes as below their reference point and less risk taking when above their reference point. AGENCY THEORY Agency theory addresses how a principal (usually seen as the firm’s shareholders, board of directors, or top management) can get an agent (seen as the managers or employees of the firm) to act in the principal’s interest (Demsetz, 1983; Fama, 1980; Jensen & Meckling, 1976). It generally assumes an honest, risk-neutral principal and an amoral, riskaverse agent. The principal has perfect rationality, except it cannot tell if the agent is working fully in the principal’s interest. Monitoring the agent is costly. Given a chance, the agent will lie to the principal and act in his or her own interest. Agents do so by taking less risk than the principal would want and obtaining direct benefits, such as excessive pay or perquisites. For example, Jensen and Meckling (1976) applied agency theory to problems arising from a separation of ownership and control (Berle & Means, 1932). By assuming shareholders can diversify their risk through multiple investments, Jensen and Meckling proposed that shareholders want managers to maximize the value of the firm. In contrast, much of management’s wealth comes from the specific firm, so risk-averse managers will take insufficient risk to maximize 167

Copyright American Psychological Association. Not for further distribution.

Bromiley and Rau

shareholder wealth. Managers can increase their utility by avoiding risk, working less (or less hard), increasing their perquisites, and so forth. Shareholders must design incentive and control systems to encourage managers to maximize shareholder value. Given this problem, the theory predicts principals will use monitoring and incentives to control agents. Agency costs refer to the costs of these measures, as well as any residual losses that occur from contracting (Lubatkin, Lane, & Schulze, 2001). According to agency theory, contracting parties will incur agency costs up to the point where the cost of fully eliminating agency costs exceeds the benefits (Jensen & Smith, 1985). Researchers have applied agency theory to issues ranging from executive compensation to ownership structure to corporate diversification and acquisitions. Researchers disagree on the appropriate domain of agency theory (Amihud & Lev, 1999; Denis, Denis, & Sarin, 1999; Lane, Cannella, & Lubatkin, 1998, 1999; Lubatkin et al., 2001). Recent research has used agency theory to examine managers’ commitment to innovation (Gamble, 2000), the extent to which an employee will make firm-specific investments (H. C. Wang & Barney, 2006), and the effect of monitoring on chief executive compensation (Kerr & Kren, 1992). Other research has used agency theory to explain organizational choices, including franchising decisions (Michael, 1996), planning and control decisions (Natarajan, Sethuraman, & Suryasekar, 2005), outsourcing (Tiwana & Bush, 2007), assignment of decision responsibilities, hierarchical levels and information systems (Mosakowski, 1998), and the alliances a firm forms (Reuer & Ragozzino, 2006). Several caveats merit mention. First, none of the studies actually test whether the agent behaves optimally. A very long history of psychological research on incentives offers predictions that strongly resemble agency theory—basically people respond sensibly to incentives. Second, the empirical literature on monitoring corporate management (termed “corporate governance”) has had inconsistent empirical results. The theory does not predict which of various mechanisms (including number of outside directors, having a majority of outside directors, having separate CEOs and board of director chairs, stock ownership 168

by corporate directors, stock ownership by mutual funds, etc.) should work. Research usually includes multiple measures of governance, but the results differ across studies. Although it seems obvious that some features of governance should influence corporate behavior, apart from incentives, research has not demonstrated that particular governance features influence behavior across the range of issues relevant to management. Third, the predictions come directly from arbitrary assumptions (e.g., the combination of risk-neutral principals and risk-averse agents implies agents take fewer risks than principals want) that may or may not apply across different decisionmaking situations. TOP MANAGEMENT TEAMS Beginning with Hambrick and Mason (1984), many studies have examined the relations between top management team characteristics (which are assumed to reflect top managers’ cognitive bases) and organizational outcomes, such as strategic change and innovation (e.g., Bantel & Jackson, 1989; Cho & Hambrick, 2006; Finkelstein & Hambrick, 1990; Grimm & Smith, 1991; Murray, 1989; Smith et al., 1994; Wally & Becerra, 2001; Wiersema & Bantel, 1992). The earliest studies used readily available demographic measures (e.g., age, education, functional background) of the members of the top management team. The results, however, were not consistent. West and Schwenk (1996), for instance, failed to replicate significant results from previous studies relating demographic variables with top managers’ perceptions and firm performance. Using experienced managers in a simulated decision situation, Kilduff, Angelmar, and Mehra (2000) found that demographic diversity had little association with cognitive diversity. Across highly selected groups like top manager management teams, knowing team average age, education, and so forth may not tell us that much about what decisions the group will make. Recent research has shifted from team demographics to examining the influence of top management team processes on strategic decisions and organizational performance (Amason & Schweiger, 1994; Michel & Hambrick, 1992; C. C. Miller, Burke,

Copyright American Psychological Association. Not for further distribution.

Strategic Decision Making

& Glick, 1998; Simons, Pelled, & Smith, 1999). One stream of research focuses on team information processing as a determinant of organizational performance and strategic choice (Anand, Manz, & Glick, 1998; Rau, 2005; Rulke, Zaheer, & Anderson, 2000). Some studies on this topic examine how transactive memory, defined as the set of knowledge possessed by members of the team, combined with an awareness of who knows what within the team (Faraj & Sproull, 2000; Wegner, 1987), influences organizational performance. For example, Rau (2005) found that the level of transactive memory in the top management teams of small banks positively influenced bank performance. Bolstering these findings, earlier work by Eisenhardt (1989) found that firms with extremely dominant CEOs performed substantially worse than management teams in which the functional experts have the greatest influence on decisions in their functional areas. Other studies have examined the performance effects of different forms of mental models, including knowledge structures (Walsh, 1995), strategic schemas (Fiske & Taylor, 1991; Nadkarni & Narayanan, 2007), shared cognition (Klimoski & Mohammed, 1994), dominant logic (Prahalad & Bettis, 1986), or shared consensus (D. Knight et al., 1999). Still other research has proposed a meta construct called behavioral integration, which tries to explain organizational outcomes better than team information processing or team mental models alone (Hambrick, 1994; Mooney, 2000; Simsek, Veiga, Lubatkin, & Dino, 2005). Behavioral integration captures three key elements of top management team decision processes: level of collaborative behavior, the quantity and quality of information exchanged, and emphasis on joint decision making. A related stream of work looks at how consensus and conflict among members of top management teams influences the quality of their strategic decisions and subsequent firm performance (Bourgeois, 1985; Dess & Priem, 1995). Frederickson and Mitchell (1984) and Frederickson (1984) posed the question not as a top management team issue per se, but rather as one related to the effects of comprehensiveness (i.e., exhaustiveness or inclusiveness) in organizational decision making. Frederickson and Mitchell (1984) and Frederickson (1984) demon-

strated that the association of comprehensiveness and performance differed across industries. Research on top management teams is moving toward a coherent set of findings and theories but has not achieved it yet. We know that gross demographic measures appear insufficient to explain differences in team behavior. We know that interactions among team members appear to matter, but which of several measures of interaction matter most is unclear. Although research has begun to recognize the possibility that good top management team processes depend on the firm’s environment (and potentially other factors), we do not yet have contingency theory and empirical results on the subject. That said, some strong results have appeared. These results generally agree with the extant research on team decision making. Though using various terms, the literature appears to show that how top management teams interact significantly influences corporate behavior and performance. In particular, teams that take advantage of the abilities of their members by deferring to members in their areas of expertise and openly sharing information appear to have higher performance than other firms. COGNITION The cognition paradigm on strategic decision making has taken two very different directions. One stream of work applies a cognitive view to top management teams’ or CEOs’ decisions. A second stream of work attempts to use measures of firm cognitive structures or knowledge structures to explain firm learning and other behaviors (e.g., Lyles & Schwenk, 1992; Prahalad & Bettis, 1986). Work on cognition overlaps with work on heuristics and biases. Scholars have identified many different cognitive biases that may affect decision makers (Bazerman, 1994; Das & Teng, 1999; Hogarth, 1980; Lyles & Thomas, 1988; Schwenk, 1984, 1985, 1995; Tversky & Kahneman, 1974). Das and Teng (1999) identified four basic types of cognitive biases (prior hypotheses and focusing on limited targets, exposure to limited alternatives, insensitivity to problem outcomes, and the illusion of manageability) and considered the impact of these biases under different modes of decision making (rational, avoidance, logical 169

Copyright American Psychological Association. Not for further distribution.

Bromiley and Rau

incrementalist, political, and garbage can). Schwenk (1995) identified three major themes in the empirical and theoretical literature on cognitive biases and strategic decision making—biases in causal attributions, biases in recollection, and escalating commitment. Some recent studies have focused on escalation of commitment, its relations with risk, and methods of attenuating it (He & Mittal, 2007; Keil, Depledge, & Rai, 2007; McNamara, Moon, & Bromiley, 2002; Sharp & Salter, 1997). Keil et al. (2007), for instance, examined whether problem recognition and the cognitive biases of selective perception and illusion of control promote escalation behavior. He and Mittal (2007) found that decision risk has the strongest effect on escalation of commitment at the intermediate stage of project completion. McNamara, Moon, and Bromiley (2002) found that increased monitoring attenuates escalation but has undesirable unintended effects. Other research (e.g., Hodgkinson, Brown, Maule, Glaister, & Pearman, 1999) has examined cognitive mapping as a means of overcoming cognitive biases arising from the framing of strategic decision problems. Cognitive mapping involves developing a diagram that represents an individual’s beliefs about the causal situation—making that individual’s beliefs explicit. Cognitive mapping has been extended as a prescriptive tool by Eden and Ackerman (1998). Some research has demonstrated that such beliefs influence firm behavior. Barr, Stimpert, and Huff (1992) developed maps of the causal arguments inherent in two railroads’ letters to shareholders over many years. This let them understand how the railroads changed (or did not change) their beliefs, finding that a firm that blamed its problems on external factors did not adapt as well as a firm that looked to internal explanations for problems. Other work has looked at bank managers’ beliefs about competitors and how such beliefs influence bank behavior (Reger & Huff, 1993; McNamara, Luce, & Thompson, 2002). Research on the framing of strategic decision problems and the effects of this framing on organizational outcomes constitutes a major stream of inquiry in the cognition paradigm. Staw, Sandelands, and Dutton (1981) proposed a threat-rigidity hypothesis, wherein performance below an aspiration level decreases risk taking if decision makers perceive the 170

gap between performance and aspirations as a threat to survival. This hypothesis contrasts with the predictions of both the BTOF and prospect theory, which suggest that below-target performance leads to increased risk taking (Mone, McKinley, & Barker, 1998; Ocasio, 1995). However, March and Shapira (1987, 1992) offered a revision of the BTOF that may explain these results. Firms with extremely low performance may switch their aspiration level from social comparison to survival. Then, if the firms exceed the default level slightly, they may make few changes to avoid the risk of bankruptcy. Empirical results related to the threat-rigidity hypothesis are mixed. Chattopadhyay et al. (2001), for instance, found results that support both the threat-rigidity hypothesis and prospect theory. In their study, consistent with the threat-rigidity hypothesis, control-reducing threats led to more conservative, internally directed actions, whereas, consistent with prospect theory, likely losses led to riskier, externally directed actions. Similarly, Audia and Greve (2006) examined the contrasting predictions made by prospect theory and the threat-rigidity hypothesis in the context of decisions related to factory expansion in shipbuilding firms. They found that performance below aspiration level reduced risk taking in small firms but did not affect risk taking or increased risk taking in large firms. Studying strategic actions taken by hospitals, Ketchen and Palmer (1999) found support for the BTOF but no support for the threat-rigidity hypothesis. Chattopadhyay et al. (2001) proposed that these conflicting results stem from differences in the dimensions of threat identified by prospect theory and the threat-rigidity hypothesis. The study argued that prospect theory focuses on events directly related to the loss or gain of tangible resources, whereas the threat-rigidity hypothesis focuses on the reduction of control. Highhouse and Yuce (1996) found empirical support for this argument. In an experimental study, decision makers distinguished threat and opportunity dimensions from gain and loss dimensions. Thus, for instance, decision makers perceived a risky alternative as an opportunity when they were in the loss domain but as a threat when they were in the gain domain. Overall, as with the BDT work, cognition research on strategic decision making has alerted researchers

Copyright American Psychological Association. Not for further distribution.

Strategic Decision Making

and managers to the different biases that decision makers may be subject to during the decision-making process. In addition, cognitive work has demonstrated that managers’ beliefs about causal structures influence their actions and firm outcomes. Prescriptive work aims to clarify managers’ causal assumptions so that they can assess those assumptions directly and can more easily trace the logic from their assumptions to conclusions. Strategic management researchers have yet to explore how the many other aspects of cognition studied by cognitive psychologists may influence strategic decision making, though some recent research has made a start in this area by examining whether emotions, such as fear, and motivational orientations, such as curiosity, are linked to judgments of negative versus positive decision outcomes (Maner & Gerend, 2007). RISK IN STRATEGIC DECISION MAKING RESEARCH Parallel to the interest in psychology, much of the decision-making work in strategy emphasizes choices under uncertainty. Few, if any, important choices come to management with absolute certainty about outcomes contingent on management decisions. A central element in the five approaches discussed previously, therefore, is risk. Some of the five approaches incorporate the notion of risk explicitly (the BTOF, BDT, and agency theory), and others incorporate risk implicitly (top management teams and cognition). The latter two approaches (with the exception of the threat-rigidity hypothesis in the cognition paradigm) examine decisions that are strategic and therefore, by implication, risky but may not emphasize risk throughout. Given the centrality of risk to the different decision-making approaches, in this section, we focus on how the BTOF, BDT, and agency theory explicitly treat and measure risk. We begin by providing an overview of the different ways in which risk is defined and measured in different areas of scholarly inquiry.

Meanings and Measures of Risk Different scholarly areas use risk to mean different things. F. H. Knight (1921) defined risk as the state where decision makers have accurate knowledge of

the probability distribution of outcomes contingent on actions. He defined uncertainty as cases where the decision maker does not know the probability distributions of outcomes. Finance researchers often measure risk by the association between a firm’s stock price and some general index for the stock market but also discuss the probability of bankruptcy. Management researchers talk of risky substantive decisions (e.g., higher R&D spending) and the uncertainty of a firm’s income stream. Research suggests managers understand risk as potential losses or the potential of not reaching targets rather than as F. H. Knight’s probability distribution (Baird & Thomas, 1990; March & Shapira, 1987). Although researchers use the term risk, the decisions they generally study involve uncertainty or ambiguity rather than risk as defined by F. H. Knight (1921). Managers seldom face choices where they can accurately identify all the potential outcomes, let alone put probability distributions on each outcome. Researchers often use the term risk in conjunction with the term uncertainty, which some researchers on strategic decision making define as the perceived unpredictability of environmental and organizational contingencies (Duncan, 1972; Miles & Snow, 1978; Pfeffer & Salancik, 1978). Milliken (1987) identified three types of perceived uncertainty—state, effect, and response uncertainty— which refer to an inability to predict the future state of the environment, its impact on the organization, and potential organizational actions and their outcomes, respectively. March and Shapira (1987, 1992) interviewed numerous managers about their understandings of risk. Consistent with how we use the term in everyday life, and inconsistent with most academic work, they found that managers think of risk in terms of the possibility of outcomes below their target levels and the amount or degree to which low outcomes can hurt them. These managers would not apply the term risk to activities involving very little funds or consequence. Managers analyze outcomes relative to some target, with potential outcomes below the target contributing to risk and outcomes above the target being irrelevant to risk. This matches common usage. People with moderate incomes would not see buying a lottery ticket 171

Copyright American Psychological Association. Not for further distribution.

Bromiley and Rau

as risky, particularly because the amount of money is small relative to the individual’s worth. They would see wagering $100,000 on the flip of a coin as risky, despite the probability of success being much higher than the lottery. The amount to be won in either case has nothing to do with the individual’s perception or description of risk—someone buying a lottery ticket does not perceive winning the lottery on that ticket as a risk. Given the massive literature on risk, we attempt to give some view of the territory rather than an indepth discussion. Our focus is on the treatment of risk by strategic management researchers.

Risk in Strategic Management Research Much of the interest in risk in strategic management started with some relatively simple articles by Bowman (1980, 1982, 1984) that demonstrated that (a) across firms in an industry, a negative association exists between variance in performance (risk) and performance (return), both measured as return on equity for firms; and (b) firms with low performance appear to undertake more risky actions as indicated by mentioning new initiatives in their letters to the shareholders, engaging in litigation, and R&D spending. Work soon after Bowman (1980) continued to emphasize the association of variation in risks and returns. Fiegenbaum (1990) and Fiegenbaum and Thomas (1986, 1988, 1990) found that the association varies by industry and time but often remains negative. Work by Singh (1986), Bromiley (1991), and others used different measures of risk but still found that corporate risk increases for firms with low performance relative to their reference group. Low-performance-drives-risk findings appear in research using a diverse set of measures for risk, including variance in actual outcomes, variance in analysts forecast of earnings, R&D spending, commercial lending, and management responses to questionnaires (Singh, 1986; Larraza-Kintana, Wiseman, Gomez-Mejia, & Welbourne, 2007).

Risk in the BTOF and Prospect Theory Reflecting the mixing that is common in this area, articles addressing performance versus risk offer very similar predictions claiming they come from 172

the BTOF or prospect theory. The difficulty comes because, as interpreted, both theories predict increased risk taking for low-performance firms and reduced risk taking for high-performance firms. Thus, Fiegenbaum (1990) and Fiegenbaum and Thomas (1986, 1988, 1990) emphasized prospect theory, whereas Bromiley (1991), Singh (1986), Wiseman and Bromiley (1996), Wiseman and Catanach (1997), and others emphasized the BTOF. Some research considers risk behavior for firms with extremely low performance. K. D. Miller and Chen (2004) found that, contrary to predictions, organizations showed increased risk (measured as standard deviation of return on assets) as they approached bankruptcy. Gomez-Mejia et al. (2007), in contrast, found that family owned firms take fewer risks (measured as the coefficient of variation) as the threat to survival increases. Wiseman and Bromiley’s (1996) results suggest a cycle in which organizational decline and a reduction in slack increases risk. Increased risk, in turn, reduces performance and results in further declines in slack. Audia and Greve (2006) empirically examined the contrasting predictions of the threatrigidity hypothesis on one hand and prospect theory and the BTOF on the other, with respect to risk taking by firms with low performance. They argued that the fundamental difference between the two sets of theories lies in whether managers view low performance as repairable. In a study of shipbuilding firms, they found that performance below aspirations reduced risk taking in small firms but had no effect or a positive effect on risk taking in large firms. Although the coarseness of the hypotheses generally does not allow empirical differentiation between the BTOF and prospect theory, the empirical results strongly support arguments that organizations take more risks when their performance falls below their aspiration levels. The literature is less clear on the impact of high performance on risk taking.

Treatment of Risk at the Level of the Individual Another stream of research has looked at the risk propensities of individual managers (MacCrimmon

Copyright American Psychological Association. Not for further distribution.

Strategic Decision Making

& Wehrung, 1985, 1990; Nicholson, Soane, FentonO’Creevy, & Willman, 2005; Pennings & Smidts, 2000). These studies typically develop scales for measuring risk attitudes or propensity and examine correlations between risk propensity with other personality and demographic measures (MacCrimmon & Wehrung, 1990; Nicholson et al., 2005) and outcome measures, such as innovativeness and performance (MacCrimmon & Wehrung, 1990; Pennings & Smidts, 2000). Although studies often assume individuals or organizations have a general risk preference that applies to many different domains, the evidence indicates this is incorrect. A substantial literature demonstrates individuals differ and demonstrate different risk propensities in different situations (Bromiley & Curley, 1992). In perhaps the largest and most complete study of managerial risk preferences, MacCrimmon and Wehrung (1986, 1990) elicited risk propensities from over 500 executives using a variety of techniques. They offered the managers in their study three variations of simple risk experiments of the BDT sort (a choice among gambles in standard unadorned fashion, ranking the desirability of gambles, and gambles embedded in a short case). They also asked about actual behaviors—proportion of the managers’ investments in risky activities, how much debt the manager has taken on, whether the manager gambles recreationally, and so forth. Finally, they used standard self-reporting of attitudes and sensation seeking. The results are shocking in their lack of association. MacCrimmon and Wehrung (1986, 1990) found that risk propensity measures correlated somewhat with similar risk propensity measures (e.g., answers to the two in-basket exercises correlated positively) but almost not at all with the others. For example, none of the 28 correlations between the in-basket exercises and other measures of risk propensity reached statistical significance. None of the 55 correlations between the five self-report measures of risk taking and the other 11 measures was statistically significant. The natural situation measures often correlated negatively—people with risky assets tended to have less debt, people with debt tended to have insurance, and so forth. MacCrimmon and Wehrung

(1986, 1990) strongly counted against the possibility of a general risk propensity. Sitkin and Pablo (1992) proposed that risk propensity and risk perception mediate the influence of a number of variables on risk taking. Sitkin and Weingart (1995) partially tested Sitkin and Pablo’s (1992) model and found that both risk perception and risk propensity mediated the relation between outcome history and decision-making behavior and that risk perception mediated the relation between problem framing and decision making. Sitkin and Weingart (1995) argued that individuals have risk propensity dispositions, but those dispositions can change. Researchers continue to look for general risk propensities. Early work on entrepreneurs attempted to demonstrate that they differ in their risk preferences from managers but could not do so (Brockhaus, 1980, 1982; Kogan & Wallach, 1964). More recently, Stewart and Roth (2001) found that the risk propensity of entrepreneurs significantly differed from that of managers. Nicholson et al. (2005) suggested that risk takers are of three nonexclusive types, namely, stimulation seekers, goal achievers, and risk adapters, with only the first group being truly risk seeking. The sensitivity of risk propensity measures to context raises questions about experimental studies such as Mullins et al. (1999), which attempted to use prospect theory to identify organizational and individual factors that influence managers’ decisions regarding new product investments with equal expected values but differing degrees of risk. They measured risk propensities and manipulated risk levels in an experimental setting. Bromiley, Miller, and Rau (2001) raised concerns about using prospect theory to examine organizational decisions because prospect theory was developed to explain individual behavior; hence, its assumptions may not make sense for firms. Shimizu (2007) attempted to address this concern by complementing prospect theory with organizational level behavioral theory and the threat-rigidity hypothesis. Shimizu examined how individual level tendencies predicted by prospect theory interact with organizational level factors to influence organizational level decisions related to retaining or divesting formerly acquired units. 173

Bromiley and Rau

Copyright American Psychological Association. Not for further distribution.

Risk in Agency Theory Although researchers often do not clearly differentiate the predictions of the BTOF and prospect theory, incentive arguments associated with agency theory clearly differ. Studies have associated management incentives with risks reflected as credit risk (Wiseman & Catanach, 1997); types of investment spending, including R&D, capital, and acquisition investment (Beckman & Haunschild, 2002; Sanders & Hambrick, 2007); international risk measured as the total number of risk factors listed in a prospectus (Carpenter, Pollock, & Leary, 2003); riskiness in supplier relations (Camuffo, Furlan, & Rettore, 2007); management reports of risk taking, compensation risk, and unemployment risk; and overall variation in accounting returns and returns to shareholders (Wright, Kroll, Krug, & Pettus, 2007). Many studies develop independent hypotheses under each theory to explain the same phenomena, without attempting to integrate the theories. Markovitch, Steckel, and Yeung (2005), for instance, used concepts from agency theory and prospect theory to explain the role of stock price variation in managerial decision making. Wiseman and Catanach (1997) examined the relative contributions of prospect and agency theory in explaining operational risks and firm performance in regulated and unregulated environments, whereas D’Aveni (1989) used both theories to examine organizational bankruptcy. Harris and Bromiley (2007) tested hypotheses from the BTOF and agency theory in explaining firm financial misrepresentation. In contrast to the BTOF and agency theory, which historically ignored firm impropriety, Harris and Bromiley found that executives are more likely to misrepresent the state of their firm’s finances when firm performance falls below aspiration levels and when CEOs have high proportions of their compensation in stock options. Overall, the research demonstrates several things. Although not demonstrated on top management, massive literatures in psychology suggest that without formal statistical analysis, individuals do not handle uncertainty or derive expectations in ways consistent with rational models. That is, their expectations will generally be biased, and their risk-related choices will not be optimal. Strategy research has demonstrated that performance below aspiration 174

levels generally drives organizations to take greater risks. In addition, top management incentive structures influence risk taking. Finally, risk-related behavior changes when the firm switches its attention from conventional aspiration levels to survival. INTEGRATIONS AND EXTENSIONS OF THE STRATEGIC DECISION MAKING APPROACHES Some research has begun to integrate the different approaches discussed in this chapter. Behavioral agency theory represents perhaps the most advanced effort at integration. Behavioral agency theory integrates agency theory and prospect theory to explain executive risk-taking behavior (Wiseman & Gomez-Mejia, 1998). Although continuing to emphasize the importance of incentives and monitoring, behavioral agency theory considers the influence of the framing of strategic problems on the agent’s choices under risk. Research based on behavioral agency theory includes work by Larazza-Kintana et al. (2007), Sanders and Carpenter (2003), and C. X. Wang and Webster (2007). Carpenter et al. (2003) also attempted to integrate agency and behavioral perspectives to develop a theory of reasoned risk taking, where the risks executives take depend on the interaction of governance mechanisms and stakeholder characteristics. Some research has begun to examine the applicability of these strategic decision making approaches, particularly the BTOF, prospect theory, and agency theory, to international contexts. Jegers (1991), for instance, replicated and extended Fiegenbaum and Thomas’s (1988) application of prospect theory to a Belgian context. Gomez-Mejia et al. (2007) examined predictions from the BTOF and prospect theory in Spanish olive oil firms, Greve (2003) examined the BTOF in the Japanese shipbuilding industry, and Camuffo et al. (2007) examined agency theory in the Italian air-conditioning industry. Although these studies are set in different countries, Lubatkin, Lane, Collin, and Very (2007) examined the role of national context more actively. They argued that the fundamental assumptions of agency theory, namely, opportunistic and boundedly rational actors, may not consistently apply in all countries because differences

Copyright American Psychological Association. Not for further distribution.

Strategic Decision Making

in national institutions and cultures may result in differences in the boundaries and constraints on individual and firm behavior. Lubatkin, Lane, Collin, and Very (2005) tested these ideas in three countries (United States, Sweden, and France). They reviewed the institutions of each country to explore how their governance mechanisms had evolved. Sharp and Salter (1997) explored whether agency theory and prospect theory could explain escalation of commitment to failing projects not just in North America but also in Asia. They found that agency theory explains project escalation decisions in North America but not in Asia. Framing effects predicted by prospect theory explained project escalation decisions in both North America and Asia. Das and Teng (2001) extended the idea of individual risk perceptions to an organizational level. They proposed a framework to explain how partner firms in a strategic alliance can effectively manage the risk inherent in that alliance through trustbuilding techniques and control mechanisms. They proposed that trust (which has two dimensions, goodwill trust and competence trust) and control (which has three modes, behavioral, output, and social control) are antecedents to risk. Risk itself has two dimensions, relational risk and performance risk. Several risk-reduction approaches (consisting of combinations of trust and control) can reduce these two types of risk. Research indicates some empirical support for this model (Sengun & Wasti, 2007). Kaufmann and O’Neill (2007) applied Das and Teng’s (2001) model to an international context and found that cultural distance influenced the type of international strategic alliance that was formed. Specifically, when cultural distance was high (increasing the probability of friction and reduced trust), managers reduced risk of failure by choosing joint ventures that they could more easily monitor and manage. Other scholars have begun to use decisionmaking approaches such as prospect theory to examine group decision making in organizations. Whyte (1998), for instance, proposed that prospect theory, combined with group polarization, could explain the excessive risk seeking observed in group decisions resulting in fiascoes. This stream of research,

however, is still undeveloped, particularly with respect to strategic decision processes. This is in sharp contrast to the rich stream of research in social psychology that examines group decision making, particularly group polarization effects (Hartwick, Sheppard, & Davis, 1982; Janis, 1982; Laughlin & Ellis, 1986; Libby, Trotman, & Zimmer, 1987; Stasser, 1992). APPLIED CONCLUSIONS The work on strategic decision making is extremely diverse and often complex. However, it has generated a number of reasonably reliable findings suitable for application. Here are a few. 1. Organizational routines structure the information and the potential actions of management. Where information from current routines focuses attention may not align with where the firm should pay attention. Actions that can fit within current routines will be easier to sell and to implement than actions that do not. 2. Firms tend to take more risks when performance is low, and such risk taking is often unwise. 3. Risk taking is extremely sensitive to domain and framing. 4. Individuals do not deal well with risk (relative to the prescriptions of statistics). Individuals evidence numerous biases both in their perceptions of risk and in their actions given risk. Where possible, real statistical analysis may improve decision making over expert judgment. 5. Organizations that take responsibility for their performance will learn more quickly than organizations that blame their problems on the environment. This learning results in higher performance and improved survival. 6. Management beliefs about the causal relations in the environment directly influence their behavior. This suggests efforts to clarify such beliefs and to check such beliefs objectively may benefit firms. 7. Management responds to incentives, sometimes too well. Excessively high incentives may cause undesirable outcomes, such as financial misrepresentation. 175

Bromiley and Rau

Copyright American Psychological Association. Not for further distribution.

GENERAL CONCLUSIONS Research has made significant progress in identifying the different types of strategic decision making processes and developing approaches that explain the nature of the strategic decision making process and its resultant outcomes. This chapter described some of the most influential strategic decision making approaches and discussed some methodological issues associated with each of these approaches. Researchers currently working in the area of strategic decision making frequently draw upon these approaches to present compelling explanations for organizational outcomes. Much work, however, remains to be done in this area. As we discussed in this chapter, many of the approaches depend on arbitrary (and often implicit) assumptions. One way to advance the field might be to examine these assumptions more closely, particularly by drawing on relevant research in psychology and other fields. For instance, research in the five strategic decision making approaches we discussed in this chapter typically assumes no gender effects— men and women are expected to respond in similar ways to risky choices. Some research in entrepreneurship, however, indicates that women may perceive situations and make decisions differently from men. Other researchers suggest that women differ from men in their risk awareness, not in their risk aversion (Stacey, 2008). This idea of differences in risk taking between men and women also extends to everyday life. Iceland, for instance, recently appointed two women to head their newly nationalized big banks, under the assumption that women would manage the banks without taking disastrous, aggressive risks (Stacey, 2008). Some researchers, however, suggest that this may not be particularly wise because senior female CEOs may take even more risks than male CEOs to compensate for gender disadvantages (Stacey, 2008). Another way to advance the field of strategic decision making might be to examine how well the approaches discussed in this chapter, either singly or jointly, help us understand real-life events and help decision makers develop practical, effective responses to these events. The current financial crisis and its precursor, the collapse of the subprime 176

market, have led many to question the compensation provided to top managers of failing financial institutions. Although agency theory certainly informs us as to the role of incentives in influencing risky decision making by senior managers, a closer examination of the factors leading to the crisis indicates that incentives alone cannot explain the crisis completely. Bromiley, Cordero, and Gong (2008) examined evidence from the subprime mortgage market and suggested that industry competition and organizational learning (but with firms learning the wrong thing) may have contributed significantly to the crisis in this market. They argued that when the salience of feedback does not match its importance, firms will adapt toward the salient items with potentially negative results. Further, they noted the role of competition in forcing firms to match the offerings of other firms even when they perceive the offerings as unwise in the long run. A consideration of these kinds of findings may lead researchers to revisit some the assumptions underpinning some of the strategic decision making approaches, potentially advancing theory and improving predictions regarding strategic decision making.

References Abdellaoui, M., Bleichrodt, H., & Parschiv, C. (2007). Loss aversion under prospect theory: A parameter-free measurement. Management Science, 53, 1659–1674. Amason, A. C., & Schweiger, D. M. (1994). Resolving the paradox of conflict, strategic decision making, and organizational performance. International Journal of Conflict Management, 5, 239–253. Amihud, Y., & Lev, B. (1999). Does corporate ownership structure affect its strategy towards diversification? Strategic Management Journal, 20, 1063–1069. Anand, V., Manz, C. C., & Glick, W. H. (1998). An organizational memory approach to information management. Academy of Management Review, 23, 796–809. Argote, L., & Greve, H. R. (2007). A behavioral theory of the firm—40 years and counting: Introduction and impact. Organization Science, 18, 337–349. Arkes, H. R., Hirshleifer, D., Jiang, D., & Lim, S. (2008). Reference point adaptation: Tests in the domain of security trading. Organizational Behavior and Human Decision Processes, 105, 67–81. Audia, P. G., & Greve, H. R. (2006). Less likely to fail: Low performance, firm size, and factory expansion in the shipbuilding industry. Management Science, 52, 83–94.

Strategic Decision Making

Baird, I. S., & Thomas, H. (1990). What is risk anyway? Using and measuring risk in strategic management. In R. A. Bettis & H. Thomas (Eds.), Risk, strategy, and management (Vol. 5, pp. 21–54). Greenwich, CT: JAI Press. Bamberger, P., & Fiegenbaum, A. (1996). The role of strategic reference points in explaining the nature and consequences of human resource strategy. Academy of Management Review, 21, 926–958.

Copyright American Psychological Association. Not for further distribution.

Bantel, K. A., & Jackson, S. E. (1989). Top management and innovations in banking: Does the composition of the top team make a difference? Strategic Management Journal, Summer Special Issue, 10, 107–124. Barr, P., Stimpert, J. L., & Huff, A. S. (1992). Cognitive change, strategic action, and organizational renewal. Strategic Management Journal, 13, 15–36. Bazerman, M. H. (1994). Judgment in managerial decision making (3rd ed.). New York: Wiley. Beckman, C. M., & Haunschild, P. R. (2002). Network learning: The effect of partners’ heterogeneity of experience on corporate acquisitions. Administrative Science Quarterly, 47(1), 92–124. Berle, A., & Means, G. (1932). The modern corporation and private property. New York: MacMillan. Bourgeois, L. J., III. (1985). Strategic goals, perceived uncertainty and economic performance in volatile environments. Academy of Management Journal, 28, 548–573. Bowman, E. H. (1980). A risk/return paradox for strategic management. Sloan Management Review, 21(3), 17–31. Bowman, E. H. (1982). Risk seeking by troubled firms. Sloan Management Review, 23(4), 33–42. Bowman, E. H. (1984). Content analysis of annual reports for corporate strategy and risk. Interfaces, 14, 61–71. Brockhaus, R. H. (1980). Risk taking propensity of entrepreneurs. Academy of Management Journal, 23, 509–520. Brockhaus, R. H. (1982). The psychology of the entrepreneur. In C. A. Kent, D. L. Sexton, & K. H. Vesper (Eds.), Encyclopedia of entrepreneurship (pp. 39–71). Englewood Cliffs, NJ: Prentice Hall. Bromiley, P. (1987). Do forecasts produced by organizations reflect anchoring and adjustment? Journal of Forecasting, 6, 201–210. Bromiley, P. (1991). Testing a causal model of corporate risk taking and performance. Academy of Management Journal, 34, 37–59. Bromiley, P. (in press). A prospect theory model of resource allocation. Decision Analysis. Bromiley, P., Cordero, J., & Gong, Y. (2008). Learning the wrong thing: Organizational adaptation, competition,

and disaster. Irvine: University of California, Merage School of Business. Bromiley, P., & Curley, S. P. (1992). Personality and individual differences in risk-taking. In J. Frank Yates (Ed.), Risk-taking behavior (pp. 87–132). Chichester, England: John Wiley & Sons. Bromiley, P., Miller, K., & Rau, D. (2001). Risk in strategic management research. In M. A. Hitt, R. E. Freeman, & J. S. Harrison (Eds.), The Blackwell handbook of strategic management (pp. 259–288). Oxford, England: Blackwell. Butler, R. J., Davies, L., Pike, R., & Sharp, J. (1993). Strategic investment decisions. London: Routledge. Camuffo, A., Furlan, A., & Rettore, E. (2007). Risk sharing in supplier relations: An agency model for the Italian air-conditioning industry. Strategic Management Journal, 28, 1257–1266. Carpenter, M. A., Pollock, T. G., & Leary, M. M. (2003). Testing a model of reasoned risk-taking: Governance, the experience of principals and agents, and global strategy in high-technology IPO firms. Strategic Management Journal, 24, 803–820. Chattopadhyay, P., Glick, W. H., & Huber, G. P. (2001). Organizational actions in response to threats and opportunities. Academy of Management Journal, 44, 937–955. Chen, W. R., & Miller, K. D. (2007). Situational and institutional determinants of firms’ R&D search intensity. Strategic Management Journal, 28, 369–381. Cho, T. S., & Hambrick, D. C. (2006). Attention as the mediator between top management team characteristics and strategic change: The case of airline deregulation. Organization Science, 17, 453–469. Cray, D., Mallory, G. R., Butler, R. J., Hickson, D. J., & Wilson, D. C. (1988). Sporadic, fluid, and constricted decision-making in organizations. Journal of Management Studies, 25, 13–39. Cray, D., Mallory, G. R., Butler, R. J., Hickson, D. J., & Wilson, D. C. (1991). Explaining decision processes. Journal of Management Studies, 28, 227–251. Cyert, R. M., & March, J. G. (1963). A behavioral theory of the firm. Englewood Cliffs, NJ: Prentice-Hall. Das, T. K., & Teng, B. (1999). Cognitive biases and strategic decision processes: An integrative perspective. Journal of Management Studies, 36, 757–778. Das, T. K., & Teng, B. (2001). Strategic risk behavior and its temporalities: Between risk propensity and decision context. Journal of Management Studies, 38, 515–534. D’Aveni, R. A. (1989). Dependability and organizational bankruptcy—An application of agency and prospecttheory. Management Science, 35, 1120–1138. 177

Bromiley and Rau

Demsetz, H. (1983). The structure of ownership and the theory of the firm. Journal of Law and Economics, 26, 375–390. Denis, D. J., Denis, D. K., & Sarin, A. (1999). Agency theory and the influence of equity ownership structure on corporate diversification strategies. Strategic Management Journal, 20, 1071–1076. Dess, G. D., & Priem, R. L. (1995). Consensus-performance research: Theoretical and empirical extensions. Journal of Management Studies, 32, 401–417.

Copyright American Psychological Association. Not for further distribution.

Duncan, R. B. (1972). Characteristics of organizational environments and perceived environmental uncertainty. Administrative Science Quarterly, 17, 313–327. Eden, C., & Ackerman, F. (1998). Making strategy. London, England: Sage. Eisenhardt, K. M. (1989). Making fast strategic decisions in high-velocity environments. Academy of Management Journal, 32, 543–576.

Frederickson, J. W., & Mitchell, T. R. (1984). Strategic decision processes: Comprehensiveness and performance in an industry with an unstable environment. Academy of Management Journal, 27, 399–423. Gamble, J. E. (2000). Management commitment to innovation and ESOP stock concentration. Journal of Business Venturing, 15, 433–447. Gigerenzer, G. (2000). Adaptive thinking: Rationality in a real world. New York: Oxford University Press. Gigerenzer, G., & Goldstein, D. G. (1996). Reasoning the fast and frugal way: Models of bounded rationality. Psychological Review, 103, 650–669. Gomez-Mejia, L. R., Haynes, K. T., Nunez-Nickel, M., Jacobson, K. J. L., & Moyano-Fuentes, J. (2007). Socioemotional wealth and business risks in familycontrolled firms: Evidence from Spanish olive oil mills. Administrative Science Quarterly, 52, 106–137.

Fama, E. (1980). Agency problems and the theory of the firm. Journal of Political Economy, 88, 288–307.

Greve, H. R. (2003). A behavioral theory of R&D expenditures and innovations: Evidence from shipbuilding. Academy of Management Journal, 46, 685–702.

Faraj, S., & Sproull, L. (2000). Coordinating expertise in software development teams. Management Science, 46, 1554–1568.

Grimm, C. M., & Smith, K. G. (1991). Management and organizational change: A note on the railroad industry. Strategic Management Journal, 12, 557–562.

Fiegenbaum, A. (1990). Prospect theory and the risk– return association: An empirical examination in 85 industries. Journal of Economic Behavior and Organization, 14, 187–203.

Hambrick, D. C. (1994). Top management groups: A conceptual integration and reconsideration of the “team” label. Research in Organizational Behavior, 16, 171–213.

Fiegenbaum, A., & Thomas, H. (1986). Dynamic and risk measurement perspectives on Bowman’s risk–return paradox for strategic management: An empirical study. Strategic Management Journal, 7, 395–407. Fiegenbaum, A., & Thomas, H. (1988). Attitudes toward risk and the risk–return paradox—Prospect theory explanations. Academy of Management Journal, 31, 85–106. Fiegenbaum, A., & Thomas, H. (1990). Strategic groups and performance: The U.S. insurance industry, 1970–1984. Strategic Management Journal, 11, 197–215. Fiegenbaum, A., & Thomas, H. (1995). Strategic groups as reference groups: Theory, modeling, and empirical examinations of industry and competitive strategy. Strategic Management Journal, 16, 461–476. Finkelstein, S., & Hambrick, D. (1990). Top management team tenure and organizational outcomes: The moderating role of managerial discretion. Administrative Science Quarterly, 35, 484–503. Fiske, S., & Taylor, S. (1991). Social cognition (2nd ed.). New York: McGraw Hill. Frederickson, J. W. (1984). The comprehensiveness of strategic decision processes: Extensions, observations, future directions. Academy of Management Journal, 27, 445–466. 178

Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a reflection of its top managers. Academy of Management Review, 9, 193–206. Harris, J., & Bromiley, P. (2007). Incentives to cheat: The influence of executive compensation and firm performance on financial misrepresentation. Organization Science, 18, 350–367. Hartwick, J., Sheppard, B. H., & Davis, J. H. (1982). Group remembering: Research and implications. In R. A. Guzzo (Ed.), Improving group decision making in organizations (pp. 41–72). San Diego, CA: Academic Press. He, X., & Mittal, V. (2007). The effect of decision risk and project stage on escalation of commitment. Organizational Behavior and Human Decision Processes, 103, 225–237. Hickson, D. J., Butler, R. J., Cray, D., Mallory, G. R., & Wilson, D. C. (1986). Top decisions: Strategic decisionmaking in organizations. San Francisco: Jossey-Bass. Highhouse, S., & Yuce, P. (1996). Perspectives, perceptions, and risk-taking behavior. Organizational Behavior and Human Decision Processes, 65, 159–167. Hodgkinson, G. P., Brown, N. J., Maule, A. J., Glaister, K. W., & Pearman, A. D. (1999). Breaking the frame: An analysis of strategic cognition and decision making under uncertainty. Strategic Management Journal, 20, 977–985.

Strategic Decision Making

Hogarth, R. (1980). Judgement and choice: The psychology of decision. Chichester, England: John Wiley & Sons. Janis, I. L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes. Boston: Houghton Mifflin. Jegers, M. (1991). Prospect theory and the risk return relation—Some Belgian evidence. Academy of Management Journal, 34, 215–225.

Copyright American Psychological Association. Not for further distribution.

Jensen, M. C., & Meckling, W. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. The Journal of Financial Economics, 3, 305–360. Jensen, M. C., & Smith, C. L. (1985). Stockholder, manager, and creditor interests: Applications of agency theory. In E. I. Altman & M. G. Subrahmanyam (Eds.), Recent advances in corporate finance (pp. 95–131). Homewood, IL: Irwin. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–291. Kaufmann, J. B., & O’Neill, H. M. (2007). Do culturally distant partners choose different types of joint ventures? Journal of World Business, 42, 435–448. Keil, M., Depledge, G., & Rai, A. (2007). Escalation: The role of problem recognition and cognitive bias. Decision Sciences, 38, 391–421. Kerr, J. L., & Kren, L. (1992). Effect of relative decision monitoring on chief executive compensation. Academy of Management Journal, 35, 370–397. Ketchen, D. J., Jr., & Palmer, T. B. (1999). Strategic responses to poor organizational performance: A test of competing perspectives. Journal of Management, 25, 683–706. Kilduff, M., Angelmar, R., & Mehra, A. (2000). Top management-team diversity and firm performance: Examining the role of cognitions. Organization Science, 11(1), 35–57. Klimoski, R., & Mohammed, S. (1994). Team mental model: Construct or metaphor? Journal of Management, 20, 403–437. Knight, D., Pearce, C. L., Smith, K. G., Olian, J. D., Sims, H. P., Smith, K. A., & Flood, P. (1999). Top management team diversity, group process, and strategic consensus. Strategic Management Journal, 20, 445–465. Knight, F. H. (1921). Risk, uncertainty, and profit. Chicago: University of Chicago Press. Kogan, N., & Wallach, M. A. (1964). Risk taking: A study in cognition and personality. New York: Holt, Rinehart and Winston. Lane, P. J., Cannella, A. A., & Lubatkin, M. H. (1998). Agency problems as antecedents to unrelated mergers and diversification: Amihud and Lev reconsidered. Strategic Management Journal, 19, 555–578.

Lane, P. J., Cannella, A. A., & Lubatkin, M. H. (1999). Ownership structure and corporate strategy: One question viewed from two different worlds. Strategic Management Journal, 20, 1077–1086. Larraza-Kintana, M., Wiseman, R. M., Gomez-Mejia, L. R., & Welbourne, T. M. (2007). Disentangling compensation and employment risks using the behavioral agency model. Strategic Management Journal, 28, 1001–1019. Laughlin, P. R., & Ellis, A. L. (1986). Demonstrability and social combination processes on mathematical intellective tasks. Journal of Experimental Social Psychology, 22, 177–189. Lauriola, M., Levin, I. P., & Hart, S. S. (2007). Common and distinct factors in decision making under ambiguity and risk: A psychometric study of individual differences. Organizational Behavior and Human Decision Processes, 104, 130–149. Lehner, J. M. (2000). Shifts of reference points for framing of strategic decisions and changing risk–return associations. Management Science, 46, 63–76. Libby, R., Trotman, K. T., & Zimmer, I. (1987). Member variation, recognition of expertise, and group performance. Journal of Applied Psychology, 72, 81–97. Lubatkin, M. H., Lane, P. J., Collin, S. O., & Very, P. (2005). Origins of corporate governance in the USA, Sweden and France. Organization Studies, 26, 867–888. Lubatkin, M. H., Lane, P. J., Collin, S. O., & Very, P. (2007). An embeddedness framing of governance and opportunism: Towards a cross-nationally accommodating theory of agency. Journal of Organizational Behavior, 28, 43–58. Lubatkin, M. H., Lane, P. J., & Schulze, W. S. (2001). A strategic management model of agency relationships in firm governance. In M. A. Hitt, R. E. Freeman, & J. S. Harrison (Eds.), The Blackwell handbook of strategic management (pp. 229–258). Oxford, England: Blackwell. Lyles, M., & Schwenk, C. R. (1992). Top management, strategy, and organizational knowledge structures. Journal of Management Studies, 29, 155–174. Lyles, M., & Thomas, H. (1988). Strategic problem formulation: Biases and assumptions embedded in alternative decision making models. Journal of Management Studies, 25, 131–146. MacCrimmon, K. R., & Wehrung, D. A. (1985). A portfolio of risk measures. Theory and Decision, 19, 1–29. MacCrimmon, K. R., & Wehrung, D. A. (1986). Taking risks: The management of uncertainty. New York: Free Press. MacCrimmon, K. R., & Wehrung, D. A. (1990). Characteristics of risk taking executives. Management Science, 36, 422–435. 179

Bromiley and Rau

Maner, J. K., & Gerend, M. A. (2007). Motivationally selective risk judgments: Do fear and curiosity boost the boons or the banes? Organizational Behavior and Human Decision Processes, 103, 256–267.

Milliken, F. J. (1987). Three types of perceived uncertainty about the environment: State, effect, and response uncertainty. Academy of Management Review, 12, 133–143.

March, J. G. (1994). A primer on decision making: How decisions happen. New York: Free Press.

Mintzberg, H. (1990). The design school: Reconsidering the basic premises of strategic management. Strategic Management Journal, 11, 171–195.

March, J. G., & Shapira, Z. (1987). Managerial perspectives on risk and risk taking. Management Science, 33, 1404–1418.

Copyright American Psychological Association. Not for further distribution.

March, J. G., & Shapira, Z. (1992). Variable risk preferences and the focus of attention. Psychological Review, 99, 172–183. March, J. G., & Simon, H. (1958). Organizations. New York: John Wiley & Sons. Markovitch, D. G., Steckel, J. H., & Yeung, B. (2005). Using capital markets as market intelligence: Evidence from the pharmaceutical industry. Management Science, 51, 1467–1480. Massini, S., Lewin, A. Y., & Greve, H. R. (2005). Innovators and imitators: Organizational reference groups and adoption of organizational routines. Research Policy, 34, 1550–1569. McNamara, G., & Bromiley, P. (1997). Decision making in an organizational setting: Cognitive and organizational influences on risk assessment in commercial lending. Academy of Management Journal, 40, 1063–1088. McNamara, G., Luce, R. A., & Thompson, G. H. (2002). Examining the effect of complexity in cognitive strategic group knowledge structures on firm performance. Strategic Management Journal, 23, 153–170.

Mintzberg, H., Raisinghani, D., & Theoret, A. (1976). The structure of “unstructured” decision processes. Administrative Science Quarterly, 21, 246–275. Mintzberg, H., & Waters, J. (1985). Of strategies, deliberate and emergent. Strategic Management Journal, 6, 257–272. Mone, M. A., McKinley, W., & Barker, V. L. (1998). Organizational decline and innovation: A contingency framework. Academy of Management Review, 23, 115–132. Moon, H., Conlon, D. E., Humphrey, S. E., Quigley, N., Devers, C. E., & Nowakowski, J. M. (2003). Group decision process and incrementalism in organizational decision making. Organizational Behavior and Human Decision Processes, 92, 67–79. Mooney, A. C. (2000). The antecedents to conflict during strategic decision making: The importance of behavioral integration. Unpublished doctoral dissertation, University of Georgia, Athens. Mosakowski, E. (1998). Entrepreneurial resources, organizational choices, and competitive outcomes. Organization Science, 9, 625–643. Mullins, J. W., Forlani, D., & Walker, O. C. (1999). Effects of organizational and decision-maker factors on new product risk taking. Journal of Product Innovation Management, 16, 282–294.

McNamara, G., Moon, H., & Bromiley, P. (2002). Banking on commitment: Intended and unintended consequences of an organization’s attempt to attenuate irrational commitment. Academy of Management Journal, 45, 443–452.

Murray, A. I. (1989). Top management group heterogeneity and firm performance. Strategic Management Journal, Summer Special Issue, 10, 125–141.

Michael, S. C. (1996). To franchise or not to franchise: An analysis of decision rights and organizational form shares. Journal of Business Venturing, 11, 57–71.

Nadkarni, S., & Narayanan, V. K. (2007). Strategic schemas, strategic flexibility, and firm performance: The moderating role of industry clockspeed. Strategic Management Journal, 28, 243–270.

Michel, J. G., & Hambrick, D. C. (1992). Diversification posture and top management team characteristics. Academy of Management Journal, 35, 9–37.

Natarajan, R., Sethuraman, K., & Suryasekar, K. (2005). Reducing incremental costs associated with the delegation of incentive contracting decisions: An analysis of planning and control decisions in a manufacturing setting in the presence of moral hazard. European Journal of Operational Research, 162, 816–838.

Miles, R. E., & Snow, C. C. (1978). Organizational strategy, structure, and process. New York: McGraw-Hill. Miller, C. C., Burke, L. M., & Glick, W. H. (1998). Cognitive diversity among upper-echelon executives: Implications for strategic decision processes. Strategic Management Journal, 19, 39–58. Miller, K. D., & Chen, W. R. (2004). Variable organizational risk preferences: Tests of the March–Shapira model. Academy of Management Journal, 47, 105–115. 180

Newell, B. R., Weston, N. J., & Shanks, D. R. (2003). Empirical tests of a fast and frugal heuristic: Not everyone “takes the best.” Organizational Behavior and Human Decision Processes, 91, 82–96. Nicholson, N., Soane, E., Fenton-O’Creevy, M., & Willman, P. (2005). Personality and domain-specific risk taking. Journal of Risk Research, 8, 157–176.

Strategic Decision Making

Nutt, P. C. (1984). Types of organizational decision processes. Administrative Science Quarterly, 29, 414–450.

Schwenk, C. (1984). Cognitive simplification processes in strategic decision making. Strategic Management Journal, 5, 111–128.

Ocasio, W. (1995). The enactment of economic adversity— A reconciliation of theories of failure-induced change and threat-rigidity. Research in Organizational Behavior: An Annual Series of Analytical Essays and Critical Reviews, 17, 287–331.

Schwenk, C. (1985). Management illusions and biases: Their impact on strategic decisions. Long Range Planning, 18(5), 74–80.

Copyright American Psychological Association. Not for further distribution.

Park, K. M. (2007). Antecedents of convergence and divergence in strategic positioning: The effects of performance and aspiration on the direction of strategic change. Organization Science, 18, 386–402. Pennings, J. M. E., & Smidts, A. (2000). Assessing the construct validity of risk attitude. Management Science, 46, 1337–1348. Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations. New York: Harper & Row.

Schwenk, C. R. (1995). Strategic decision making. Journal of Management, 21, 471–493. Sengun, A. E., & Wasti, S. N. (2007). Trust, control, and risk: A test of Das and Teng’s conceptual framework for pharmaceutical buyer–supplier relationships. Group and Organization Management, 32, 430–464. Sharp, D. J., & Salter, S. B. (1997). Project escalation and sunk costs: A test of the international generalizability of agency and prospect theories. Journal of International Business Studies, 28, 101–121.

Prahalad, C., & Bettis, R. (1986). The dominant logic: A new linkage between diversity and performance. Strategic Management Journal, 7, 485–502.

Shimizu, K. (2007). Prospect theory, behavioral theory, and the threat-rigidity thesis: Combinative effects on organizational decisions to divest formerly acquired units. Academy of Management Journal, 50, 1495–1514.

Quinn, J. B. (1978). Strategic change: Logical incrementalism. Sloan Management Review, 20, 7–21.

Simon, H. A. (1957). Models of man: Social and rational. New York: Wiley.

Quinn, J. B. (1980). Strategies for change: Logical incrementalism. Homewood, IL: Irwin.

Simon, H. A. (1960). The new science of management decision. New York: Harper & Row.

Rau, D. (2005). The influence of relationship conflict and trust on the transactive memory—Performance relationship in top management teams. Small Group Research, 36, 746–771.

Simons, T., Pelled, L. H., & Smith, K. A. (1999). Making use of difference: Diversity, debate, and decision comprehensiveness in top management teams. Academy of Management Journal, 42, 662–673.

Reger, R. K., & Huff, A. S. (1993). Strategy groups: A cognitive perspective. Strategic Management Journal, 14, 103–124.

Simsek, Z., Veiga, J. F., Lubatkin, M. H., & Dino, R. N. (2005). Modeling the multilevel determinants of top management team behavioral integration. Academy of Management Journal, 48, 69–84.

Reuer, J. J., & Ragozzino, R. (2006). Agency hazards and alliance portfolios. Strategic Management Journal, 27, 27–43.

Singh, J. V. (1986). Performance, slack, and risk taking in organizational decision making. Academy of Management Journal, 29, 562–585.

Rulke, D. L., Zaheer, S., & Anderson, M. H. (2000). Sources of managers’ knowledge of organizational capabilities. Organizational Behavior and Human Decision Processes, 82, 134–149.

Sitkin, S. B., & Pablo, A. L. (1992). Reconceptualizing the determinants of risk behavior. Academy of Management Review, 17, 9–38.

Sanders, W. G., & Carpenter, M. A. (2003). Strategic satisficing? A behavioral-agency theory perspective on stock repurchase program announcements. Academy of Management Journal, 46, 160–178.

Sitkin, S. B., & Weingart, L. R. (1995). Determinants of risky decision-making behavior: A test of the mediating role of risk perceptions and propensity. Academy of Management Journal, 38, 1573–1592.

Sanders, W. G., & Hambrick, D. C. (2007). Swinging for the fences: The effects of CEO stock options on company risk taking and performance. Academy of Management Journal, 50 1055–1078.

Slattery, J. P., & Ganster, D. C. (2002). Determinants of risk taking in a dynamic uncertain context. Journal of Management, 28, 89–106.

Schmidt, U., & Zank, H. (2008). Risk aversion in cumulative prospect theory. Management Science, 54, 208–216. Schoemaker, P. J. H. (1993). Strategic decisions in organizations—Rational and behavioral views. Journal of Management Studies, 30, 107–129.

Smith, K. G., Smith, K. A., Olian, J. D., Sims, H. P., O’Bannon, D. P., & Scully, J. A. (1994). Top management team demography and process: The role of social integration and communication. Administrative Science Quarterly, 39, 412–438. Stacey, K. (2008). Recession dents image of macho management. Financial Times, November 16. 181

Copyright American Psychological Association. Not for further distribution.

Bromiley and Rau

Stasser, G. (1992). Pooling of unshared information during group discussion. In S. Worchel, W. Wood, & J. A. Simpson (Eds.), Group process and productivity (pp. 48–67). Newbury Park, CA: Sage.

Wang, H. C., & Barney, J. B. (2006). Employee incentives to make firm-specific investments: Implications for resource-based theories of corporate diversification. Academy of Management Review, 31, 466–476.

Staw, B. M., Sandelands, L. E., & Dutton, J. E. (1981). Threat-rigidity effects in organizational behavior: A multilevel analysis. Administrative Science Quarterly, 26, 501–524.

Wegner, D. M. (1987). Transactive memory: A contemporary analysis of the group mind. In B. Mullen & G. R. Goethals (Eds.), Theories of group behavior (pp. 185–208). New York: Springer-Verlag.

Stewart, W. H., Jr., & Roth, P. L. (2001). Risk propensity differences between entrepreneurs and managers: A meta-analytic review. Journal of Applied Psychology, 86, 145–153.

West, C. T., & Schwenk, C. R. (1996). Top management team strategic consensus, demographic homogeneity and firm performance: A report of resounding nonfindings. Strategic Management Journal, 17, 571–576.

Tamura, H. (2005). Behavioral models for complex decision analysis. European Journal of Operational Research, 166, 655–665.

Wezel, F. C., & Saka-Helmhout, A. (2006). Antecedents and consequences of organizational change: “Institutionalizing” the behavioral theory of the firm. Organization Studies, 27, 265–286.

Tiwana, A., & Bush, A. A. (2007). A comparison of transaction cost, agency, and knowledge-based predictors of IT outsourcing decisions: A U.S.–Japan cross-cultural field study. Journal of Management Information Systems, 24, 259–300.

Whyte, G. (1998). Recasting Janis’s groupthink model: The key role of collective efficacy in decision fiascoes. Organizational Behavior and Human Decision Processes, 73, 185–209.

Tversky, A., & Kahneman, D. (1974, September 27). Judgement under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131.

Wiersema, M. F., & Bantel, K. A. (1992). Top management team demography and corporate strategic change. Academy of Management Journal, 35, 91–121.

Tversky, A., & Kahneman, D. (1981, January 30). The framing of decisions and the psychology of choice. Science, 211(4481), 453–458.

Wiseman, R. M., & Bromiley, P. (1991). Risk–return associations: Paradox or artifact? An empirically tested explanation. Strategic Management Journal, 12, 231–241.

Voss, G. B., Sirdeshmukh, D., & Voss, Z. G. (2008). The effects of slack resources and environmental threat on product exploration and exploitation. Academy of Management Journal, 51, 147–164. Wally, S., & Becerra, M. (2001). Top management team characteristics and strategic changes in international diversification: The case of U.S. multinationals in the European community. Group and Organization Management, 26, 165–188.

Wiseman, R. M., & Bromiley, P. (1996). Toward a model of risk in declining organizations: An empirical examination of risk, performance, and decline. Organization Science, 7, 524–543. Wiseman, R. M., & Catanach, A. H. (1997). A longitudinal disaggregation of operational risk under changing regulations: Evidence from the savings and loan industry. Academy of Management Journal, 40, 799–830.

Walsh, J. P. (1995). Managerial and organizational cognition: Notes from a trip down memory lane. Organization Science, 6, 280–321.

Wiseman, R. M., & Gomez-Mejia, L. R. (1998). A behavioral agency model of managerial risk taking. Academy of Management Review, 23, 133–153.

Wang, C. X., & Webster, S. (2007). Channel coordination for a supply chain with a risk-neutral manufacturer and a loss-averse retailer. Decision Sciences, 38, 361–389.

Wright, P., Kroll, M., Krug, J. A., & Pettus, M. (2007). Influences of top management team incentives on firm risk taking. Strategic Management Journal, 28, 81–89.

182

CHAPTER 7

LEADERSHIP

Copyright American Psychological Association. Not for further distribution.

Julian Barling, Amy Christie, and Colette Hoption

Leadership is a fascinating and controversial topic, about which much is known and much remains to be learned. Leadership has long captivated the attention of scholars, practitioners, and the public. For that reason, numerous journals are devoted to the topic of leadership, such as Leadership Quarterly, Leadership and Organization Development Journal, Journal of Leadership Studies and Leader to Leader. Nonetheless, some questions remain unresolved, such as: Are leaders born or made? Can we teach leadership in organizations? Can leaders be as influential as we might hope? Can leaders wreak as much harm as we might fear? Not only are these questions intriguing, but their answers also contribute substantially to our understanding of, and have implications for, the psychology of both leaders and their followers and behavior in organizations more generally. In this chapter, we address these and other questions. In doing so, we provide a review of what is known and what remains to be understood about leadership. Throughout this chapter, we follow what is now referred to as an evidence-based approach (Pfeffer & Sutton, 2006; Rousseau, 2006). Broadly speaking, the evidence-based approach mandates that management practices should be based on the best available empirical data, whenever such data exist. In a similar manner, we believe that the development of knowledge about leadership must also rest on the best available empirical evidence. Although this approach is by no means new—there were earlier indications that evaluating the effects of organizational interventions on the basis of “soft” versus “hard” outcomes would lead

to different conclusions (Bass, 1983; Terpstra, 1981)—concerns about methodological rigor in leadership research remain (e.g., Hunter, BedellAvers, & Mumford, 2007). Following from this evidence-based approach, the chapter includes discussions of leadership theories that have been subject to empirical scrutiny (e.g., leader–member exchange theory [LMX], transformational leadership) and excludes those that have largely escaped such scrutiny (e.g., Covey’s principle-centered leadership; Greenleaf’s servant leadership). We also choose this approach because the sheer number of leadership theories precludes the possibility of discussing them all. The remainder of this chapter presents an overview of the leadership literature guided by the evidenced-based approach, beginning with a review of leadership theories, then summarizing key findings within the field, and concluding with a path for future leadership research. A BRIEF HISTORY OF LEADERSHIP THEORIES As the topic of scholarly debate for centuries and the subject of systematic theoretical and empirical research for much of the past 100 years, leadership has a long tradition in the social sciences. Not surprisingly, with such an extensive history, the leadership literature has demonstrated ebbs and flows of prevailing wisdom. Although many ideas of the past have fallen from popular favor, the evolution of leadership perspectives is both reflected

Writing of this chapter was supported by grants from the Social Sciences and Humanities Research Council of Canada.

http://dx.doi.org/10.1037/12169-007 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

183

Barling, Christie, and Hoption

in and critical to the understanding of the dominant leadership theories of the present day. In this preliminary section we provide a brief history of the progression of leadership theories that form the foundation of current thinking and research in the field of leadership today. To complement this section, a visual depiction of leadership trends over the past quarter of a century is provided in Table 7.1, which we will reflect on in greater detail at the end of this section.

Copyright American Psychological Association. Not for further distribution.

Trait Theories of Leadership The very first theories of leadership sought to identify the physical characteristics or psychological traits that differentiated leaders from nonleaders or good leaders from poor leaders (House & Aditya, 1997). Some of the traits identified and studied included height (for a review, see Judge & Cable, 2004) and physical appearance (e.g., Cherulnik, Turns, & Wilderman, 1990), gender (for a review, see Eagly & Karau, 1991), authoritarianism (e.g., Tarnopol, 1958), intelligence (for a review, see Judge, Colbert, & Ilies, 2004), and self-confidence (e.g., Richardson & Hanawalt, 1952). Although selected evidence supported the notion that traits could be used to predict leadership emergence, and numerous studies showed a link between intelligence and leadership emergence and effectiveness (e.g., see Judge et al., 2004), most results lacked consistency across time, setting, and studies. Thus, even though the majority of leadership research conducted during the 1940s was devoted to understanding leader traits, by the 1950s very little consistent or conclusive empirical support emerged for any universal traits that could predict leader emergence and effectiveness. Research then progressed toward the identification of effective leader behaviors (Stogdill, 1950)— an endeavor that continues today. Nevertheless, a renewed interest in trait theories of leadership emerged in the 1970s when evidence for complex theories and measures of personality traits (e.g., the Big Five) emerged within the field of psychology, providing the leadership literature with better theories and measures with which to investigate potential leadership traits (House & Aditya, 1997). Based on these advances, more conclusive evidence has been found for the existence of personality traits that char184

acterize leaders, a topic that will be discussed in more detail later in this chapter.

Early Behavioral Theories of Leadership Similar to trait perspectives of leadership, early behavioral approaches attempted to uncover and verify leadership behaviors that were universally effective. The most extensively studied behaviors in these early endeavors were two dimensions identified in the Ohio State studies: Initiating Structure and Consideration. These two behaviors broadly represent task-focused and people-focused leadership behaviors, respectively. Initiating Structure describes leadership behaviors that create clear guidelines and procedures to facilitate the achievement of specified goals (e.g., Kerr, Schriesheim, Murphy, & Stogdill, 1974). Accordingly, three items from the Leader Behavior Description Questionnaire (LBDQ), the instrument often used to measure Initiating Structure, include “[My leader] maintains definite standards of performance,” “[My leader] encourages the use of uniform procedures,” and “[My leader] schedules the work to be done.” Whereas Initiating Structure focuses on what today might be more consistent with performance appraisal, Consideration depicts leadership behaviors that are centered on reciprocal trust, respect, and a concern for the welfare of followers (e.g., Kerr et al., 1974), and is represented by LBDQ items such as “[My leader] helps people in the work group with their personal problems” and “[My leader] backs up what people under him/her do.” Originally, leadership scholars hypothesized that the ideal leadership style, the one that would most positively affect follower attitudes and performance, would incorporate high levels of both Initiating Structure and Consideration (see Kerr et al., 1974, for a review). Such a leader would be able to guide followers toward the accomplishment of organizational goals and provide them with the emotional support necessary to help them perform at the highest level. Researchers concluded that Initiating Structure behaviors would have a stronger positive relationship with follower performance, whereas Consideration behaviors would be more closely related to follower attitudes, such as satisfaction and morale (see Judge, Piccolo, & Illies, 2004, for a review). Nevertheless, meta-analytic evidence suggests that

Consideration-initiating structure and leadership Situational leadership theory Path–goal theory Substitutes for leadership Leader–member exchange or vertical dyad linkage Transformational Leadership Charismatic Leadership Romance of leadership Implicit leadership theories Prototypicality Followership Shared leadership Authentic leadership

Leadership theory 4 5 3 0 3 0 2 2 0 3 1

2 2 0 1 0 0 1 0 0 0

1983

1982

8

1981

1980

0 0 1 4

0 2

2 2 1 8

7

1985

1984

0 0 0 0

0 2

3 1 4 5

10

1987

1986

0 0 2 1

1 3

4 1 2 6

5

1989

1988

0 0 1 3

0 7

4 2 2 11

6

1991

1990

Frequencies With Which Leadership Theories Have Been Studied, 1980–2007

TABLE 7.1

0 0 3 0

1 8

3 0 3 10

3

1993

1992

0 0 1 1

2 8

0 2 4 12

3

1995

1994

1 2 2 2

7 8

3 2 8 23

4

1997

1996

11 22 1 1 1 1 2

7 0 1 27

2

1999

1998

Copyright American Psychological Association. Not for further distribution.

30 15 1 6 4 4 4

1 0 4 24

0

2001

2000

52 12 1 2 0 2 5

0 0 3 24

3

2003

2002

75 28 3 6 4 2 7 8

0 0 7 35

3

2005

2004

105 18 8 7 5 5 19 7

3 4 1 57

0

2007

2006

284 135 14 26 16 27 49 15

37 19 40 246

58

Total

Leadership

185

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

the correlation between Initiating Structure and Consideration is .17 and that sometimes they are, in fact, negatively correlated with each other, depending on the measurement instrument used (see Judge et al., 2004). Thus, although high Initiating Structure and high Consideration may be the optimal expression of these behaviors from a theoretical perspective, this combination may not accurately capture patterns of leader behavior in organizations. The inconclusive relationship between Initiating Structure and Consideration was not the only challenge to the Ohio State leadership framework; the theory was also criticized for its inability to generate consistent and reliable results across multiple samples during the 1950s and 1960s. In some cases, studies found negative relationships between Initiating Structure and follower satisfaction and between Consideration and performance (see Kerr et al., 1974). Although a later study, using robust statistical techniques to summarize the approximately 100 empirical studies examining these behaviors, showed that inconsistencies were largely the result of methodological contradictions (Judge et al., 2004), researchers in the 1950s and 1960s turned to situational moderators to explain discrepancies between studies and, ultimately, to identify the situations in which Initiating Structure and Consideration would be most effective. This pursuit began with a focus on organizational context, such as time pressures and physical demands, and characteristics of followers, such as intrinsic motivation and role ambiguity. By doing so, the leadership literature evolved from seeking the discovery of universally successful leadership behaviors to acknowledging that effective leadership behaviors may be situation-specific and then to examining the interaction between leadership behavior, situational contingencies, and leader effectiveness. (The progression toward perspectives that emphasized the importance of situational constraints on leadership is discussed in the following section.) Although Initiating Structure and Consideration were largely abandoned and deemed conceptually and methodologically invalid by the 1970s, there has been renewed interest in the Ohio State leadership dimensions. Judge et al. (2004) argued that deserting these two leadership dimensions was premature, 186

calling them the “forgotten ones.” In Judge et al.’s. meta-analytic review, they showed a significant relationship (all corrected for unreliability of measures and measurement error) between Initiating Structure (average r ≈ .29) and Consideration (average r ≈ .48), and job satisfaction (Initiating Structure r = .22; Consideration r = .46), satisfaction with the leader (Initiating Structure r = .33; Consideration r = .78), follower motivation (Initiating Structure r = .40; Consideration r = .50), leader job performance (Initiating Structure r = .24; Consideration r = .25), group-organization performance (Initiating Structure r = .30; Consideration r = .28), and leader effectiveness (Initiating Structure r = .39; Consideration r = .52). Furthermore, as expected, Consideration showed a stronger relationship with attitudinal outcomes (e.g., satisfaction with leader, job satisfaction) than did Initiating Structure; on the other hand, Initiating Structure related more to grouporganization performance than did Consideration. These findings suggest that despite the disappearance of the Ohio State leadership dimensions from the literature, more research attention is still needed to verify their contribution to our understanding of leadership, especially because researchers (e.g., Keller, 2006) have argued that they may be largely independent of the behavioral theory that dominates contemporary leadership research—transformational leadership theory.

Contingency Theories of Leadership Situational or contingency approaches to leadership build on the early behavioral theories previously described. Situational theories contend that the effectiveness of leadership traits or behaviors is dependent on characteristics of the situation, including features of the organization, the workplace, and the followers. The next sections discuss three situational theories of leadership that have received considerable theoretical and empirical attention: Fiedler’s contingency theory, the path–goal theory of leadership, and substitutes for leadership. Fiedler’s contingency theory. Fiedler’s (1967) theory is often credited as the first true contingency theory of leadership. The theory categorizes leaders as either task-motivated or relationship-motivated.

Copyright American Psychological Association. Not for further distribution.

Leadership

These categories are similar to Initiating Structure and Consideration, respectively, but Fiedler’s theory goes further by suggesting that the effectiveness of each varies by situation. In particular, contingency theory draws attention to three dimensions that may characterize a situation: (a) leader–follower relations, (b) performance goal clarity, and (c) formal authority or power. Each dimension influences the extent to which leaders are afforded a sense of control over their jobs. For example, regarding authority, leaders are more likely to feel a sense of control when they have the formal authority to make decisions, and less likely when they don’t. Fiedler created eight “situations” based on the various combinations of these three dimensions, which ranged on a continuum from the most favorable situation (i.e., strong leader–follower relations, high task clarity, and high power) to the least favorable situation (i.e., weak leader–follower relations, low task clarity, and low power). Fiedler assumed that leadership behaviors were more difficult to change than were situations; consequently, he stipulated that optimal outcomes are a result of a correspondence between the leader’s existing behaviors and the situation. A central tenet of Fiedler’s contingency theory is that task-motivated leaders are more effective in extreme situations (i.e., very favorable or unfavorable situations), whereas relationship-motivated leaders are more effective in moderately favorable situations. To illustrate, a task-motivated leader was hypothesized to respond best during times of crisis or disaster, which are often chaotic and characterized by little formal structure in terms of the task and authority, while working closely with unfamiliar others (i.e., a very unfavorable situation). A relationship-motivated leader may become distracted and feel overwhelmed by the needs of others, whereas a task-motivated leader may have less difficulty acting quickly and decisively. On the other hand, in some situations or organizations, more moderate situational constraints are common, as is the case in creative industries, in which the leader–follower relationship often is strong and the task and formal authority structure is loosely defined. In these cases, relationship-oriented leaders excel because they are more considerate of followers’ individual needs and foster followers’ creative

inputs, whereas task-motivated leaders are likely to undermine followers’ creativity by failing to adequately appreciate the flexibility that they require. Despite a considerable amount of research, the evidence supporting Fiedler’s contingency theory is mixed. A number of meta-analytic reviews of the literature (e.g., Peters, Hartke, & Pohlmann, 1985; Schriesheim, Tepper, & Tetrault, 1994) have supported the notion that leadership style and situation interact, such that task-motivated leaders and relationship-motivated leaders excel in different situations. However, these results were strongest when tested in laboratory settings (Ayman, Chemers, & Fiedler, 1995). For instance, Peters et al. found that in laboratory studies (r ranging from an absolute magnitude of .01 to .33, corrected for sampling error), the more effective leadership style was correctly predicted in six of the eight contingency theory situations, compared to only half of the situations for field studies (r ranging from an absolute magnitude of .01 to .50, corrected for sampling error). The less persuasive findings in field settings could derive from improper measurement techniques. Thus, although the predictions of contingency theory are promising, advances in the theory are still needed to understand why discrepancies emerge between laboratory and field settings if it is to regain prominence in the current leadership literature. Path–goal leadership theory. A second contingency theory of leadership, path–goal leadership theory (House, 1971), also emerged in response to the disappointing findings of the Ohio State studies. Similar to Fiedler’s contingency theory, path–goal leadership theory had two objectives: (a) identify the role and behaviors of effective leaders, and (b) explore the situational contingencies that modify those behaviors. First, the theory posits that a leader’s role is to align the goals of followers with those of the organization. Next, a leader must facilitate the achievement of those goals. This is accomplished by helping followers realize that they have the capabilities to meet their goals, clarifying the path between the effort that they exert and goal attainment, and ensuring that the goals are valuable to followers. Unlike the Ohio State model (i.e., Initiating Structure and Consideration), path–goal leadership 187

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

theory identifies four categories of leadership behaviors that motivate followers to achieve their goals (House & Mitchell, 1974). Participative leadership behavior involves including followers in decision making and soliciting follower feedback. Consistent with the focus on participation in decision making that was already prevalent in the 1970s (e.g., Alutto & Belasco, 1972), this leadership behavior was hypothesized to enhance motivation by fostering overlap between follower and organizational goals and by providing followers with more appreciation and understanding of the pathway between effort and goal achievement. Directive path–goal-clarifying leadership behavior mimics Initiating Structure, motivating followers by providing task structure, feedback, and procedures that reduce role ambiguity, linking follower effort to performance and goal attainment, and communicating the rewards contingent on performance. By contrast, supportive leadership behavior is similar to Consideration, whereby leaders demonstrate their concern for the needs and best interests of followers and, by doing so, remove some of the potential obstacles that may prevent followers from obtaining their goals. The final leadership behavior identified in the theory is achievement-oriented leadership behavior, which involves creating challenging and high-standard performance goals and expressing confidence in followers’ abilities to meet such challenges. Followers should then respond with greater self-efficacy and effort toward goal attainment. Path–goal leadership theory’s second focus is on situational factors that render leadership behaviors more or less effective. These factors relate to the organizational environment, job design, and follower characteristics. However, empirical tests have not yielded conclusive support for the situational factors. For example, in a meta-analysis of 120 studies, Wofford and Liska (1993) found support for only 6 of 16 moderation hypotheses predicted by path– goal leadership theory (zr ranged from .31 to .51 corrected for unreliability of measures and sampling error). Schriesheim and Neider’s (1996) qualitative summary of the empirical literature suggested that the most consistent results for the situational focus of the theory have been for the relationship between directive clarifying behavior (most often measured 188

using the LBDQ Initiating Structure scale) and follower satisfaction when task characteristics (e.g., autonomy, task variety, and feedback) promote intrinsic motivation. We found the relationships between the remaining categories of leadership behaviors and the performance outcomes far less conclusive. It has been suggested that improper measurement and incomplete or inappropriately specified testing may account for the theory’s empirical shortcomings, as a result of which future research should fairly test the tenets of path–goal leadership theory (Schriesheim, Castro, Zhou, & DeChurch, 2006). Nonetheless, the complexity of the model may inhibit the possibility of subjecting the theory to an omnibus test. Substitutes for leadership. Kerr and Jermier’s (1978) substitutes for leadership theory further extends the situational perspective of leadership, in particular path-goal leadership theory, by illuminating additional situational contingencies of leadership behaviors (House, 1996). Specifically, Kerr and Jermier identified numerous situational variables that influence the relationship between leadership and its outcomes. These variables fit within one of two primary categories: neutralizers and substitutes. Neutralizers of leadership are situational factors that block the effects of leadership, rendering leadership behaviors inconsequential. One example of a neutralizer of leadership is spatial distance between the leader and the follower. By contrast, substitutes for leadership both neutralize leadership and positively influence attitudinal and performance outcomes. Examples of leadership substitutes include the follower’s intrinsic interest in the task, ability, training, and experience. These are substitutes because they relate positively to follower satisfaction, morale, and performance and therefore eliminate the need for leadership. Because the substitutes for leadership theory has intuitive appeal, the inadequate substantiating evidence for the theory across multiple rigorous study designs and samples remains a surprise (e.g., Dionne, Yammarino, Atwater, & James, 2002). Most past research suggests that substitutes for leadership have an additive effect; that is, they influence important outcomes irrespective of leadership behaviors without diminishing or negating the influence of leader-

Leadership

ship itself (Keller, 2006; Podsakoff, MacKenzie, & Bommer, 1996). Accordingly, substitutes for leadership theory has largely disappeared from current leadership research.

Copyright American Psychological Association. Not for further distribution.

Emergence of a Relational Theory of Leadership Fiedler’s contingency theory, path–goal leadership theory, and substitutes for leadership theory all emphasize the importance, if not the primacy, of situational factors. However, by the late 1970s and 1980s, focus shifted to a relational perspective of leadership as evidenced by the solidification of LMX (e.g., Dansereau, Graen, & Haga, 1975; Gerstner & Day, 1997). LMX is a theory of neither leader nor follower traits or behaviors; instead, it focuses primarily on the leader–follower dyad. Proponents of LMX argue that most leadership theories assume that there is unidirectional influence flowing from leader to follower and, furthermore, that a leader’s behaviors and effectiveness are both consistent across followers (e.g., Schriesheim, Castro, & Cogliser, 1999). To address these assumptions, LMX studies the unique leader–follower relationships that are fostered through leadership behaviors. These dyadic relationships are inherently unique because LMX contends that leaders develop different exchange relationships with different followers and, furthermore, that mutual influence occurs within each dyad (Gerstner & Day, 1997). In addition, LMX examines the quality of leader–follower relationships and in turn posits that higher quality relationships between leaders and followers will result in more positive organizational outcomes than lower quality relationships. High-quality LMX relationships are defined by mutual support, trust, liking, provision of latitude, attention, and loyalty (Schriesheim et al., 1999). It follows that the opposite qualities, such as downward influence, role distinctions, social distance, contractual obligations, and distrust, have been used to define low-quality LMX. These are general definitions of good and poor LMX, but realistically the components that constitute the good or poor quality of a relationship are likely to vary between individuals (House & Aditya, 1997). Accordingly, defining and measuring (universally) high- or low-quality LMX is challenging.

Nonetheless, based on the generalized features of high-quality LMX relationships provided above, research has revealed that high-quality LMX enjoys a number of positive implications for organizations. For instance, high-quality LMX has been positively associated with follower satisfaction, commitment, role clarity, and performance and negatively associated with follower turnover intentions and role conflict, across numerous empirical studies (see Gerstner & Day, 1997, for a review). Traditionally, there has been a concentrated focus on the positive effects of high-quality LMX, and it is implied that low-quality LMX simply fails to produce those positive effects. However, Townsend, Phillips, and Elkins (2000) showed that low-quality LMX may be more harmful than previously appreciated. These authors showed that high-quality LMX relationships lead to superior follower performance and increased organizational citizenship behaviors, whereas low-quality LMX relationships were not only unconnected to these outcomes but were also linked to retaliatory behaviors from followers. LMX theory has helped to advance our understanding of the leadership literature by considering not only the leader but also the leader–follower relationship within the leadership process. However, although the aim of LMX theory is to understand leader–follower relationships, it has been critiqued for its inability to adequately capture the dyadic nature of these relationships. A meta-analytic review of numerous published studies showed only a moderate correlation (r = .37 corrected for measurement unreliability and sampling error) between leaders’ perceptions of LMX quality and followers’ perceptions of LMX quality (Gerstner & Day, 1997), a disappointing finding given LMX theory’s presumption of dyadic associations. Furthermore, LMX theory has attracted criticism for its minimal attention to practical applications; the theory predicts associations between high-quality LMX and positive organizational outcomes, but it reveals little about how leaders can differentially generate these relationships with followers. Despite its shortcomings, LMX theory continues to be studied in current leadership research and has contributed greatly to our appreciation of and progression toward relational notions of leadership. 189

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

In summary, over the course of almost a century of research, the leadership literature has evolved from differentiating leaders from nonleaders on the basis of personal traits to recognizing the intricacies of leadership behaviors, situational contingencies, leader–follower relationships, and mutual, dyadic influences. As indicated throughout this section, some of these theories of leadership continue to attract scholarly attention today. However, we now turn our attention to the theory that undoubtedly dominates the leadership literature, namely, transformational leadership theory (e.g., Judge & Bono, 2000).

The Transformational Leadership Framework Although it is often difficult to pinpoint the intellectual origins of a theory, two seminal books stand out in the development of transformational leadership. First, Burns (1978) laid the groundwork for much subsequent thinking about transformational leadership in differentiating transformational leadership from other forms of leadership. Second, Bass (1985) further conceptualized the transformational framework (which includes both transformational and transactional leadership behaviors) and extended its focus to the organizational context. Both Burns and Bass stimulated others to conduct research on transformational leadership around the world; Bass himself was a major figure in empirical research on transformational leadership until his death in 2007. It is now generally accepted that four different behaviors (idealized influence, inspirational motivation, intellectual stimulation, and individualized consideration) constitute transformational leadership, whereas laissez-faire, management-by-exception, and contingent reward fall under the rubric of transactional leadership. We first discuss the latter three behaviors involved in transactional leadership and then move on to a discussion of the four behaviors involved in transformational leadership. In doing so, we note that our descriptions are based on discussions of these behaviors presented extensively elsewhere (e.g., Avolio, 1999a; Bass, 1985; Bass & Riggio, 2006) and that we limit our descriptions accordingly. 190

Transactional leadership. Transactional leadership comprises three different behaviors. First, a style of nonmanagement and nonleadership characterizes laissez-faire leader behaviors. These behaviors include avoiding and denying responsibility and neglecting to take any action even in dire situations. In its most extreme form, laissez-faire managers do nothing most of the time. Second, active management-by-exception takes place when leaders focus vigorously on followers’ mistakes and failures to meet standards. These leaders consistently look for errors at the expense of, rather than in addition to, a focus on positive events. Upon encountering these errors, leaders are likely to yell at, embarrass, punish, or discipline followers for their mistakes. Kelloway, Sivanathan, Francis, and Barling (2005) suggested that this negative focus, coupled with punitive action by the leader, results in employees experiencing active management-byexception as abusive. In contrast to active management-by-exception style, passive management-by-exception also describes managers who focus on errors. However, these managers do not actively monitor for mistakes; instead, they wait until the mistakes are of such consequence that they can no longer be ignored. Therefore, passive management-byexception is closer in nature to laissez-faire than active management-by-exception behaviors, and it also results in similar outcomes (Avolio, 1999b; Bass & Riggio, 2006). Third, while laissez-faire reflects the absence of management, and management-by-exception reflects “poor” management, contingent reward reflects “good” management. Contingent reward involves managers setting goals, providing feedback, and ensuring that employee behaviors have consequences, both positive and negative. As the name suggests, these leaders make rewards contingent on followers’ meeting a specified performance target. Nonetheless, the meaning of contingent reward is somewhat ambiguous, because in some analyses (e.g., the factor intercorrelations in Bycio, Hackett, & Allen’s [1995] study; their Table 3), contingent reward is significantly more highly correlated with elements of transformational leadership (such as charismatic leadership, intellectual

Copyright American Psychological Association. Not for further distribution.

Leadership

stimulation, individualized consideration) than with management-by-exception (rs of .79, .81, and .83, respectively, vs. −.26). To summarize Judge and Piccolo’s (2004) recent meta-analytic findings, active management-byexception is related to follower satisfaction with the leader (r = .24), follower motivation (r = .14), leader job performance (r = .13), group or organizational performance (r = −.09; 90% confidence interval includes zero), and leader effectiveness (r = .21). The overall relationships between passive managementby-exception and laissez-faire leadership and various leadership criteria are negative: specifically, passive management-by-exception is negatively related to follower satisfaction with the leader (r = −.14; 90% confidence interval includes zero), follower motivation (r = −.27), group or organizational performance (r = −.17), and leader effectiveness (r = −.19), while laissez-faire leadership is negatively related to job satisfaction (r = −.28), follower satisfaction with the leader (r = −.58), follower motivation (r = −.07; 90% confidence interval includes zero), leader job performance (r = −.01; 90% confidence interval includes zero), and leader effectiveness (r = −.54). By contrast, contingent reward has been associated positively with various criteria for leader effectiveness, such as job satisfaction (r = .64), satisfaction with the leader (r = .55), follower motivation (r = .59), leader job performance (r = .45), group or organizational performance (r = .16), and leader effectiveness (r = .55; all correlations corrected for measure unreliability, measurement error, and sampling error; Judge & Piccolo, 2004). Upon review of the transactional leadership framework, the term transactional leadership could be seen as an oxymoron because the behaviors (i.e., contingent reward and management-by-exception) are responses to employees’ behaviors and are based on the formal power accorded to managers. In contrast, leadership transcends situational needs and is based more on informal than formal sources of power. This might lead to the conceptualization of transactional leadership as consistent with “management” rather than leadership (Bass & Riggio, 2006). Transformational leadership. At the conceptual level, transformational leadership comprises four separate behaviors (see Bass & Riggio, 2006).

Idealized influence centers on leaders’ behaviors that are motivated by what is best for the organization and its members, rather than what is easy and expedient; these behaviors include providing a vision for the future and creating a collective sense of mission. Leaders who exhibit idealized influence are guided by their moral commitment to their followers and the collective good; they go beyond self-interest. Leaders who manifest idealized influence are able to resist organizational pressures for short-term financial outcomes and instead focus their efforts on the long-term well-being of their employees, themselves, and their organizations. Thus, a hallmark of idealized influence is that these leaders act with integrity. In this respect, Burns (1978) insisted that transformational leadership extends beyond Maslow’s (1965) self-actualization theory because a self-oriented focus is eschewed. Transformational leadership theory posits that leaders who manifest inspirational motivation encourage their employees to achieve more than what was thought possible, either by themselves or by those around them. They do so by setting high but realistic standards, which they transmit through interpersonal interactions; they tell stories and use symbols. These leaders inspire employees to surmount psychological setbacks and external obstacles, and they instill in employees the belief that they can confront and overcome current and future hurdles. Inspirationally motivating leaders’ interactions with subordinates reflect the self-fulfilling prophecy, as they help foster resilience and self-efficacy in their followers. Intellectual stimulation is the third facet of transformational leadership behavior. Earlier understandings of leadership assumed the technical expertise of the leader who could answer all questions posed by subordinates. In contrast, leaders who display intellectual stimulation encourage employees to think for themselves, question their own commonly held assumptions, reframe problems, and approach matters in innovative ways. Given the encouragement and opportunity to develop their own personal strategies to tackle setbacks, employees become more confident and adept in work-related and personal issues. Fourth and finally, individualized consideration characterizes leaders who pay special attention to employees’ personal needs for achievement and 191

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

development and act as mentors. These leaders provide the necessary caring, compassion, and empathy that influence employees’ well-being, and their instrumental and emotional support helps employees develop their potential and skills. In doing so, leaders establish a relationship with followers. The majority of the studies on transformational leadership use the Multi-factor Leadership Questionnaire (MLQ) to assess its four components, a measurement tool that we discuss more in depth later in this chapter. Although the four dimensions of transformational leadership are presented as conceptually distinct, it has become apparent from meta-analyses (Bycio et al., 1995; Tepper & Percy, 1994) that measurement issues make assessing them separately rather difficult. The recent development of several new scales (e.g., Alimo-Metcalfe & AlbanMetcalfe, 2001; Herold, Fedor, Caldwell, & Liu, 2008; Rafferty & Griffin, 2004) offers some promise for future research to answer remaining questions about whether the presence of all four behaviors is necessary for leadership to qualify as “transformational” or whether some of the four behaviors are more critical than others (Mackenzie, Podsakoff, & Jarvis, 2005). This issue is discussed in more detail later in this chapter.

Charismatic Leadership Theory The theory of charismatic leadership represents a major attempt to explain leadership. Initially proposed by the sociologist Max Weber (1947), charisma was characterized by followers’ belief that the leader possessed unusual and exceptional qualities. There are now several interpretations of this theory, with the two most prominent emphasizing either attributions that followers make about the leader (Conger & Kanungo, 1998) or actual leadership behaviors (House, 1977). These explanations of charismatic leadership evolved around the same time as transformational leadership theory (House, 1977). Although authors debate whether charismatic leadership and transformational leadership are interchangeable constructs, most suggest that the differences between the theories are minor (e.g., Conger & Kanungo, 1998; House & Podsakoff, 1994), which is in part evidenced by their similar or equivalent 192

measurement instruments (see Judge, Woolf, Hurst, & Livingston, 2008, for a review). One notable feature distinguishing the theories is their differential emphasis on follower attributions of leader behavior; the charismatic leadership tradition often argues that leadership exists “in the eyes of followers” (Conger, 1999, p. 153). What distinguishes charismatic leaders from other leaders is their ability to act in ways that encourage followers to perceive them and their visions as extraordinary. Specifically, charismatic leadership is attributed to leaders who challenge the status quo, inspire followers around a collective-focused vision of the future, show sensitivity to the needs of followers, and take personal risks to achieve their vision. Several general issues concerning transformational and charismatic leadership warrant comment at this stage. First, Judge and Bono (2000) noted that transformational and charismatic leadership attracted more research during the decade of the 1990s than all other theories of leadership combined. To understand the relative status of leadership theories today, we conducted a similar analysis of leadership topics from 1980 to 2007 using PsycINFO to search for the number of articles related to each topic. The results of this analysis appear in Table 7.1. The results for each of the theories discussed in this chapter illustrate the flux of empirical attention to leadership theories over time. Based on these data, several conclusions are appropriate. First, transformational leadership remains the most widely researched theory, with LMX and charismatic leadership the second and third most studied, respectively. Second, together, these three theories account for 63% of all leadership research between 1980 and 2007. However, recent years have also seen increasing devotion to theories of leadership that accentuate the role of followers in the leadership process; these theories are discussed in detail later in this chapter. THE DEVELOPMENT OF LEADERSHIP: ARE LEADERS BORN OR MADE? As we have already noted, leadership has long attracted attention. Still, one of the most enduring social questions remains: Are leaders born or made? A search of scholar.google.com (on May 4, 2009)

Copyright American Psychological Association. Not for further distribution.

Leadership

yielded at least 182 articles relating to this question. Understanding how leadership develops will have profound social, theoretical, and managerial implications. In this section, we address several different ways in which empirical research has confronted this question, by considering studies on genetic influences on leadership emergence, early family influences on leadership behavior, and executive development or leadership training. Before embarking on this discussion, we must draw a distinction between leadership emergence or role occupancy on the one hand, and leadership behaviors on the other. Leadership emergence reflects whether individuals become leaders by either occupying formal leadership roles or informally arising as a leader, whereas, traditionally, leadership behaviors center on how people actually behave once they have assumed a leadership role, or perhaps even when they are not in formal leadership positions (a topic discussed later in this chapter). As such, leadership emergence and leadership behaviors remain separate constructs. Despite the importance of understanding leadership development that reflects both leadership emergence and leadership behaviors, very little empirical research has addressed these critical issues.

Leadership Emergence “Some people are born to move and shake the world. Their blessings: high energy, exceptional intelligence, extreme persistence, self-confidence and a yearning to influence others” (Avolio, 1999a, p.18). Although much of the discussion in this chapter focuses on the issue of leadership behaviors and effectiveness, precisely which individuals attain positions of leadership in the first instance is of equal intrigue. As suggested by Avolio (1999a), one possibility is that some individuals are naturally predisposed to become leaders. Consistent with recent advances in the burgeoning field known as social neuroscience (Cacioppo et al., 2007), the notion that genetic and biological factors play an important role in the development of leadership should not be unexpected. Nevertheless, well-controlled studies on possible genetic and/or biological effects on leadership development remain scant. Empirical research on the role of genetic factors in behavior was initially stimulated by the classic studies

conducted by Bouchard and his colleagues on identical twins reared apart (see Bouchard, Lykken, McGue, Segal, & Tellegen, 1990). In a series of studies, they identified the extent to which genetic factors influence cognitive ability, personality, antisocial behavior, and psychopathology (e.g., Baker, Jacobson, Raine, Lozano, & Bezdijan, 2007; Bouchard et al., 1990; McGue & Bouchard, 1998). Research using the identical-twins-reared-apart paradigm has shown that, even controlling for job characteristics and physical demands, approximately 30% of the variance in intrinsic job satisfaction was a function of genetic factors (Arvey, Bouchard, Segal, & Abraham, 1989). Arvey and his colleagues have subsequently conducted several studies using this paradigm to investigate leadership emergence, operationalized as leadership role occupancy. Leadership role occupancy is a type of leadership emergence that reflects whether people actually hold positions of leadership within organizations and can also reflect an individual’s position in the organizational hierarchy (Ilies, Gerhardt, & Le, 2004). In the first of these later studies, Arvey, Rotundo, Johnson, Zhang, and McGue (2006) studied 110 identical twin pairs and 94 nonidentical twins, all males, from the Minnesota Twin Registry. Leadership role occupancy was indicated by the number of leadership roles that each twin held within work-related professional associations and the hierarchical level of the twins’ current leadership positions in their respective organizations. The results support a genetic influence on leadership role occupancy: At a descriptive level, if one twin held leadership positions in professional organizations and associations, the second was more likely to do so as well. Their heritability analyses went further to show that, after controlling for two personality variables (social potency and achievement), 30% of the variance in leadership role occupancy was explained by genetic factors, with the remaining variance, 70%, accounted for by environmental factors. In addition, the two personality variables that predict leadership role occupancy (i.e., achievement and social potency) were shown to share an important genetic component (accounting for 24% and 42% of the variance, respectively). Nonetheless, contrary to the authors’ predictions, personality did not mediate any genetic effects on role occupancy. 193

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

In a second study, Arvey, Zhang, Avolio, and Krueger (2007) intended to extend earlier findings with the inclusion of additional environmental predictors of leadership role occupancy, and they investigated 214 identical and 178 nonidentical female twins from the Minnesota Twin Registry. They again showed that heritability accounted for a significant proportion of the variance in leadership role occupancy (in this case, 32% of the variance), and their findings showed that work experience explained an additional 17% of the total variance in leadership role occupancy. A second environmental predictor, early family experiences, did not significantly relate to subsequent leadership role occupancy. However, it is possible that their three-item measure of family experience was not sufficiently comprehensive and therefore could not be fairly compared to the effects of work experiences on leadership role occupancy. These findings should remind organizational practitioners (and leaders alike) that there are aspects of leadership role occupancy that lie beyond their control; yet, the amount of variance accounted for by environmental factors in leadership role occupancy still provides organizations with substantial opportunities for involvement and potential influence. Furthermore, although intriguing, understanding more about genetic influences on leadership role occupancy offers very little in terms of selecting future leaders. Although these studies have begun to isolate the heritability influence on leadership role occupancy, it remains for future research to isolate genetic influences on leadership behaviors.

Early Family Influences The notion that early developmental influences affect leadership emergence and behavior are neither new nor implausible (see Bass, 1960; Karnes & D’Ilio, 1989). Elder (1974) noted that children whose fathers were unemployed during the Great Depression had to deal with external challenges at a young age. However, later in life, these children had achieved more educationally (they did better at school and were more likely to pursue higher education) and were more satisfied with their lives. Similarly, Cox and Cooper’s (1989) retrospective analysis demonstrated that a disproportionate number of successful British CEOs had experienced early 194

familial adversity (either loss of a parent or separation from parents). As a result of such adversity, they learned to take responsibility for themselves from an early age. Certainly the Cox and Cooper study justifies further research on early family influences on subsequent leadership emergence and behaviors. There has been interest in the specific nature of the early environment that might influence subsequent leadership. Generally, studies have shown that parents’ warmth and acceptance and their achievement demands predict predispositions to leadership behaviors in 10th-grade adolescents and in boys aged 16.5 years and girls aged 15.6 years (Bronfenbrenner, 1961, and Klonsky, 1983, respectively). Similarly, Towler (2005) showed that young adults (18– 25 years of age) with fathers who exercised high levels of psychological control were less likely to exhibit charismatic leadership. Hartman and Harris (1992) also showed that college students who subsequently held management positions modeled the leadership of individuals whom they admired early in their lives; most of these individuals were their parents. A more recent study of 196 pairs of twins who were part of the ongoing Minnesota Twin Family Study (Avolio, Rotundo, & Walumbwa, 2009) refines our understanding of the role of early experiences on later leadership role occupancy in several important ways. First, Avolio et al. (2009) focused on Baumrind’s (1971) notion of authoritative parenting, which is a combination of psychological autonomy, acceptance, and supervision, reflects positive parenting, and is not to be mistaken for authoritarian parenting. Avolio et al. (2009) showed that authoritative parenting was associated with lower levels of pre–high school children’s modest (delinquency and family/school offenses) and serious (serious crime, drug use) rule-breaking behavior. Second, modest and serious rule-breaking behaviors predicted subsequent leadership role occupancy differently: Specifically, modest rule-breaking behavior was positively, and serious rule-breaking behavior negatively, associated with leadership role occupancy. Avolio et al. (2009) suggested that early experiences with rule breaking (a) enable parents to guide their children to learn from these experiences and (b) reflect the same qualities necessary for leadership role occupancy. In contrast, just as early diagnoses of conduct

Copyright American Psychological Association. Not for further distribution.

Leadership

problem disorders predict subsequent counterproductive behaviors in organizations (Roberts, Harms, Caspi, & Moffitt, 2007), early contact with the law and drug use are unlikely to result in positive organizational outcomes. Recent research conducted within the context of sports teams provides additional information about the early development of leadership. Zacharatos, Barling, and Kelloway (2000) studied high school sports teams and assessed whether perceptions of parents’ behaviors influenced athletes’ leadership behaviors during sports events. This study was especially interested in the extent to which parents modeled transformational leadership in the home. The authors showed that adolescents’ perceptions of parents’ transformational parenting behaviors predicted their own enactment of these behaviors as rated by their coaches and peers. However, other findings (e.g., Hartman & Harris, 1992; Harris, 1995; Tucker, Turner, Barling, & McEvoy, 2008) suggested that research needs to focus on multiple social influences simultaneously. In summary, leadership behaviors may be learned during adolescence, the “impressionable years” (Krosnick & Alwin, 1989), or between the ages of 18 and 25, the period of emerging adulthood (Arnett, 2000), and they can be linked to subsequent leader behavior, making this a fertile issue for future research.

Executive and Leadership Development Having discussed early influences on leader development, we now turn our attention to the question of whether leadership can be taught. However, if we are to answer this question and be consistent with our evidence-based approach, there need to be rigorous experimental evaluations of the effects of leadership interventions using interpretable experimental or quasi-experimental designs (Cook & Campbell, 1979). Without methodological rigor, unstable conclusions are likely to be drawn. In an area in which experimental rigor is not the norm, our literature search uncovered four published studies that exemplify the importance of rigorous control, inasmuch as they used control group designs, conventional statistical testing, random assignment of participants to groups or treatments to groups, and/or

pretest/posttest measurement. As such, these four studies used interpretable designs (Cook & Campbell, 1979) and satisfied most of Terpstra’s (1981) criteria. This is important if valid inferences are to be drawn: An analysis of 52 articles on organizational interventions with varying levels of methodological sophistication showed that positive results were more likely with weaker methodological designs, whereas disconfirming evidence was more probable with methodologically rigorous designs (indicated by the use of census or representative samples, more than 30 participants, pretest—posttest control group designs with random assignment, and conventional significance testing; Terpstra, 1981; Bass, 1983)— a situation that potentially characterizes current leadership research (Hunter at al., 2007). In the first experimental study of the effects of transformational leadership training in a field setting, Barling, Weber, and Kelloway (1996) randomly assigned 9 branches of a regional bank to the experimental group and 11 to the wait-list control group. The managers of the nine branches then received 1 full day of transformational leadership training (Kelloway & Barling [2000] provide a more detailed description of the group-based training). A day after the training, leaders in the experimental group met individually with a coach, who provided the leaders with individual feedback based on a recently completed 360-degree exercise of transformational leadership. Based on this feedback, goals were set for making improvements to their transformational leadership behaviors. Individual meetings between the coach and each of the nine managers were held each month for the next 3 months, both to review performance and to boost the transformational leadership training. Barling et al.’s (1996) findings support the notion that transformational leadership can be taught: First, following the full training program, there were significant differences in transformational leadership ratings (as reported by subordinates) between the experimental and control groups. Second, levels of organizational commitment were significantly higher among subordinates who were led by leaders in the experimental training than among subordinates who were led by leaders in the control group, who had not received any training. Third, banks run by 195

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

managers who had received transformational leadership training sold significantly more personal loans and credit cards—two indices of special relevance to the banks—in comparison to the banks run by managers in the control group. The generalizability of this phenomenon is supported: Kelloway, Barling, and Helleur (2000) showed that transformational leadership could be taught to a sample of managers of a health care facility. Dvir, Eden, Avolio, and Shamir (2002) also focused on the development of transformational leadership in their study of infantry soldiers. Seven individuals were randomly assigned to the transformational leadership condition, which included 5 days of training, consisting of role playing exercises, simulations, video presentations, and group, peer, and trainer feedback. Akin to the booster sessions described in Barling et al.’s (1996) study, the leaders participated in a 3-hour session prior to a leadership assignment to reinforce the lessons of leadership training. Although the specific effects of transformational leadership in this context are discussed later in this chapter, it is important to note here that the training was effective: Both knowledge of transformational leadership theory and transformational leadership behaviors rated by subordinates were significantly enhanced by the training, whereas there were no such changes in the control group. Methodologically, it is also worth noting that although the Barling et al. (1996), Dvir et al. (2002), Kelloway et al. (2000), and Mullen and Kelloway (2009) studies focused on fewer than 30 leaders, all their evaluations were based on samples of more than 30 individuals (employees) per group, thereby fulfilling one of Terpstra’s (1981) criteria for a rigorous evaluation. Most recently, following the earlier studies showing that transformational leadership behaviors can be developed through training and that safety-specific transformational leadership predicted employee safety behaviors (Barling, Loughlin, & Kelloway, 2002), Mullen and Kelloway (2009) conducted an experiment and showed that safety-specific transformational leadership could be developed in leaders. Leaders’ attitudes toward safety were influenced by the training, as were employees’ safety behaviors. More intriguingly, Mullen and Kelloway also 196

included a group that received general transformational leadership training, and no significant effects on safety emerged for this group, raising the question of how specific leadership training needs to be in order to influence desired subordinate outcomes. It also remains to be seen whether safety training by itself would yield the same results as those from the safety-specific transformational leadership training; if not, then stronger support for the effectiveness of transformational leadership would result. Skarlicki and Latham (1996, 1997) directed their training toward meeting the specific demands that leaders in their sample faced, such as enhancing citizenship behavior within a union. In these studies, leaders (union shop stewards) in the experimental group were provided with four 3-hour sessions focused on training behaviors that influence follower perceptions of procedural and interactional justice. The results strongly supported the effectiveness of the leadership training. Perceptions of union fairness were higher among union members whose shop stewards had attended the training. Evidence for a “downstream effect” was also found in both studies: Changes in the shop stewards’ behaviors resulted in increases in rank-and-file members’ citizenship behaviors on behalf of the union. A possible opportunity lost in executive and leadership development deserves mention. Training in organizations has long been studied, with robust lessons learned from decades of research (Aguinis & Kraiger, 2009; Salas & Cannon-Bowers, 2001). To date, however, executive and leadership development initiatives have largely failed to benefit from the literature on training. One consequence of this is that, as is apparent from the prior discussion, the research on executive and leadership development interventions has not considered the types of issues considered central within training, such as needs assessment, different delivery modes, and the transfer of training. Similarly, any potential indirect benefits of leadership development (e.g., leader self-efficacy, commitment; Tannenbaum, Mathieu, Salas, & Cannon-Bowers, 1991) have also been ignored. Parenthetically, the training literature has perhaps not devoted sufficient attention to leadership development. It is likely that our understanding of executive and leadership development and the effectiveness of intervention initia-

Copyright American Psychological Association. Not for further distribution.

Leadership

tives would be enhanced in the future by a greater rapprochement between the two. To summarize what is known about leadership development, we not only considered the role of genetic factors and early family influences but also discussed several studies showing that leadership can be taught. It is also imperative that we pause to acknowledge that timing is important to leader development. Even when holding a leadership position, leaders need time to adopt, learn, and demonstrate leader behaviors (Day, Sin, & Chen, 2004), even after relatively brief leadership development initiatives (Barling et al., 1996; Mullen & Kelloway, 2009). How long outcomes take to emerge, and how long training effects last, are just two inquiries worthy of further attention. The dearth of studies on the development of leadership means that we have only scratched the surface of the question, Are leaders born or made? SOME CORRELATES OF LEADERSHIP Other topics in leadership research have simultaneously attracted empirical attention. Research conducted on leadership over the past several decades has tended to emphasize topics that were of substantial importance within the social sciences in general. As a result, much is known about the intersections of leadership and personality, gender, and ethnic and cross-cultural differences.

Personality and Leadership In the search for answers to the fundamental questions raised in thinking about leadership (e.g., why some people emerge as leaders, why some people are more effective than others in leadership roles), the role of personality looms large. In this section, we extend our earlier discussion of trait approaches to leadership by discussing three ways in which personality pertains to leadership. First, we consider personality and leader emergence and effectiveness; second, we explore the relationship between personality and transformational leadership behaviors; and third, we examine how follower personality influences perceptions and evaluations of leadership. (See also Vol. 2, chap. 5, this handbook.) Personality and general leadership. Two central questions are traditionally asked to identify “leader-

ship”—who emerges as a leader and who is an effective leader—and research has focused on whether and how personality influences both leadership emergence and effectiveness. For example, individuals are more likely to occupy leadership roles or be perceived as leaders when they are achievement-oriented and socially potent, meaning that they are hard-working and thrive when they are in charge of others (Arvey et al., 2006). Self-monitoring behaviors also predict emergent leadership, particularly because high selfmonitors are more likely to exhibit task-oriented behaviors in groups (Eby, Cader, & Noble, 2003). Furthermore, results from two meta-analyses reviewing the accumulated empirical studies link personality traits to leadership. First, Lord, de Vader, and Alliger (1986) showed that leader intelligence (r = .52), masculinity-femininity (r = .34), and dominance (r = .17) predicted perceptions of leadership (correlations corrected for unreliability of measures, sampling error, and range restriction). Second, researchers Judge, Bono, Ilies, and Gerhardt (2002) showed that the Big Five model of personality (i.e., extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience) explained 28% and 15% of the variance in leader emergence and leader effectiveness, respectively. These reviews suggest that personality trait perspectives of leadership emergence and effectiveness offer considerable credibility. In a refined analysis of the intersection between personality and leadership, Judge et al.’s (2002) quantitative review showed that of the Big Five personality traits, openness to experience and extraversion were significantly related (all correlations corrected for average reliability of measures and measurement error) to both leadership emergence (r = .24 and r = .33, respectively) and effectiveness (r = .24 and r = .24, respectively), while conscientiousness was significantly related to leadership emergence (r = .33) and neuroticism significantly related to leadership effectiveness (r = −.22). Openness to experience is defined by creativity and risk-taking, which the authors argue are relevant to both leader selection and performance, and extraverts are social, energetic, and dominant and therefore may more easily hold others’ attention and have social influence (Judge et al., 2002). Similarly, conscientious individuals characteristically demonstrate discipline in and diligence 197

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

toward their work (McCrae & John, 1992); it follows that conscientious personalities are likely to be selected for leadership roles. Likewise, the relative insecurity and emotional instability of individuals high on neuroticism can limit their leadership effectiveness. Despite knowledge about average tendencies, little is known about the processes through which leader personality traits influence organizational outcomes. Peterson, Smith, Martorana, and Owens (2003) offered one explanation, showing that CEO personality affects team dynamics at the senior-management level. For example, dynamics such as corrupt behaviors, risk-taking propensity, flexibility, and cohesiveness are affected by CEO personality and that each of these dynamics relates to organizational performance. Personality and transformational leadership behaviors. As the popularity of the transformational leadership framework persists (see Table 7.1), extensive research has been invested into understanding the connections between personality and transformational leadership behaviors. A meta-analysis analyzing 384 correlations from 26 samples (Bono & Judge, 2004) showed consistent relationships (correlations corrected for measure reliability and measurement error) between transformational leadership and extraversion (r = .22) and neuroticism (r = −.17). Individuals who were outgoing and optimistic (i.e., extraverted personality) tended to display more idealized influence and inspirational motivation behaviors. Conversely, those who were distressed, anxious, and prone to insecurities (i.e., neurotic personality) were unlikely to have the confidence needed to take on transformational roles. There were, however, inconsistent findings between agreeableness and openness to experience and the transformational leadership behaviors, and the authors also reported a relatively weak relationship between personality and transactional leadership, which was unexpected. The authors did not predict a link between conscientiousness and transformational leadership because, as they noted, “There is no particular reason to expect that conscientious individuals will exhibit vision, enthusiasm, or creativity” (Bono & Judge, 2004, p. 903). Overall, the associations established in Bono and Judge’s (2004) study were relatively modest– 198

specifically, the Big Five explain twice as much variance in leadership emergence as in leader charisma, and almost five times more in emergence than in intellectual stimulation and individualized consideration. The leader development section of this chapter highlights the malleability of leadership behaviors, and thus leadership training may partially explain this discrepancy (Bono & Judge, 2004); leadership behaviors may be less stable than personality characteristics. Consequently, personality traits can tell us more about who is likely to attain leadership positions than how individuals might lead once they must fulfill those roles. The Big Five are not the only characteristics associated with transformational behaviors. Proactive personality (Crant & Bateman, 2000), histrionic personality (Khoo & Burch, 2008), secure attachment style (Popper, Mayseless, & Castelnovo, 2000), and positive affectivity (Rubin, Munz, & Bommer, 2005) are also positive correlates of transformational leadership. Of particular interest in recent years is the relationship between narcissism (defined as immense self-love; Judge, LePine, & Rich, 2006) and transformational leadership. Judge et al.’s study confirmed that narcissistic personalities possess inflated selfevaluations as evidenced by their high self-ratings of transformational leadership; importantly, Judge et al.’s study also demonstrated that followers rated narcissistic leaders lower on transformational leadership than did the leaders themselves. These results highlight the role of perceptions of leadership—a topic of growing interest. Accordingly, our discussion now shifts to the topic of followers’ personalities and how those personalities might determine followers’ perceptions of leadership. Follower personality and leadership. Any discussion of leadership and personality would be incomplete without accounting for the personality of followers. A growing research interest suggests that follower personality is integral to the leadership process. For example, Bernerth, Armenakis, Feild, Giles, and Walker (2007) related perceptions of LMX to both follower and leader dimensions of the Big Five—leaders’ conscientiousness and agreeableness, and followers’ conscientiousness, extraversion, openness, and neuroticism were associated with follower

Copyright American Psychological Association. Not for further distribution.

Leadership

perceptions of LMX. Schyns and colleagues’ work has been particularly helpful to our understanding of follower personality and transformational leadership. These authors have shown that followers high on agreeableness (Schyns & Felfe, 2006), conscientiousness, neuroticism, honesty/humility (Schyns & Sanders, 2007), and particularly, extraversion (Felfe & Schyns, 2006) tend to perceive leaders as more transformational. Furthermore, follower personality has been shown to predict leader preferences: Ehrhart and Klein (2001) showed that followers with higher self-esteem and achievement orientation preferred charismatic leaders to noncharismatic leaders. All of these studies suggest that follower personality influences perceptions of leadership; what remains to be seen is whether or not follower personality directly affects leadership behavior, which would occur if leaders adjust their behaviors to better suit the personalities of their followers. By contrast, leadership style may predict personality. Hofmann and Jones (2005) showed that transformational leadership predicted collective personality; collective-level agreeableness, conscientiousness, extraversion, and openness to experience were all positively related to transformational leadership and negatively related to passive leadership across 68 collectives and 448 employees. One cannot rule out the possibility, however, that transformational leaders may select followers with distinct personality traits. Our knowledge about the relationship between personality and leadership is enhanced with refined analyses, such as studies showing that personality is more strongly linked with leader emergence than behavior and that some personality dimensions (e.g., conscientiousness) might not be expected to predict certain leadership behaviors (e.g., charisma; Bono & Judge, 2004). Our knowledge also stands to be enriched by answering questions about how followers’ personalities in tandem with leaders’ personalities influence the leadership process. Exploring these issues will be of importance to future research.

Gender and Leadership As perhaps the most apparent individual difference, gender has captured the attention of scholars in

numerous domains, and leadership is no exception. The topic of gender inequalities in prominent leadership roles has been actively debated in the public domain and thoroughly explored by researchers. We need look no further than the 2008 Democratic Convention, which highlighted the role of gender in leadership as Senator Hillary Clinton aimed for the nomination to be the first female U.S. president. The media raised numerous gender issues that apply to organizational research, such as masculinity in leadership (e.g., Romano, 2007) and the interaction between follower and leader gender, in frequently questioning whether women would be more supportive of a female president than would men (Sullivan, 2008). These and other issues have fueled much research in this area. Many gender-based discussions in leadership revolve around the question of whether men and women are, or can be, equally effective leaders. Certainly, no shortage of research exists asking this question. In a quantitative review of the literature, Eagly, Karau, and Makhijani (1995) found that overall leadership effectiveness is not dependent on leader gender, but men and women perform differentially better or worse under certain conditions. Male leaders were more effective (effect sizes corrected for unreliability) when their role was regarded as more “masculine” (β = .19),1 when the majority of their subordinates were male (β = .22), and in military settings (d = .42). In contrast, women performed better as leaders in roles that were defined as more “feminine” (e.g., required interpersonal ability; β = .20). On aggregate, men were only marginally favored as leaders in terms of leadership evaluations (e.g., in 56% of the studies comparing men to women, men were rated more favorably than women, a percentage not significantly different from the hypothesized 50%), yet when women used stereotypically masculine leadership styles (e.g., autocratic leadership; d = .30), were in male-dominated positions (d = .09), or were evaluated by men (d = .15), this imbalance was much greater and in favor of men (all effect sizes corrected for unreliability; Eagly et al., 1995). Past research has also explored the predictive value of gender on differential leadership behaviors.

1

The effect size statistics reported in this section on gender are positive when in favor of males and negative when in favor of females.

199

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

Meta-analytic results have suggested that women have a tendency to demonstrate more democratic and less autocratic leadership behaviors (d = .27 corrected for unreliability; Eagly & Johnson, 1990) and were perceived by followers as slightly more transformational (d = −.10) and less transactional (d = .12 for management by exception active, d = .27 for management by exception passive; Eagly, Johannesen-Schmidt, & van Engen, 2003) than their male counterparts. Yet, when comparing senior-level leaders, men reported higher levels of emotional expression (which pertains to charismatic leadership [Conger, 1989]) than women. Thus, although some evidence suggests that men and women are perceived differently on various leadership styles, extending this line of research will be necessary if any strong conclusions are to be reached. Despite the seemingly similar performance of female and male leaders, men remain far more likely to hold leadership positions in organizations, whereas women are stereotyped to have a lower aptitude for leadership (see Ryan & Haslam, 2007, for a review). In leaderless groups, men emerge as the leader more often than do women (e.g., for direct task leadership, d = .37; Eagly & Karau, 1991). What is also notable is that women may emerge as “social leaders” more so than men (e.g., for direct social leadership, d = −.12; Eagly & Karau, 1991), and that gender role identity may be a stronger predictor of emergent leadership than biological sex (Kent & Moss, 1994). Although some evidence suggests that the stereotype of women being less effective leaders is beginning to change (Duehr & Bono, 2006), gender inequality in management roles remains prevalent. Numerous explanations for this disparity have been offered. Women may hit “glass ceilings” that prevent their rise to high-level management positions, whereas men ride “glass escalators,” quickly advancing through the organizational hierarchy (e.g., Maume, 1999). More recently, Ryan and Haslam (2007) proposed that as women begin to occupy more prominent managerial roles, they are now more likely to fall over the “glass cliff.” From this perspective, women are most often selected for high-level leadership positions when those positions are associated with greater risk, thereby setting women up to be unsuccessful. To illustrate, initial empirical evidence shows that women are preferred leadership 200

candidates for projects that have failed in the past or are projected to fail in the future (see Ryan & Haslam, 2007, for a review). Unfortunately, inequality begets further inequality because such failure is easily used to legitimize the women-as-less-effective-leaders stereotype. It is also possible that gender biases in organizations are learned or modeled. Phillips (2005) showed that in a sample of Silicon Valley law firms, if the firm founders had parent firms with few women in prominent leadership positions, then the firm founders were less likely to place women in leadership roles in the new firm, and vice versa. Another reason for the scarcity of women in the upper echelons of the organizational hierarchy may be women’s own motivation to fulfill those roles. Eagly, Karau, Miner, and Johnson’s (1994) metaanalysis of 51 studies concluded that although the difference was slight, men were more motivated to obtain management roles (d = .22 corrected for unreliability). Stereotype threat may be one explanation for this finding. Stereotype threat occurs when knowledge of a salient negative stereotype (e.g., women are not very good leaders) causes the stigmatized individual to fear confirming the stereotype, creating anxiety and lowering expectations and/or performance (Steele & Aronson, 1995). Indeed, Davies, Spencer, and Steele (2005) showed that stereotype threat limited women’s aspirations to become leaders. In two experiments, women were exposed to gender stereotypical television commercials (subtly displaying traditional female stereotypes); in both studies, those women were unwilling to take on leadership roles during a task. However, removing the stereotype threat by telling the women that there were no gender differences associated with the task restored their motivation. Field evidence also suggests that women are less likely to draw on these stereotypical beliefs about women’s leadership abilities when they are more frequently exposed to women in counterstereotypical leadership positions (Dasgupta & Asgari, 2004). In light of the growing number of women who are obtaining prominent leadership roles in organizations (Ryan & Haslam, 2007), it is possible that downward biases on women’s interest in leadership positions and potential performance within leadership positions, may be minimized in the future.

Copyright American Psychological Association. Not for further distribution.

Leadership

Topics in gender and leadership will no doubt continue to attract the interest of management, organizational researchers, and society at large. Certainly, although gender differences in leadership emergence and behaviors have been identified, no consensus has been reached as to why gender inequalities in leadership prevail or how they should be addressed. However, the changing demographics of the workforce suggest that these questions will be of increasing importance to organizations and to organizational scholars. To extend our knowledge, and in light of Kent and Moss’s (1994) finding (noted earlier), comparing the predictive value of gender role identity to biological sex on leader emergence is one promising avenue for future inquiry.

Cross-Cultural Leadership and Ethnicity With broad-based pressures for social equality and the rush to globalization of the past two decades, it comes as no surprise that researchers have turned their attention to how leadership and culture, nationality, and ethnicity intersect; among these, culture has received the most attention. As defined by Hofstede (2001), culture defines members of a nation, region, or group in a way that determines members’ core values. In his classic research, Hofstede distinguished national cultures along the dimensions of power distance, uncertainty avoidance, individualism/collectivism, masculinity/ femininity, and future orientation. Many organizational psychologists adopted Hofstede’s framework to understand intercultural relations in the workplace, and in this section we review some of that research as it pertains to leadership. To begin, we describe the Global Leadership and Organizational Behavior Effectiveness (GLOBE) project that has spearheaded a great deal of crosscultural research in the field. The GLOBE project involves researchers from all over the world and diverse cultures, all of whom seek to understand the interplay between leadership and culture (House, Javidan, Hanges, & Dorfman, 2002). Specifically, for this 10-year research program, quantitative data were collected from 17,000 managers in 951 organizations across 62 different societies (House, 2004), a mammoth undertaking. As testaments to the project’s effectiveness, many research studies have used

GLOBE data, and their varied topics demonstrate the broad spectrum of research questions spawned from the intersection of leadership and culture. For example, Dickson, Resick, and Hanges (2006) used these data to investigate different organizational climates and organizationally shared leadership prototypes, and Den Hartog and colleagues (1999) used GLOBE data to study cultural-specific implicit leadership theories. Considerable attention has been given to the cross-cultural validity of transformational leadership. This attention should perhaps not be surprising after Bass (1997) wrote “in whatever the country, when people think about leadership, their prototypes and ideals are transformational” (p. 135). In support of this view, Singer and Singer’s (1990) studies of police officers in New Zealand and Taiwanese employees both paralleled results from American samples: Transformational leadership was the preferred leadership style in comparison with transactional leadership. Furthermore, in that same study, the New Zealand police officers displayed more transformational behaviors than transactional behaviors, and the Taiwanese sample reported that transformational leadership behaviors were displayed in their organizations. Together with studies discussed throughout this chapter demonstrating the utility of transformational leadership in other countries (e.g., Canada, Germany, Israel, Singapore, Tanzania, Turkey), credibility emerges for the notion that transformational leadership applies to contexts outside of North America. As mentioned, Den Hartog and colleagues (1999) tested whether there were certain leader characteristics that were universally endorsed. They found that trustworthiness, fairness, honesty, and being encouraging and positive were some of the universally endorsed characteristics of leaders, and they further rationalized that these universally endorsed characteristics reflected transformational/charismatic leadership. However, the application of universal negative leader attributes (e.g., being noncooperative and nonexplicit) to transformational leadership was not discussed, and yet there may be valuable connections. For example, as a consequence of encouraging followers to think creatively, leaders may appear nonexplicit because they refrain from imposing their own 201

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

views on others. Den Hartog and colleagues are not alone in supporting the universal appeal of transformational leadership; the numerous MLQ translations (e.g., Felfe, 2006; Shao & Webber, 2006) are another indicator of the cross-national popularity and utility of transformational leadership. Nonetheless, conceptually and empirically, there are reasons to question the universality of transformational leadership theory. Specifically with regard to its applicability in China, Shao and Webber (2006) noted that leaders, “even those who are open to experience[,] will not exhibit intellectual stimulation behavior, [because it] is incompatible with high uncertainty-avoidance Chinese culture” (p. 942). Similarly, Fukushige and Spicer (2007) argued that leadership preferences in Japan are distinct from preferences documented in the majority of the leadership literature based on Western standards. In their study, Japanese workers preferred contingent-reward leadership over idealized influence and inspirational motivation. The authors argued that inspirational motivation elicited skepticism in Japanese followers, and idealized influence demonstrated immodesty. Insights such as these illuminate cultural influences on leadership perceptions; what is seen as optimism and confidence in one context might be interpreted as insincerity and immodesty in another. Not only have there been questions about the cultural boundaries of leader preferences, but also cultural influences on typical leadership behaviors have spurned research. As Gerstner and Day (1994) argued, “The most characteristic traits of a leader in one culture may be very different from prototypical traits in another culture” (p. 123). Their results showed that typical traits of business leaders varied across countries. For example, in India, the top three typical traits of business leaders reported were industriousness, competitiveness, and determination. In contrast, in France, “determined,” “open-minded,” and “informed” described the top three typical traits of business leaders. Furthermore, Rosette, Leonardelli, and Phillips (2008) concluded that leader race is critical to that leader’s prototypicality in the United States. More specifically, being White was significantly associated with a leader role, and the authors argued that one reason for this is the frequent exposure to White leaders in North America. 202

Collectively, these results have obvious implications for leaders working in a national culture other than their own, as leaders need to be constantly mindful of local employees’ implicit expectations of their leaders. These results also pertain to dynamics within a national culture, because people from different cultural backgrounds, such as ethnicities, interact and therefore may hold different expectations of each other. Supporting the notion that within-nation cultural differences also influence leadership and leadership perceptions, Ah Chong and Thomas (1997) examined two ethnic groups in New Zealand, namely the Pakeha (New Zealanders who are predominantly of European descent) and Pacific Islanders, and their findings showed the complexity of cultural differences in leadership. They found support for ethnic differences in leadership perceptions and also a significant interaction between leader ethnicity and follower ethnicity: Pacific Islanders reported more satisfaction with Pacific Islander leaders than with Pakeha leaders. These studies appear to question Bass’s (1997) statement that people from all countries hold prototypes of leadership that are transformational. Gertsner and Day (1994) showed that different cultures valued different leader traits, and Ah Chong and Thomas (1997) demonstrated that, even within the same national culture, preferences for leadership varied as a function of leader and follower ethnicity. Such findings raise questions about the universality or cross-national validity of transformational leadership. Nonetheless, one explanation consistent with Bass’s (1997) assertion regarding the universality of transformational leadership is that, despite these inter- and intracultural differences, ultimately the characteristics preferred are associated with transformational leadership. For example, open-mindedness (favored in France) is an expression of intellectual stimulation, and determination (favored in India) contributes to inspirational motivation. Moreover, Bass had suggested that transformational leadership behaviors may be manifested differently across cultures; for example, in Indonesia, boastfulness contributes to inspirational motivation, whereas boastfulness is eschewed in Japan, but this does

Copyright American Psychological Association. Not for further distribution.

Leadership

not mean the Japanese do not display inspirational motivation in other ways. Therefore, with regard to Ah Chong and Thomas’s (1997) study, Pacific Islanders may have preferred leaders of the same ethnicity because they were more attuned to those leaders’ displays of transformational leadership than to those of Pakeha leaders. On the other hand, accepting such a flexible conceptualization of transformational leadership would mean that a fair test of its validity would be difficult. Clearly, more research is needed that seeks to understand the differential expression of leadership behaviors and the outcomes of such behaviors, within and across cultures. At present, research tends to focus more on demonstrating the cross-cultural equivalence of transformational leadership. One problem inherent in all cross-cultural or cross-national research is ensuring measurement equivalence; measures of leader behavior are culturespecific because, as mentioned previously, behaviors (e.g., boastfulness) can take on different meanings in different cultures. Indeed, Ayman and Chemers (1983) suggested that Euro-American measures of leader behavior are not adequate to describe leader behavior for Iranians. For example, they added new items to the LBDQ such as “is like a kind father” to reflect the paternalistic society of Iran, and this item contributed to a new factor called Benevolent Paternalism. Therefore, Ayman and Chemers concluded that “the use of ‘universal’ measures and constructs [is] likely to lead to uninterpretable research findings” (p. 341). This is a potentially limiting issue because, in many cross-cultural or cross-national studies, research instruments measuring leadership are based on items developed in and for North American cultures, hence the numerous translations of the MLQ noted earlier. In addition, measurement equivalence is necessary for research focused on group differences, such as differences across nations and ethnicities; “if one set of measures means one thing to one group and something different to another group, a group mean comparison may be tantamount to comparing apples and spark plugs” (Vandenberg & Lance, 2000, p. 9). Certainly, future research on leadership across cultures needs to be more sensitive to this issue.

A second problem concerns the ability to generalize from cross-cultural studies over time. Thus, although Hofstede’s (2001) work has been fruitful in guiding many cross-cultural studies of leadership, the characteristics inherent in the national cultures studied may change over time as a result, for example, of economic recessions and globalization (e.g., Ralston, Holt, Terpstra, & Kai-Cheng, 1997). Indeed, Fukushige and Spicer’s (2007) qualitative research suggested that Japanese work culture was in the process of becoming a meritocracy. To account for these and other changes, contemporary research should take the time to reevaluate Hofstede’s dimensions and ensure that the cultures studied are accurately depicted in their present-day state. Finally, a third challenge to cross-culture leadership studies is complying with the various national ethical standards for research. Aguinis and Henle (2002) reviewed some of the studies which described the differing ethical standards across countries and, in particular, drew attention to Leach and Harbin’s (1997) research. Although there are some universal ethical standards (e.g., privacy, avoiding harm, remuneration for participations), there are also noticeable differences, such as China’s ethics codes being most divergent from the American Psychological Association’s (APA) guidelines, as well as the United States’ standards for sharing and duplicating data. OUTCOMES OF LEADERSHIP Jack Welch is a figurehead in contemporary business leadership (Amernic, Craig, & Tourish, 2007), and this is largely due to performance outcomes attributed to him. Byrne (1998) suggests that if leadership is an art, then surely Welch has proved himself a master painter. Few have personified corporate leadership more dramatically. Fewer still have so consistently delivered on the results of that leadership. For 17 years, while big companies and their chieftains tumbled like dominoes in an unforgiving global economy, Welch has led GE to one revenue and earnings record after another. (p. 90) 203

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

Undoubtedly, the long-standing interest in leadership in organizations derives from the widespread belief that leaders like Jack Welch have the potential to affect important organizational outcomes. Numerous studies have demonstrated the beneficial consequences of positive leadership, and transformational leadership in particular, on follower attitudes and states and performance in organizations. More recently, there has also been more interest into leadership’s effect on follower health, well-being, and safety. In this section of the chapter, we review these leadership outcomes, explore the mechanisms through which leadership exerts its effects, and discuss the conditions that moderate these effects (see Figure 7.1 for an overview). First, however, several observations flowing from the tendency to attribute so much success to one leader (e.g., Jack Welch) require elaboration. First, the widespread myths linking single leaders like Welch to the fate of organizational behemoths like GE, with its operations in Prelaunch Phase IndividualLevel Variables

GroupLevel Variables

62 countries, derive not from sound empirical research but rather from impressionistic, subjective accounts. Second, such attributions ignore the fact that the success (or failure) of any one organization will have multiple determinants. The tendency to attribute success to one leader is consistent with the “romance of leadership” phenomenon (Meindl, 1998; Meindl, Ehrlich, & Dukerich, 1985), which will be discussed in more detail later in this chapter.

Direct Outcomes Numerous quantitative reviews have now established that transformational leadership relates to various metrics of organizational effectiveness, such as satisfaction with the leader, job satisfaction, motivation, follower perceptions of effective leadership, leader performance, and group-organizational performance. Judge and Piccolo (2004) showed that across 87 studies, transformational leadership (r ≈ .44) was positively related to these leadership effectiveness Postlaunch Phase

Launch Phase SocietalLevel Variables

Sample Activities Identification of opportunities Initial opportunity evaluation Assembly of required resources Gathering pertinent information

IndividualLevel variables

GroupLevel Variables

SocietalLevel Variables

Sample Activities Choosing legal form of new venture Obtaining intellectual property protection Developing initial business model and strategies

Dependent Measures

Dependent Measures

Number and quality of opportunities identified Capital raised Success in attracting highquality partners, employees

Time until first sale Time until “break-even” Time until first employee hired Number, strength of patents acquired

IndividualLevel Variables

GroupLevel Variables

Sample Activities Building customer base Hiring key employees Improving product design Conducting negotiations Influencing, motivating others

Dependent Measures Financial Measures (growth in sales, earnings, number of employees; value of initial public offering) Success in Obtaining Required Resources Attitudinal Measures (e.g., personal and life satisfaction) Measures of Entrepreneurs’ Personal Health and WellBeing

FIGURE 7.1. Outcomes, moderators, and mediators of leadership. LMX = leader–member exchange; OCBs = organizational citizenship behaviors. 204

SocietalLevel Variables

Copyright American Psychological Association. Not for further distribution.

Leadership

criteria, as was contingent reward leadership (r ≈ .39). These results are in sharp contrast to the validity found for laissez-faire leadership (r ≈ −.37), and the less consistent correlations between managementby-exception and organizational outcomes (all correlations corrected for unreliability of measures, measurement error, and sampling error). Judge and Piccolo’s pattern of findings coincides with earlier meta-analyses (e.g., Lowe, Kroeck, & Sivasubramaniam, 1996). However, Judge and Piccolo go beyond past research by showing that transformational leadership relates to effectiveness measures even after controlling for the effects of transactional and laissez-faire leadership behaviors. Generally, studies have considered leadership effectiveness at the level of the follower (including follower attitudes and performance), the group, and the organization. Beyond those examined by Judge and Piccolo (2004), a number of additional follower attitudes show relationships with transformational leadership, including trust in the leader (for a review see Burke, Sims, Lazzara, & Salas, 2007), commitment to the organization (e.g., Barling, Weber, & Kelloway, 1996), responsiveness to change initiatives (Herold, Fedor, Caldwell, & Liu, 2008), turnover intentions (e.g., Bycio et al., 1995), psychological safety (Detert & Burris, 2007), cynicism (Bommer, Rich, & Rubin, 2005), and identification with the leader, group, and organization (e.g., Epitropaki & Martin, 2005a; Kark, Shamir, & Chen, 2003). There is also support for the influence of transformational leadership and charisma on follower performance. For instance, in an experiment using a trained actor to elicit charismatic leader behaviors, Kirkpatrick and Locke (1996) found that charisma positively affected follower task performance. Consistent with the “falling dominoes” analogy (e.g., Avolio, 1999b), Dvir et al.’s (2002) findings also suggest that leadership can have indirect performance effects, such that the transformational leadership of higher-level leaders positively influenced the performance of followers who were not their direct reports. Moreover, Bono and Anderson (2005) showed that transformational leaders and their followers played more central roles in advice and influence networks.

Transformational leadership exerts influence on more discretionary forms of follower performance as well. Organizational citizenship behaviors (i.e., positive organizational behaviors that go beyond the formal requirements of one’s job) are more common among followers of a transformational leader (e.g., Piccolo & Colquitt, 2006). By enhancing intrinsic motivation, sparking intellectual stimulation, and energizing followers, transformational leadership can also foster follower creativity and ingenuity (e.g., Shin & Zhou, 2003). Finally, transformational leadership has been related to the development of followers in terms of follower morality, motivation, and empowerment (Dvir et al., 2002). As described earlier, Dvir et al.’s study of leader development indicated that leaders who received transformational leadership training had a stronger influence on follower development, particularly on follower selfefficacy and collective orientation. Evidence for the relationship between leadership and performance extends beyond the individual follower to group-level outcomes, as was the case in the Barling et al. (1996) study that found a significant relationship between transformational leadership training and sales performance in bank branches. Lim and Ployhart (2004) focused specifically on transformational leadership and performance in a team setting and also found a significant positive effect. Moreover, Bass, Avolio, Jung, and Berson (2003) found a strong relationship between transactional leadership and unit performance in army platoons, and a positive relationship emerged between empowering leader behaviors and team performance in Srivastava, Bartol, and Locke’s (2006) study of management teams. Less understood is the relationship between CEO charisma and transformational leadership and firm-level performance, possibly because obtaining leadership ratings of CEOs and top-level managers is difficult. Of the studies that investigated the effects of CEO leadership on firm financial performance, findings were inconclusive (see Agle, Nagarajan, Sonnenfeld, & Srinivasan, 2006, for a summary). Agle et al.’s recent study found little support for the relationship between CEO charisma and firm financial performance, even in uncertain market conditions. Instead, their findings are more consistent 205

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

with a “romance of leadership” approach; CEOs included in the study were rated as more charismatic when prior firm performance was stronger, but CEO charisma did not predict firm performance prospectively. With mixed support, Waldman, Ramírez, House, and Puranam (2001) reported an important moderating effect: CEO charisma predicted firm financial performance under conditions of market uncertainty, when CEO discretion is heightened, thereby increasing the potential for CEO leadership to have an effect. Ling, Simsek, Lubatkin and Veiga’s (2008) findings suggest that focusing on the initial effects of CEOs’ transformational leadership on top management teams may be important in linking CEO leadership and financial outcomes. To summarize, ample evidence suggests that leadership relates to both follower perceptions of leader effectiveness and to individual-, group-, and in some cases, firm-level performance criteria. We now shift our attention away from outcomes that evaluate leadership effectiveness based on its relation to performance and toward a focus on employee safety and well-being as metrics of effectual leadership. Leaders have the opportunity to embrace and improve the well-being of their followers by promoting safe working practices. For example, high-quality leader–follower relationships have been related to safety communication and commitment, both of which improve safety behaviors (Hofmann & Morgeson, 1999). In another example, Barling et al. (2002) developed a safety-specific measure of transformational leadership, and they found that employees of safety-specific transformational leaders were more conscious of safety concerns and perceived a stronger climate of safety in the organization. In turn, higher safety consciousness and climate were associated with fewer safety-related incidents and occupational injuries. This model was later extended to contrast transformational leadership with passive leadership styles (Kelloway, Mullen, & Francis, 2006). Results confirmed that transformational leaders can effectively communicate safety beliefs and raise safety consciousness and climate, yet safetyspecific passive leadership can be detrimental in that it is negatively related to safety awareness. 206

Outside of a safety context, there is a growing interest in the relationship between leadership and followers’ psychological well-being. Effective leadership may reduce workplace stressors, while enhancing followers’ moods and experiences. Epitropaki and Martin (2005a) found a direct link between high-quality leader–follower relationships and follower well-being, and van Dierendonck, Haynes, Borrill, and Stride (2004) related leadership to followers’ work-related and general well-being across four time periods. Support for the transformational and charismatic leadership framework has also associated this leadership style with follower well-being, in part through mood contagion. Leader charisma is associated with more positive and less negative affect in followers (e.g., Cherulnik, Donley, Wiewel, & Miller, 2001; Erez, Misangyi, Johnson, LePine, & Halverson, 2008). Transformational leaders also contribute to followers’ psychological well-being by helping them find meaning in their work (Arnold, Turner, Barling, Kelloway, & McKee, 2007). These findings suggest that meaningful work stimulates intrinsic reasons for working and, as a result, promotes resiliency in the face of challenges, as well as higher quality of life. Even in its infancy, this area of leadership studies presents compelling evidence for the effect of leadership not only on followers’ attitudes and behaviors but also on their psychological health. Before shifting our attention toward mediating effects, it is important to note that these outcomes of leadership are not necessarily static. For example, Shao and Webber (2006) argued that “in the long run, intellectual stimulation may produce desirable effects. Yet, in the short run, leaders who continually urge followers to search for new and better methods of doing things may create ambiguity, conflict, or other forms of stress in the minds of followers” (p. 937). It would be imprudent, then, to assume that either positive or negative effects will surface immediately or be enduring.

Mediating Effects: Explaining How Leadership Affects Outcomes As reviewed, much is known about the outcomes of leadership, but less is known about how and why these effects occur. In response, an emergent body of

Copyright American Psychological Association. Not for further distribution.

Leadership

literature explores the various means through which leadership exerts its influence. In general, three categories of mediators have been proposed: those relating to follower perceptions of the leader, of themselves, and of the job. Demonstrative of the first category of mediators, we know that there is a relationship between transformational leadership and followers’ positive perceptions of their leaders. Arguably, these followers feel a greater sense of commitment to the leader as a result of those positive perceptions and therefore demonstrate a willingness to exert effort that ultimately benefits an organization. Wang, Law, Hackett, Wang, and Chen (2005) argued further that followers are more likely to reciprocate transformational behaviors as a mechanism of social exchange. They sampled 162 leader–follower dyads and found that followers’ perceptions of high-quality LMX (e.g., mutual respect and consideration) mediated the relationship between transformational leadership and follower performance. Supporting the notion that followers’ perceptions of their leaders’ attributes mediate leader effectiveness are studies that show a relationship between perceived trust in the leader (e.g., Dirks, 2000; Jung & Avolio, 1999; Pillai, Schriesheim, & Williams, 1999), leader fairness (for a review, see van Knippenberg, De Cremer, & van Knippenberg, 2007), and identification with the leader (e.g., Kark et al., 2003) and desired follower outcomes. For example, transformational leaders empower their followers, invite follower contributions to decisionmaking, are respectful of followers, and are willing to forgo their own interests for the good of the group. As a result, followers of transformational leaders are more likely to trust that their leader will act in good faith and will treat them fairly. Pillai et al. illustrated these relationships in a path model, showing that followers of transformational leaders perceived greater procedural justice, which predicted higher trust in the leader and subsequent organizational citizenship behaviors on the part of the follower. In contrast, followers who do not trust their leader are unlikely to accept the leader’s vision or commit themselves on the leader’s behalf (Jung & Avolio, 2000). Farmer and Aguinis (2005) echoed these ideas in a conceptual model relating

follower perceptions of leader power to various follower outcomes. Followers may also be more likely to resist a leader’s vision when they perceive the leader’s values to be inconsistent with their own. Jung and Avolio (2000) suggested that through communication and inspirational motivation, transformational leaders transmit their collective-focused values to followers; thus, followers of transformational leaders will be more likely to internalize the leader’s mission, aligning their own values with those of the leader. These authors show that transformational leadership is positively related to value congruence between leaders and followers and that value congruence mediates the relationship between transformational leadership and follower performance. Likewise, Kark et al. (2003) found that transformational leadership was positively related to identification with the leader and with the group overall. However, the nature of the identification with the leader was critical: Personalized identification with the leader mediated the relationship between transformational leadership and follower dependence on the leader, whereas socialized identification with the work group mediated the relationship between transformational leadership and follower empowerment. Thus, further research is needed to fully understand the conditions under which transformational leadership is related to follower development and when followers may become overly reliant on the leader for guidance or motivation. The second category of mediators prominent in the leadership literature considers how leaders alter follower or group internal states. Of particular interest has been empowerment and efficacy at the individual and group levels. To illustrate, transformational leaders show intellectual stimulation, encouraging followers to voice their concerns and to “speak up,” by creating an environment of psychological safety (Detert & Burris, 2007). In this way, followers are empowered by transformational leaders who stimulate their interest and involve them in organizational visions. Consequently, psychological empowerment plays a role in mediating the relationship between transformational leadership and organizational commitment (Avolio, Zhu, Koh, & Bhatia, 2004) as well as LMX and follower performance 207

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

(Chen, Kirkman, Kanfer, Allen, & Rosen, 2007). Similar findings have been reported at the team level: Team empowerment mediates the effects of leadership climate on team performance (Chen et al., 2007). Another example of this second category of mediators is intrinsic motivation. Intrinsic motivation mediates the relationship between transformational leadership and sports performance (e.g., Charbonneau, Barling, & Kelloway, 2001). Moreover, empowering leadership behaviors also enhance intrinsic motivation and have been linked to knowledge sharing within teams (Srivastava, Bartol, & Locke, 2006). Srivastava et al. supported the presence of a second mediator in their study: team efficacy. Indeed, efficacy and empowerment are intimately intertwined (e.g., Chen et al., 2007; Spreitzer, 1995). Giving followers opportunities to share their ideas and voice their opinions is just one way in which leaders can influence followers’ confidence in their own abilities, and this may be the route through which empowering leadership garners such positive outcomes (e.g., Srivastava et al., 2006.). The relationship between follower and group efficacy and transformational leadership is stimulated through transformational leaders’ individualized mentorship and coaching, engagement of followers in the task and organizational goals, and expressions of confidence in followers’ abilities to achieve group goals (Schaubroeck, Lam, & Cha, 2007). Accordingly, group efficacy has been shown to mediate the relationship between transformational leadership and team performance (Bass et al., 2003; Schaubroeck et al., 2007.). The final category of mediating mechanisms used to explain the positive effects of leadership relate to the way in which followers perceive their jobs. As suggested previously, transformational leaders can influence the meaningfulness that followers find in their jobs (Arnold et al., 2007). Piccolo and Colquitt (2007) called this phenomenon the “management of meaning” and showed that followers of transformational leaders find more meaning in their work as measured by core job characteristics: task significance, autonomy, task variety, task identity, and feedback. In their study, more meaningful work stimulated intrinsic sources of motivation and goal 208

commitment, which were directly related to organizational citizenship behaviors and task performance. Similar empirical evidence suggests that followers’ self-concordance, or “the extent to which activities such as job-related tasks or goals express individuals’ authentic interests or values,” partially mediates the relationship between transformational leadership and follower attitudes (Bono & Judge, 2003, p. 556). Future research would benefit from further explaining the ways through which leadership can influence follower perceptions of meaningful work and, correspondingly, their attitudes, performance, and well-being.

Moderating Effects: Isolating the Boundary Conditions of Leadership Effectiveness One question that is frequently asked is whether leadership matters; Nye (2008) refined this popular question, asking not just whether leadership matters, but when it matters—a question of theoretical and practical importance. Research has identified factors relating to the follower, group, leader– follower relationship, and context that moderate the relationship between leadership and organizational outcomes. Next we consider these influences in greater depth and point to areas that warrant future research. At the follower level, leadership effectiveness may be conditional on followers’ levels of positive and negative affect. Epitropaki and Martin (2005a) showed that the positive effects of transformational leadership on follower identification with the organization were dependent on the follower’s positive and negative affect. Followers with high positive affect and low negative affect reported more favorable impressions of their work, and, as a result of such a positive outlook, these followers were more likely to identify with their organization. The reverse was true for individuals with low positive affect and high negative affect. However, in this latter circumstance, transformational leaders had a relatively greater impact on followers’ identification, perhaps through the “management of meaning.” These results suggest that the followers most likely to benefit from transformational leadership are those who need leadership the most.

Copyright American Psychological Association. Not for further distribution.

Leadership

This interpretation is consistent with the findings of Whittington, Goodwin, and Murray (2004). In their study, transformational leadership had a stronger effect on follower affective commitment and performance when followers faced difficult or challenging goals and thus needed guidance. Conversely, transformational leadership exerted less influence on followers when job enrichment was high (Whittington et al., 2004.). Similar moderated effects were also apparent at the team level where inclusion of a team leader was more relevant to information gathering and performance for teams with an external locus of control than for those with an internal locus of control (Boone, van Olffen, & Van Witteloostuijn, 2005). Leadership is more instrumental to external teams who lack team potency and require greater direction and motivation to be successful. Leadership effectiveness is contingent on aspects of the leader–follower relationship. Piccolo and Colquitt (2006) proposed that followers are more open to and accepting of transformational leadership under conditions of higher quality LMX, and they found that transformational leadership related more strongly to follower performance and citizenship behaviors under these circumstances. Positive leadership behaviors may also be more successful when leaders and followers share a sense of common identity or in situations that highlight their shared identity (Ellemers, de Gilder, & Haslam, 2004). Despite these propositions, conceptualizations of high-quality leader–follower relationships as both mediators (as described in the previous section) and moderators of the leadership process remain to be reconciled. Research on characteristics of leader–follower relationships as moderators of leader effectiveness has also considered proximal relationships between leaders and followers, but the findings warrant cautious interpretation. For example, Howell and Hall-Merenda (1999) showed that transformational leadership predicted follower performance over a 1-year period for followers within close physical proximity of the leader but did not have an effect when followers were physically distant. To the contrary, others have argued that structural distance between leaders and followers (i.e., when

leaders have an indirect relationship with followers through a middle leader) may enhance the relationship between transformational leadership and both performance and commitment (e.g., Avolio et al., 2004; Dvir et al., 2002). To add to the complexity, structural and physical distances are typically highly correlated (Avolio et al., 2004). Perhaps acknowledging that leader hierarchical “level” differentially predicts effectiveness can shed light on this contradiction. Quantitative reviews of the transformational and charismatic leadership literatures have suggested that transformational leaders higher in the organizational hierarchy are somewhat more effective than those at lower levels (Fuller, Patterson, Hester, & Stringer, 1996; Judge & Piccolo, 2004). Thus, future research will be necessary to disentangle the relative influences of physical distance, structural distance, and hierarchical level on leader effectiveness. As mentioned, there are suggestions that culture can direct the effectiveness of leadership (Shamir & Howell, 1999). For instance, experimental evidence suggests that transactional leaders, who focus on self-directed, short-term goals, elicited stronger performance from individualists (Jung & Avolio, 1999). Conversely, collectivists may perform better when their leader is transformational because such leaders focus on the goals of the group and a shared purpose. Based on a cross-national study of financial services teams, Schaubroeck et al. (2007) found results at the team level that coincide with those of Jung and Avolio’s (1999) experiment: Transformational leadership affected team potency more for groups with shared collectivist values. In addition, shared power distance values also enhanced the effects of transformational leadership on team potency. Individuals with strong power distance values have a greater respect for leaders and thus are more likely to internalize a transformational leader’s confidence in the team’s capabilities. However, evidence for the moderating effects of culture extends beyond followers’ values to the leaders’ values as well. Spreitzer, Perttula, and Xin (2005), for example, showed that leaders’ traditional cultural values moderated the relationship between transformational leadership and effectiveness. Taken together, these studies suggest that culture is a significant, yet perhaps underrepresented 209

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

and under-explored, component of the leadership process. The final category of moderators that has garnered empirical support relates to organizational context. Shamir and Howell (1999) proposed that charismatic leadership emerges more often and has stronger effects in weak psychological situations, which are characterized by ambiguity, whereas in strong psychological situations, which constrict behavior, the reverse is true. In contrast is Lim and Ployhart’s (2004) exploration of transformational leadership in typical and maximum contexts. A maximum context consists of three components: (a) followers are aware that their performance is under evaluation, (b) followers are committed to exerting maximum effort, and (c) the task is short in duration (e.g., SWAT teams). Lim and Ployhart argued that the potential of transformational leaders is amplified in maximum contexts (which should seemingly be “stronger situations”) and showed that the relationship between transformational leadership and performance is weaker in typical performance contexts as compared with these maximum contexts. Another example of a contextual moderator of leadership stems from Lowe et al.’s (1996) metaanalytic review, which concluded that charisma was significantly more highly correlated with effectiveness in public (r = .74) versus private (r = .59) organizations (correlations corrected for unreliability of measures, measurement error, and sampling error). One possible reason for Lowe et al.’s finding is that public-sector organizations tend to be more consultative than their private-sector counterparts, which are more hierarchical in nature. Sizeable proportions of the employed workforce in virtually all countries are employed within the public sector, making the study of leadership in this context more than a mere passing curiosity. Also motivating the need for this focus is the frequent stereotype of differences in leadership and motivation between the private- and public-sector environments (e.g., leadership is “better” in the private sector). Instead of denigrating leadership within public-sector organizations, a more constructive approach might be to learn from them. In this case, given the demonstrated effectiveness of transformational leadership, one lesson might be that changing 210

to a more consultative and less hierarchical nature might help private-sector organizations enhance their transformational leaders’ effectiveness. The foregoing discussion of contextual moderators suggests that the results of charismatic and transformational leadership in one context may be different in another context. Along the same lines, Judge and Piccolo (2004) showed that transformational leadership validities differed between organizational and military contexts. This result is important because many leadership studies take place in military contexts, and their findings have been used in support of obtaining similar results in nonmilitary organizations. Therefore, a more detailed discussion of these studies and their findings is warranted. First and foremost, unlike business or similar organizational contexts, military contexts provide an opportunity to understand the nature, development, and consequences of leadership in life-and-death situations. Second, the fact that military personnel who are often engaged in highly dangerous work could be volunteers, regular employees, or involuntarily conscripted provides a rich context in which to test the intersection between leadership effectiveness and follower characteristics. Third, given the norms and prescriptive behavior in the military, leadership behaviors may be more evident in individuals who hold relatively high status in the chain of command (Kane, Tremble, & Trueman, 2000). Two studies based in a military context provide useful insights. First, Bass et al. (2003) studied the longitudinal effects of transformational leadership in four different brigades undergoing tactical mission exercises in the United States. Like others, they showed that the effects of transformational leadership and contingent reward on platoon performance were not necessarily direct; they were partially mediated by the direct effects of leadership on platoon cohesion and potency. In addition, their results revealed that passive or laissez-faire leadership style exerted negative effects on group cohesion, potency, and performance. Second, the implications of Dvir et al.’s (2002) findings for the development of leadership, discussed earlier in this chapter, were that cadets’ performance (e.g., knowledge and use of light weapons and physical performance on an obstacle

Copyright American Psychological Association. Not for further distribution.

Leadership

course) was highest in the transformational leadership group. Continuing the discussion on moderators of leadership, an understanding of the labor-union context, in which union officials exercise little formal power and democracy is integral to union dynamics, would be instructive. Unions differ in many ways from private- and public-sector organizations. Specifically, although membership is sometimes mandatory, participation is voluntary. Leadership occurs at all levels of organizations—and unions are no exception; a great deal of research has focused on the lowest levels of unions (i.e., on shop stewards; Barling et al., 2002). In addition, because obtaining participation from members is critical for union survival and success, understanding leadership in unions is both a means to an end and an end in itself. Beginning in the 1950s, leadership research emphasized typologies of union leaders’ behavior. Gouldner (1947) first differentiated between union leaders who were focused on bread-and-butter issues and those who were governed by more progressive priorities. Batstone, Boraston, and Frenkel (1977) extended this model, suggesting four “types” of union leaders, namely, the leader, cowboy, nascent leader, and populist leader. Later, researchers applied existing leadership theories to the unionized context. Consistent with LMX theory, McClane (1991) initially showed the importance of the interaction between union leaders and their members and identified some critical personality factors that moderated this relationship. However, most of the subsequent research has focused on transformational leadership, showing that transformational leadership influences union attitudes and union participation (e.g., Catano, Pond, & Kelloway, 2001; Fullagar, McCoy, & Shull, 1992; Kelloway & Barling, 1993). Nonetheless, given the democratic nature of unions, it remains for research to investigate directly just how the unionized context might moderate the effects of transformational leadership. In summary, although a great deal of knowledge has emerged from studying the outcomes of leadership, further conceptual and empirical attention is needed to fully appreciate the leadership process, particularly in terms of its mechanisms and contin-

gencies. We hope that this relatively brief review stimulates thought toward such endeavors. EXPLORING THE ROLE OF FOLLOWERS IN LEADERSHIP So far, we have focused our attention exclusively on leaders and leadership. However, any act of leadership requires the active involvement of, and agreement by, followers. In this next section, we reverse this bias by introducing followers into our understanding of leadership. There are many reasons why followers look to their leaders, one of which is to help them to make sense of organizational life (Pfeffer, 1977). To understand how followers engage in sensemaking, we turn our discussion to topics involving the social construction of leadership, focusing on two frameworks: the “romance of leadership” (Meindl et al., 1985) and implicit leadership theories (ILTs; Epitropaki & Martin, 2005b). We then consider the social identity analysis of leadership, which relates both leader and follower properties to the emergence and effectiveness of leadership. All of these frameworks help to accomplish our goal of shifting the discussion away from leader behaviors and traits and toward followers’ perceptions of those behaviors and traits. Accordingly, this section raises some critical questions for future research on followership. As mentioned earlier in this chapter, an emphasis on leader behaviors—or what might appropriately be referred to as a leader-centric focus—dominates leadership studies. Consequently, we know much about leader behaviors and how they are associated with follower performance and satisfaction; in contrast, much less is known about followers’ perceptions of leader behaviors and how they might be influenced, or be a function of, leaders’ behaviors. Consistent with a follower-centric focus, in this section we ask how followers perceive and understand leadership behaviors.

The “Romance of Leadership” Perhaps nowhere might the concept of the romance of leadership be better understood than during times of abject crises. During crises, leadership almost always induces follower perceptions of leader charisma; where recovery from the crisis is critical 211

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

to the future well-being of followers, such recovery is often attributed to that leader’s capabilities (Pillai, 1996; Pillai & Meindl, 1998). This process was evident in Rudi Giuliani’s popularity after the attacks of September 11: “He had created ardent critics and foes along the way, but all were silenced by the stern-jawed competence that characterized the outgoing mayor’s statesmanlike response to the worst tragedy in his city’s history” (Fiorina, 2001, p. 8). Fiorina’s observation illustrates how a successful response to a grave crisis can overwrite past “critics and foes,” with much more favorable leadership perceptions taking their place. Similarly, Bligh, Kohles, and Meindl (2004) wrote of President Bush, Prior to the events of 9/11, President George W. Bush was generally not seen as a strong, charismatic leader that people would place their faith in during times of crisis or external threat. . . . the media often characterized the President as oratorically challenged. . . . Seemingly overnight, however, Americans embraced the President and his leadership. Before the terrorist attacks, 51% of Americans approved of Bush’s job performance, whereas after the attacks, his approval ratings jumped to 86%. (p. 213) In both of these cases, one plausible explanation is that followers are retrospectively romanticizing the role that their leaders played in their recovery, attributing their safety and well-being to the perceived extraordinary qualities and behaviors of their leaders. The romance of leadership notion has spurred many research questions that differ significantly from the traditional leader-centric paradigm. To illustrate: In the traditional focus, ratings of leadership (e.g., to what degree the leader exudes power and confidence) are interpreted as actual accounts of leader behavior; consistent with the romance of leadership notion, ratings are more likely to be seen as information about followers’ constructions of leadership (Meindl, 1995). Moreover, after gathering leader ratings from different followers, researchers 212

typically aggregate those ratings. Aggregation of the ratings disregards the uniqueness of each follower’s construction of leadership. Inspired by the romance of leadership notion, researchers would seek to explain why there are differences or similarities in followers’ ratings. Assuming a romance of leadership perspective raises several issues. First, formal leadership status, rank, and position in the organization assume less importance. When followers’ constructions of leadership are critical to a comprehensive understanding of leadership, the importance of informal leaders (e.g., peer leaders) who attract followers’ attention and commitment increases (e.g., Loughead & Hardy, 2005; Neubert, 1999), even in the presence of a formal leader (e.g., Manz & Sims, 1987). Evidently, for followers, leadership means more than official status or ranking in an organization. Second, although in theory we are all sensemakers, some individuals are more prone to romanticizing leadership than others (Felfe, 2005), hence the development and usefulness of the Romance of Leadership Scale (RLS; Meindl, 1998). The RLS suggests how leader ratings could be over- or underestimated: Individuals who score high on the RLS tend to overestimate leader competencies, whereas those low on the RLS tend to underestimate leaders (Felfe & Petersen, 2007). These findings should prompt researchers and practitioners to interpret follower ratings of leadership cautiously. Third, leaders also have romantic notions about their own leadership style and may encourage the romance of their leadership so as to garner follower support and approval (Gray & Densten, 2007). One way of romanticizing one’s own leadership is to deemphasize one’s faults, which may be done unconsciously (e.g., leaders deceive themselves into believing their own rhetoric) or consciously (e.g., they deliberately manage a favorable impression; Gray & Densten). Fourth, the “romance” of leadership does not imply only positive constructions of leadership. If a positive halo exists, people may see even obviously poor performance in a somewhat positive light if they attribute it to the efforts and activities of top

Leadership

Copyright American Psychological Association. Not for further distribution.

management. But it is more likely that the halo is negative when people evaluate poor performance—when leadership appears to have produced poor outcomes, observers may view those outcomes in an exaggeratedly negative way. (Meindl & Ehrlich, 1987, p. 105) In other words, followers can attribute both organizational successes and organizational failures to leaders during sense-making (e.g., Bligh, Kohles, Pearce, Justin, & Stovall, 2007; Boeker, 1992; Cameron, Kim, & Whetten, 1987). Therefore, what is being romanticized is the power of leaders (to bring about success or failure), not the leaders themselves.

Implicit Leadership Theories Implicit leadership theory (ILT) originates from leadership categorization theory, which suggests that expectations and beliefs about the “ideal leader” serve as standards against which we compare our actual leaders (Lord & Maher, 1993). ILT research has investigated the distinctions among expectations and beliefs about “ideal leaders,” “leaders in general,” “effective leaders,” “supervisors,” and “leaders worthy of influence” (e.g., Kenney, Schwartz-Kenney, & Blascovich, 1996; Offermann, Kennedy, & Wirtz, 1994). Much of the initial research was conducted in laboratories, but more recent research has applied ILT to the field. Offermann et al. conducted a series of studies using university students and employed adults. The results of these studies yielded eight dimensions that were frequently used to describe leaders: sensitivity, dedication, tyranny, charisma, attractiveness, masculinity, intelligence, and strength. Ever since these dimensions were identified, subsequent research has asked questions such as “What are the implications for leaders, leadership, and organizations of matching or not matching a follower’s implicit notions of leadership with specific leaders?” and “Where do these ILTs originate?” To answer these and other questions, Epitropaki and Martin (2005b) investigated the extent to which followers’ actual leaders matched those followers’ ILTs and whether the degree of matching significantly predicted relationship quality, work attitudes, and well-being. As predicted, closer matches were

significantly related to ratings of leader–follower relationship quality, and relationship quality related to attitudes (e.g., job satisfaction) and well-being. Additionally, Epitropaki and Martin (2005a) collected data a year later from the same participants and used cross-lagged modeling analyses to test for the feedback loop originally proposed by Lord and Maher (1993). If a feedback loop exists, then individuals should update their ILTs when they are confronted with contrary evidence. A research question that tests the feedback loop could be, “What happens to the structure of one’s ILTs if the characteristic ‘male’ is part of the ILTs, and yet one’s male leader fails at a business venture?” Epitropaki and Martin (2005a) sought to better understand whether the feedback loop existed and how it worked. However, no support for the feedback loop was found: Followers’ perceptions of leaders tended to be stable over time despite any disconfirming evidence. However, Epitropaki and Martin (2005a) maintained that a feedback loop may still exist; one year of exposure to disconfirming evidence may not have been sufficient to have an effect on follower perceptions. Overall, Epitropaki and Martin’s (2005a) research demonstrated that despite the abstract nature of ILTs, they have practical implications for organizations and work outcomes. Keller (1999) provided insight into the origin of ILTs. Although Offermann et al.’s (1994) initial research specified eight leader dimensions, Keller (1999) related each dimension to the Big Five dimensions of personality mentioned earlier in this chapter. For example, agreeableness and sensitivity were correlated. Expanding our understanding of the development of leadership, Keller also examined the role of parenting style in this study. She concluded that an individual’s ideal leader images were significantly related to that individual’s perceptions of his/her parents; if one’s parents were perceived as tyrannical, then tyrannical was integral to his/her perception of an ideal leader. Keller’s research goes beyond understanding the structure of ILTs and toward understanding the source of them. As presented here, research on the social construction of leadership (whether focusing on the romance of leadership or ILTs) accentuates the role of follower cognitions in attributing leadership and 213

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

identifying leaders. Despite progress in this area, important questions remain. Although the conceptual role of ILTs is obvious—they influence people’s expectations about their leaders—empirical research has yet to establish how extensive ILTs are in the first place. Looking forward, attempts should also be made to track the stability of ILTs over time. In addition, how much others’ (e.g., parents’, first supervisors’, peers’) expectations and beliefs about leaders affect one’s own ideal leader traits needs to be better understood. A rich understanding of followership awaits a social constructionist view of leadership.

Prototypicality In the predominant leader-centric tradition, heroic images of leaders overshadow the role of the group that the leader belongs to and leads, yet many leader behaviors are targeted toward mobilizing followers, emphasizing group goals, and uplifting group morale (Chemers, 2001). Accordingly, group characteristics are critical for leader effectiveness. The social identity analysis of leadership postulates that the congruence between group characteristics and leader characteristics is critical to understanding evaluations of leader effectiveness and leader endorsement. The term leader group prototypicality is used throughout this literature to describe the extent to which leaders represent group norms, values, and standards, also known as group prototypes. Group prototypes are “fuzzy sets of characteristics that in a given context define the group” and they “describe and prescribe group membership appropriate attributes and behavior in a specific context” (Giessner & van Knippenberg, 2008, p. 15); this effect is heightened as the salience of the group increases (Hogg, 2001). The social identity analysis of leadership that underlies group prototypicality describes the mechanisms through which leaders emerge and gain follower endorsement. Unlike LMX and transformational leadership theory, which focus on the nature of leadership, the social identity analysis of leadership is concerned with identifying the features of leaders and followers that critically define a leader’s emergence and development (van Knippenberg & van Knippenberg, 2005). In addition, unlike implicit leadership theories, which provide a within-person understanding 214

of leadership, prototypicality addresses relational issues between followers, groups, and leaders. The social identity analysis of leadership addresses itself to questions such as “To what extent does group composition relate to leader effectiveness?” “How do groups select their leaders?” and “Why do some leaders gain significantly more follower support than other leaders?” A major tenet of the social identity analysis of leadership is that the group member who most accurately embodies the group’s values and norms is the most likely to emerge as the group’s leader. One example of this derives from the fashion industry, where leadership succession is frequent and group prototypes are not only readily observable but are also potential sources for competitive advantage. In that industry, successors are pressured to maintain a fashion house’s values and standards while simultaneously injecting their own uniqueness into each collection. Valentino Garavani’s announced departure from his fashion house in 2007 led to a search for his successor, with Alessandra Facchinetti eventually named as his replacement. The CEO of Valentino Fashion Group, Stefano Sassi, remarked, “Facchinetti is the designer who can interpret and continue the legacy of Valentino’s core values at their best” (Barnett, 2007), suggesting that Facchinetti’s successful emergence as the house’s leader was based, at least in part, on her ability to accurately represent the group prototype (i.e., “Valentino’s core values”). In a second example, after CEO Robert Meers announced his retirement from the primarily women’s clothing retailer Lululemon Athletica Inc., Christine Day was named his successor in April 2008. Clearly expressing an interest in preserving the group prototype, founder and chairman Chip Wilson had said earlier in February, “I have a quest now to turn Lululemon into truly a women’s company. I believe the next CEO will be a woman” (Shaw, 2008, p. FP7). Having a female CEO reflects and reinforces Lululemon’s interest in women’s health and lifestyle issues; indeed, Day admitted to participating in yoga, Pilates, and running—activities to which Lululemon caters with its fitness-brand merchandise (Constantineau, 2008). Meers further affirmed the importance of Day’s representation of

Copyright American Psychological Association. Not for further distribution.

Leadership

the organization: “She also happens to be our target customer—she lives and breathes it, and that was exactly what I was looking for” (Constantineau, 2008). Wilson intimated that Day will have a hand in creating “female-friendly policies,” underscoring how much the organization values being a “strong women’s company” (Shaw). Moreover, being female, Day enhances her leader group prototypicality, as the majority of the organization’s employees are women. The social identity analysis of leadership can be applied to both examples; leader role occupancy and emergence were dependent on the congruence between the organization’s and the successor’s values, norms, and standards, thereby enhancing leader group prototypicality. Beyond leader emergence (e.g., van Knippenberg, van Knippenberg, & van Dijk, 2000), research findings support the notion that leader group prototypicality is associated with leader endorsement (e.g., Platow & van Knippenberg, 2001), ratings of effectiveness (e.g., Hains, Hogg, & Duck, 1997), and persuasiveness (e.g., van Knippenberg, Lossie, & Wilke, 1994), and it affords prototypical leaders latitude in their actions (Giessner & van Knippenberg, 2008). Likewise, the effects of leader group prototypicality are more pronounced when group members strongly identify with the group (e.g., Hains et al.; Hogg, Hains, & Mason, 1998; Platow & van Knippenberg). The underlying rationale is that leader group prototypicality conveys to followers the notion that the leader is aligned with collective interests and is truly a member of the group, and it reinforces the group’s identity (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). Despite such findings, there is no empirical support for the relationship between leader prototypicality and objective measures of performance. Nonetheless, group-prototypical leaders fare better with respect to evaluations of effectiveness. Giessner and van Knippenberg (2008) reported that in times of failure, leaders who were prototypical of group norms were judged less harshly than leaders who were less prototypical of group norms. Leader prototypicality, then, provided a “license to fail.” Similarly, leader prototypicality also provides what may be thought of as a “license to eccentricity.” The positive feelings

surrounding a prototypical leader (e.g., who is presumed to have the collective’s interests at heart) could afford the leader leeway in his or her actions (including actions that lead to failure) that should extend to that individual’s unconventional or risky behaviors (van Knippenberg et al., 2000; van Knippenberg & van Knippenberg, 2005). The situation of low group prototypicality has also been studied. Van Knippenberg and van Knippenberg (2005) found a significant relationship between leader self-sacrifice and evaluations of leader effectiveness, and this was most pronounced for leaders who were low in group prototypicality. This finding suggests that leaders’ self-sacrificing behaviors may compensate for the lack of group prototypicality: Self-sacrificing behavior, not unlike group prototypicality, signals that a leader is genuinely interested in the collective good. Future directions for research may take this a step further and posit situations in which low leader group prototypicality is actually beneficial for followers, such as during times of change or when the group’s culture is counterproductive. Relatively new to leadership studies, the social identity analysis of leadership is predominantly studied in laboratory experiments. It follows that the challenge is to apply these concepts to real-world situations or to experiments that maximize ecological validity, where group prototypes and leader prototypicality may not be as salient as they appear in experiment designs. In this vein, Hogg and colleagues (2006) conducted an experiment in which the group prototype was not explicit. To create an ambiguous group prototype, they manipulated the salience of either a stereotypically feminine (e.g., creative) or masculine (e.g., rational) group norm rather than explicitly stating that the group prototype was “male” or “female.” They also manipulated leader gender and measured participants’ sex-role orientation (i.e., beliefs about female and male behavioral norms), and they found that even when the group prototype was ambiguous, male leaders were judged as more effective when the group prototype was stereotypically masculine. Similarly, female leaders were judged as more effective when the group prototype was stereotypically feminine. It follows that even under conditions of inexplicitness the significance 215

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

of a group prototype for leadership endorsement is not diminished. To further improve upon the generalizability of findings from this area of research, it would be useful to investigate leader emergence, endorsement, and perceptions in groups without any group prototypes. The absence of a group prototype may be more representative of the real-world context in which (new and temporary) groups operate. Following the earlier examples of Valentino and Lululemon, future field research may also explore the phenomena of leader emergence on group norms (i.e., Day’s contribution to a “women’s company”) as well as group norms contributing to leader emergence (i.e., Facchinetti’s fit with the established Valentino norms) in tandem. The social identity analysis of leadership provides a fresh perspective that depicts a relational model of leadership: The characteristics that support a leader’s endorsement and emergence are specific to the group being led. Reflecting its social identity and selfcategorization foundations, the social identity analysis of leadership promotes research where the variables of interest are followers’ perceptions, such as their perceptions of a leader’s representativeness of the group, and curiosity lies in the cognitive processes in which followers engage to inform their perceptions. DESTRUCTIVE LEADERSHIP Until now, we have focused primarily on high-quality leadership. Unfortunately, not all leadership is either ethical or productive. In this section, we review common conceptualizations of undesirable leadership styles, including passive, abusive, and unethical leadership. (See also Vol. 3, chap. 15, this handbook.)

Neglectful and Abusive Leadership For the most part, the effects of passive leadership have been compared with those of active leadership (e.g., by comparing laissez-faire with transformational leadership). It is widely held that passive leadership fails to produce the positive outcomes of more active leadership styles (Skogstad, Einarsen, Torsheim, Aasland, & Hetland, 2007), but it is not necessarily presumed to result in negative effects. Only rarely has passive leadership been linked to 216

detrimental results. A meta-analysis of the full-range model of transformational leadership showed that laissez-faire leadership was negatively related to leader effectiveness (r = −.54), and that followers of laissez-faire leaders tended to be dissatisfied with their jobs (r = −.28) and leaders (r = −.58; all correlations corrected for unreliability of measures, measurement error, and sampling error; Judge & Piccolo, 2004). Skogstad et al. (2004) extended these findings to show that passive, indirect leadership also predicted role ambiguity, role conflict, conflict with coworkers, bullying behaviors, and, indirectly, psychological distress. Similarly, passive leadership may also be conceptualized as neglectful, particularly with respect to employee safety. Earlier we described how transformational leaders may act as role models of safe behaviors, lowering incidents of workplace injuries. Conversely, passive leaders are detached from their leadership responsibilities and unlikely to be involved in promoting safety behaviors (Kelloway et al., 2006). As a result, instead of acting as role models, passive leaders signal that safety is unimportant, and such neglect can have adverse effects on safety outcomes. Kelloway et al. (2006) provided empirical support for this proposition, finding that safety-specific passive leadership negatively predicted safety-related outcomes even after accounting for the positive effects of safety-specific transformational leadership. More recently, Hinkin and Schriesheim (2008) showed in a series of four studies that when leaders avoid giving appropriate rewards and punishment, follower perceptions (satisfaction and perceptions of effectiveness) of the leader are negatively influenced, as are subordinates’ perceptions of role clarity and supervisors’ perceptions of their subordinates’ performance. Kelloway et al. (2006) made a clear distinction between the passive leadership that they studied and directly abusive leadership, suggesting that, in the case of safety, leaders are more likely to overlook safety issues than to purposefully or maliciously obstruct and compromise the safety of their employees. A separate line of research has clarified the behaviors of abusive supervision, despite its relative infrequency (e.g., Tepper, 2000; Zellars, Tepper, & Duffy, 2002; for a detailed review of the abusive

Copyright American Psychological Association. Not for further distribution.

Leadership

supervision literature, see Tepper, 2007). Tepper (2000) defined abusive supervisions as “subordinates’ perceptions of . . . the extent to which supervisors engage in the sustained display of hostile verbal and nonverbal behaviors, excluding physical contact” (p. 178, emphasis in original). Example behaviors include criticizing followers in front of others, yelling at followers, belittling followers, lying, and unjustifiably blaming followers for mistakes. Abusive supervision predicts a host of negative outcomes, including follower deviance, poor attitudes and performance, turnover, diminished psychological health, and poorer work-family functioning (see Tepper, 2007, for a review). Abusive supervision is also met with greater follower resistance (Tepper, Duffy, & Shaw, 2001). Tepper, Moss, Lockhart, and Carr (2007) reported that followers of abusive supervisors tended to use “regulative maintenance communication tactics” that involved maintaining the leader–follower relationship through avoidance and evasion tactics (e.g., distorting negative reports to prevent a punitive reaction from the leader). Ironically, Tepper et al. (2007) found that such tactics actually exacerbate the relationship between abusive supervision and follower psychological distress. Two primary mediators relate abusive supervision to these dysfunctional outcomes, namely, perceptions of injustice and lack of control. Followers of abusive supervisors are likely to feel unfairly treated, explaining both their resultant negative attitudes and states and also their retaliatory behaviors (e.g., Tepper, 2000; Zellars et al., 2002). Followers may retaliate against an abusive supervisor not only to restore equity but also to regain a sense of personal control over the situation. Mitchell and Ambrose (2007) showed that followers of abusive supervisors were more likely to display deviant behaviors toward their supervisor and that this relationship was stronger when followers held negative, “an eye for an eye,” reciprocity beliefs. In addition, followers acted more defiantly toward others in the organization as well, suggesting that their inability to restore justice and personal control may have initiated displaced aggression. An equally interesting stream of research has considered how displaced aggression can explain abusive supervision (Tepper, 2007). In a study of

supervisors, followers, and followers’ family members, Hoobler and Brass (2006) found evidence consistent with the notion that supervisors displaced their anger on followers, as evidenced by abusive supervisory behaviors, and that followers displaced their aggression onto their family members. Specifically, when supervisors perceived a breach in their psychological contract with their organizations, then followers were more likely to report abusive supervisory behaviors; this was particularly the case for leaders with a hostile attribution bias or the tendency to overly blame others. Moreover, family members of abused followers reported higher levels of family undermining committed by the abused followers. A second empirical study also lends some support to the displaced aggression theory of abusive supervision. In this study, leaders who experienced interactional injustice from their immediate supervisors were more likely to be perceived by their followers as abusive. However, this relationship only held for supervisors with an authoritarian leadership style (Aryee, Chen, Sun, & Debrah, 2007). Although concern about the predominant negative focus in much of psychology is not new, such negativity may be less prevalent in the field of leadership, with its strong emphasis on the benefits of positive leadership. In fact, further research is required to more fully understand the antecedents of abusive supervision and how follower attributes and behaviors moderate these relationships (Tepper, 2007). For example, leader depression is one empirically tested predictor of abusive supervision (Tepper, Duffy, Henle, & Lambert, 2006), while other proposed antecedents, such as leader personality and organizational culture, remain unexplored (Tepper, 2007). Perhaps one explanation for the overall lack of studies on negative leadership is that leaders’ behaviors are not always consistent; for example, they could be charismatic, yet also display bouts of hostile behavior (Pfeffer, 2007). Steve Jobs of Apple Computer might reflect this: He has been credited with reviving Apple and painted as charismatic in the process (e.g., Harvey, 2001), but he also has a reputation for “sadistic perfectionism, often without discernible provocation” (Berglas, 1999, p. 29). These complexities and inconsistencies are difficult to capture in single studies but must be reflected in 217

Barling, Christie, and Hoption

research to avoid simplistic interpretations of “purely evil” (or “purely good”) leaders. We now turn our attention to unethical leadership, which goes beyond behaviors and spotlights leaders’ values, beliefs, and morals.

Copyright American Psychological Association. Not for further distribution.

Unethical Leaders Despite considerable concern devoted to this issue by the lay public, the ethics of leadership has all too often escaped systematic study by organizational scholars (Brown, Treviño, & Harrison, 2005), although this critical omission is now being reversed. Greater attention is being accorded to ethical issues in leadership from an array of different approaches, such as personality (e.g., House & Howell, 1992; Judge et al., 2006), values (e.g., Bass & Steidlmeier, 1999), moral reasoning (Turner, Barling, Epitropaki, Butcher, & Milner, 2002), moral orientation (Simola, Barling & Turner, in press), generalized ethical leadership (Brown et al.), follower attributions (Dasborough & Ashkanasy, 2002), and behavioral integrity (Dineen, Lewicki, & Tomlinson, 2006). Brown et al. (2005) defined ethical leadership as “the demonstration of normatively appropriate conduct through personal actions and interpersonal relationships, and the promotion of such conduct to followers through two-way communication, reinforcement, and decision-making” (p. 120). The authors conceptually distinguished this concept of ethical leadership from other similar constructs, such as transformational leadership and leader honesty, and showed that ethical leadership was positively associated with follower satisfaction with the leader, job dedication, willingness to speak up about problems, and leader effectiveness. More frequently, scholarly attention has been drawn to depictions of leaders as immoral or unethical, which should describe leaders who fail to uphold the behaviors described by Brown et al. (2005). One of the most prominent distinctions is between socialized charismatic and personalized charismatic leaders (e.g., House & Howell, 1992; Howell & Avolio, 1992) or, in parallel, transformational, and pseudotransformational leaders (e.g., Bass & Steidlmeier, 1999). Personalized charismatic, or pseudotransformational, leaders offer the illusion of transformational leadership through their strong inspirational 218

appeal. As Bass described pseudotransformational leadership, “it looks like a transformational leader, it acts like a transformational leader, but in fact it is not” (Hooijberg & Choi, 2000, p. 298). In contrast to transformational leaders, these leaders place their own self-interested goals above the collective good; a typical example, according to Bass, “would be the executive who cries crocodile tears when downsizing, but then gives himself a big bonus” (Hooijberg & Choi, 2000, p. 298). Pseudotransformational leadership is thought to be especially destructive to followers because these leaders have a powerful ability to motivate others, while largely ignoring their welfare. They may also select followers who provide them unwavering support: “As one former disciple of Michael Milken, the junk bond king, said, ‘If he walked off the cliff, everyone in that group would have followed him’ ” (Howell & Avolio, 1992, p. 47). Related to unethical leadership is self-focused leadership, as depicted in the study of leader narcissism. By nature, narcissism may provoke leaders to overly attribute organizational successes to their own virtues and efforts and accordingly undervalue the contributions of followers. This is an important issue, as anecdotal evidence suggests that narcissistic individuals may be more likely to emerge as leaders (see Rosenthal & Pittinsky, 2006, for a review). However, Paunonen, Lönnqvist, Verkasalo, Leikas, and Nissinen (2006) distinguished between what they deemed as the more positive elements (i.e., egotism and self-esteem) and the more negative elements (i.e., impression management and manipulation) of narcissism, and they showed that the positive elements of narcissism predicted leadership only in the absence of the negative elements. Research is needed to explain the relationship between narcissism and leadership emergence and thus further understand self-focused leadership styles. Resisting the temptation to draw conclusions about leadership based on outcomes, this section on the dark side of leadership clearly describes deleterious leadership based on underlying motivations, intentions, values, and behaviors. Moving forward, research may benefit most from understanding which conditions are more amenable to destructive leadership styles, and why, and whether some individuals are more predisposed to the “dark side” of leadership

Leadership

(Hogan & Hogan, 2001), as well as ways to mitigate any harmful effects (Sutton, 2007).

Copyright American Psychological Association. Not for further distribution.

MEASURING LEADERSHIP The typical approach to investigating leadership is not without its flaws. More precisely, by conducting a study that focuses on subordinate perceptions and uses pre-developed measures and approaches to leadership, we are making a number of assumptions—assumptions that appear, in many cases, misguided. (Hunter et al., 2007, p. 436) In a quest for “true meaning, substance, and practical utility” (p. 443), Hunter et al. challenged leadership researchers to move away from a number of fundamental assumptions, including (a) that leadership is equally important for all followers, (b) that followers witness leader behaviors and therefore can evaluate them, (c) that instruments to measure leadership are psychometrically sound, and (d) that current leadership instruments capture critical leader behaviors. It is tempting to regard Hunter et al.’s observations as a critique of the leadership field, but we regard their insights as directing future research. In this vein, the present section reviews select assumptions and debates about common methods in leadership research. As noted earlier in this chapter, transformational leadership remains the most widely researched leadership theory, and to date the most widely used instrument to measure transformational leadership is the MLQ (Bass & Avolio, 1990). The MLQ has also been used to assess transactional leadership and laissez-faire leadership. Despite its extensive use in organizational research, many concerns remain, primarily relating to the factor structure of the MLQ. Bycio et al. (1995) noted that the four facets of transformational leadership are so highly correlated that, in practice, it is unlikely that a leader would score high on one facet and low on the others. Moreover, the high correlation between transactional and transformational leadership has also been raised, and together these concerns have resulted in much

research on the psychometric properties of the MLQ (e.g., Bycio et al., 1995; Carless, 1998; Den Hartog, Van Muijen, & Koopman, 1997; Heinitz, Liepmann, & Felfe, 2005; Tejeda, Scandura, & Pillai, 2001; Tepper & Percy, 1994). Interpreting precisely what these findings mean is difficult. From a practical perspective it is evident that it is challenging to separate the four transformational components (Heinitz et al., 2005). The four behaviors are held to be conceptually separate, but the consistent factor analytic findings could point to measurement problems (which is frequently the position taken) or might indicate that the four behaviors are not conceptually separate after all. To complicate matters further, Bass (1998) himself suggested that “transformational leaders . . . behave in ways to achieve superior results by employing one or more of the four components of transformational leadership” (p. 5). If a leader has to use only one component of transformational leadership to be considered transformational, then despite any conceptual differences, all four behaviors are equally indicative of transformational leadership, and substitutable. Clearly, research will need to isolate the most appropriate conceptualization and measurement of transformational leadership (MacKenzie et al., 2005). Accordingly, alternative measures for transformational leadership have been proposed. Carless, Wearing, and Mann (2000) developed the seven-item Global Transformational Leadership scale (GTL; e.g., “encourages thinking about problems in new ways and questions assumptions,” “communicates a clear and positive vision of the future”) rated on a 5-point Likert-type scale (e.g., 1 = Rarely or never, 5 = Very frequently, if not always). More recently, in Herold at al.’s (2008) study, transformational leadership was assessed using 22 items (e.g., “I believe my leader encourages employees to be ‘team players’ ” and “I believe my leader shows respect for individuals’ feelings”) that were based on the earlier work of Rubin et al. (2005). In addition, Alimo-Metcalfe and AlbanMetcalfe (2001) developed the Transformational Leadership Questionnaire (TLQ), as well as a specific version for government employees. The TLQ yielded nine factors, including “genuine concern for others,” “accessible and approachable,” and “encourages critical and strategic thinking.” The comparative 219

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

advantages and disadvantages of all these scales, including the MLQ, need to be investigated further. To complement this discussion of transformational leadership measures, Table 7.2 contains a list of leadership measures that cover many of the leadership concepts in this chapter. Although not exhaustive, this list illustrates the advance of relational leadership measures (e.g., how the leader treats the follower). This trend has prompted studies on leader–follower agreement, and it presents an opportunity for the development of a measure of leadership that requires information from both the leader and the follower to capitalize on the unique information about the relationship that both members possess. Departing from how to measure of leadership, we now turn our attention to the question of when to measure leadership, and this discussion identifies three perceptual biases that may significantly affect leadership assessment: (a) honeymoon effect, (b) hangover effect, and (c) halo error. Honeymoon biases occur at the start of one’s tenure in the organization; the enthusiasm of starting a new job, combined with the organization putting its best foot forward for new hires, may solicit (overly) positive attitudes toward the organization; the hangover effect describes the decline and eventual stability in positive attitudes after the honeymoon period (Boswell, Boudreau, & Tichy, 2005). Both of these biases could influence leadership ratings, with followers rating their leaders more highly at the start of their relationship (i.e., during the honeymoon period). Over time, perceptions of the leaders may decline as followers gain more information about the organization, the job, and the leaders’ behaviors (i.e., during the hangover period). It follows that the length of time that the leader and follower have spent together can influence leadership ratings, and researchers might be well advised to account for honeymoon and hangover biases. Approaches based on employees’ ratings of leadership rest on the assumption that followers have continued interactions with their leaders. In reality, that is likely not the case; Hunter et al. (2007) asked researchers to consider the extent to which followers are actually privy to leader behaviors. One way to satisfy this query is to study specific incidents or 220

episodes of leadership. Within this framework, the most appropriate time to measure leader effectiveness would be after an episode of leadership rather than relying on retrospective accounts. Self-defined episodes of leadership may be the most appropriate, given the established role of critical moments in (mis)shaping memories (e.g., Redelmeier, Katz, & Kahneman, 2003); on the other hand, focusing on episodes may promote a so-called halo effect. A halo effect results from judgments that are based on a general impression, thereby neglecting specific acts that may disconfirm one’s general impression, and “[a]s a result of the halo effect, individuals are rated as consistently good or consistently poor performers, regardless of their variable strengths and weaknesses” (Nathan & Lord, 1983, p. 102). To underscore the importance of this bias, consider that measures of leadership rely on follower perceptions and that previous research has established that follower ratings are especially vulnerable to halo error (Frone, Adams, Rice, & Instone-Noonan, 1986). Therefore, when measuring leadership, it is important to note that followers’ ratings of leadership may contain followers’ general impressions of the leader and may not necessarily be based upon what a leader actually does. We have concentrated on measurement issues in this section, but methodological issues in leadership research are also pervasive. Hunter et al. (2007) provided important recommendations to address such concerns. These included the need for multiple sources of information about leader behavior, paying equal attention to positive and negative aspects of leadership, accounting for the context in which leadership occurs, and engaging in multilevel and/or longitudinal research. Collectively, confronting the measurement and methodological challenges that pervade the field of leadership studies will enable the further development of comprehensive understanding of leadership. LEADERSHIP IN RELATED CONTEXTS Thus far, we have reviewed the leadership literature from a multitude of contexts and have considered how leadership effects can be minimized or maximized in different contexts. In this next section, we

Meindl, 1998; Meindl and Ehrlich, 1988 Houghton and Neck, 2002 Hersey and Blanchard, 1988 Dupré and Barling, 2006 Bass and Avolio, 1990 Carless, Wearing, and Mann, 2000

LMX-7 Supervisory Behavior Description Questionnaire (SBDQ) Ohio State Leader Behavior Description Questionnaire (LBDQ) Need for Supervision Scale Romance of Leadership Scale (RLS)

Revised self-leadership questionnaire (RSLQ) Leadership Effectiveness and Adaptability Description (LEAD) instrument Supervisory Control Over Work Performance Multifactor Leadership Questionnaire (MLQ) Global Transformational Leadership (GTL)

Leader–member exchange Leadership style

Romance of leadership

Self-leadership Situational leadership/leader adaptability Supervisory control

Transformational leadership

Need for supervision

de Vries, Roe, and Taillieu, 1998

Stodgill, 1963

Epitropaki and Martin, 2004; Offermann, Kennedy, and Wirtz (1994) Graen and Uhl-Bien, 1995 Fleishman, 1953

“My supervisor does not give me the freedom to do things that I want to do in my work.” “My leader displays a sense of power and confidence.” “My leader communicates a clear and positive vision of the future.”

“The region manager has a marked influence on my performance.” “When the top leaders are good, the organization does well; when the top leaders are bad, the organization does poorly.” “I establish specific goals for my own performance.”

“How well does your leader recognize your potential?” “He ‘needles’ people under him for greater effort.”

Implicit leadership theories

Implicit leadership

Brown, Treviño, and Harrison, 2005 Kelley, 1992

“My boss tells me my thoughts or feelings are stupid.” “Appears to be a skillful performance when presenting to a group.” “Discusses business ethics or values with employees.” “Are your personal work goals aligned with the organization’s priority goals?” How characteristic is each trait of a business leader? (e.g., dominant, sensitive)

Tepper, 2000 Conger and Kanungo, 1994

Abusive Supervision Conger-Kanungo (C-K) scale of charismatic leadership Ethical Leadership Scale Followership Questionnaire

Abusive supervision Charismatic leadership

Ethical leadership Follower style

Example item (if available)

Source

Measure

Construct

Measures of Leadership Styles, Behaviors, and Perceptions

TABLE 7.2

Copyright American Psychological Association. Not for further distribution.

Leadership

221

Barling, Christie, and Hoption

take a special interest in the studies conducted in the sports and education contexts. By doing so, we illustrate how organizational leadership frameworks are being applied to other settings and how we can potentially advance organizational leadership studies.

Copyright American Psychological Association. Not for further distribution.

The Sports Context Sports psychologists have likened coaches to business leaders (see Kellett, 1999), and as a result the application of organizational leadership frameworks to a sports context has garnered interest. In particular, several empirical tests of the role of transformational leadership in sports performance exist. For example, Rowold (2006) studied 200 martial arts students and showed that leaders’ idealized influence, inspirational motivation, and individualized consideration predicted athletes’ extra effort, perceptions of coaches’ effectiveness, and satisfaction with the coach. Inspirational motivation also predicted the frequency of training each month. Zacharatos et al.’s (2000) study showed the generalizability of these findings: Although the average age of Rowold’s sample was 32 years, Zacharatos et al. showed that the same pattern of results was obtained among adolescent athletes. Charbonneau et al. (2001) studied elite university athletes, finding that the effects of transformational leadership on sports performance were indirectly related through intrinsic motivation. Given these performance outcomes, it follows that transformational leadership in sports teams may be a competitive advantage for sports teams. Day et al. (2004) did not focus their study on a particular style of leadership, but they found that occupying a leadership role (i.e., team captain) in the National Hockey League (NHL) was related to performance—even after controlling for performance prior to assuming the leadership role. To our knowledge, there has yet to be an empirical test of whether leadership titles in organizations would yield similar results, and Day et al. also made an important distinction between the sports and organizational contexts. In sports teams, there are followers—the athletes—and leaders such as coaches, managers, and team captains, but there is some overlap; that is, a team captain can be considered one of the athletes. In organizations, there is a distinction between leaders and followers—often 222

defining each role against the other and arguably placing more emphasis on hierarchy and status. This structural difference between contexts warrants consideration before organizational researchers (or sports psychologists) transplant leadership frameworks developed for organizations to the sports context. Revealing another difference between these contexts, Kellett (1999) interviewed coaches and concluded that “leadership” was not an integral part of coaches’ self-defined job description, even though researchers are quick to make such a parallel. Kellett went on to argue that elements of transformational leadership were rarely referenced in his interviews, yet Kellett also argued that coaches described their work as “facilitating the development of others” (p. 165), which is consistent with individualized consideration. Furthermore, not confining their analysis to coaches, Hoption, Phelan, and Barling (2007) interviewed professional athletes and found evidence of transformational leadership in sports. To explain some of the inconsistency, future research may explore sports-specific transformational leadership.

Education The educational or school context also applies to the study of leadership. Intriguingly, effective leadership in this context may have long-term consequences for student attitudes and performance. Whether the focus is on the influence of principals on teachers or on the influence of teachers on students, many of the behaviors under consideration (e.g., student performance) are largely discretionary, increasing the potential influence of leadership. Like other contexts we have showcased thus far, studies in the education milieu have focused on transformational leadership, showing, for example, that teachers’ transformational leadership is related to students’ perceptions of teacher performance and students’ involvement in their own studies (Harvey, Royal, & Stout, 2003). Nguni, Sleegers, and Denessen’s (2006) study also highlighted the effectiveness of transformational leadership but reiterated that any leadership effects are often indirect. In addition, their results supported the “augmentation hypothesis” (Bass, 1998) demonstrating that teachers’ transformational leadership provided unique variance after accounting for the effects of

Copyright American Psychological Association. Not for further distribution.

Leadership

transactional leadership. In another example of transformational leadership in this context, Sahin’s (2004) study documented the association between school principals’ transformational leadership and school culture. Arguably the most comprehensive study in an educational context is Koh, Steers, and Terborg’s (1995) multilevel study within 89 schools in Singapore. They studied the effects of school principals’ transformational leadership on teachers’ satisfaction, organizational commitment, and citizenship behaviors and their students’ academic performance. Their results both replicated and extended other findings on transformational leadership. First, they found no compelling evidence that transformational leadership affected student performance directly; instead, like others studying transformational leadership and performance, they found evidence for an indirect effect, thereby highlighting the role of critical mediating variables (e.g., teacher attitudes). Second, Koh et al.’s data replicated the “augmentation hypothesis”; however, the authors go a step further, showing that transactional leadership did not account for any significant variance after the effects of transformational leadership were controlled. Third, the importance of cross-cultural nuances is demonstrated in Singaporean context because of one finding that appears to be at odds with prior research, namely, that principals’ transformational leadership did not predict teachers’ citizenship behaviors. However, this is consistent with the notion that many of the teachers’ measured citizenship behaviors were not discretionary but instead were contractual obligations in the Singaporean context. Last, like findings in the military context (Dvir et al., 2002), Koh et al. went beyond this dyadic focus, showing that leaders can indeed affect the performance of the followers of their followers. WHAT WE STILL NEED TO KNOW As indicated at the outset of this chapter, although much is known about leadership, much remains to be learned. In this final section of the chapter, we consider what we believe are some of the major lessons that need to be learned.

Followership Perhaps Napoleon Bonaparte said it best: “Soldiers generally win battles; generals get credit for them.” The leader-centric studies that dominate research on leadership result in an incomplete knowledge about “leadership”; however, advocating an exclusive follower-centric research agenda would result in similarly unbalanced knowledge. Thus, we advocate a relational view of leadership, one in which leaders and followers together produce leadership. Inserting followers into the leadership equation is not novel, but, as reflected in the opening quote to this section, the contribution of followers is underappreciated and, moreover, often restricted to “obeying orders” and “taking direction” (Baker, 2008). We join the call for research to combat such a passive stereotype of followers; it diminishes the role of followers in organizational success and leader effectiveness and followers’ ability to motivate leadership change (e.g., Deluga, 1987), and by default it exaggerates the role and importance of leaders. Suggestive support for followers as active contributors to leadership emerged in Dvir and Shamir’s (2003) longitudinal study in which leaders were rated as less transformational as followers developed their own leadership skills. One possible explanation is that with follower development, a transformational relationship emerged, meaning that the responsibility to motivate, inspire, stimulate, and nurture are shared between leader and follower. Although the possibility for bidirectional socialization remains to be investigated directly in a leadership context, evidence of such effects exists in other hierarchical relationships, including parent–child relationships (e.g., Glass, Bengtson, & Dunham, 1986). Developing implicit followership theories (i.e., expectations and beliefs about followers) might be especially useful in understanding followership. With expectations on leaders and followers, dyadic data become essential. Each member of the leader–follower dyad gains meaning from and through the other.

Shared Leadership Moving beyond leader–follower distinctions, what about the situation in which many individuals in the same group demonstrate leadership behaviors? 223

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

When this occurs, it is termed shared leadership. Although varied definitions of shared leadership exist, in general, it is a team property that depicts the extent to which team members mutually influence each other with or without formal leadership (e.g., Carson, Tesluk, & Marrone, 2007; Ensley, Hmieleski, & Pearce, 2006). Essentially, the presence of shared leadership flattens an organization’s hierarchy and harnesses the power of leadership by distributing it among team members so that peerto-peer influence flourishes. Shared leadership is associated with performance outcomes in problem-solving groups (e.g., Carson et al., 2007), new ventures (e.g., Ensley et al., 2006), and medical teams (Klein, Ziegert, Knight, & Xiao, 2006). In addition to performance, shared leadership is linked to group cohesion (e.g., Perry, Pearce, & Sims, 1999), collective vision (e.g., Ensley, Pearson, & Pearce, 2003), collective identity (e.g., Shamir & Lapidot, 2003), and creativity and intrinsic motivation (e.g., Hooker & Csikszentmihalyi, 2003). Nonetheless, Perry and colleagues (1999) cautioned against the assumption that shared leadership is always beneficial: “When the inherent benefits of working in a [team] are not necessary for the [task], the costs of the shared leadership process may actually decrease effectiveness” (p. 45). Additionally, Barry (1991) argued that effective shared leadership becomes problematic if members do not have the skills to demonstrate those behaviors and if teams are not purposefully composed to address a team’s various leadership needs. Equally important, the effectiveness of shared leadership depends on team members’ enjoying shared goals (Conger & Pearce, 2003). Qualitative and conceptual manuscripts outnumber the empirical tests of shared leadership. One reason for this could be the varied operationalizations of shared leadership, each of which expresses shared leadership differently. Conger and Pearce (2003) and Carson et al. (2007) identified aggregation techniques, network analysis, and group measures as possible ways to quantify shared leadership. Each methodological approach has its advantages (e.g., richness of the data) and disadvantages (e.g., complicated statistical methods), so it would be helpful for future research to compare the use of different 224

methodological approaches (Conger & Pearce, 2003) in empirically advancing our understanding of shared leadership. Having acknowledged that there are limitations to shared leadership and its operationalization, its benefits are still intriguing for many scholars and practitioners. The following suggestions have been raised for establishing shared leadership: provide feedback on leader behaviors (Shamir & Lapidot, 2003): offer leadership training for all team members (Houghton, Neck, & Manz, 2003); ensure that the organizational culture is aligned with team leadership, such as replacing individual rewards with team-based rewards (e.g., Hooker & Csikszentmihalyi, 2003); and use an external leader to monitor the rate and progress of shared leadership (Perry et al., 1999). There is a strong likelihood that team members who know each other well will be more likely to engage in shared leadership (e.g., Barry, 1991; Morgan, Salas, & Glickman, 1993; Perry et al., 1999), so it is important to keep in mind that shared leadership emergence requires time and long-term planning.

Authentic Leadership Consistent with the attention given to some major ethical lapses by leaders, the notion of authentic leadership has attracted much interest since the turn of the century. Following early debate, Avolio and his colleagues included the following components in their definition of authentic leadership (see Avolio & Luthans, 2006; Walumbwa, Avolio, Gardner, Wernsing, & Peterson, 2008): self-awareness, relational transparency, balanced processing of all relevant data in an objective manner before decisions are made, and finally, an internalized moral perspective. Although the relative newness of this perspective means that few empirical studies have been reported, Walumbwa et al. have provided a reliable and valid measure, and their data supported a higher-order multidimensional model of the construct. One benefit of these early data is that they were collected not just in the United States but also in Kenya and China (Walumbwa et al., 2008), allaying concerns that the concept of authentic leadership may be culturally bound. Important questions remain for this nascent theory. On a conceptual level, there are differing views as

Copyright American Psychological Association. Not for further distribution.

Leadership

to whether authenticity requires a grounding in moral values (Avolio & Gardner, 2005) or whether authenticity requires leaders be true to their values and beliefs no matter how socially unacceptable (Shamir & Eilam, 2005). Shamir and Eilam (2005) and Sparrowe (2005) went further, questioning the advisability of authenticity among leaders who are narcissistic or who display questionable values. Second, the theory underlying the concept of authentic leadership suggests that the four dimensions are related but separate, but the available data suggest strong intercorrelations among the dimensions. Third, given the core characteristics of authentic leadership, it remains for research to demonstrate that it is empirically separate from conceptually similar constructs. Fourth, it remains to be seen whether authentic leadership can be developed in leaders.

Leadership in Critical Moments Leadership is often operationalized as the frequency of leadership behaviors. For example, using the MLQ, followers rate how often their leaders engage in transformational behaviors, and it is assumed that followers construct their ratings by accurately averaging those behaviors over a specified time period. However, it may be more likely that leadership has relatively greater meaning in times of crisis (Pillai, 1996; Pillai & Meindl, 1998) and that followers construct their perceptions of leadership in pivotal moments when leadership is either highlighted or critical (Tucker, Turner, Barling, Reid, & Elving, 2006). Understanding leadership in the midst of these critical moments is thus essential to a complete characterization of leadership. How leaders respond to critical moments can be vital. Dutton, Frost, Worline, Lilius, and Kanov (2002) highlighted the need to lead with compassion in times of organizational and individual crisis, which could include incidents such as the events of September 11, a natural disaster, the death of an employee, or an employee’s unexpected illness. When this occurs, employees are likely to feel a greater sense of commitment to the organization, often enhancing performance. Accordingly, leadership behaviors in times of crisis can extend beyond those critical moments, shaping leader–follower relationships and the organization into the future.

Supporting this notion is the aforementioned finding that leaders’ apologies following transgressions are related to transformational leadership ratings (Tucker et al., 2006); Tucker et al. argued that apologies are critical moments in the leader–follower relationship and defined such moments as “distinct interactions that, while occurring relatively infrequently, serve to punctuate or reinforce the status quo” (p. 197). Similarly, whether or not charismatic leaders are more likely to emerge in times of crisis has been debated (for a review. see Shamir & Howell, 1999). Although some studies suggest that charismatic leaders emerge in times of crisis (e.g., Roberts & Bradley, 1988), evidence to the contrary is not uncommon (e.g., Pillai & Meindl, 1998). Alternatively, it is possible for leaders to frame situations in a way that is perceived as critical or extraordinary, suggesting reverse causality (Shamir & Howell, 1999). Nevertheless, recent research suggests that transformational leaders may be more effective under critical performance conditions than in more ordinary situations (Lim & Ployhart, 2004). Clarifying these issues is required if researchers are to elucidate the role of leadership in critical moments. Preliminary evidence points to potential differences in leadership behaviors and leadership perceptions in ordinary versus extraordinary situations. However, many questions remain. Is leadership best conceptualized as an average frequency of behaviors over time, or do some situations enhance the salience of certain leadership behaviors? Do any benefits that accrue to leaders because of their behaviors during a crisis carry over to an everyday basis? Researchers have the opportunity to provide new knowledge by focusing on questions such as these.

Humility One way to understand leadership is to through the core role of humility. Morris et al. (2005) defined humility as “a personal orientation founded on a willingness to see the self accurately and a propensity to put oneself in perspective” (p. 1331). While acknowledging that other leadership theories also accord a critical role to humility (e.g., level 5 leadership; Collins, 2001), the behaviors involved in transformational leadership make it as relevant to humility. 225

Barling, Christie, and Hoption

Humility has been identified as important to understanding transformational leadership (Bass & Riggio, 2006) in that humility would restrain leaders from becoming entranced with public adulation and would influence leaders to be other-focused (Morris et al, 2005). Despite this recognition of the importance of humility in transformational leadership, there has not yet been a focus on how the specific humility-related behaviors engaged in by leaders underlie transformational leadership.

Copyright American Psychological Association. Not for further distribution.

Leadership Meets Social Neuroscience Unquestionably, some of the most substantial and intriguing advances in our understanding of individual behavior over the past 2 decades originated with merging social and biological knowledge. The field of social neuroscience uses broadly based biological processes to explain social interactions and behaviors (Cacioppo et al., 2007). This breadth and depth of the knowledge awaits those wishing to expand our understanding of leadership. As noted earlier in this chapter, an intriguing question that continues to bedevil social and behavioral scientists is whether leadership is learned or inherited. Several studies based on a behavioral genetic approach have addressed the relative role of environmental and genetic factors in leadership emergence (Arvey et al., 2006, 2007). Despite those advances, genetic and biological effects on the development of leadership behaviors warrant attention and robust empirical examination. For example, testosterone is consistently linked with behaviors (e.g., social control and aggression) that characterize negative leadership (e.g., van Honk & Schutter, 2007; White, Thornhill, & Hampson, 2006). There is also evidence linking counterproductive organizational behaviors to the early emergence of childhood conduct problem disorders (Roberts et al., 2007). Incorporating leadership into the field of social neuroscience would respond to calls for more integrative theory building with respect to leadership (Avolio, 2007) and open up the field of leadership development as a viable opportunity for future research on social neuroscience (Cacioppo et al., 2007). Nonetheless, the success of these endeavors requires knowledge of the social and the biological 226

bases of behavior and the research methods underlying both. With the majority of leadership researchers being trained in the social sciences, collaboration across fields traditionally seen as unrelated will be a necessity.

Humor Although there is some research on humor in the workplace (e.g., Fleming, 2005; Francis, 1994; Yovetich, Dale, & Hudak, 1990), its prevalence in organizations is widespread and diverse in nature. Leaders have certainly been exhorted to provide a so-called fun workplace culture (Pfeffer, 1998), including the use of humor (Fleming, 2005). Such appeals are based on the notion that fun workplace cultures are associated with many benefits, including increased worker motivation (Crawford, 1994), commitment, and performance (Avolio, Howell, & Sosik, 1999). In this sense, humor is instrumental and strategic; as explained by Fleming, “humor is ultimately a serious business. It is unsurprisingly driven by very sober corporate motives” (p. 288). Although humor can sometimes be beneficial, positive outcomes do not always ensue. Avolio and colleagues (1999) found that the frequent use of humor does not always lead to better performance; they speculated that some issues (e.g., setting target objectives) were not amenable to humor because of their seriousness. Additionally, positive outcomes depend on the type of humor used. Decker and Rotondo (2001) distinguished between negative (e.g., sexual and insult humor) and positive (e.g., nonoffensive humor) forms of humor and concluded that leaders who used positive forms of humor received more positive leader ratings than leaders who used negative forms of humor. One question that immediately emerges is: Why would leaders choose to use negative forms of humor? According to the superiority theory of humor, negative forms of humor (such as insults) reinforce the hierarchy between leaders and followers (Westwood, 2004). Leaders who want to maintain power distance may be especially prone to using humor in this manner. In contrast, positive forms of humor should minimize the distance between leaders and followers (Barsoux, 1996). In essence, humor can be used to communicate a leader’s values, especially

Copyright American Psychological Association. Not for further distribution.

Leadership

the type of leader–follower relationship that he or she desires. Thus far, we have considered the strategic use of humor, but we also acknowledge that humor could be an individual difference: People have general tendencies when using humor. To this end, Martin, Puhlik-Doris, Larsen, Gray, and Weir (2003) developed an instrument to measure individuals’ humor style, the Humor Styles Questionnaire (HSQ). They defined four styles of humor usage: affiliative (e.g., “I enjoy making people laugh”), self-enhancing (e.g., “If I am feeling upset or unhappy I usually try to think of something funny about the situation to make myself feel better”), aggressive (e.g., “If I don’t like someone, I often use humor or teasing to put them down”), and self-defeating (e.g., “If I am having problems or feeling unhappy, I often cover it up by joking around, so that even my closest friends don’t know how I really feel.”). To our knowledge, empirical investigations into leadership and Martin et al.’s forms of humor have yet to be explored. However, because the HSQ focuses on the joketeller’s motivations (e.g., to feel better versus to make others laugh), it would be useful to explain why leaders choose to use some forms of humor over others. Furthermore, it is possible that leaders use more than one form of humor, such as both affiliative and aggressive humor, with equal frequency; how would this impact their leader ratings? It would be interesting to investigate how followers reconcile this seeming inconsistency, as well as the situational conditions that may be more (or less) conducive to certain forms of humor and the stability of humor style over time.

Corporate Social Responsibility Organizations’ attempts to recognize social issues are usually encapsulated within “corporate social responsibility” (CSR), which “reflect[s] the organization’s status and activities with respect to its perceived societal obligations” (Dacin & Brown, 1997, p. 68). We would be remiss if we did not comment on the potential role for leadership in this area, as one might reasonably expect that this will become an increasingly important issue for organizations in the future (see Vol. 3, chap. 24, this handbook). In general, most research on CSR has been conducted at the organiza-

tional level (e.g., Bird, Hall, Momentè & Reggiani, 2007); it is extremely unusual to encounter research on this topic at the individual employee level. This remains surprising, as the success or failure of any CSR initiative presumably rests largely on whether employees identify with the values inherent in the CSR initiative and are willing to go beyond normal job expectations to ensure that the initiative succeeds. Given its value-based focus and empirical research showing that it can motivate employees’ discretionary behaviors, charismatic leadership might be especially appropriate in this context. Previous research has highlighted the effectiveness of charismatic leaders in selling a message and producing rhetoric to justify a cause (e.g., Den Hartog & Verburg, 1997; Seyranian & Bligh, 2008), and the societal impact of CSR initiatives could be consistent with idealized influence’s emphasis on the collective good and collective goals. In the future, it will be critical for leaders to move beyond a philanthropic approach to one which emphasizes the many ways in which organization and their members can benefit from organizational responsibility interventions (see Vol. 3, chap. 24, this handbook). How organizations respond to, and raise awareness of social issues (e.g., the environment) further illustrates the potential role for and impact of leadership both within and beyond the organization.

Leadership Selection While genetic and early family environment influences leadership role occupancy and behaviors, the selection of leaders is in the hands of organizations. To facilitate leadership selection, questionnaires and different assessment tools (e.g., structured interviews; Krajewski, Goffin, McCarthy, Rothstein, & Johnston, 2006) are frequently used. Nonetheless, the basic conceptual assumption, that performance on a questionnaire can predict performance “on the battlefield,” is debatable (Gladwell, 2004, 2008). Use of personality tests to aid in selecting leaders rests on the premise that certain personality traits are empirically associated with effective (and noneffective; Hogan & Hogan, 2001) leadership behaviors. Perhaps the most widely used personality inventory 227

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

for leadership selection is the Myers-Briggs Type Instrument (MBTI; Gardner & Martinko, 1996; Roush & Atwater, 1992), a self-report measure of personality preferences or “types” developed by the mother-and-daughter team of Katharine and Isabel Briggs, that is based on the theory of Carl Jung. The use of the MBTI for selection in general, and leadership selection in particular, is clouded in considerable controversy owing to lingering concerns about theory, reliability and validity, and appropriate administration and scoring procedures (e.g., Jackson, Parker, & Dipboye, 1996; Pittenger, 1993). One straightforward prediction is that, given the importance of leadership selection and the practical and financial efficiency of online questionnaires, there will be growing pressure to move toward the use of online leader selection tools. Nonetheless, practical problems need to be confronted, including whether online methods can assess the interactive and dynamic functions of leadership. The future for leadership selection tests depends on the availability of reliable and valid tests that can be used appropriately and that predict the outcomes of interest. It would be premature to discard leadership selection tests in their entirety and risk throwing the baby out with the bathwater (e.g., Paunonen, Rothstein, & Jackson, 1999). However, an updated perspective of leadership selection and selection tests is warranted. At the same time, nontechnological techniques such as leaderless group tasks (Hooijberg & Choi, 2000) and unstructured interviews might still prove useful in leadership selection, and subordinates need to be engaged in the leadership selection process. Moreover, given that a large portion of the leadership function now involves team processes and that teams hold a central role in organizations, team-based selection will likely be a part of the future of leadership selection.

Contrasting Different Leadership Theories As is no doubt evident from the preceding discussions, there is clearly no dearth of available leadership theories. Although Hunter et al. (2007) have described the content of the typical leadership study, it is also possible to characterize the approach taken within the typical leadership study: Specifically, Hunter et al. described research—be it correlational 228

or experimental, cross-sectional or longitudinal— that tests the central tenets of a single, specific theory. One of the ultimate goals of all such research would be to make informed judgments about the relative utility and validity of the different leadership theories. However, because of the meaningful differences across all these studies—for example, in the measures used, the outcomes addressed, and the time needed for interventions to exert any effects— any such judgments would be difficult at best. Instead, like the research conducted on goalsetting theory some 2 decades ago by Latham, Erez, and Locke (1988) that was designed to disentangle seemingly contradictory findings about the same phenomena from different studies, substantially more can be gained at this stage from conducting research that a priori sets up fair comparisons (Cooper & Richardson, 1986) between competing leadership theories that can produce “winners” and “losers.” Moreover, the process described by Latham et al., which engages the theories’ disputants in the design of the research, remains underutilized yet extremely promising. The results of such research would have the potential to help researchers and practitioners alike make sense of the existing leadership theories and enumerable empirical tests. CONCLUSION Our understanding of leadership has come a long way since physical features such as height and attractiveness were thought to be prime determining factors of leadership emergence and effectiveness. Despite the wealth of knowledge that has accumulated, opportunities to answer new questions promise to expand our knowledge of leadership to new and different directions in the near future. After almost a century of social and behavioral science research on leadership, we stand on the brink of a new and expanded knowledge of organizational leadership.

References Agle, B. R., Nagarajan, N. J., Sonnenfeld, J. A., & Srinivasan, D. (2006). Does CEO charisma matter? An empirical analysis of the relationships among organizational performance, environmental uncertainty, and top management team perceptions of CEO charisma. Academy of Management Journal, 49, 161–174.

Leadership

Aguinis, H., & Henle, C. A. (2002). Ethics in research. In S. G. Rogelberg (Ed.), Handbook of research methods in industrial and organizational psychology (pp. 34–56). Malden, MA: Blackwell. Aguinis, H., & Kraiger, K. (2009). Benefits of training and development for individuals, teams, organizations and society. Annual Review of Psychology, 60, 451–474.

Copyright American Psychological Association. Not for further distribution.

Ah Chong, L. M., & Thomas, D. C. (1997). Leadership perceptions in cross-cultural context: Pakeha and Pacific Islanders in New Zealand. Leadership Quarterly, 8, 275–293. Alimo-Metcalfe, B., & Alban-Metcalfe, R. J. (2001). The development of a new Transformational Leadership Questionnaire. Journal of Occupational and Organizational Psychology, 74, 1–27. Alutto, J. A., & Belasco, J. A. (1972). A typology for participation in organizational decision making. Administrative Science Quarterly, 17, 117–125. Amernic, J., Craig, R., & Tourish, D. (2007). The transformational leader as pedagogue, physician, architect, commander, and saint: Five root metaphors in Jack Welch’s letters to stockholders of General Electric. Human Relations, 60, 1839–1872. Amit, K., Popper, M., Gal, R., Miskal-Sinai, M., & Lissak, A. (2006). The potential to lead: The difference between “leaders” and “non-leaders.” Megamot, 44, 277–296. Arnett, J. J. (2000). Emerging childhood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469–480. Arnold, K. A., Turner, N. A., Barling, J., Kelloway, E. K., & McKee, M. (2007). Transformational leadership and well-being: The mediating role of meaningful work. Journal of Occupational Health Psychology, 12, 193–203. Arvey, R. D., Bouchard, T. J., Segal, N. L., & Abraham, L. M. (1989) Job satisfaction: Environmental and genetic components. Journal of Applied Psychology, 74, 187–192. Arvey, R. D., Rotundo, M., Johnson, W., Zhang, Z., & McGue, M. (2006). The determinants of leadership role occupancy: Genetic and personality factors. Leadership Quarterly, 17, 1–20. Arvey, R. D., Zhang, Z., Avolio, B. J., & Krueger, R. F. (2007). Developmental and genetic determinants of leadership role occupancy among women. Journal of Applied Psychology, 92, 693–705. Aryee, S., Chen, Z. X., Sun, L., & Debrah, Y. A. (2007). Antecedents and outcomes of abusive supervision: Test of a trickle-down model. Journal of Applied Psychology, 92, 191–201. Avolio, B. J. (1999a). Are leaders born or made? Psychology Today, 32, 18.

Avolio, B. J. (1999b). Full leadership development: Building the vital forces in organizations. Thousand Oaks, CA: Sage Publications. Avolio, B. J. (2007). Promoting more integrative strategies for leader theory-building. American Psychologist, 62, 25–33. Avolio, B. J., & Gardner, W. L. (2005). Authentic leadership development: Getting to the root of positive forms of leadership. Leadership Quarterly (June 2005), 16, 315–338. Avolio, B. J., Howell, J. M., & Sosik, J. J. (1999). A funny thing happened on the way to the bottom line: Humor as a moderator of leadership style effects. Academy of Management Journal, 42, 219–227. Avolio, B. J., & Luthans, F. W. (2006). The high impact leader: Authentic, resilient leadership that gets results and sustains growth. New York: McGraw Hill. Avolio, B. J., Rotundo, M., & Walumbwa, F. O. (2009). Early life experiences as determinants of leadership role occupancy: The importance of parental influence and rule breaking behaviour. Leadership Quarterly, 20, 324–342. Avolio, B. J., Zhu, W., Koh, W., & Bhatia, P. (2004). Transformational leadership and organizational commitment: Mediating role of psychological empowerment and moderating role of structural distance. Journal of Organizational Behavior, 2, 951–968. Ayman, R., & Chemers, M. M. (1983). Relationship of supervisory behavior ratings to workgroup effectiveness and subordinate satisfaction among Iranian managers. Journal of Applied Psychology, 68, 338–341. Ayman, R., Chemers, M. M., & Fiedler, F. (1995). The contingency model of leadership effectiveness: Its level of analysis. Leadership Quarterly, 6, 147–167. Baker, L. A., Jacobson, K. C., Raine, A., Lozano, D. I., & Bezdijan, S. (2007). Genetic and environmental bases of childhood antisocial behavior: A multi-informant twin study. Journal of Abnormal Psychology, 116, 219–236. Baker, S. D. (2008). Followership: The theoretical foundation of a contemporary construct. Journal of Leadership & Organizational Studies, 14, 50–60. Barling, J., Loughlin, C., & Kelloway, E. K. (2002). Development and test of a model linking safetyspecific transformational leadership and occupational safety. Journal of Applied Psychology, 87, 488–496. Barling, J., Weber, T., & Kelloway, E. K. (1996). Effects of transformational leadership training on attitudinal and financial outcomes: A field experiment. Journal of Applied Psychology, 81, 827–832. Barnett, L. (2007, September 5). Breaking news: Alessandra Facchinetti at Valentino. Retrieved May 15, 2008, from http://www.vogue.co.uk/news/daily/2007–09/070905229

Barling, Christie, and Hoption

breaking-news-alessandra-facchinetti-at-valentino. aspx Barry, D. (1991). Managing the bossless team: Lessons in distributed leadership. Organizational Dynamics, 20, 31–47. Barsoux, J. (1996). Why organizations need humor. European Management Journal, 14, 500–508. Bass, B. M. (1960). Leadership, psychology, and organizational behavior. New York: Harper & Brothers.

Copyright American Psychological Association. Not for further distribution.

Bass, B. M. (1983). Issues involved in relations between methodological rigor and reported outcomes in evaluations in evaluations of organizational development. Journal of Applied Psychology, 68, 197–199.

Bligh, M. C., & Kohles, J. C., Pearce, C. L., Justin, J. E., & Stovall, J. F. (2007). When the romance is over: Follower perspectives of aversive leadership. Applied Psychology: An International Review, 56, 528–557. Boeker, W. (1992). Power and managerial dismissal: Scapegoating at the top. Administrative Science Quarterly, 37, 400–421. Bommer, W. H., Rich, G. A., & Rubin, R. S. (2005). Changing attitudes about change: Longitudinal effects of transformational leader behavior on employee cynicism about organizational change. Journal of Organizational Behavior, 26, 733–753.

Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press.

Bono, J. E., & Anderson, M. H. (2005). The advice and influence networks of transformational leaders. Journal of Applied Psychology, 90, 1306–1314.

Bass, B. M. (1997). Does the transactional–transformational leadership paradigm transcend organizational and national boundaries? American Psychologist, 52, 130–139.

Bono, J. E., & Judge, T. A. (2003). Self-concordance at work: Toward understanding the motivational effects of transformational leaders. Academy of Management Journal, 46, 554–571.

Bass, B. M. (1998). Transformational leadership: Industrial, military, and educational impact. Mahwah, NJ: Erlbaum.

Bono, J. E., & Judge, T. A. (2004). Personality and transformational and transactional leadership: A meta-analysis. Journal of Applied Psychology, 89, 901–910.

Bass, B. M., & Avolio, B. J. (1990). Manual for the Multifactor Leadership Questionnaire. Palo Alto, CA: Consulting Psychologists Press. Bass, B. M., Avolio, B. J., Jung, D. I., & Berson, Y. (2003). Predicting unit performance by assessing transformational and transactional leadership. Journal of Applied Psychology, 88, 207–218. Bass, B. M., & Riggio, R. E. (2006). Transformational leadership (2nd ed.). Mahwah, NJ: Erlbaum. Bass, B. M., & Steidlmeier, P. (1999). Ethics, character, and authentic transformational leadership behavior. Leadership Quarterly, 10, 181–217. Batstone, E., Boraston, I., & Frenkel, S. (1977). Shop stewards in action: The organization of workplace conflict and accommodation. Oxford, England: Blackwell. Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology Monographs, 4, 1–102. Berglas, S. (1999). What you can learn from Steve Jobs. Inc, 21, 29–31. Bernerth, J. B., Armenakis, A. A., Feild, H. S., Giles, W. F., & Walker, H. J. (2007). Is personality associated with perceptions of LMX? An empirical study. Leadership & Organization Development Journal, 28, 613–631. Bird, R., Hall, A., Momentè, F., & Reggiani, F. (2007). What corporate responsibility activities are valued by the market? Journal of Business Ethics, 76, 189–206. Bligh, M. C., Kohles, J. C., & Meindl, J. R. (2004). Charisma under crisis: Presidential leadership, rhetoric, and media responses before and after the September 11 terrorist attacks. Leadership Quarterly, 15, 211–239. 230

Boone, C., Van Olffen, W., & Van Witteloostuijn, A. (2005). Team locus-of-control composition, leadership structure, information acquisition, and financial performance: A business simulation study. Academy of Management Journal, 48, 889–909. Boswell, W. R., Boudreau, J. W., & Tichy, J. (2005). The relationship between employee job change and job satisfaction: The honeymoon-hangover effect. Journal of Applied Psychology, 90, 882–892. Bouchard, T. J., Lykken, D. T., McGue, M., Segal, N. L., & Tellegen, A. (1990, October 12). Sources of human differences: The Minnesota Study of Twins Reared Apart. Science, 250, 223–228. Bronfenbrenner, U. (1961). Some familial antecedents of responsibility and leadership in adolescents. In L. Petrullo & B. M. Bass (Eds.), Leadership and interpersonal behavior (pp. 239–271). New York: Holt, Rinehart & Winston. Brown, M. E., Treviño, L. K., & Harrison, D. A. (2005). Ethical leadership: A social learning perspective for construct development and testing. Organizational Behavior and Human Decision Processes, 97, 117–134. Burke, C. S., Sims, D. E., Lazzara, E. H., & Salas, E. (2007). Trust in leadership: A multi-level review and integration. Leadership Quarterly, 18, 606–632. Burns, J. M. (1978). Leadership. New York: Harper. Bycio, P., Hackett, R. D., & Allen, J. S. (1995). Further assessments of Bass’s (1985) conceptualization of transactional and transformational leadership. Journal of Applied Psychology, 80, 468–478.

Leadership

Byrne, J. A. (1998). Jack: A close-up look at how America’s number 1 manager runs GE. Business Week, 3581 (8 June), 90.

Conger, J. A. (1989). The charismatic leader: Behind the mystique of exceptional leadership. San Francisco: Jossey-Bass.

Cacioppo. J. T., Amaral, D. G., Blanchard, J. J., Cameron, J. L., Carter, C. S., Crews, D., et al. (2007). Social neuroscience: Progress and implications for mental health. Perspectives on Psychological Science, 2, 99–123.

Conger, J. A. (1999). Charismatic and transformational leadership in organizations: An insider’s perspective on these developing streams of research. Leadership Quarterly, 10, 145–179.

Copyright American Psychological Association. Not for further distribution.

Cameron, K. S., Kim, M. U., & Whetten, D. A. (1987). Organizational effects of decline and turbulence. Administrative Science Quarterly, 32, 222–240. Carless, S. A. (1998). Assessing the discriminant validity of transformational leader behaviour as measured by the MLQ. Journal of Occupational and Organizational Psychology, 71, 353–358. Carless, S. A., Wearing, A. J., & Mann, L. (2000). A short measure of transformational leadership. Journal of Business and Psychology, 14, 389–405. Carson, J. B., Tesluk, P. E., Marrone, J. A. (2007). Shared leadership in teams: An investigation of antecedent conditions and performance. Academy of Management Journal, 50, 1217–1234. Catano, V. M., Pond, M., & Kelloway, E. K. (2001). Exploring commitment and leadership in voluntary organizations. Leadership and Organization Development Journal, 22, 256–263. Charbonneau, D., Barling, J., & Kelloway, E. K. (2001). Transformational leadership and sports performance: The mediating role of intrinsic motivation. Journal of Applied Social Psychology, 31, 1521–1534. Chemers, M. M. (2001). Leadership effectiveness: An integrative view. In M. Hogg and S. Tindale (Eds.), Blackwell handbook of social psychology: Group processes. (pp. 376–399). Oxford, England: Blackwell. Chen, G., Kirkman, B. L., Kanfer, R., Allen, D., & Rosen, B. (2007). A multilevel study of leadership, empowerment, and performance in teams. Journal of Applied Psychology, 92, 331–346. Cherulnik, P. D., Donley, K. A., Wiewel, T. S. R., & Miller, S. R. (2001). Charisma is contagious: The effect of leaders’ charisma on observers’ affect. Journal of Applied Social Psychology, 31, 2149–2159. Cherulnik, P. D., Turns, L. C., & Wilderman, S. K. (1990). Physical appearance and leadership: Exploring the role of appearance-based attribution in leader emergence. Journal of Applied Social Psychology, 20, 1530–1539. Clark, P. F., Barling, J., Carroll, A., Fullagar, C., & Gallagher, D. G. (2007). The Influence of Transformational Leadership on Newcomers’ Union Attitudes: A Longitudinal Study. Annual Meeting of the Academy of Management, Philadelphia, PA, August. Collins, J. (2001). Good to great: Why some companies make the leap . . . and others don’t. New York: Harper Collins.

Conger, J. A., & Kanungo, R. N. (1994). Charismatic leadership in organizations: Perceived behavioral attributes and their measurement. Journal of Organizational Behavior, 15, 439–452. Conger, J. A., & Kanungo, R. N. (1998). Charismatic leadership in organizations. Thousand Oaks, CA: Sage. Conger, J. A., & Pearce, C. L. (2003). A landscape of opportunities: Future research on shared leadership. In C. L. Pearce & J. A. Conger (Eds.) Shared leadership: Reframing the hows and whys of leadership. Thousand Oaks, CA: Sage. Constantineau, B. (2008, April 2). Profits soar at Lululemon. The Vancouver Sun. Retrieved May 3, 2008, from http://www.canada.com/vancouversun/ news/business/story.html?id=7d4a6589-01e7-4d07bb24-5b324eb651e4&k=62474 Cook, T. D., & Campbell, D. T. (1979). Quasiexperimentation: Design and analysis issues for field settings. Boston, MA: Houghton Mifflin. Cooper, W. H., & Richardson, A. H. (1996). Unfair comparisons. Journal of Applied Psychology, 71, 179–184. Cox, C. J., & Cooper, C. L. (1989). The making of the British CEO: Childhood, work experience, personality, and management style. Academy of Management Executive, 3, 353–358. Crant, J. M., & Bateman, T. S. (2000). Charismatic leadership viewed from above: The impact of proactive personality. Journal of Organizational Behavior, 21, 63–75. Crawford, C. B. (1994). Theory and implications regarding the utilization of strategic humor by leaders. Journal of Leadership Studies, 1, 53–67. Dacin, P. A., & Brown, T. J. (1997). The company and the product: Corporate associations and consumer product responses. Journal of Marketing, 61, 68–84. Dansereau, F. J., Graen, G., & Haga, W. J. (1975). A vertical dyad linkage approach to leadership within formal organizations: A longitudinal investigation of the role making process. Organizational Behavior and Human Decision Processes, 13, 46–78. Dasborough, M. T., & Ashkanasy, N. M. (2002). Emotion and attribution of intentionality in leader-member relationships. Leadership Quarterly, 13, 615–634. Dasgupta, N., & Asgari, S. (2004). Seeing is believing: Exposure to counterstereotypic women leaders and its effect on the malleability of automatic gender stereotyping. Journal of Experimental Social Psychology, 40, 642–658. 231

Barling, Christie, and Hoption

Davies, P. G., Spencer, S. J., & Steele, C. M. (2005). Clearing the air: Identity safety moderates the effects of stereotype threat on women’s leadership aspirations. Journal of Personality and Social Psychology, 88, 276–287. Day, D. V., Sin, H., & Chen, T. T. (2004). Assessing the burdens of leadership: Effects of formal leadership roles on individual performance over time. Personnel Psychology, 57, 573–605.

Copyright American Psychological Association. Not for further distribution.

Decker, W. H., and Rotondo, D. M. (2001). Relationships among gender, type of humor and perceived leader effectiveness. Journal of Managerial Issues, 13, 450–465. Deluga, R. J. (1987). Corporate coups d’état: What happens and why. Leadership and Organization Development Journal, 8, 9–15. Den Hartog, D. N., House, R. J., Hanges, P. J., RuizQuintanilla, S. A., Dorfman, P. W., Abdalla, I. A., et al. (1999). Culturally-specific and cross-culturally generalizable implicit leadership theories: Are attributes of charismatic leadership theories universally endorsed? Leadership Quarterly, 10, 219–256.

Dupré, K. E., & Barling, J. (2006). Predicting and preventing supervisory workplace aggression. Journal of Occupational Health Psychology, 11, 13–26. Dutton, J. E., Frost, P. J., Worline, M. C., Lilius, J. M., & Kanov, J. M. (2002). Leading in times of trauma. Harvard Business Review, 80, 54–61. Dvir, T., Eden, D., Avolio, B. J., Shamir, B. (2002). Impact of transformational leadership on follower development and performance: A field experiment. Academy of Management Journal, 45, 735–744. Dvir, T., & Shamir, B. (2003). Follower developmental characteristics as predicting transformational leadership: A longitudinal field study. Leadership Quarterly, 14, 327–344. Eagly, A. H., Johannesen-Schmidt, M. C., & van Engen, M. L. (2003). Transformational, transactional, and laissez-faire leadership styles: A meta-analysis comparing women and men. Psychological Bulletin, 129, 569–591. Eagly, A. H., & Johnson, B. T. (1990). Gender and leadership style: A meta-analysis. Psychological Bulletin, 108, 233–256.

Den Hartog, D. N., Van Muijen, J. J., & Koopman, P. L. (1997). Transactional versus transformational leadership: An analysis of the MLQ. Journal of Occupational and Organizational Psychology, 70, 19–34.

Eagly, A. H., & Karau, S. J. (1991). Gender and the emergence of leaders: A meta-analysis. Journal of Personality and Social Psychology, 60, 685–710.

Den Hartog, D. N., & Verburg, R. M. (1997). Charisma and rhetoric: Communicative techniques of international business leaders. Leadership Quarterly, 8, 355–391.

Eagly, A. H., Karau, S. J., & Makhijani, M. G. (1995). Gender and the effectiveness of leaders: A metaanalysis. Psychological Bulletin, 117, 125–145.

Detert, J. R., & Burris, E. R. (2007). Leadership behavior and employee voice: Is the door really open? Academy of Management Journal, 50, 869–884.

Eagly, A. H., Karau, S., Miner, J. B., & Johnson, B. T. (1994). Gender and motivation to manage in hierarchical organizations: A meta-analysis. Leadership Quarterly, 5, 135–159.

de Vries, R. E., Roe, R. A., & Taillieu, T. C. B. (1998). Need for supervision: Its impact on leadership effectiveness. The Journal of Applied Behavioral Science, 34, 486–501. Dickson, M. W., Resick, C. J., & Hanges, P. J. (2006). Systematic variation in organizationally-shared cognitive prototypes of effective leadership based on organizational form. Leadership Quarterly, 17, 487–505. Dineen, B. R., Lewicki, R. J., & Tomlinson, E. C. (2006). Supervisory guidance and behavioral integrity: Relationships with employee citizenship and deviant behavior. Journal of Applied Psychology, 91, 622–635. Dionne, S. D., Yammarino, F. J., Atwater, L. E., & James, L. R. (2002). Neutralizing substitutes for leadership theory: Leadership effects and common-source bias. Journal of Applied Psychology, 87, 454–464.

Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. (1992). Gender and the evaluation of leaders: A meta-analysis. Psychological Bulletin, 111, 3–22. Eby, L. T., Cader, J., & Noble, C. L. (2003). Why do high self-monitors emerge as leaders in small groups? A comparative analysis of the behaviors of high versus low self-monitors. Journal of Applied Social Psychology, 33, 1457–1479. Ehrhart, M. G., & Klein, K. J. (2001). Predicting followers’ preferences for charismatic leadership: The influence of follower values and personality. Leadership Quarterly, 12, 153–179. Elder, G. H. (1974). Children of the Great Depression. Chicago: University of Chicago Press.

Dirks, K. T. (2000). Trust in leadership and team performance: Evidence from NCAA basketball. Journal of Applied Psychology, 85, 1004–1012.

Ellemers, N., de Gilder, D., & Haslam, S. A. (2004). Motivating individuals and groups at work: A social identity perspective on leadership and group performance. Academy of Management Review, 29, 459–478.

Duehr, E. E., & Bono, J. E. (2006). Men, women, and managers: Are stereotypes finally changing? Personnel Psychology, 59, 815–846.

Ensley, M. D., Hmieleski, K. M., & Pearce, C. L. (2006). The importance of vertical and shared leadership within new venture top management teams:

232

Leadership

Implications for the performance of startups. Leadership Quarterly, 17, 217–231. Ensley, M. D., Pearson, A., & Pearce, C. L. (2003). Top management team process, shared leadership, and new venture performance: A theoretical model and research agenda. Human Resource Management Review, 13, 329–346.

Copyright American Psychological Association. Not for further distribution.

Epitropaki, O., & Martin, R. (2004). Implicit leadership theories in applied settings: Factor structure, generalizability, and stability over time. Journal of Applied Psychology, 89, 293–310. Epitropaki, O., & Martin, R. (2005a). From ideal to real: A longitudinal study of the role of implicit leadership theories on leader-member exchanges and employee outcomes. Journal of Applied Psychology, 90, 659–676. Epitropaki, O., & Martin, R. (2005b). The moderating role of individual differences in the relation between transformational/transactional leadership perceptions and organizational identification. Leadership Quarterly, 16, 569–589. Erez, A., Misangyi, V. F., Johnson, D. E., LePine, M. A., & Halverson, K. C. (2008). Stirring the hearts of followers: Charismatic leadership as the transferal of affect. Journal of Applied Psychology, 93, 602–616. Farmer, S. M. and Aguinis, H. (2005). Accounting for subordinate perceptions of supervisor power: An identity-dependence model. Journal of Applied Psychology, 90, 1069–1083. Felfe, J. (2005). Personality and romance of leadership. In B. Schyns & J. R. Meindl (Eds.), Implicit leadership theories: Essays and explorations (pp. 199–225). Greenwich, CT: Information Age. Felfe, J. (2006). Validation of a German version of the “Multifactor Leadership Questionnaire” (MLQ Form 5x Short) by Bass and Avolio (1995). Zeitschrift für Arbeits und Organisationspsychologie, 50, 61–78. Felfe, J., & Petersen, L. (2007). Romance of leadership and management decision making. European Journal of Work and Organizational Psychology, 16, 1–24. Felfe, J., & Schyns, B. (2006). Personality and the perception of transformational leadership: The impact of extraversion, neuroticism, personal need for structure, and occupational self-efficacy. Journal of Applied Social Psychology, 36, 708–739. Fiedler, F. E. (1967). A theory of leadership effectiveness. New York: McGraw-Hill.

Francis, L. E. (1994). Laughter, the best mediation: Humor as emotion management in interaction. Symbolic Interaction, 17, 147–163. Frone, M. R., Adams, J., Rice, R. W., & Instone-Noonan, D. (1986). Halo error: A field study comparison of self- and subordinate evaluations of leadership process and leader effectiveness. Personality and Social Psychology Bulletin, 12, 454–461. Fukushige, A. & Spicer, D. P. (2007). Leadership preferences in Japan: An exploratory study. Leadership & Organization Development Journal, 28, 508–530. Fullagar, C., McCoy, D., & Shull, C. (1992). The socialization of union loyalty. Journal of Organizational Behavior, 13, 13–26. Fuller, J. B., Patterson, C. E. P., Hester, K., & Stringer, D. Y. (1996). A quantitative review of research on charismatic leadership. Psychological Reports, 78, 271–287. Gardner, W. L. & Martinko, M. J. (1996). Using the Myers-Briggs Type Indicator to study managers: A literature review and research agenda. Journal of Management, 22, 45–83. Gerstner, C. R., & Day, D. V. (1994). Cross-cultural comparison of leadership prototypes. Leadership Quarterly, 5, 121–134. Gerstner, C. R., & Day, D. V. (1997). Meta-analytic review of leader–member exchange theory: Correlates and construct issues. Journal of Applied Psychology, 82, 827–844. Giessner, S. R., & van Knippenberg, D. (2008). “License to fail”: Goal definition, leader group prototypicality, and perceptions of leadership effectiveness after leader failure. Organizational Behavior and Human Decision Processes, 105, 14–35. Gladwell, M. (2004, September 20). Personality plus. The New Yorker, 42–48. Gladwell, M. (2008, December 15). Most likely to succeed. The New Yorker, 36–42. Glass, J., Bengtson, V. L. & Dunham, C. C. (1986). Attitude similarity in three-generation families: Socialization, status inheritance, or reciprocal influence? American Sociological Review, 51, 685–698. Gouldner, A. W. (1987). Attitudes of “progressive” trade union leaders. American Journal of Sociology, 52, 389–393.

Fleishman, E. A. (1953). The description of supervisory behavior. Journal of Applied Psychology, 37, 1–6.

Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of leader– member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. Leadership Quarterly, 6, 219–247.

Fleming, P. (2005). Workers’ playtime? Boundaries and cynicism in a “Culture of Fun” program. The Journal of Applied Behavioral Science, 41, 285–303.

Gray, J. H., & Densten, I. L. (2007). How leaders woo followers in the romance of leadership. Applied Psychology: An International Review, 56, 558–581.

Fiorina, C. (2001, December 29). People of the year. The Financial Times, p. 8.

233

Barling, Christie, and Hoption

Hains, S. C., Hogg, M. A., & Duck, J. M. (1997). Selfcategorization and leadership: Effects of group prototypicality and leader stereotypicality. Personality and Social Psychology Bulletin, 23, 1087–1099. Harris, J. R. (1995). Where is the child’s environment? A group socialization theory of development. Psychological Review, 102, 458–489. Hartman, S. J., & Harris, O. J. (1992). The role of parental influences on leadership. The Journal of Social Psychology, 132, 153–167.

Copyright American Psychological Association. Not for further distribution.

Harvey, A. (2001). A dramaturgical analysis of charismatic leader discourse. Journal of Organizational Change Management, 14, 253–265. Harvey, S., Royal, M., Stout, D. (2003). Instructor’s transformational leadership: University student attitudes and ratings. Psychological Reports, 92, 395–402. Heinitz, K., Liepmann, D., & Felfe, J. (2005). Examining the factor structure of the MLQ: Recommendation for a reduced set of factors. European Journal of Psychological Assessment, 21, 182–190. Hersey, P., & Blanchard, K. (1988). Management of organization behavior: Utilizing human resources, 5th ed., Englewood Cliffs, NJ: Prentice-Hall. Herold, D. M., Fedor, D. B., Caldwell, S., & Liu, Y. (2008). The effects of transformational and change leadership on employees’ commitment to a change: A multilevel study. Journal of Applied Psychology, 93, 346–357. Hinkin, T. R. & Schriesheim, C. A. (2008). An examination of “nonleadership”: From laissez-faire leadership to leader reward omission and punishment omission. Journal of Applied Psychology, 93, 1234–1248. Hofmann, D. A., & Jones, L. M. (2005). Leadership, collective personality, and performance. Journal of Applied Psychology, 90, 509–522. Hofmann, D. A., & Morgeson, F. P. (1999). Safety-related behavior as a social exchange: The role of perceived organizational support and leader-member exchange. Journal of Applied Psychology, 84, 286–296. Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions, and organizations across nations (2nd ed.). Thousand Oaks, CA: Sage. Hogan, R., & Hogan, J. (2001). Assessing leadership: A view from the dark side. International Journal of Selection and Assessment, 9, 40–51. Hogg, M. A. (2001). A social identity theory of leadership. Personality and Social Psychology Review, 5, 184–200. Hogg, M. A., Fielding, K. S., Johnson, D., Masser, B., Russell, E., & Svensson, A. (2006). Demographic category membership and leadership in small groups: A social identity analysis. Leadership Quarterly, 17, 335–350. Hogg, M. A., Hains, S. C., & Mason, I. (1998). Identification and leadership in small groups: 234

Salience, frame of reference, and leader stereotypicality effects on leader evaluations. Journal of Personality and Social Psychology, 75, 1248–1263. Hoobler, J. M., & Brass, D. J. (2006). Abusive supervision and family undermining as displaced aggression. Journal of Applied Psychology, 91, 1125–1133. Hooijberg, R., & Choi, J (2000). From selling peanuts and beer in Yankee Stadium to creating a theory of transformational leadership: An interview with Bernie Bass. Leadership Quarterly, 11, 291–306. Hooker, C., & Csikszentmihalyi, M. (2003). Flow, creativity, and shared leadership: Rethinking the motivation and structuring of knowledge work. In C. L. Pearce & J. A. Conger (Eds.), Shared leadership: Reframing the hows and whys of leadership. Thousand Oaks, CA: Sage. Hoption, C., Phelan, J., & Barling, J. (2007). Transformational leadership in sport. In M. R. Beauchamp & M. A. Eys (Eds.), Group dynamics advances in sport and exercise psychology: Contemporary themes (pp. 45–60). Abingdon, England: Routledge. Houghton, J. D., & Neck, C. P. (2002). The revised selfleadership questionnaire: Testing a hierarchical factor structure for self-leadership. Journal of Managerial Psychology, 17, 672–691. Houghton, J. D., Neck, C. P., & Manz, C. C. (2003). Selfleadership and superleadership: The heart and art of creating shared leadership in teams. In C. L. Pearce & J. A. Conger (Eds.), Shared leadership: Reframing the hows and whys of leadership (pp. 123–140). Thousand Oaks, CA: Sage. House, R., & Howell, J. M. (1992). Personality and charismatic leadership. Leadership Quarterly, 3, 81–108. House, R., Javidan, M., Hanges, P., & Dorfman, P. (2002). Understanding cultures and implicit leadership theories across the globe: An introduction to project GLOBE. Journal of World Business, 37, 3–10. House, R. J. (1971). A path goal theory of leader effectiveness. Administrative Science Quarterly, 16, 321–339. House, R. J. (1977). A 1976 theory of charismatic leadership. In J. G. Hunt & L. L. Larsen (Eds.), Leadership: The cutting edge (pp. 189–207). Carbondale, IL: Southern Illinois University Press. House, R. J. (1996). Path–goal theory of leadership: Lessons, legacy and a reformulated theory. Leadership Quarterly, 7, 323–352. House, R. J. (2004). Illustrative examples of GLOBE findings. In R. J. House, P. J. Hanges, M. Javidan, P. W. Dorfman, & V. Gupta (Eds.), Culture, leadership and organizations: The GLOBE study of 62 societies (pp. 3–8). Thousand Oaks, CA: Sage. House, R. J., & Aditya, R. N. (1997). The social scientific study of leadership: Quo vadis? Journal of Management, 23, 409–473.

Copyright American Psychological Association. Not for further distribution.

Leadership

House, R. J., & Mitchell, T. R. (1974). Path-goal theory of leadership. Contemporary Business, 3, 81–98.

structure in leadership research. Journal of Applied Psychology, 89, 36–51.

House, R. J., & Podsakoff, P. M. (1994). Leadership effectiveness: Past perspectives and future directions for research. In J. Greenberg (Ed.), Organizational behavior: The state of the science (pp. 45–82). Hillsdale, NJ: Erlbaum.

Judge, T. A., Woolf, E. A., Hurst, C., & Livingston, B. (2008). Leadership. In J. Barling & C. Cooper (Eds.), Handbook of organizational behavior (pp. 334–352). Thousand Oaks, CA: Sage.

Howell, J. M., & Avolio, B. J. (1992). The ethics of charismatic leadership: Submission or liberation? The Executive, 6, 43.

Jung, D. I., & Avolio, B. J. (1999). Effects of leadership style and followers’ cultural orientation on performance in groups and individual task conditions. Academy of Management Journal, 47, 208–218.

Howell, J. M., & Hall-Merenda, K. E. (1999). The ties that bind: The impact of leader–member exchange, transformational and transactional leadership, and distance on predicting follower performance. Journal of Applied Psychology, 84, 680–694.

Jung, D. I., & Avolio, B. J. (2000). Opening the black box: An experimental investigation of the mediating effects of trust and value congruence on transformational and transactional leadership. Journal of Organizational Behavior, 21, 949–964.

Ilies, R., Gerhardt, M. W., & Le, H. (2004). Individual differences in leadership emergence: Integrating metaanalytic findings and behavioral genetics estimates. International Journal of Selection and Assessment, 12, 207–219.

Kane, T. D., Tremble, T. R., & Trueman, R. (2000). Transformational leadership effects at different levels of the army. Military Psychology 12, 137–150.

Hunter, S. T., Bedell-Avers, K. E., & Mumford, M. D. (2007). The typical leadership study: Assumptions, implications, and potential remedies. Leadership Quarterly, 18, 435–446. Jackson, S. L., Parker, C. P., & Dipboye, R. L. (1996). A comparison of competing models underlying responses to the Myers-Briggs Type Indicator. Journal of Career Assessment, 4, 99–115. Judge, T. A., & Bono, J. E. (2000). Five-factor model of personality and transformational leadership. Journal of Applied Psychology, 85, 751–765. Judge, T. A., Bono, J. E., Ilies, R., & Gerhardt, M. W. (2002). Personality and leadership: A qualitative and quantitative review. Journal of Applied Psychology, 87, 765–780. Judge, T. A., & Cable, D. M. (2004). The effect of physical height on workplace success and income: Preliminary test of a theoretical model. Journal of Applied Psychology, 89, 428–441. Judge, T. A., Colbert, A. E., & Ilies, R. (2004). Intelligence and leadership: A quantitative review and test of theoretical propositions. Journal of Applied Psychology, 89, 542–552.

Kark, R., Shamir, B., & Chen, G. (2003). The two faces of transformational leadership: Empowerment and dependency. Journal of Applied Psychology, 88, 246–255. Karnes, F. A., & D’Ilio, V. R. (1989). Student leaders’ and their parents’ perceptions of the home environment. Gifted Child Quarterly, 33, 165–168. Keller, R. T. (2006). Transformational leadership, initiating structure, and substitutes for leadership: A longitudinal study of research and development project team performance. Journal of Applied Psychology, 91, 202–210. Keller, T. (1999). Images of the familiar: Individual differences and implicit leadership theories. Leadership Quarterly, 10, 589–607. Kellett, P. (1999). Organisational leadership: Lessons from professional coaches. Sport Management Review, 2, 150–171. Kelley, R. E. (1992). The power of followership: How to create leaders people want to follow and followers who lead themselves. New York: Doubleday. Kelloway, E. K., & Barling, J. (1993). Member’s participation in local union activities: Measurement, prediction and replication. Journal of Applied Psychology, 78, 262–279.

Judge, T. A., LePine, J. A., & Rich, B. L. (2006). Loving yourself abundantly: Relationship of the narcissistic personality to self- and other perceptions of workplace deviance, leadership, and task and contextual performance. Journal of Applied Psychology, 91, 762–776.

Kelloway, E. K., & Barling, J. (2000). What we have learned about developing transformational leaders. Leadership and Organization Development Journal, 21, 355–363.

Judge, T. A., & Piccolo, R. F. (2004). Transformational and transactional leadership: A meta-analytic test of their relative validity. Journal of Applied Psychology, 89, 755–768.

Kelloway, E. K., Barling, J., & Helleur, J. (2000). Enhancing transformational leadership: the roles of training and feedback. Leadership and Organization Development Journal, 21, 145–149.

Judge, T. A., Piccolo, R. F., & Ilies, R. (2004). The forgotten ones? The validity of consideration and initiating

Kelloway, E. K., Mullen, J., & Francis, L. (2006). Divergent effects of transformational and passive 235

Barling, Christie, and Hoption

leadership on employee safety. Journal of Occupational Health Psychology, 11, 76–86. Kelloway, E. K., Sivanathan, N., Francis, L., & Barling, J. (2005). Poor leadership. In J. Barling & E. K. Kelloway (Eds.) Handbook of work stress (pp. 89–112). Thousand Oaks, CA: Sage. Kenney, R. A., Schwartz-Kenney, B. M., & Blascovich, J. (1996). Implicit leadership theories: Defining leaders described as worthy of influence. Personality and Social Psychology Bulletin, 22, 1128–1143.

Copyright American Psychological Association. Not for further distribution.

Kent, R. L., & Moss, S. E. (1994). Effects of sex and gender role on leader emergence. Academy of Management Journal, 37, 1335–1346. Kerr, S., & Jermier, J. M. (1978). Substitutes for leadership: Their meaning and measurement. Organizational Behavior and Human Performance, 22, 375–403. Kerr, S., Schreisheim, C. A., Murphy, C. J., & Stogdill, R. M. (1974). Toward a contingency theory of leadership based upon the consideration and initiating structure literature. Organizational Behavior and Human Performance, 12, 62–82. Khoo, H. S., & St. J. Burch, G. (2008). The ‘dark side’ of leadership personality and transformational leadership: An exploratory study. Personality and Individual Differences, 44, 86–97. Kirkpatrick, S. A., & Locke, E. A. (1996). Direct and indirect effects of three core charismatic leadership components on performance and attitudes. Journal of Applied Psychology, 81, 36–51. Klein, K. J., Ziegert, J. C., Knight, A. P., Xiao, Y. (2006). Dynamic delegation: Shared, hierarchical, and deindividualized leadership in extreme action teams. Administrative Science Quarterly, 51, 590–621. Klonsky, B. G. (1983). The socialization and development of leadership ability. Genetic Psychology Monographs, 108, 97–135. Koh, W. L., Steers, R. M., & Terborg, J. R. (1995). The effects of transformational leadership on teacher attitudes and student performance in Singapore. Journal of Organizational Behavior, 16, 319–333. Krajewski, H. T., Goffin, R. D., McCarthy, J. M., Rothstein, M. G., Johnston, N. (2006). Comparing the validity of structured interviews for managerial-level employees: Should we look to the past or focus on the future? Journal of Occupational and Organizational Psychology, 79, 411–432. Krosnick, J. A., & Alwin, D. F. (1989). Aging and susceptibility to attitude change. Journal of Personality and Social Psychology, 57, 416–425. Latham, G. P., Erez, M., & Locke, E. A. (1988). Resolving scientific disputes by the joint design of crucial experiments by the antagonists: Application to the Erez-Latham dispute regarding participation in goal setting. Journal of Applied Psychology, 73, 753–772. 236

Leach, M. M., & Harbin, J. J. (1997). Psychological ethics codes: A comparison of twenty-four countries. International Journal of Psychology, 32, 181–192. Lim, B. C., & Ployhart, R. E. (2004). Transformational leadership: Relations to the five-factor model and team performance in typical and maximum contexts. Journal of Applied Psychology, 89, 610–621. Ling, Y., Simsek, Z., Lubatkin, M. H., & Veiga, J. F. (2008). Transformational leadership’s role in promoting corporate entrepreneurship: Examining the CEO–TMT interface. Academy of Management Journal, 51, 557–576. Lord, R. G., & Maher, K. J. (1993). Leadership and information processing: Linking perceptions and performance. Boston, MA: Rutledge. Lord, R. G., de Vader, C. L., & Alliger, G. M. (1986). A meta-analysis of the relation between personality traits and leadership perceptions: An application of validity generalization procedures. Journal of Applied Psychology, 71, 402–410. Louhead, T. M., & Hardy, J. (2005). An examination of coach and peer leader behaviors in sport. Psychology of Sport and Exercise, 6, 303–312. Lowe, K. B., Kroeck, K. G., & Sivasubramaniam, N. (1996). Effectiveness correlates of transformation and transactional leadership: A meta-analytic review of the MLQ literature. Leadership Quarterly, 7, 385–425. Mackenzie, S. B., Podsakoff, P. M., & Jarvis, C. B. (2005). The problem of measurement model misspecification in behavioural and organizational research and some recommended solutions. Journal of Applied Psychology, 90, 710–730. Manz, C. C., & Sims, H. P. (1987). Leading workers to lead themselves: The external leadership of selfmanaging work teams. Administrative Science Quarterly, 32, 106–129. Martin, R. A., Puhlik-Doris, P., Larsen, G., Gray, J., & Weir, K. (2003). Individual differences in uses of humor and their relation to psychological well-being: Development of the Humor Styles Questionnaire. Journal of Research in Personality, 37, 48–75. Maslow, A. H. (1965). Eupsychian management: A journal. Homewood IL: Dorsey. Maume, D. J., Jr. (1999). Glass ceilings and glass escalators: Occupational segregation and race and sex differences in managerial promotions. Work and Occupations, 26, 483–509. McClane, W. E. (1991). The interaction of leader and member exchange characteristics in the leader– member exchange (LMX) model of leadership. Small Group Research, 22, 283–300. McCrae, R. R & John, O. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60, 174–214.

Leadership

McGue, M., & Bouchard, T. J. (1998). Genetic and environmental influences on human behavioral differences. Annual Review of Neuroscience, 21, 1–24. Meindl, J. R. (1995). The romance of leadership as a follower-centric theory: A social constructionist approach. Leadership Quarterly, 6, 329–341. Meindl, J. R. (1998). Appendix: Measure and assessments for the romance of leadership approach. In F. Dansereau & F. J. Yammarino (Eds.), Leadership: The multiple-level approaches—Part B: Contemporary and alternative (p. 199–302). Stamford, CT: JAI.

Copyright American Psychological Association. Not for further distribution.

Meindl, J. R., & Ehrlich, S. B. (1987). The romance of leadership and the evaluation of organizational performance. Academy of Management Journal, 30, 91–109. Meindl, J. R., Ehrlich, S. B., & Dukerich, J. M. (1985). The romance of leadership. Administrative Science Quarterly, 30, 78–102. Mitchell, M. S., & Ambrose, M. L. (2007). Abusive supervision and workplace deviance and the moderating effects of negative reciprocity beliefs. Journal of Applied Psychology, 92, 1159–1168. Morris, J. A., Brotheridge, C. M., & Urbanski, J. C. (2005). Bringing humility to leadership: Antecedents and consequences of leader humility. Human Relations, 58, 1323–1350. Morgan, B. B., Jr., Salas, E., & Glickman, A. (1993). An analysis of team evolution and maturation. Journal of General Psychology, 120, 277–291. Mullen, J., & Kelloway, E. K. (2009). Safety leadership: A longitudinal study of the effects of transformational leadership on safety outcomes. Journal of Occupational and Organizational Psychology, 2, 253–272. Nathan, B. R., & Lord, R. G. (1983). Cognitive categorization and dimensional schemata: A process approach to the study of halo in performance ratings. Journal of Applied Psychology, 68, 102–114. Neubert, M. J. (1999). Too much of a good thing or the more the merrier? Exploring the dispersion and gender composition of informal leadership in manufacturing teams. Small Group Research, 30, 635–646. Nguni, S., Sleegers, P., & Denessen, E. (2006). Transformational and transactional leadership effects on teachers’ job satisfaction, organizational commitment, and organizational citizenship behavior in primary schools: The Tanzanian case. School Effectiveness and School Improvement, 17, 145–177. Nye, J. S. (2008). The powers to lead. New York: Oxford University Press. Offermann, L. R., Kennedy, J. K., & Wirtz, P. W. (1994). Implicit leadership theories: Content, structure, and generalizability. Leadership Quarterly, 5, 43–58. Paunonen, S. V., Lönnqvist, J., Verkasalo, M., Leikas, S., & Nissinen, V. (2006). Narcissism and emergent

leadership in military cadets. Leadership Quarterly, 17, 475–486. Paunonen, S. V., Rothstein, M. G., & Jackson, D. N. (1999). Narrow reasoning about the use of broad personality measures for personnel selection. Journal of Organizational Behavior, 20, 389–405. Perry, M. L., Pearce, C. L., & Sims, H. P., Jr. (1999). Empowered selling teams: How shared leadership can contribute to selling team outcomes. Journal of Personal Selling and Sales Management, 19, 35–51. Peters, L. H., Hartke, D. D., & Pohlmann, J. T. (1985). Fiedler’s contingency theory of leadership: An application of the meta-analysis procedures of Schmidt and Hunter. Psychological Bulletin, 97, 274–285. Peterson, R. S., Smith, D. B., Martorana, P. V., & Owens, P. D. (2003). The impact of chief executive officer personality on top management team dynamics: One mechanism by which leadership affects organizational performance. Journal of Applied Psychology, 88, 795–808. Pfeffer, J. (1977). The ambiguity of leadership. Academy of Management Review, 2, 104–112. Pfeffer, J. (1998). The Human Equation: Building Profits by Putting People First. Boston: Harvard Business School Press. Pfeffer, J. (2007). In defense of the boss from hell. Business 2.0, 8, 70. Pfeffer, J., & Sutton, R. I. (2006). Hard facts, dangerous half-truths and total nonsense. Cambridge, MA: Harvard Business School Press. Phillips, D. J. (2005). Organizational genealogies and the persistence of gender inequality: The case of Silicon Valley law firms. Administrative Science Quarterly, 50, 440–472. Piccolo, R. F., & Colquitt, J. A. (2006). Transformational leadership and job behaviors: The mediating role of core job characteristics. Academy of Management Journal, 49, 327–340. Pillai, R. (1996). Crisis and the emergence of charismatic leadership in groups: An experimental investigation. Journal of Applied Social Psychology, 26, 543–562. Pillai, R., & Meindl, J. R. (1998). Context and charisma: A “meso” level examination of the relationship of organic structure, collectivism, and crisis to charismatic leadership. Journal of Management, 24, 643–671. Pillai, R., Schriesheim, C. A., & Williams, E. S. (1999). Fairness perceptions and trust as mediators for transformational and transactional leadership: A two-sample study. Journal of Management, 25, 897–933. Pittenger, D. J. (1993). The utility of the Myers-Briggs Type Indicator. Review of Educational Research, 63, 467–488. 237

Copyright American Psychological Association. Not for further distribution.

Barling, Christie, and Hoption

Platow, M. J., & van Knippenberg, D. (2001). A social identity analysis of leadership endorsement: The effects of leader ingroup prototypicality and distributive intergroup fairness. Personality and Social Psychology Bulletin, 27, 1508–1519.

Rubin, R. S., Munz, D. C., & Bommer, W. H. (2005). Leading from within: The effects of emotion recognition and personality on transformational leadership behavior. Academy of Management Journal, 48, 845–858.

Podsakoff, P. M., MacKenzie, S. B., & Bommer, W. H. (1996). Meta-analysis of the relationships between Jerr and Jermier’s substitutes for leadership and employee job attitudes, role perceptions, and performance. Journal of Applied Psychology, 81, 380–399.

Ryan, M. K., & Haslam, S. A. (2007). The glass cliff: Exploring the dynamics surrounding the appointment of women to precarious leadership positions. Academy of Management Review, 32, 549–572.

Popper, M., Mayseless, O., & Castelnovo, O. (2000). Transformational leadership and attachment. Leadership Quarterly, 11, 267–289.

Sahin, S. (2004). The relationship between transformational and transactional leadership styles of school principals and school culture. Kuram ve Uygulamada Egitim Bilimleri, 42, 387–396.

Rafferty, A. E., & Griffin, M. A. (2004). Dimensions of transformational leadership: Conceptual and empirical extension. Leadership Quarterly, 15, 329–354.

Salas, E., & Cannon-Bowers, J. A. (2001). The science of training: A decade of progress. Annual Review of Psychology, 52, 471–499.

Ralston, D., Holt, D., Terpstra, R., & Kai-Cheng, Y. (1997). The impact of national culture and economic ideology on managerial work values: A study of the United States, Russia, Japan, and China. Journal of International Business Studies, 28, 177–207.

Schaubroeck, J., Lam, S. S. K., & Cha, S. E. (2007). Embracing transformational leadership: Team values and the impact of leader behavior on team performance. Journal of Applied Psychology, 92, 1020–1030.

Redelmeier, D. A., Katz, J., & Kahneman, D. (2003). Memories of colonoscopy: A randomized trial. Pain, 104, 187–194. Richardson, H. M., & Hanawalt, N. G. (1952). Leadership as related to the Bernreuter personality measures: V. Leadership among adult women in social activities. Journal of Social Psychology, 36, 141–153.

Schriesheim, C. A., Castro, S. L., & Cogliser, C. C. (1999). Leader–member exchange (LMX) research: A comprehensive review of theory, measurement, and dataanalytic practices. Leadership Quarterly, 10, 63–113. Schriesheim, C. A., Castro, S. L., Zhou, X., & DeChurch, L. A. (2006). An investigation of path–goal and transformational leadership theory predictions at the individual level of analysis. Leadership Quarterly, 17, 21–38.

Roberts, B. W., Harms, P. D., Caspi, A., & Moffitt, T. E. (2007). Predicting the counterproductive employee in a child-to-adult prospective study. Journal of Applied Psychology, 92, 1427–1436.

Schriesheim, C. A., & Neider, L. L. (1996). Path–goal leadership theory: The long and winding road. Leadership Quarterly, 7, 317–321.

Roberts, N. C., & Bradley, R. T. (1988). Limits of charisma. In J. A. Conger & R. N. Kanungo (Eds.), Charismatic leadership: The elusive factor in organizational effectiveness (pp. 253–275). San Francisco: Jossey-Bass.

Schriesheim, C. A., Tepper, B. J., & Tetrault, L. A. (1994). Least preferred co-worker score, situational control, and leadership effectiveness: A meta-analysis of contingency model performance predictions. Journal of Applied Psychology, 79, 561–573.

Romano, A. (2007). Bracing for the gender neutral test. Newsweek, 150, 13.

Schyns, B., & Felfe, J. (2006). The personality of followers and its effect on the perception of leadership: An overview, a study, and a research agenda. Small Group Research, 37, 522–539.

Rosenthal, S. A., & Pittinsky, T. L. (2006). Narcissistic leadership. Leadership Quarterly, 17, 617–633. Rossette, A. S., Leonardelli, G. J., & Phillips, K. W. (2008). The White standard: Racial bias in leader categorization. Journal of Applied Psychology, 93, 758–777. Roush, P. E. & Atwater, L. E. (1992). Using the MBTI to understand transformational leadership and selfperception accuracy. Military Psychology, 4(1), 17–34. Rousseau, D. M. (2006). Is there such a thing as “evidencebased management”? Academy of Management Review, 31, 256–269. Rowold, J. (2006). Transformational and transactional leadership in martial arts. Journal of Applied Sport Psychology, 18, 312–325. 238

Schyns, B., & Sanders, K. (2007). In the eyes of the beholder: Personality and the perception of leadership. Journal of Applied Social Psychology, 37, 2345–2363. Seyranian, V., & Bligh, M. C. (2008). Presidential charismatic leadership: Exploring the rhetoric of social change. Leadership Quarterly, 19, 54–76. Shamir, B., & Eilam, G. (2005). “What’s your story?” A life-stories approach to authentic leadership development. Leadership Quarterly, 16, 395–417. Shamir, B., & Howell, J. M. (1999). Organizational and contextual influences on the emergence and effective-

Leadership

ness of charismatic leadership. Leadership Quarterly10, 257–283. Shamir, B., & Lapidot, Y. (2003). Shared leadership in the management of group boundaries: A study of expulsions from officers’ training courses. In C. L. Pearce & J. A. Conger (Eds.), Shared leadership: Reframing the hows and whys of leadership. Thousand Oaks, CA: Sage. Shao, L., & Webber, S. (2006). A cross-cultural test of the ‘five-factor model of personality and transformational leadership.’ Journal of Business Research, 59, 936–944.

Copyright American Psychological Association. Not for further distribution.

Shaw, H. (2008, February 15). The making of Lululemon: Superheroines inspired company’s founder to revolutionize women’s yoga wear. National Post, p. FP7. Shin, S. J., & Zhou, J. (2003). Transformational leadership, conservation, and creativity: Evidence from Korea. Academy of Management Journal, 46, 703–714. Simola, S., Barling, J., & Turner, N. (in press). Transformational leadership and moral orientation: Contrasting an ethic of justice and an ethic of care. Leadership Quarterly. Singer, M. S., & Singer, A. E. (1990). Situational constraints on transformational versus transactional leadership behaviour, subordinates’ leadership preference, and satisfaction. Journal of Social Psychology, 130, 385–396. Skarlicki, D. P., & Latham, G. P. (1996). Increasing citizenship behavior within a labor union: A test of organizational justice theory. Journal of Applied Psychology, 81, 161–169. Skarlicki, D. P., & Latham, G. P. (1997). Leadership training in organizational justice to increase citizenship behavior within a labor union: A replication. Personnel Psychology, 50, 617–633. Skogstad, A., Einarsen, S., Torsheim, T., Aasland, M. S., & Hetland, H. (2007). The destructiveness of laissezfaire leadership behavior. Journal of Occupational Health Psychology, 12, 80–92. Sparrowe, R. N. (2005). Authenticity and the narrative self. Leadership Quarterly, 16, 419–439. Spreitzer, G. (1995). Psychological empowerment in the workplace: Dimensions, measurement, and validation. Academy of Management Journal, 38, 1442–1465. Spreitzer, G. M., Perttula, K. H., & Xin, K. (2005). Traditionality matters: An examination of the effectiveness of transformational leadership in the United States and Taiwan. Journal of Organizational Behavior, 26, 205–227. Srivastava, A., Bartol, K. M., & Locke, E. A. (2006). Empowering leadership in management teams: Effects on knowledge sharing, efficacy, and performance. Academy of Management Journal, 49, 1239–1251. Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African

Americans. Journal of Personality and Social Psychology, 69, 797–811. Stogdill, R. M. (1950). Leadership, membership and organization. Psychological Bulletin, 47, 1–14. Stodgill, R. M. (1963). Manual for the leader behavior description questionnaire. Columbus: Bureau of Business Research, Ohio State University. Sullivan, A. (2008). Gender bender. Time, 171, 36. Sutton, R. I. (2007). The no arsehole rule: Building a civilized workplace and surviving one that isn’t. New York: Warner Books. Tannenbaum, S. I., Mathieu, J. E., Salas, E., & CannonBowers, J. A. (1991). Meeting trainees’ expectations: The influence of training fulfillment on the development of commitment, self-efficacy, and motivation. Journal of Applied Psychology, 76, 759–769. Tarnopol, L. (1958). Personality differences between leaders and non-leaders. Personnel Journal, 37, 57–60. Tejeda, M. J., Scandura, T. A., & Pillai, R. (2001). The MLQ revisited: Psychometric properties and recommendations. Leadership Quarterly, 12, 31–52. Tepper, B. J. (2000). Consequences of abusive supervision. Academy of Management Journal, 43, 178–190. Tepper, B. J. (2007). Abusive supervision in work organizations: Review synthesis, and research agenda. Journal of Management, 33, 261–289. Tepper, B. J., Duffy, M. K., Henle, C. A., & Lambert, L. S. (2006). Procedural injustice, victim precipitation, and abusive supervision. Personnel Psychology, 59, 101–123. Tepper, B. J., Duffy, M. K., & Shaw, J. D. (2001). Personality moderators of the relationship between abusive supervision and subordinates’ resistance. Journal of Applied Psychology, 86, 974–983. Tepper, B. J., Moss, S. E., Lockhart, D. E., & Carr, J. C. (2007). Abusive supervision, upward maintenance communication, and subordinates’ psychological distress. Academy of Management Journal, 50, 1169–1180. Tepper, B. J., & Percy, P. M. (1994). Structural validity of the Multifactor Leadership Questionnaire. Educational and Psychological Measurement, 54, 734–744. Terpstra, D. (1981). Relationship between methodological rigor and reported outcomes in organization development evaluation research. Journal of Applied Psychology, 66, 541–543. Towler, A. (2005). Charismatic leadership development: Role of parental attachment style and parental psychological control. Journal of Leadership and Organizational Studies, 11(4), 15–25. Townsend, J., Phillips, J. S., & Elkins, T. J. (2000). Employee retaliation: The neglected consequence of poor leader-member exchange relations. Journal of Occupational Health Psychology, 5, 457–463. 239

Barling, Christie, and Hoption

Tucker, S., Turner, N., Barling, J., & McEvoy, M. (in press). Transformational leadership and children’s aggression in team setting: A short-term, longitudinal study. Leadership Quarterly. Tucker, S., Turner, N., Barling, J., Reid, E. M., & Elving, C. (2006). Apologies and transformational leadership. Journal of Business Ethics, 63, 195–207.

Copyright American Psychological Association. Not for further distribution.

Turner, N., Barling, J., Epitropaki, O., Butcher, V., & Milner, C. (2002). Transformational leadership and moral reasoning. Journal of Applied Psychology, 87, 304–311. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987). Rediscovering the social group: A self-categorization theory. Oxford, England: Blackwell. Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4–70. van Dierendonck, D., Haynes, C., Borrill, C., & Stride, C. (2004). Leadership behavior and subordinate wellbeing. Journal of Occupational Health Psychology, 9, 165–175. van Honk, J., & Schutter, D. J. L. G. (2007). Testosterone reduces conscious detection of signals serving social correction: Implications for antisocial behavior. Psychological Science, 18, 663–667. van Knippenberg, D., De Cremer, D., & van Knippenberg, B. (2007). Leadership and fairness: The state of the art. European Journal of Work and Organizational Psychology, 16, 113–140. van Knippenberg, D., Lossie, N., & Wilke, H. (1994). In-group prototypicality and persuasion: Determinants of heuristic and systematic message processing. British Journal of Social Psychology, 33, 289–300. van Knippenberg, B. & van Knippenberg, D. (2005). Leader self-sacrifice and leadership effectiveness: The moderating role of leader prototypicality. Journal of Applied Psychology, 90, 25–37. van Knippenberg, D., van Knippenberg, B., & van Dijk, E. (2000). Who takes the lead in risky decision mak-

240

ing? Effects of group members’ risk preferences and prototypicality. Organizational Behavior and Human Decision Processes, 84, 213–234. Waldman, D. A., Ramírez, G. G., House, R. J., & Puranam, P. (2001). Does leadership matter? CEO leadership attributes and profitability under conditions of perceived environmental uncertainty. Academy of Management Journal, 44, 134–143. Wang, H., Law, K. S., Hackett, R. D., Wang, D., & Chen, Z. X. (2005). Leader–member exchange as a mediator of the relationship between transformational leadership and followers’ performance and organizational citizenship behavior. Academy of Management Journal, 48, 420–432. Weber, M. (1947). The theory of social and economic organizations (T. Parsons, Trans.). New York: Free Press. Westwood, R. (2004). Comic relief: Subversion and catharsis in organizational comedic theatre. Organization Studies, 25, 775–795. White, R. E., Thornhill, S., & Hampson, E. (2006). Entrepreneurs and evolutionary biology: The relationship between testosterone and new venture creation. Organizational Behavior and Human Decision Processes, 100, 21–34. Whittington, J. L., Goodwin, V. L., & Murray, B. (2004). Transformational leadership, goal difficulty, and job design: Independent and interactive effects on employee outcomes. Leadership Quarterly, 15, 593–606. Wofford, J. C., & Liska, L. Z. (1993). Path–goal theories of leadership: A meta-analysis. Journal of Management, 19, 857–876. Yovetich, N. A., Dale, J. A., & Hudak, M. A. (1990). Benefits of humor in reduction of threat-induced anxiety. Psychological Reports, 66, 51–58. Zacharatos, A., Barling, J., & Kelloway, E. K. (2000). Development and effects of transformational leadership in adolescents. Leadership Quarterly, 11, 211–226. Zellars, K. L., Tepper, B. J., & Duffy, M. K. (2002). Abusive supervision and subordinates’ organizational citizenship behavior. Journal of Applied Psychology, 87, 1068–1076.

CHAPTER 8

ENTREPRENEURSHIP: THE GENESIS OF ORGANIZATIONS

Copyright American Psychological Association. Not for further distribution.

Robert A. Baron and Rebecca A. Henry

Do you ever purchase books or other items from Amazon.com? Drink coffee at Starbucks? Buy or sell on eBay? Use Microsoft products in your work (e.g., Word, Excel, PowerPoint)? Eat at Panera Bread? Shop at Wal-Mart? If so, you probably realize that none of these companies existed (or at least were not large or well known) 20 or 25 years ago. This fact underscores a key point with which we wish to begin: Entrepreneurs—individuals who launch new ventures to exploit business opportunities, often by developing new products or services (Shane & Venkataraman, 2000)—do not merely create wealth for themselves and their close associates; they also provide a major engine of economic growth and development for local communities and, sometimes, entire societies. For example, in the United States, firms with fewer than 500 employees (a large proportion of which are relatively young businesses) account for 51% of private-sector output and constitute 99% of all employers (Small Business Association, 2001). From the perspective of industrial and organization (I/O) psychology, entrepreneurs are of key importance for another basic reason: All businesses are new ventures at one point in time, so studying them provides an invaluable window into the origins and early development of organizations and of crucial organizational processes (Baum, Frese, & Baron, 2007). It is also important to note that by bringing new products or services to market, entrepreneurs change the lives of many millions of persons who come to rely on these products or services in their daily lives and work. For instance, currently, many young people below the age of 30 would find

life without their “smart” phones or MP3 players to be almost inconceivable. These points lead us to a key issue we wish to consider before proceeding further: Why, in essence, should a separate chapter on entrepreneurship be included in this important handbook? From the perspective of mainstream I/O psychology, it could be noted that entrepreneurs are simply a particular (albeit distinct) occupational group and that other occupational groups (e.g., executives, accountants) are not considered in separate chapters. Similarly, entrepreneurial firms (new ventures, as they are usually termed) can be perceived simply as one specific type of organization. True, they are smaller, newer, and typically have less formal management and human resources systems in place than the organizations in which most I/O research is conducted. However, as just one of many types of organizations, why are they worthy of special attention? Although we fully appreciate the reasonable nature of such suggestions, we believe that there are, indeed, strong grounds for the presence of this chapter—and for linking entrepreneurship closely to mainstream I/O psychology. First, and most important, choosing to view entrepreneurs and the new ventures they create simply as specific occupations or types of organizations overlooks or minimizes the importance of several truly unique elements of the process of new venture creation. Nearly half of this chapter (and well over half of the research cited) focuses on aspects of the new venture creation process that are seldom the focus of I/O research—topics such as opportunity

http://dx.doi.org/10.1037/12169-008 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

241

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

recognition and acquiring essential resources (both human and financial). In essence, the field of entrepreneurship focuses on the process through which new ventures are created and subsequently grow (or fail), whereas I/O research, in contrast, generally focuses on individual behavior and processes occurring in organizations that already exist and often have a long history of active operation. Metaphorically, one can view this difference of perspective as parallel to that between giving birth and being a parent. To the extent these are distinct processes (and we suggest that, to a degree, they are), then entrepreneurship may well merit separate treatment in this volume because, in a sense, it is the source from which organizations and the processes that occur within them originate (hence the subtitle of this chapter). Another strong justification for including a chapter on entrepreneurship in this handbook lies in the fact that the field of entrepreneurship offers the potential for strong contributions to theory development in I/O psychology. Knowledge of how specific individuals develop ideas for new and useful products or services and then actually convert these ideas into profitable companies may offer valuable insights into several basic processes and topics long of interest to I/O psychologists, processes such as motivation, leadership, organizational commitment, influence, team development and functioning, and conflict, to mention just a few (e.g., Cummings, 2007; Vecchio, 2003). Links to these processes are made explicit throughout this chapter. Finally, knowledge of the entrepreneurial process can offer insights into effective techniques for training potential entrepreneurs, providing them with the skills and knowledge they need to achieve the goals they seek. It addition, it may offer information useful to accurately predict which new ventures will succeed and which will quickly disappear. Such knowledge would link closely to central themes in I/O psychology (e.g., selection, training) and could, therefore, help to enrich these active areas of theory and research. To provide broad coverage of the rapidly growing field of entrepreneurship, we proceed as follows. First, an initial section offers a definition of entrepreneurship both as a field of scholarly research and as a key business activity, plus a very brief overview of 242

the intellectual foundations of this field—its roots in existing, older disciplines. As part of this discussion, we also examine the process perspective in entrepreneurship—the view that this activity is a continuing process that changes and develops continuously over relatively long periods of time (e.g., Shane, 2003). Next, we turn to entrepreneurial motivation and the key question of why some individuals make the decision to “take the plunge,” accepting the high levels of risk often associated with launching new companies. In fact, these risks are substantial: 80% to 85% of all new ventures disappear within 3 years of their founding (Shane, 2003). Although a complete answer to the basic question Why do entrepreneurs do it? is not yet available, ongoing research offers many insights into the motives behind such actions (e.g., Locke & Baum, 2007). Third, we examine a key aspect of entrepreneurship—opportunity recognition. In a sense, this is the core activity in the new venture creation process—the font from which the entire process flows—because entrepreneurs almost always found new companies to develop specific opportunities they have recognized or discovered. Basic psychological processes of perception appear to play an important role in this key aspect of entrepreneurship (e.g., Baron, 2006), but many other variables (e.g., the size and scope of entrepreneurs’ social networks; Ozgen & Baron, 2007) play a crucial role. Once entrepreneurs decide to proceed—that is, to take overt action to start new companies—they must identify, and then obtain, the specific resources needed for this activity. These include, among others, human resources (e.g., appropriate partners, employees) and financial resources (i.e., the capital funds necessary to proceed). How do entrepreneurs obtain these resources, and what personal skills and characteristics are relevant in this respect? After this, we turn briefly to the question of whether, and in what ways, entrepreneurs differ from other persons. Almost by definition, entrepreneurs are not “typical” individuals. The initial act of launching a new venture suggests that they may differ in important ways from most persons, who choose the security and predictability of working for existing organizations. Recent research—much of it based on theories and measures developed by I/O psychologists—has addressed this issue, with the

Entrepreneurship

Copyright American Psychological Association. Not for further distribution.

result providing new insights into the personal characteristics of entrepreneurs (e.g., their skills, attitudes, values, motives) and how these relate to their success (e.g., Rauch & Frese, 2006). Following this, we present a brief discussion of the issues involved in measuring entrepreneurial success and, finally, close with suggestions for future research. Overall, we strive to provide a broad and multidisciplinary overview of the process through which entrepreneurs create and operate viable new companies through vigorous application of their ideas, skills, knowledge, and talents (Baum et al., 2007). ENTREPRENEURSHIP: A FIELD AND A BUSINESS ACTIVITY Defining any field, especially one that is relatively new, is a difficult task. Entrepreneurship, which has emerged as an independent field only in the past 20 or 25 years, is no exception to this rule. Thus, it is not surprising that at present, there is no single, universally accepted definition of its scope or mission. With that point clearly stated, however, we note that a definition offered by Shane and Venkataraman (2000) has gained increasing acceptance. This definition, slightly paraphrased, suggests that entrepreneurship, as a field, seeks to understand how opportunities to create something new (e.g., new products or services, new markets, new production processes or uses of raw materials, new ways of organizing existing technologies) arise and are discovered (or created) by specific persons, who then use various means (especially, launching new business ventures) to exploit or develop them, thus producing a wide range of effects. By implication, this definition suggests that entrepreneurship, as a business activity, involves several key actions: identifying an opportunity—one that is potentially valuable in economic terms (i.e., it is capable of yielding profits)—obtaining resources essential to developing this opportunity, and then participating in a broad range of activities involved in actually exploiting or developing this opportunity (e.g., launching a new venture, obtaining intellectual property protection, developing effective business models and strategies, operating the new company so as to achieve growth and profitability; Baron & Markman, 2005). Further, this definition

implies that entrepreneurs are individuals who engage in these activities, typically by launching new ventures. We adopt this general framework in the present chapter because the definition offered by Shane and Venkataraman encompasses both the academic discipline of entrepreneurship and the business activity this field seeks to study. Although this definition appears to be very useful, it has been extended and clarified by McMullen and Shepherd (2006), who noted that in essence, entrepreneurship involves two key phases or activities. In the first, individuals (potential entrepreneurs) use their existing knowledge and skills to recognize that an opportunity exists. In the second, they evaluate this opportunity to determine whether they, personally, possess the knowledge and skills needed to actually exploit it (e.g., by obtaining the resources required for this task). In other words, they try to determine whether the opportunity is one that they, as individuals, can actively pursue. These extensions are fully consistent with the definition offered by Shane and Venkataraman (2000) but help to emphasize the importance of individual motivation, skills, and knowledge in entrepreneurial action, plus the additional point (to which we return in a later section) that recognizing an opportunity is often quite distinct from—and only preliminary to—actually doing something concrete about it. Before turning to other topics, we should note, briefly, that recognizing opportunities for developing something new can occur within existing organizations as well as outside of them (Lumpkin, 2007; see also chap. 9, this volume). In fact, many successful companies are deeply concerned with encouraging innovation and take active steps to provide an environment in which it can flourish. Individuals who respond to such environments and behave like entrepreneurs within an existing company are often termed intrapreneurs. They recognize opportunities and take various actions to develop them but do so inside the boundaries of an existing company rather than by founding a new venture. Unfortunately, they often face formidable barriers or obstacles because not all organizations are equally committed to or accepting of innovation. As noted by many authors, however, innovation is truly 243

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

essential for gaining and sustaining competitive advantage, so it is something all organizations should seek (Rauch, Wiklund, Frese, & Lumpkin, 2005). Thus, although the focus of this chapter is firmly on entrepreneurs and their efforts to launch new ventures, it is important to note that individuals can, and often do, act entrepreneurially within large, existing companies. Indeed, growing evidence suggests that it is precisely such companies that tend to gain and retain competitive advantage in their respective markets (e.g., Lumpkin, 2007). One additional point that should be noted at the outset is that entrepreneurs are not in any sense a homogeneous group. Rather, individuals choose to launch new ventures for many different reasons, ranging from the desire to obtain increased autonomy, task significance, and task completion (e.g., Baron, in press) to the necessity of finding employment because their current jobs have vanished in economic downturns, downsizing, or related events and trends. We should also note that although a large majority of entrepreneurs are men (only 27% of business owners are women; Shane, 2008), a growing number of women are currently choosing this career path (Baron, Markman, & Hirsa, 2001; Eddleston & Powell, 2008). Similarly, the representation of minorities among the ranks of entrepreneurs, too, has increased somewhat in recent years. However, in the United States, African Americans are only about half as likely as European Americans to start their own businesses (Shane, 2008). Many entrepreneurs are relatively young (in their 20s or 30s), but a sizeable group launch new ventures during mid-life, especially when declining economic conditions result in loss of what appeared to be secure, long-term employment (Shane, 2003).

The Field of Entrepreneurship: Its Roots in Existing Disciplines All domains of scholarly activity derive from existing ones, and entrepreneurship is no exception to this general rule. Entrepreneurship, as a field of inquiry, has important roots in several fields that, together, contributed to its development. Consider, again, our definition of entrepreneurship as a scholarly domain—a field that seeks to understand how opportunities to create new products or services 244

arise and are discovered by specific persons, who then use various means to develop them. This definition implies that to understand entrepreneurship as a process—and as an activity in which entrepreneurs engage—it is essential to consider (a) the economic, technological, governmental, and social conditions from which opportunities arise; (b) the people who recognize these opportunities (entrepreneurs) and their various characteristics; (c) the business methods and legal structures they use to develop opportunities; and (d) the economic and social effects that stem from such development. All of these elements play a role in entrepreneurship, and all must be carefully considered if we are to fully understand this complex process. This, in turn, implies that the field of entrepreneurship is closely linked to older disciplines such as economics, behavioral science (psychology, I/O psychology, cognitive science), and sociology. The findings and principles of these disciplines helped the field of entrepreneurship to pose its most basic questions (e.g., How and when do new business opportunities arise? Why do some persons but not others recognize these opportunities? What are the economic and social effects of new ventures on their local communities and on regional or national economies? Why do some new ventures succeed whereas others—the vast majority—fail?). These questions—as well as methods for answering them and theoretical frameworks for interpreting the findings of such research—derive from the basic disciplines that underlie entrepreneurship as an academic field. It is interesting to note that when the intellectual roots of entrepreneurship are traced to the early 20th century (the period during which, most scholars agree, it first emerged), its dual roots in economics and psychology become readily apparent. Economists such as Schumpeter (1911) and Knight (1921) were interested in the question of how economic growth occurs and quickly reached the conclusion that individuals—entrepreneurs—are the agents or focal points of both change and growth. Although economists did not typically focus on the actions of individuals (this situation has changed considerably in recent years), it was clear to many theorists that to fully understand the foundations of economic growth, attention to some aspects of individual

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

behavior was essential. In other words, influential economists adopted some aspects of a “great person” theory of entrepreneurship, concluding that specific persons who possessed certain characteristics played a key role in this regard. It remained for psychologists, with their greater sophistication concerning human characteristics, skills, and motives, to describe these characteristics and relate them to important forms of behavior. Such research was begun by David McClelland (1961), whose work on the need for achievement was crucial in linking economic growth to human behavior and especially to achievement motivation and related factors. Together, these two streams of research—one from economics and the other from psychology—combined to encourage early efforts to identify the key or central characteristics of entrepreneurs (e.g., Mayer & Goldstein, 1964). Although this work continued for several decades, it generally failed to yield conclusive or lucid findings, primarily because it was not based on carefully derived theoretical frameworks and often used measures of individual characteristics (i.e., personality) that were relatively low in reliability and validity (e.g., Shaver & Scott, 1991). These disappointing results led to a sharp decrease in the influence of the psychological perspective in the field of entrepreneurship—a decline that was only reversed in the late 1990s, as the scope of such research was extended beyond the search for central personality variables to the role of cognitive and social factors and processes in entrepreneurship (e.g., Baum et al., 2007). This renewed interest in applying a psychological perspective to entrepreneurship has led to rapid progress on many fronts and is fully represented in this chapter. In short, the modern field of entrepreneurship has important roots in economics, psychology, and, more recently, several additional fields. For instance, sociology has provided important insights into the importance and formation of entrepreneurs’ social networks as well as the key benefits they derive from these networks (e.g., Aldrich & Kim, 2007; Aldrich, & Ruef, 2006). In a sense, then, entrepreneurship as a field is truly a hybrid that adopts the methods, theories, and levels of analysis (micro, macro) most appropriate to the specific issues and topics it addresses. We call attention to

this fusion of different perspectives at several points in the present chapter but focus primarily on the micro–behavioral science perspective in entrepreneurship and the ways in which I/O psychology has enriched (and continues to enrich) the development of this relatively young field.

Entrepreneurship: A Process Perspective As defined by Shane and Venkataraman (2000), entrepreneurship is definitely a continuing process and not a single event or action. This process involves many actions and does not unfold in a simple or straightforward manner (Shane, 2003). However, there is general agreement in the field that it involves a number of crucial steps that occur, although not necessarily in any fixed order. We have mentioned several of these previously, but it is useful to recapitulate them here to clearly illustrate the nature of the emerging process perspective: (a) the emergence of opportunities (which, in turn, derive from changing economic, technological, and social conditions); (b) the recognition of these opportunities by specific persons; (c) the evaluation of the economic potential of these opportunities coupled with an active decision to pursue or not pursue them; (d) the attainment of required resources (human, financial, informational); (e) the development of a strategy or business model for using these resources to exploit the opportunity; and (f) the actual exploitation, which involves launching and operating a new organization—one that can take many different legal forms. Note again that these steps do not necessarily occur in this order; for instance, exploitation of an identified opportunity may occur simultaneously with strategy development and refinement in a recursive manner. Although each of these activities is multifaceted, for purposes of both theory development and research, they are often viewed as occurring in three major phases of the overall process: (a) prelaunch phase (identification of opportunities, initial evaluation of them), (b) launch phase (e.g., choosing a legal form for the new venture, obtaining intellectual property protection, securing essential financial resources), and (c) postlaunch phase (e.g., developing a customer base, refining and improving products). Obviously, so many activities are involved in 245

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

this process that any single model of it must, necessarily, represent an oversimplification. However, the model shown in Figure 8.1 is consistent with many current views of entrepreneurship as a process and is offered here primarily because it provides a framework for further, detailed discussion of key topics. It should also be noted that during each of these phases, many variables influence the activities being performed (e.g., Katz & Shepherd, 2003). Despite their great diversity, however, these variables can be viewed as falling under three major headings: individual-level variables, group- or interpersonallevel variables, and societal-level (i.e., macrolevel) variables. Included among individual-level variables are the cognitions, knowledge, skills, and characteristics of entrepreneurs (Baum & Locke, 2004). Group- or interpersonal-level variables encompass entrepreneurs’ social networks (e.g., Aldrich & Kim, 2007), their social capital (the benefits they derive from such networks; Nahapiet & Ghoshal, 1998),

Prelaunch Phase IndividualLevel Variables

GroupLevel Variables

Postlaunch Phase

Launch Phase SocietalLevel Variables

IndividualLevel variables

Sample Activities Identification of opportunities Initial opportunity evaluation Assembly of required resources Gathering pertinent information

GroupLevel Variables

SocietalLevel Variables

Sample Activities Choosing legal form of new venture Obtaining intellectual property protection Developing initial business model and strategies

Dependent Measures

Dependent Measures

Number and quality of opportunities identified Capital raised Success in attracting highquality partners, employees

Time until first sale Time until “break-even” Time until first employee hired Number, strength of patents acquired

FIGURE 8.1. Process model of entrepreneurship. 246

and many factors that influence group performance (Kozlowski & Ilgen, 2006). Finally, societal-level (macro) variables include industry-related factors, such as environmental dynamism (the pace of change in a given industry), technological advances, government policies, economic and financial markets, and various actions and strategies adopted by competitors (e.g., Patzelt, Shepherd, Deeds, & Bradley, 2008). Variables in the first two categories are ones of major interest to I/O psychologists, but in recent years there has been growing interest in understanding the interface between individualand group-level variables on the one hand and macrolevel (environmental, societal) variables on the other (e.g., Hmieleski & Baron, 2009). These factors represent only a small sample of the large number of variables that have been found to play a role in the entrepreneurship process; they are offered here primarily to illustrate the vast scope of these variables.

IndividualLevel Variables

GroupLevel Variables

SocietalLevel Variables

Sample Activities Building customer base Hiring key employees Improving product design Conducting negotiations Influencing, motivating others

Dependent Measures Financial Measures (growth in sales, earnings, number of employees; value of initial public offering) Success in Obtaining Required Resources Attitudinal Measures (e.g., personal and life satisfaction) Measures of Entrepreneurs’ Personal Health and WellBeing

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

Two final points should be noted. First, a corollary of the process perspective on entrepreneurship outlined earlier and shown in Figure 8.1 suggests that because entrepreneurs focus on somewhat different activities during phases of new venture development, the relative impact of different variables may change appreciably from one phase to the next. For instance, consider the impact of one variable long viewed as relevant to entrepreneurship: a propensity to accept risk (or engage in “risky” behavior). Many studies have been performed to investigate the role of this variable in entrepreneurship, and results have been highly variable (e.g., Begley & Boyd, 1987; Simon, Houghton, & Aquino, 2000). In fact, two careful meta-analyses published in the Journal of Applied Psychology (Miner & Raju, 2004; Stewart & Roth, 2001) reached very different conclusions concerning the role of risk taking in entrepreneurship. The first (Stewart & Roth, 2001) concluded that entrepreneurs are indeed more accepting of risk than other persons (effect size = .30), whereas the second (Miner & Raju, 2004) reached the opposite conclusion (effect size = −.42). Although many differences between the two analyses may account for these contrasting conclusions, another explanation closely linked to the process perspective and its corollary is as follows: Perhaps the propensity to accept risk has different effects during different phases of new venture creation. Early (during the prelaunch or launch phases) acceptance of relatively high levels of risk is virtually required; entrepreneurs must accept high levels of risk to get started—to launch a new company. They often underestimate such risks, but it is clear that the risks are present (e.g., Ariely, 2008). In contrast, during later (postlaunch) phases, conserving and stretching existing resources often becomes crucial, with the result that entrepreneurs (and certainly successful ones) strive actively to limit or manage risk. This basic point—that overall and relative importance of certain variables may shift across various phases of the entrepreneurial process—has been increasingly recognized in ongoing research (e.g., Lévesque, Shepherd, & Douglas, 2002) and, of course, comes as no surprise to readers of this chapter.

Second, it is clear that new ventures that survive and prosper ultimately become mature organizations—ones with an established structure, past history, relatively clear business strategies, and recognized norms governing behavior within them. Although no single point in time can be identified as the dividing line between new ventures and mature organizations, most researchers in the field of entrepreneurship consider the first 3 to 4 years of a new venture’s experience as the central focus of their domain (e.g., Shane, 2003). In sum, modern views of entrepreneurship perceive it as a continuing, ever-evolving process and one, moreover, that can only be fully understood through a multilevel perspective that considers individual-, group-, and macro- (societal-) level variables simultaneously, as well as the complex interactions between them (Hitt, Beamish, Jackson, & Mathieu, 2007). In the sections that follow, we turn to reviews of existing evidence concerning key aspects of the entrepreneurial process: entrepreneurial motivation, opportunity recognition, and resource acquisition. ENTREPRENEURIAL MOTIVATION Why do individuals launch new ventures? This has long been a key question in the field of entrepreneurship, and for good reason. The risks, effort, and uncertainty involved in launching a new venture are all great, and the rewards, if any, are far from guaranteed (e.g., Cooper, Woo, & Dunkelberg, 1988). In view of these facts, it is reasonable to suggest that the persons who choose this activity have strong motives for doing so. However, what, precisely are these motives, and how do they influence not only the decision to become an entrepreneur but also the subsequent survival and success of new ventures? Many researchers have examined these issues, but one framework for understanding them that is especially useful from the point of view connecting such work closely to theory and research in I/O psychology has been proposed by Locke and Baum (2007). These scholars noted that motivation, which energizes, directs, and sustains action, is truly a central concept in entrepreneurship for, as they succinctly put it, “A person may have sufficient technical skill 247

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

and money to start a business, but without motivation [to do so] nothing happens” (Locke & Baum, 2007, p. 93). In the context of entrepreneurship, then, motivation refers to efforts directed toward attaining entrepreneurial goals—recognition and exploitation of new business opportunities. As Locke and Baum (2007) further noted, motivation cannot be considered independent of cognition, for knowledge (information) in the absence of motivation leads nowhere, whereas motivation in the absence of knowledge leads to random or unproductive action (Locke, 2000). Several major theories of motivation have been applied to understanding entrepreneurs’ decision to take the plunge and start new ventures. (See also Vol. 3, chap. 3, this handbook.) For instance, Gartner, Bird, and Starr (1991) suggested that expectancy theory is useful in this context (e.g., Vroom, 1964). They noted that one reason entrepreneurs choose to start new ventures is that they anticipate a close link between effort and performance (i.e., expectancies are strong). Similarly, they believe that excellent performance on the tasks involved in starting a new business (e.g., success in attaining financial resources) will result in positive outcomes—the rapid growth and rising profits they seek (i.e., instrumentalities, too, are strong). Finally, persons who choose to become entrepreneurs find the excitement and uncertainty involved in launching and operating new ventures to be enjoyable; so for them, the third component of expectancy theory—valence—also is positive. Together, the combination of high expectancy, instrumentality, and valence generate high levels of entrepreneurial motivation. Goal setting also offers important insights into such motivation. As noted by Locke and Latham (2002), specific, challenging goals (assuming the presence of adequate commitment, feedback, and knowledge) result in high levels of performance. Extending this basic principle to entrepreneurship, Locke and Baum (2007) proposed that persons who decide to start new ventures are ones who set especially high goals in this respect—higher than persons who do not decide to take this action. Thus, their motivation to exploit the business opportunities they have identified is also strong. Other approaches have also helped clarify the complex nature of entrepreneurial motivation. For 248

instance, proceeding from an economic perspective, Choi, Lévesque, and Shepherd (2008) proposed that entrepreneurs shift from exploring business opportunities (obtaining information about them and their feasibility) to exploitation—actually developing them—at the point where they have acquired sufficient knowledge to assess the viability of the opportunity but also realize that (a) the costs of acquiring further knowledge are rising sharply and (b) prospective competitors, too, may benefit from this growing knowledge, thus reducing the potential value of the opportunity to the initial entrepreneur. Translated into motivational terms (specifically, basic tenets of expectancy theory and Naylor Pritchard Ilgen theory; Naylor, Pritchard, & Ilgen, 1980), entrepreneurs should choose to proceed when they reach the point at which they recognize that further effort (i.e., gathering additional knowledge) will not significantly improve their performance or the success of their new ventures and that delay will actually reduce the likelihood or magnitude of such success (Choi et al., 2008). Finally, it has recently been proposed that individuals often become entrepreneurs because they are seeking the kind of working conditions identified by research on job design as ones that generate high levels of motivation, satisfaction, and performance (e.g., Baron, in press; also see chap. 13, this volume). In other words, they become entrepreneurs because they are seeking meaningfulness in their work, as provided by high levels of autonomy, task significance, task identity, and skill variety. Thus, although extant theories of motivation developed by I/O psychologists certainly apply to entrepreneurs and offer important insights into why they assume the risks involved in launching new ventures, entrepreneurs may constitute a unique occupational group in one sense: Their career choices may reflect an active and vigorous search for the “perfect job” or career—one that provides exceptionally high levels of personal meaningfulness. In addition to efforts to apply widely accepted theories of motivation to entrepreneurship, a considerable amount of empirical research has adopted a different strategy, focusing instead on identifying specific situational and personal variables that play a

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

role in the decision to launch a new venture. In other words, there has been continuing interest in identifying personal characteristics that may predispose individuals to become entrepreneurs or enable them to continue in this role even in the face of major setbacks and adversity. With respect to situational factors, existing evidence suggests that markets, government policies, and industry characteristics all play a role (e.g., Shane, 2003, 2008). Specifically, individuals are more likely to vigorously pursue entrepreneurial actions when financial markets are favorable, thus providing ready access to essential financial resources when government policies favor the founding and operation of new ventures (e.g., through favorable tax policies; Tang, Walls, & Frese, 2007) and when markets for specific new products or services are growing rapidly (Shane, 2008). These factors are all closely related to opportunity recognition (a topic we consider in detail in the following section), such that the potential economic value of identified opportunities is influenced by these situational conditions. For instance, consider one company, Second Time Around, which was founded in 2005. The entrepreneurs who founded this business observed that a rapidly growing number of individuals were being married for the second or even third time and often wanted help in planning their weddings. The needs of such persons, however, were radically different from those of younger individuals getting married for the first time. However, no existing business served these needs. Recognizing this fact—and the business opportunity it provided—these entrepreneurs launched a company to fill this niche. They quickly obtained considerable success, and this encouraged many other entrepreneurs to start similar, competing companies. Several individual-level factors, too, have been found to be closely linked to entrepreneurial motivation, thus confirming the informal observation that individuals who choose to become entrepreneurs do indeed differ, in measurable ways, from persons who do not pursue this activity. (Please note that we return to the role of individual characteristics in entrepreneurship in the Personal Characteristics of Entrepreneurs section.) Here, we consider personal characteristics in terms of their relationships to entre-

preneurial motivation.) Among these variables, evidence concerning general self-efficacy—individuals’ belief that they can successfully accomplish various tasks they undertake (Bandura, 1997)—is perhaps the strongest. Research findings indicate a positive relationship between self-efficacy and both the tendency to actually start new ventures (Markman, Baron, & Balkin, 2005; Zhao, Seibert, & Hills, 2005; average r = .25) and achieving financial success through such activities (Chandler & Jansen, 1992). In addition, other findings (Zhao et al., 2005) indicate that selfefficacy fully mediates the impact of several other variables on entrepreneurial intentions—interest in performing various entrepreneurial activities, such as starting a new venture. Specifically, these researchers found that the effects on entrepreneurial intentions of (a) perceived learning from entrepreneurship-related courses, (b) previous entrepreneurial experience, and (c) risk propensity were all fully mediated by entrepreneurial self-efficacy. It is important to note that excessively high levels of self-efficacy can sometimes exert inimical effects. When individuals experience overconfidence— excessively high levels of self-efficacy combined with extremely high levels of optimism—the result may be hubris—a tendency to believe so strongly in one’s abilities and talents that the possibility of failure seems remote (Hayward & Hambrick, 1997). A framework proposed by Hayward, Shepherd, and Griffin (2006) suggests that when this is coupled with arrogance—contempt for others and the input they provide—the results may be devastating for new ventures. Individuals who suffer from such hubris tend to be overconfident with respect to their abilities to obtain needed resources and also overconfident in their decisions concerning the use of such resources. As a result, they greatly underestimate the resources needed for a successful new venture launch and start with less than they actually require. Insufficient initial resources are a major cause of new venture failure (Suchman, 1995). Empirical support for these suggestions has been obtained in a study by Hmieleski and Baron (2008), who found that entrepreneurs’ self-efficacy is positively related to firm performance when coupled with moderate levels of personal optimism but negatively related to firm performance when coupled with very high levels 249

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

of optimism. Apparently, the combination of high self-efficacy and high optimism tends to generate overconfidence, which can be detrimental to firm performance in many ways (e.g., overconfident entrepreneurs are unwilling to recognize that business strategies they have adopted are failing and thus become trapped in “sunk costs”). Other personal factors that have been found to be related to the decision to launch and operate new ventures include need for achievement (individuals who become entrepreneurs are higher in this characteristic than ones who do not; Collins, Hanges, & Locke, 2004), need for independence (again, entrepreneurs tend to be high on this dimension; Locke & Baum, 2007), and what has been termed passion, or powerful commitment to one’s work (Cardon, Wincent, Singh, & Drnovsek, 2009; Locke, 2000). Entrepreneurial motivation is also positively related to two of the Big Five dimensions of personality— Conscientiousness and Openness to Experience (Zhao & Seibert, 2006). Again, it should be emphasized that the variables discussed here are ones that have been found to be significantly related to entrepreneurs’ motivation—their desire to take overt actions to actually launch new ventures (Locke & Baum, 2007). Additional individual-level variables that play a role in other aspects of the entrepreneurial process are considered in detail in a later discussion. Because motivation is generally defined as involving persistence in a course of action as a well as the initiation of that action, an additional question relating to entrepreneurial motivation concerns perseverance, or the degree to which entrepreneurs persist in their efforts to develop new ventures despite initial setbacks (e.g., Markman, Balkin, & Baron, 2002). Much less research has been directed to this issue than to the question of why specific individuals choose to become entrepreneurs, but some suggestive findings have been obtained. First, it appears that self-efficacy is significantly related to perseverance among entrepreneurs (e.g., Markman et al., 2005). The higher entrepreneurs are on this dimension, the more likely they are to move forward with activities such as patenting their inventions, acquiring financial resources, and actually launching a new venture (correlations in the range of .20–.25). Similarly, recent findings indicate that cognitive fac250

tors, too, influence this aspect of motivation (i.e., perseverance). Astebro, Jeffrey, and Adomdza (2008) observed that the greater entrepreneurs’ tendency to underestimate the time and resources required to complete a project (i.e., launch their new venture), the more likely they were to continue and to successfully commercialize their inventions. This tendency to underestimate the time or effort required to complete a given task is known as the planning fallacy and is generally viewed as an instance of optimism, or, in this case, excessive, unjustified optimism. Thus, given the findings reported by Astebro et al. (2008), it might be anticipated that entrepreneurs’ optimism would be positively related to the financial success of their companies as well as their tendency to persevere in the face of adversity. However, other evidence indicates that the relationship between such optimism and new venture performance may be curvilinear in nature, such that entrepreneurs who are very high in optimism tend to adopt actions and strategies that impair their new ventures’ growth and profitability (e.g., Hmieleski & Baron, 2009). Moreover, it appears that this downturn in new venture performance is more pronounced and more likely to occur in dynamic rather than stable industry environments (Hmieleski & Baron, 2009). Overall, these recent findings offer support for the suggestion that efforts to fully understand new venture creation and the factors that influence subsequent new venture performance must, of necessity, adopt the multilevel perspective mentioned earlier— one that recognizes not only the role of diverse individual-, group-, and societal-level variables, but also the complex interplay between them (i.e., various moderating and mediating effects; Hitt et al., 2007). OPPORTUNITY RECOGNITION All fields possess a set of basic terms, or concepts or ideas that play a central role in that field’s key activities and in its efforts to understand the phenomena on which it focuses. In the field of entrepreneurship, two of the most important are opportunity and opportunity recognition. These terms are very basic in entrepreneurship because, in a key sense, the entire entrepreneurial process often (although not always) begins with identification of an opportunity

Entrepreneurship

by one or more persons. Indeed, as Shane, Locke, and Collins (2003, p. 259) stated, “The entrepreneurial process occurs because people act to pursue opportunities. . . .” In other words, everything else then stems from this initial step. In this discussion, we first define these terms and then outline key theoretical perspectives on opportunity recognition within the field of entrepreneurship.

Copyright American Psychological Association. Not for further distribution.

Basic Definitions Many definitions of the term opportunity have been proposed (e.g., Herron & Sapienza, 1992; Kirzner, 1979), but careful examination of these proposals suggests that most include reference to three central characteristics: (a) potential economic value (i.e., the potential to generate profit), (b) newness (i.e., some product, service, technology, and so forth that did not exist previously), and (c) perceived desirability (e.g., moral and legal acceptability of the new product or service in the society in which it is introduced). For purposes of the present discussion, therefore, a definition derived from these views but also consistent with the emerging cognitive perspective (Baron, 2004, 2006) on this issue is offered. Opportunity is defined as perceived means of generating economic value (i.e., profit) that have not previously been exploited and are not currently being exploited by others. The means of generating profit include the creation of something new (e.g., new products, services, technologies, markets, production processes) as well as utilization of existing products, services, technologies, and so forth in ways that offer increased potential for generating economic value. Note that there is no assumption that perceived opportunities will in fact create profit; rather, opportunities involve the perception by one or more persons that such outcomes can be obtained by certain means. The third criterion listed previously suggests that opportunities have a social or societal component: To be a bona fide opportunity, the means of generating profit must not only be new, but also acceptable to members of the society in which the new ventures are launched and operate. This requirement is included to avoid labeling as opportunities activities that involve new products or services that can potentially generate huge profits but are inconsistent with, or even inimical to, the norms and values of a society.

Although existing evidence suggests that experience in operating ventures that violate societal norms (new ventures defined as illegal or immoral in a particular society) can contribute to entrepreneurs’ human capital (Aidis & van Praag, 2007), it is important to exclude from the concept of business opportunities actions that result in actual harm to large number of persons or that violate a society’s well-established values. For instance, a new drug that is highly addicting and ultimately harmful could, potentially, generate huge profits for the entrepreneurs who develop it. It would not, however, meet the criterion of acceptability because development of this “opportunity” would involve tremendous social costs and serious harm to many persons. Given this definition of opportunity, the term opportunity recognition refers to the cognitive process (or processes) through which individuals conclude that they have identified an opportunity. The processes that lead to such cognitions can be relatively automatic in nature (i.e., rapid, requiring minimal effort, difficult to express in words) or relatively controlled in nature (i.e., slower, more effortful, and readily expressed verbally; e.g., Schneider & Shiffrin, 1977). However, ultimately, the outcome is the conscious thought that a specific and describable business opportunity has been identified. It is important to note, as emphasized by Ardichvili, Cardozo, and Ray (2003) and McMullen and Shepherd (2006), that opportunity recognition is only the initial step in a continuing process. As such, it is distinct both from detailed evaluation of the feasibility and potential economic value of identified opportunities and from active steps to develop these opportunities through new ventures. (These aspects of the process are further discussed in a later section.) Having offered these basic definitions, we turn next to an overview of theoretical development with respect to the process of opportunity recognition.

Initial Theories of Opportunity Recognition: From Alertness and Experience to Underlying Cognitive Processes Because opportunity recognition is often the start of the entire entrepreneurial process (it occurs largely during the prelaunch phase), it is not at all surprising 251

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

that it has long been a central concept in the field of entrepreneurship. Until recently, however, little effort was made to examine it as a process. Rather, opportunities were defined largely in economic terms: Any idea for a new product, service, raw material, market, or production process that could be successfully exploited so as to generate economic benefits for stakeholders was viewed as constituting an opportunity. Further, the approach to understanding opportunity recognition was, to some extent, purely empirical in nature. Efforts to understand the processes that underlie this activity were relatively rare (e.g., Gaglio & Katz, 2001). Overall, initial theorizing concerning opportunity recognition generally centered on information—having more information or better access to information than other persons (e.g., Shane, 2000, 2003). This was viewed as the basic foundation of opportunity recognition. This focus on information developed in several different directions, which are described in the sections that follow (e.g., Ardichvili et al., 2003; McMullen & Shepherd, 2006). All these factors (search, alertness, access to previously stored information) were found to play a role in opportunity recognition. Access to information and its effective use as the basis for opportunity recognition. With respect to greater access to information, it has been suggested— and confirmed in many studies (Shane, 2003)—that specific persons gain an advantage with respect to opportunity recognition by having enhanced access to relevant information. They gain this edge in several different ways. For example, they may have jobs that provide them with information that is not widely available to others. Jobs in research and development or marketing appear to be especially valuable in this respect. Another way in which individuals gain superior access to information is through varied work and life experience—factors that, because they contribute to individuals’ knowledge base, also increase their creativity (Blanchflower & Oswald, 1998). Creativity, in turn, has often been viewed as an essential ingredient in innovation—a key component in new venture success. For instance, Amabile (1996, p. 143) noted, “All innovation begins with creative ideas . . . creativity by individuals and teams is a starting point for innovation. . . .” In addition, 252

recent findings (e.g., Baron & Tang, 2009) suggest that creativity on the part of entrepreneurs is indeed significantly related to the number and radical nature of innovations by new ventures. Finally, entrepreneurs often gain enhanced access to information through a large social network (Ozgen & Baron, 2007). Other people often serve as a valuable source of information, and frequently, the information they provide cannot be acquired easily in any other way. Greater access to valuable information is only the beginning of the process, however. It was further suggested, in initial theorizing, that entrepreneurs who recognize opportunities not only have greater access to information than other persons but also are better at using such information. In other words, cognitive skills or abilities also enter the picture. As a result of having greater access to information, some persons have richer and better integrated stores of knowledge than others—for instance, more information (retained in memory) about markets and how to serve them. This, in turn, enhances their ability to interpret and use new information because they not only have more information available, but also the information is better organized. Large quantities of well-organized information play a key role in creativity, so it is not at all surprising that persons who identify opportunities have also been found to possess richer and better organized stores of information (e.g., Sternberg, 2004). Active searches, alertness, prior knowledge, and social networks. Although access to information was found to play a key role in opportunity recognition, three other factors were soon identified as important in theoretical proposals concerning the nature of opportunity. These factors were (a) engaging in an active search for opportunities, (b) alertness to opportunities (the capacity to recognize them when they emerge), and (c) prior knowledge of a market, industry, or customers as a basis for recognizing specific opportunities. Research findings provided support for the importance of all of these factors. For instance, with respect to an active search for opportunities, many studies indicate that entrepreneurs are more likely than others (e.g., managers) to engage in such searches (Kaish & Gilad, 1991). In other words, they actively search

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

for opportunities rather than waiting passively to stumble upon them. This, in turn, suggests that opportunity recognition may be closely related to proactive work behavior—actions by individuals that involve taking control and making things happen rather than simply adjusting to events once they occur (e.g., Grant & Ashford, 2008; Parker, Williams, & Turner, 2006). To date, however, the relationship between these constructs has not been systematically explored. Alertness, in contrast, emphasizes the fact that opportunities can sometimes be recognized by individuals who are not actively searching for them but who possess “a unique preparedness to identify them” when they appear (Kirzner, 1985, p. 48). Kirzner (1985), who first introduced this term into the entrepreneurship literature, defined it as referring to “alertness to changed conditions or to overlooked possibilities” (p. 48). What are the foundations of entrepreneurial alertness? It has been suggested that alertness rests, at least in part, on cognitive capacities possessed by individuals, capacities such as high intelligence and creativity (Shane, 2003). These capacities help entrepreneurs to identify new solutions to market and customer needs in available information and to imagine new products and services that do not currently exist. A third factor suggested as important in theoretical perspectives that emphasized the role of information in opportunity recognition is prior knowledge of a market or industry. Much evidence offered support for this view, suggesting that information gathered through rich and varied life experience (especially through varied business and work experience) can be a major plus for entrepreneurs in terms of recognizing potentially profitable opportunities. For example, it has been found that prior knowledge of customer needs and ways to meet them greatly enhances entrepreneurs’ ability to provide innovative solutions to these problems—in other words, to identify potentially valuable business opportunities (Shane, 2000). Finally, the breadth of entrepreneurs’ social networks has recently been included in theoretical frameworks emphasizing the importance of information in opportunity recognition (e.g., Aldrich & Kim, 2007). Specifically, it has been suggested that

the broader entrepreneurs’ social networks (the more people they know and with whom they have relationships) and the more conferences and professional meetings they attend, the more information they obtain from these sources and, hence, the more (and better quality) opportunities they identify. This suggestion, too, has been confirmed by many studies (e.g., Ozgen & Baron, 2007). Social networks are an important source of information for entrepreneurs and, as such, often provide the raw materials on which opportunity recognition rests. In sum, considerable evidence suggested that a theoretical perspective on opportunity recognition that emphasized the importance of information— access to it, efforts to obtain it, and its effective use—was very useful. As helpful as it was, however, this perspective generally devoted very little attention to a key question: How does opportunity recognition actually occur? That is, what basic cognitive processes underlie emergence, in the minds of specific entrepreneurs, of ideas for new products or services? This question, in turn, derives from the basic assumption that recognition of specific opportunities is an event that occurs primarily in the minds of individual entrepreneurs. Although information can be, and often is, shared or exchanged and various events and trends in markets, technology, and government policies (to mention just a few variables) can be readily observed by many persons, it is often just one individual or a few individuals who move from these external conditions and changes to the formulation of cognitive representations that constitute recognized opportunities. This basic perspective suggested an important route for enriching and expanding theoretical frameworks concerning opportunity recognition, one that has been adopted by a growing number of researchers (e.g., Baron, 2006; McMullen & Shepherd, 2006): adaptation of existing and well-validated theories of human cognition to the task of understanding opportunity recognition.

Cognitive Theories: Efforts to Understand the Process of Opportunity Recognition If opportunity recognition is viewed primarily as a cognitive process, then three basic questions arise: (a) What is the process (or processes) through 253

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

which opportunities are initially identified? (b) Once individuals have perceived what they believe to be an opportunity, how do they decide whether these opportunities are real, bona fide business opportunities worthy of further consideration or potential “dead-ends” that will not yield projected economic benefits? and (c) What specific skills or cognitive frameworks play a role in opportunity recognition (i.e., what frameworks, acquired through experience, contribute to the capacity to perceive connections among highly diverse conditions and changes?)? These questions have recently been considered in the context of basic cognitive processes long investigated by cognitive scientists: (a) pattern recognition, or the processes through which complex patterns are perceived in arrays of stimuli, and (b) decision mechanisms, or the mechanisms through which individuals conclude that the patterns they have perceived are, or are not, sufficiently clear to be viewed as real (i.e., as actual opportunities). In this respect, signal detection theory (e.g., Swets, 1992) has been applied to understanding key aspects of opportunity recognition (e.g., Baron, 2002). Pattern recognition: Opportunities as complex, discernible patterns. To the extent that opportunities exist in the external world as complex patterns of observable stimuli (Baddeley, 1997), perception must, logically, play a role in opportunity recognition. Presumably, there is a pattern of observable events or stimuli in the external world that entrepreneurs can notice or perceive; whether this pattern is or is not recognized is, in a sense, one of the key questions concerning opportunity recognition. Basic research on perception refers to this task as object recognition or pattern recognition (e.g., Matlin, 2002), and it involves a process through which specific persons perceive identifiable patterns in arrays of seemingly unrelated events or stimuli. The patterns recognized then become figure instead of undifferentiated (and often overlooked) ground (i.e., background). As applied to opportunity recognition, pattern recognition involves instances in which specific individuals “connect the dots,” that is, perceive links between seemingly unrelated events and changes. The patterns they perceive then become the basis for identifying new business opportunities. 254

Several lines of evidence suggest that pattern recognition plays a key role in opportunity recognition. First, it is clear that many opportunities exist for years before they are noticed and developed. For instance, consider wheeled luggage of the type that is now used by a large majority of all travelers. Such luggage was used for decades by flight crews before it was introduced into the market for general sale. Why? Perhaps because no one connected the dots between several seemingly unrelated but pertinent trends: a large increase in the number of passengers, growing problems with checked luggage, expansion in the size of airports so that travelers had everincreasing distances to traverse, and so on. Once these trends were linked as an identifiable pattern in the minds of individual entrepreneurs, a product that would help meet the needs of a large and growing market was suggested. Acceptance of this new product was so rapid that within a few years of its appearance, it quickly dominated the market for luggage and drove earlier models to near-extinction. Second, a large body of evidence in cognitive science suggests that pattern recognition is a basic aspect of our efforts to understand the world around us. That is, we do indeed expend considerable effort searching for patterns among various events or trends in the external world. To the extent that opportunity recognition also involves perceiving links or connections between seemingly independent events or trends, it may be closely related to this basic perceptual process. But how, precisely, does pattern recognition occur? Many different models of this process exist, but all agree on the following basic point: Individuals notice various events in the external world (e.g., changes in technology, markets, government policies) and then use cognitive frameworks they have acquired through life and work experience to determine whether these events are related in any way— whether, in short, they form a discernible pattern. Different models focus on contrasting kinds of cognitive frameworks, but the process is much the same in all of them. The prototype model—one widely accepted model of pattern recognition—suggests that individuals use prototypes as a basis for recognizing patterns. Prototypes are idealized representations of the most typical member of a category (a class of

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

objects or events that seem to belong together). Newly encountered events or trends are compared with existing prototypes to determine whether they belong to specific categories or can be seen as connected in some manner. Because prototypes represent the modal or most frequently experienced combination of attributes associated with an object or pattern, the prototype of “vehicle” would, for example, probably include such attributes as a mechanism for achieving movement, some kind of control gear, some means of holding or securing passengers, and so on. Applying prototype models to opportunity recognition, entrepreneurs may use prototypes as a means for identifying patterns among seemingly unrelated events or trends. Considerable evidence suggests that individuals do indeed form prototypes and that, once they exist, these cognitive frameworks are used in many ways (Matlin, 2002). For instance, they are often used for perceiving patterns in diverse and seemingly unrelated events or trends. Applied in this manner, prototypes may well play an important role in the process of opportunity recognition. A somewhat different perspective on pattern recognition is suggested by exemplar models. These models emphasize the importance of specific knowledge rather than idealized prototypes. Such models (Hahn & Chater, 1997) suggest that as individuals encounter new events or stimuli, they compare them with examples (exemplars) of relevant concepts already stored in memory. Exemplar models seem quite relevant to opportunity recognition because they do not require the construction of prototypes. Rather, individuals simply compare newly encountered events or stimuli with examples of a given concept already present in memory. Overall, research in cognitive science suggests that both prototype and exemplar models may be necessary to fully understand how individuals notice emergent patterns in diverse and apparently unrelated events or changes (Nosofsky & Palmeri, 1998), but further detailed consideration of the evidence pointing to this conclusion is beyond the scope of the present discussion. Suffice it to say that given the powerful and pervasive role of pattern recognition in many aspects of perception and cognition, it seems reasonable to suggest that this

process plays an important role in opportunity recognition as well (Baron, 2007b). This suggestion is also supported by a growing body of empirical evidence (e.g., Baron & Ensley, 2006) indicating that prototype models of pattern recognition do indeed play an important role in opportunity recognition. More generally, current research and theory suggests that the cognitive frameworks individuals construct through experience (prototypes, exemplars) often provide the basis for identifying emergent patterns, and, in at least some instances, these patterns then suggest new business opportunities. Are the perceived patterns real? Pattern verification. Opportunities are not always as clear or straightforward as the wheeled luggage example offered earlier. Rather, entrepreneurs often face a challenging task in determining whether a perceived business opportunity is one likely to yield economic benefits if actually exploited. This fact simply reflects the basic nature of perception, which is always a stochastic process, in which individuals estimate probabilities—often at a level below conscious awareness. Some stimuli in the external world are so clear or strong that they are recognized by virtually everyone, but many others are weaker and less distinct, such that recognizing them is far less certain. Moreover, as a result of underlying biological processes, sensitivity to external stimuli varies over time, such that a specific individual may perceive a given stimulus at one time but fail to perceive the same stimulus at another time. These principles also apply to pattern recognition. Whereas some patterns are so clear as to leave little room for error, many others are far more subtle and difficult to observe. This suggests that in many instances, individuals are uncertain as to whether they have, or have not, perceived a pattern constituting a business opportunity. How do they then decide whether the pattern they have tentatively identified is indeed real, or close enough to their concept (or prototype) of business opportunity, to warrant further consideration? In other words, how do they then verify the patterns they think they have observed? Several cognitive theories are relevant to this task, but perhaps the one that is most applicable is signal detection theory (Swets, 1992), an approach that takes careful account of the fact that whenever 255

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

individuals attempt to determine whether a stimulus is present or absent, four possible outcomes exist: the stimulus is actually present and the perceiver recognizes this fact (a hit or correct identification); the stimulus is present but the perceiver fails to recognize it (a miss); the stimulus is absent and the perceiver concludes, erroneously, that it is present (a false alarm); and the stimulus is absent and the perceiver correctly concludes that it is absent (a correct negative or correct rejection). The theory further notes that many factors determine the relative rate at which individuals experience hits, misses, and false alarms in any given situation. Some of these relate to the properties of the stimuli (e.g., the stronger the stimulus, in physical terms, the easier it is to be certain that it is present). Additional factors, in contrast, relate to the current state of the perceiver (e.g., is this person fatigued? highly or weakly motivated to be alert?). Still other factors involve the subjective criteria perceivers apply to the task. Consider the situation faced by an entrepreneur who believes that he or she has identified an opportunity for a profitable new venture. The venture is one that he or she can start in his or her spare time and for which little or no capital is needed. As a result, he or she may set his or her subjective criterion for concluding “this is a good business opportunity” quite low: The costs of a false alarm are minimal (a little wasted time and effort) relative to the potential gains of a hit. In contrast, consider another entrepreneur who has recognized an opportunity that cannot be pursued on a part-time basis and for which large amounts of start-up capital are required. Under these circumstances, the entrepreneur will probably set his or her criterion for concluding “this is a good opportunity” somewhat higher: The costs of a false alarm are very high, and potential rewards are reduced by the large proportion of the business that will be owned by investors. In short, potential costs and benefits of starting a new venture determine where prospective entrepreneurs will set their criteria for concluding that an opportunity they have perceived is real and therefore worthy of further consideration. Signal detection theory also offers additional insights into the nature of this process. For example, it suggests that individuals may differ greatly with 256

respect to sensitivity—the ability to distinguish between situations in which the crucial stimulus (a pattern suggesting existence of an opportunity) is present and ones in which it is not. What are the origins of such differences? A cognitive perspective suggests that they may involve the knowledge structures on which individuals rely in identifying the complex patterns that constitute opportunities: the accuracy and clarity of their prototypes or the range and content of their exemplars for the concept “business opportunity.” Applied to entrepreneurship and the process of opportunity recognition, signal detection theory further suggests that whether entrepreneurs set their subjective criteria for concluding “this is a real opportunity” relatively low or relatively high may also be influenced by other factors, such as their motives (Baron, 2002). For example, entrepreneurs who are strongly motivated to minimize risks and to avoid pursuing false alarms may set their subjective criteria relatively high, whereas those who are relatively tolerant of risk and more concerned about overlooking bona fide opportunities may set their criteria somewhat lower (see, e.g., Busenitz & Barney, 1997; Krueger & Brazeal, 1994; Stewart & Roth, 2001). Similarly, entrepreneurs who are high in certain personal characteristics (e.g., optimism) may set their subjective criteria low, with the result that they experience many false alarms. Research findings indicate that entrepreneurs generally show a greater proclivity for an optimistic bias than other persons (e.g., de Meza & Southey, 1996). This suggests that they tend to set their subjective criteria for identifying business opportunities relatively low, unless other factors counter or reverse this tendency. Another cognitive theory that has received growing attention, regulatory focus theory (Higgins, 1998; Molden & Higgins, 2005), calls attention to additional factors that may influence individuals’ decisions concerning whether they have or have not perceived an opportunity. Briefly, this theory suggests that in regulating their own behavior to achieve desired ends, individuals adopt either a promotion focus, in which they concentrate primarily on attaining positive outcomes, or a prevention focus, in which they concentrate primarily on avoiding negative outcomes. Combined with signal detection

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

theory (Baron, 2002), regulatory focus theory suggests that when individuals adopt a promotion focus (an emphasis on positive outcomes), they tend to concentrate on attaining hits (on recognizing a stimulus when it is present) and on avoiding misses (failing to recognize a stimulus that is in fact present). In contrast, when they adopt a prevention focus (an emphasis on avoiding negative outcomes), they concentrate on avoiding errors; thus, they are especially careful to attain correct rejections and to avoid false alarms. In other words, together, signal detection theory and regulatory focus theory suggest that entrepreneurs who focus on obtaining positive outcomes will set their subjective criteria for concluding that they have recognized opportunities relatively low: They will identify many opportunities and avoid misses but will also experience many false alarms. In contrast, entrepreneurs who focus primarily on avoiding negative outcomes will set their subjective criteria relatively high, thus experiencing few false alarms but a larger number of misses (failing to notice opportunities that exist). Reliable measures of these two self-regulatory foci exist (e.g., Brockner, Higgins, & Low, 2004), so these predictions can be readily investigated in future research. In sum, efforts to understand the nature of opportunity recognition have progressed from a purely economic perspective that focused primarily on the economic value of recognized business opportunities, through efforts to understand the role of information in this process, and to, most recently, the application of well-validated theories of cognitive science to this task. The result has been an increasingly sophisticated understanding of a key process that is, in a very basic sense, the start of most, if not all, entrepreneurial activities. ACQUIRING ESSENTIAL RESOURCES: ASSEMBLING HUMAN AND FINANCIAL RESOURCES Once an opportunity has been identified, entrepreneurs move forward with another key activity of the prelaunch phase (although, this is one activity that often continues during subsequent phases as well): identifying and assembling essential resources. In other words, entrepreneurs must determine what

resources—human, financial, legal, and otherwise— they need to get started. Because this discussion is focused on links between entrepreneurship and I/O psychology, we do not consider in detail other key activities of the prelaunch and launch phases, such as legal issues faced by entrepreneurs (e.g., the processes involved in obtaining various forms of legal protection for intellectual property, such as patents, trademarks, and so forth), nor do we examine the specific actions involved in identifying key resources—actions that include detailed feasibility analyses and formulation of business models and strategies. These activities have been fully described and summarized in existing literature but have not, to date, been investigated carefully from a psychological perspective (e.g., McMullen & Shepherd, 2006). Rather, we focus here on those factors relating to the behavior, skills, and actions of individual entrepreneurs that have been found to influence the effectiveness of their efforts to acquire essential resources once they have been identified. Some new ventures are started by individual entrepreneurs, but the majority are launched by teams of several entrepreneurs who, together, are known as new venture founders (e.g., Baron & Shane, 2008). Ideally, founders bring most, if not all, of the human resources needed for the initial launch of a new venture to the emerging business. They possess the knowledge, skills, social networks, and experience needed to convert their ideas for new products or services into actual operating companies. In reality, however, this ideal is rarely achieved. Rather, most founding teams quickly discover that launching a new business is a highly complex undertaking and requires knowledge, skills, and abilities that they themselves cannot provide. As a result, they must attract persons not on the original founding team to join the new venture as additional partners or key employees. This is a daunting task because in comparison with existing, well-established organizations, new ventures have relatively little to offer to such persons in the way of extrinsic rewards. In general, they cannot offer high pay, desirable benefits, or even a modicum of job security, inducements long used by existing organizations to attract highly talented and motivated employees. In addition, they do not have the well-developed and 257

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

systematic recruitment and selection procedures that are used by established organizations. This, too, puts them at a major disadvantage. What they can offer to offset such obstacles, however, is a share in the ownership of the new venture, generally in the form of stock or stock options. In addition, they can offer various intangibles that some persons find appealing: low levels of bureaucracy, ready access to the top management team, the excitement of being in at the start of a new business, and the high levels of autonomy, skill variety, task identity, task significance, and feedback mentioned previously. Although such inducements do help new ventures recruit and retain high-quality employees, it is still clear that building an outstanding, dedicated workforce remains a formidable task. However, to survive and prosper, they must accomplish it. How do they manage to do so? Research findings suggest that one key answer involves the central role of entrepreneurs’ social ties with others—the social networks they form with many people inside and outside their own companies or bring with them into the new venture creation process (e.g., Aldrich & Kim, 2007). Another involves key political and social skills that entrepreneurs use to construct social networks and that assist them in many other ways as well (e.g., Ferris, Davidson, & Perrewé, 2005).

Social Networks, Social Capital, and Their Role in New Venture Formation Social networks—the intricate set of social ties individuals form with others—are an important source of social capital for entrepreneurs. In other words, they are the source of a wide range of benefits that help entrepreneurs secure the resources they need and, in particular, the new employees and others who add human resources to the new ventures. More specifically, social capital has been defined as (a) the ability of individuals to extract benefits from their social structures, networks, and memberships or (b) these benefits themselves—the advantages individuals gain from their relationships with others (Nahapiet & Ghoshal, 1998; Portes, 1998). Both definitions agree that social capital refers to positive outcomes accruing to individuals from their social ties with others—from being known to them, having a good reputation, and from being in estab258

lished, continuing relationships (i.e., networks) with them—although these relationships can, in fact, be short-lived in nature as well as long term (Aldrich & Kim, 2007). Among the benefits provided by such social ties are support, expertise, and encouragement on the one hand, and acquisition of tangible benefits such as equipment and financial resources on the other. Viewed in the light of these benefits, it is not surprising that some experts on network theory suggest that these social structures provide entrepreneurs with “infinite possibilities” with respect to a potential array of benefits (Simmel, 1995, p. 151). Clearly, then, social capital is worth possessing: It is an intangible asset that can yield highly beneficial outcomes for the persons (and in this case, entrepreneurs) who possess it. The social ties on which social capital rests are often divided into two major categories: close ties— the strong, intimate bonds that exist between members of a nuclear family or very close friends—and loose ties—social linkages of the type that occur outside families or intimate friendships, for example, links between persons who happen to work together or who do business on a fairly regular basis (e.g., Adler & Kwon, 2002; Putnam, 2000). Social ties can occur at either the individual level, between specific persons, or at a group or organizational level. In both cases, they often serve as the basis for trust—confidence by one or more persons in the motives and predictability of one or more others. Strong ties are often viewed as leading to, or at least being associated with, bonding social capital— they generate relationships between individuals that are based on mutual trust. Examples are found in extremely high levels of such trust that often exist between founding partners of a new venture (e.g., those between Bill Gates and Steve Ballmer at Microsoft). Weak ties, in contrast, lead to (or are associated with) bridging social capital—they are useful in providing individuals with information that would otherwise be difficult or costly for them to obtain. An example would be the information individuals or organizations acquire from membership in business networks or trade associations. Before concluding this discussion of social ties and networks, it is important to distinguish between social capital and two other closely related terms

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

that are also widely used in entrepreneurship literature: human capital and social competence. Human capital refers primarily to the knowledge and skills individuals possess—what they know and bring to any work setting, including new ventures. In contrast, social competence involves an array of skills that assist individuals in interacting effectively with others. We examine such skills and their role in entrepreneurship in more detail later but mention them here to distinguish them from both social and human capital. Whatever its precise function (e.g., bonding or bridging), social capital has been found to play an important role in the entrepreneurial process. Research findings (e.g., Davidsson & Honig, 2003) indicate that factors contributing to the development of bonding social capital (e.g., having parents or close friends or neighbors in business; receiving encouragement from friends and family) are strongly associated (correlations greater than .30) with discovering or recognizing opportunities for new ventures and also with taking the initial steps required to launch such ventures. Similarly, factors that contribute to the development of bridging social capital (e.g., membership in trade organizations; Ozgen & Baron, 2007) are related to performing the activities essential to converting recognized opportunities into viable new ventures and to measures of important outcomes, such as first sales or actual profit. In sum, the concept of social capital calls attention to the fact that entrepreneurs definitely do not operate in a social vacuum. On the contrary, they depend on the support, advice, information, and financial resources provided by others. The broader and richer the social networks to which entrepreneurs belong (i.e., the higher their social capital), the greater their chances of obtaining such benefits and, hence, the greater the likelihood that they will succeed in converting their ideas and vision into profitable new businesses.

Social and Political Skills: Beyond Social Capital The social networks entrepreneurs form are an important source of resources for them. Indeed, these networks, and the social capital they generate, often assist entrepreneurs in locating and hiring

partners and key employees and in securing access to many kinds of information they need to launch and successfully operate their new businesses (Shane, 2003). However, as noted by Baron and Markman (2000), although social capital may well be a necessary condition for the entrepreneurs’ success (high levels of social capital help them to get through the door to gain access to potential investors, customers, and employees), it is not a sufficient condition for such success. This is so because once such access has been achieved (through social networks, reputation, and related aspects of social capital), it is entrepreneurs’ social competence—their social and political skills—that then determine the outcomes of face-to-face meetings with these potential sources of support. This reasoning is based on a large body of evidence indicating that in many work settings, social or political skills influence important organizational processes. To mention just a few of these effects, persons high in social–political skills, compared with persons low in such skills, are more successful as job candidates (e.g., Riggio & Throckmorton, 1988), receive higher performance reviews from supervisors (e.g., Robbins & DeNisi, 1994), and attain faster promotions and higher salaries Similarly, individuals high in social skills generally achieve greater success in many different occupations (e.g., medicine, law, sales; e.g., Seibert, Kraimer, & Liden, in press; Wayne, Liden, Graf, & Ferris, 1997), attain better results in negotiations (e.g., Lewicki, Saunders, & Barry, 2005), and often (although not always) achieve higher levels of task or job performance (e.g., Hochwarter, Witt, Treadway, & Ferris, 2006) than do persons low in such skills. Of course, social skills also exert strong effects on outcomes in many contexts outside the world of work. For example, persons high in various social skills tend to have wider social networks than do persons low in social skills (e.g., Diener & Seligman, 2002). Social and political skills have even been found to influence the result of legal proceedings, with persons high in such skills attaining acquittals more often than persons low in such skills (e.g., Downs & Lyons, 1991). The breadth of these findings, coupled with additional evidence indicating that social or political skills measured at one point in time signifi259

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

cantly predict important outcomes at later times (e.g., job performance ratings; Ferris, Witt, & Hochwarter, 2001; Hochwarter et al., 2007), indicates that such proficiencies might also exert significant effects in the realm of entrepreneurship. In fact, a growing body of empirical findings indicates that this is so. For instance, in an initial study of this issue entrepreneurs working in two different industries (cosmetics, high tech) completed a widely used and well-validated measure of social skills (e.g., Riggio, 1986). Entrepreneurs’ scores on this measure were then related to one indicator of their financial success—the income these entrepreneurs earned from their new ventures over each of several years. Results indicated that several social skills (social perception, social adaptability, expressiveness) were significantly related to this measure of financial success. A more recent investigation (Baron & Tang, 2009) extended these findings by investigating the underlying (i.e., mediating) mechanisms through which entrepreneurs’ social skills influence the success of their new ventures. Results indicated that entrepreneurs’ effectiveness in acquiring useful information and in obtaining essential resources mediated the effects of their social or political skills on widely used measures of new venture performance, such as growth in sales, growth in profits, and growth in number of employees. The study was conducted with Chinese entrepreneurs working in many different businesses; thus, it expanded earlier results to a very different cultural context and to many additional industries. As in the earlier research by Baron and Markman (2000), specific social or political skills (e.g., social perception, social adaptability, expressiveness) were significantly related to measures of new venture success (rs ranging from .19 to .36). In sum, existing evidence suggests that in acquiring essential resources for their new companies— human, financial, or otherwise—entrepreneurs draw heavily on benefits conferred by their social networks (i.e., their social capital) and also on their individual skills in interacting effectively with others (their social and political skills). Those entrepreneurs who possess and effectively use such skills tend to experience greater levels of success in 260

starting new ventures than those who are lower in this respect. MEASURING ENTREPRENEURIAL SUCCESS By its very nature, I/O psychology tends to focus primarily on measures of individual behavior— performance, attitudes (e.g., job satisfaction, organizational commitment), health (e.g., the detrimental effects of stress), emotions experienced by individuals (e.g., positive and negative affect), and cognitions (e.g., perceptions of distributive and procedural fairness; Cropanzano, Goldman, & Folger, 2003). Group outcomes, too, are viewed as important and are strongly represented in research on leadership and team performance, to mention just a few examples (e.g., chaps. 7 and 19, this volume). However, purely organizational outcomes (e.g., organizational performance, organizational growth) are often viewed as somewhat less central to the field than outcomes and processes pertaining to individuals or groups. In one basic sense, the field of entrepreneurship represents the inverse of this perspective. Interest was, and continues to be, focused on measures of organizational-level outcomes, primarily indices of growth and profitability (e.g., Baum & Locke, 2004). Very little interest has been directed toward individual-centered outcomes (e.g., high levels of satisfaction with one’s work or career, personal happiness, or effective psychological adjustment; e.g., Schjoedt & Shaver, 2004). Given the definition of entrepreneurship offered earlier (a field that seeks to understand how opportunities to create something new arise and are discovered or created by specific persons who then use various means to exploit or develop them, thus producing a wide range of effects), this focus on organization-level dependent measures is eminently reasonable. Entrepreneurs seek to create viable, profitable companies, and a key index of their success in this endeavor is the financial outcomes achieved by these newly launched ventures (e.g., Zahra & Hayton, 2008). For this reason, most research in the field of entrepreneurship has focused on such dependent measures as (a) firm survival, (b) growth in sales, (c) growth in income, (d) growth in number of employees, and, ultimately,

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

(e) market value when a new venture reaches the point at which it can “go public” through an initial public offering (refer to Figure 8.1). Indeed, in the past, research in entrepreneurship that focused on individual entrepreneurs and their characteristics, capabilities, attitudes, skills, or individual cognitions (e.g., Mitchell et al., 2007) was seen, to some degree, as a kind of diversion along the way toward the major goal: understanding the variables (usually macrolevel variables relating to economic conditions, government policies, and markets) that strongly shape new venture success (Astebro et al., 2008). Although this perspective is still very much present in the field of entrepreneurship, it is gradually being augmented by the much broader multilevel approach, suggesting that to fully understand the entrepreneurial process, it is essential to consider variables operating at many different levels as well as the complex ways in which these variables interact (e.g., Hitt et al., 2007). Simultaneously, there has been growing recognition of the fact that variables relating to the motives, skills, actions, or attitudes of entrepreneurs almost certainly do not influence key firm-level outcomes directly. Rather, they do so indirectly, through the influence of mediating processes that are, moreover, influenced by numerous moderating variables (e.g., Baron & Tang, 2009). Further, there is also increasing recognition of the fact that entrepreneurial success cannot be measured solely in terms of financial or economic variables; rather, the effects of this activity on entrepreneurs themselves—their attitudes, personal and social adjustment, health—must be considered as well (Baron, 2002). Very little research designed to examine such issues has been conducted (e.g., Schjoedt & Shaver, 2004), but at the very least, the issue has been recognized and is now viewed as worthy of detailed investigation. Overall, the multilevel perspective that has emerged both in entrepreneurship and in management research generally (Hitt et al., 2007) suggests that to understand the complex processes involved, and their broad-reaching effects on the individuals involved (entrepreneurs or others), it is essential to examine not merely the potential effects of a very wide range of variables, but also the intricate patterns of mediation and moderation that ultimately

generate the macrolevel outcomes (e.g., firm survival, growth, economic value) that have long been prominent in the field. Another implication of the multilevel perspective is that efforts to relate individual-level variables directly to various measures of firm performance are optimistic at best and perhaps misguided at worst. This is because, as Bandura’s social cognitive theory suggests (Bandura, 1986, 1997), such variables are almost always part of a web of reciprocal relationships among behavior, cognition, and affect. According to this influential theory, the effects of personal characteristics (e.g., dispositions, personality traits, attitudes, values, skills) are often determined by interactions of these individual-level variables with additional behavioral and environmental factors. Thus, the theory blends dispositional, behavioral, and environmental perspectives, providing a more comprehensive framework for examining human action and its outcomes than could be gained by focusing on any of these levels or classes of variables independently. Social cognitive theory provides a useful framework for identifying the mechanisms through which individual-level variables ultimately influence firm-level performance—a task that has been identified as crucial by many entrepreneurship researchers (e.g., Baron, 2007a; Hmieleski & Baron, 2008). A concrete example of how this perspective is currently contributing to ongoing entrepreneurship research (and concurrently strengthening its ties to I/O psychology) is provided by a study conducted by Hmieleski and Baron (2009). This research investigated the relationship between entrepreneurs’ overall level of optimism and the financial performance of the new ventures they founded. Results indicated that overall, there was a negative relationship between entrepreneurs’ dispositional optimism and new venture performance—a somewhat surprising finding given previous research on the role of optimism in entrepreneurship (e.g., Fraser & Greene, 2006). This effect was predicted partly on the basis of the fact that entrepreneurs are generally extremely high on the dimension of dispositional optimism, and previous research indicates that the relationship between optimism and task performance is often curvilinear in form, with declines occurring at very high levels of optimism. Entrepreneurs 261

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

appear to fall under the declining portion of the optimism–performance function. This effect, however, was moderated both by entrepreneurs’ prior experience in starting new ventures and environmental dynamism, such that the negative relationship between entrepreneurs’ optimism and new venture performance was significantly stronger for experienced than for inexperienced entrepreneurs and significantly stronger in dynamic than in stable environments. These findings, which were also predicted on the basis of social cognitive theory, are consistent with the person–situation perspective generally adopted by I/O psychology, a perspective indicating that the skills, motives, experience, attitudes, and characteristics of individuals do indeed influence work-related behavior and hence important organizational outcomes but that such effects are rarely direct in nature. Rather, more typically, they are moderated by other variables relating to the tasks that individuals perform and the environments in which they work. We suggest that by adopting this broader perspective, the field of entrepreneurship will acquire a more adequate understanding of the role of individual entrepreneurs in new venture performance and the importance of dependent measures aside from ones that are purely economic or financial in nature. This broadened perspective, in turn, may facilitate progress toward the central goal that this still relatively young field seeks: acquiring accurate understanding of the complex process, involving many different factors operating at many different levels, through which new ventures are conceived, launched, and operated by enterprising entrepreneurs. In our view, such research is crucial, for it is this complex, reciprocal interchange that ultimately shapes the survival and fortunes of new ventures and hence the economic fortunes of entire societies. PERSONAL CHARACTERISTICS OF ENTREPRENEURS Having examined key aspects of the entrepreneurial process (opportunity recognition, resource acquisition) and complex issues relating to the measurement of new venture success, we turn next to a topic more closely related to the scope and interests of I/O 262

psychology: the role of individual difference variables in new venture creation. As noted in an earlier section, this was the first perspective adopted by researchers applying a psychological knowledge and theory to entrepreneurship, but it yielded inconsistent and contradictory findings. As a result, it ultimately had negative implications for the application of psychological principles and methods in the field of entrepreneurship, leading many researchers with a background in economics or strategy to conclude— prematurely, we believe—that entrepreneurs play only a relatively minor role in new venture creation. We consider this topic in more detail here, as a useful conclusion to this broad and generally processfocused discussion of the field of entrepreneurship. Folk wisdom suggests that entrepreneurs are definitely a “breed apart”—that they differ from most other persons in important ways. Widely held stereotypes suggest that entrepreneurs possess distinct characteristics, such as a high tolerance (or even desire) for risk, boundless energy, a strong taste for something new, and an equally strong desire to obtain personal autonomy or independence (Baron & Shane, 2008). However, initial efforts to examine the validity of such beliefs generally yielded negative and inconsistent results. Just as researchers in the field of leadership initially failed in their efforts to identify a small set of characteristics that distinguished leaders from other persons or that contributed to leader effectiveness, researchers in entrepreneurship, too, failed in their quest for a short list of defining traits or characteristics of entrepreneurs and highly successful entrepreneurs (e.g., Shaver & Scott, 1991). In fact, so disappointing were the findings of these early studies that several researchers concluded that variables relating to individual differences (especially personality) had little or no bearing on the entrepreneurial process (e.g., Gartner, 1989). The result was a decline in the impact of psychology and other behavioral sciences on the field of entrepreneurship, a decline that, as noted earlier, was only reversed in the late 1990s. Once it reversed, however, this downward trend in the influence of a psychological perspective did so very strongly, such that at present there is growing recognition of the basic fact that personality and other characteristics of individual entrepreneurs do

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

indeed matter—in fact, they can strongly influence many aspects of entrepreneurship and even (although usually indirectly) financial measures of new venture success (e.g., Baum & Locke, 2004; Baum, Locke, & Smith, 2001). Thus, as suggested by the model in Figure 8.1, they should be included in ongoing efforts to understand the nature of entrepreneurship as a continuing process. There are many reasons for this swing in the pendulum of scientific opinion, and they have been clearly described by Rauch and Frese (2006). These researchers noted that the revival of interest in personality and other individual-level variables in entrepreneurship derives from several factors: (a) growing recognition of the fact that such variables often interact with situational variables, (b) recognition of the distinction between general and more specific traits or characteristics, (c) understanding that the relationship between personality variables and performance in many domains (including entrepreneurship) may be curvilinear rather than linear in nature, (d) growing acknowledgment of the fact that the relationship between these variables and performance may be indirect (i.e., mediated) rather than direct, and (e) the adoption in recent research of substantially improved theoretical models (e.g., the Big Five framework) and statistical procedures (e.g., meta-analysis). Together, these and other changes have led to a situation in which significant relationships between several individual-level variables (e.g., skills, personality characteristics) and important aspects of entrepreneurship have emerged. A brief overview of the most robust of these findings follows. This review is not intended to be exhaustive in nature; rather, it is designed to highlight the most clear and consistent findings reported in existing literature. Further, we avoid repeating information concerning several variables that was provided in our earlier discussion of entrepreneurial motivation (e.g., need for achievement, risk taking, self-efficacy).

The Big Five Dimensions Whereas most research on the personal characteristics of entrepreneurs has focused on specific skills (e.g., social–political skills; Baron & Tang, in press) or characteristics (e.g., need for achievement; Rauch & Frese, 2006), a growing number of recent studies

have examined relationships between entrepreneurship and the Big Five dimensions of personality (e.g., Barrick & Mount,1991). (See also Vol. 2, chap. 5, this handbook.) The results of such research indicate that four of these factors—Conscientiousness, Openness to Experience, Agreeableness, and Neuroticism—are significantly linked to becoming an entrepreneur (Zhao & Seibert, 2006). Specifically, entrepreneurs are higher than comparison groups of managers on Conscientiousness, Openness to Experience, and Agreeableness but lower on Neuroticism. However, inconsistent findings have been reported with respect to Openness to Experience, with at least one study reporting that entrepreneurs are actually lower on this dimension than other persons (Ciavarella, Bucholtz, Riordan, Gatewood, & Stokes, 2004). Thus, at present, results relating entrepreneurship to the Big Five dimensions must be interpreted with caution.

Innovativeness Not surprisingly, entrepreneurs have been found to be higher in innovativeness, or interest in considering novel modes of action, than other persons (r = .24 in a recent meta-analysis; Rauch & Frese, 2005, as cited in Rauch & Frese, 2007). Further, a significant relationship between innovativeness and new venture success also appears to exist (r = .22; Rauch & Frese, 2005, as cited in Rauch & Frese, 2007).

Autonomy Entrepreneurs are also higher in autonomy, or the desire to act independently, free from external constraint. Persons who choose this role are higher in autonomy than others, and across many different studies, there is a significant overall difference between entrepreneurs and nonentrepreneurs (Rauch & Frese, 2005, as cited in Rauch & Frese, 2007).

Locus of Control Another difference between entrepreneurs and other persons supported by research findings involves locus of control, or the extent to which individuals believe that they can control their own destinies or outcomes. Consistent with informal observation, entrepreneurs are higher in internal locus of control than are nonentrepreneurs, and this difference is 263

Baron and Henry

reflected in the success of the new venture they found: The higher entrepreneurs’ internal locus of control, the greater the success attained by new ventures (Rauch & Frese, 2005, as cited in Rauch & Frese, 2007).

Copyright American Psychological Association. Not for further distribution.

Cognition Other differences between entrepreneurs and nonentrepreneurs relate to cognition, broadly conceived (e.g., Mitchell et al., 2007). For example, entrepreneurs have been found to demonstrate weaker tendencies to engage in counterfactual thinking (imagining what might have been) than other persons (Baron, 2000). Apparently, their strong preference for focusing on the future deters them from looking back and imagining circumstances that would have generated different outcomes than those they actually experienced. This is something of a two-edged sword. On the one hand, it reduces the amount of time and effort invested in mere speculation—a potential gain. On the other hand, it reduces entrepreneurs’ tendency to imagine ways in which their performance might have been improved, and that, in turn, can impair experiential learning (e.g., Roese & Olson, 1997). Again, we wish to emphasize that this brief review is intended to be representative rather than exhaustive in nature. Even taking this fact into account, however, it seems clear that entrepreneurs do indeed differ in several discrete ways from persons who choose other career paths and, moreover, that highly successful entrepreneurs may differ from less successful ones in various respects (although information on this latter topic is, unfortunately, currently quite limited). To fully understand the nature of these differences—and the overall impact of such variables—it is necessary to examine their interactions with a host of environmental variables as well as potential mediators and moderators of such effects (e.g., Baum & Locke, 2004; Rauch & Frese, 2005, as cited in Rauch & Frese, 2007). Overall, though, the conclusion that individual-level variables, and, by extension, entrepreneurs, do indeed matter in the domain of entrepreneurship is entirely consistent with the general perspective that such variables are part of the total picture in almost all work or business settings. This view is becoming 264

a strong and widely held one in the field of entrepreneurship, a development we view as highly productive and, moreover, as likely to link entrepreneurship more closely than ever to I/O psychology, a field that offers a wealth of empirical findings and sophisticated theoretical frameworks concerning the nature and effects of such individual-level variables (e.g., Baron, 2002). OPPORTUNITIES FOR FUTURE RESEARCH AND CONCLUDING THOUGHTS—PLUS A FEW WORDS ON RESEARCH METHODS Fields of science do not progress according to any clear or established pattern. Rather, new knowledge is acquired in what often seems to be an unpredictable and asymmetrical manner. Moreover, that is especially likely to be true in fields that emerge from diverse intellectual roots, such as entrepreneurship. Yet, at any given time, there is often widespread agreement among active researchers concerning gaps in current knowledge. The field of entrepreneurship, by virtue of its roots in several different disciplines (economics, psychology, strategy, cognitive science), is highly diverse in nature. However, despite this fact, several topics and issues are currently recognized as ones worthy of future research and theoretical development. We briefly summarize a few of these here. One such area is the issue of work–family conflict in entrepreneurship. It is well known that entrepreneurs often devote large portions of their time and effort to their new ventures; in fact, existing data suggest that many work on their ventures for more than 100 hr each week (Shane, 2008). To succeed in these labor-intense, life-absorbing activities, entrepreneurs need the full support and cooperation of key people in their personal lives—spouses, families, close friends, and others. However, currently, there is relatively little information on how the long hours, high levels of stress, and deep commitment to their new ventures affect the quality of entrepreneurs’ relationships and family life. Given the large extant literature concerning the intimate links between working life and personal life (Frone, 2003; Greenhaus & Beutell, 1985), this appears to be an important issue deserving of systematic attention in future research.

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

A second area for future research concerns the role of affect in entrepreneurship, both dispositional (trait) affect and event-generated (state) affect (Baron, 2008). Three basic facts suggest that affect may be especially relevant in the domain of entrepreneurship. First, the environments in which entrepreneurs operate are often highly unpredictable and marked by rapid change (e.g., Lichtenstein, Dooley, & Lumpkin, 2006). Research on the influence of affect suggests that it is most likely to exert powerful effects on cognition and behavior in precisely this type of situation (e.g., Forgas, 1995, 2000; Forgas & George, 2001). Second, affect has been found to exert strong effects on many of the tasks that entrepreneurs perform in starting new ventures (e.g., see Forgas, 2000; Lyubomirsky, King, & Diener, 2005). For instance, affect has been shown to strongly influence creativity (which may play an important role in opportunity recognition; e.g., Isen, 1999), persuasion (which may influence entrepreneurs’ success in acquiring essential resources), decision making and judgments (which play a key role in the formation of effective business models and strategies; e.g., Ireland, Hitt, & Sirmon, 2003), and the development of productive working relationships with others (e.g., Diener & Seligman, 2002). Third, entrepreneurs generally show high levels of enthusiasm, optimism, energy, and perseverance in the face of adversity, reactions with strong emotional components (Cardon et al., 2009). Moreover, these emotion-based reactions are often recognized, by entrepreneurs themselves, as playing an important role in their success. For instance, in commenting on the basis of Microsoft’s early and rapid success, Bill Gates—a famous entrepreneur not especially known for showing high levels of affect— once remarked, “We were young, but we had good advice and good ideas and lots of enthusiasm” (italics added). Research findings offer indirect support for these informal observations, indicating that entrepreneurs do frequently experience strong emotions (e.g., passion for their ideas and work; Baum et al., 2001; Baum & Locke, 2004; Fraser & Greene, 2006; Simon et al., 2000). Thus, future investigations designed to examine the role of affect in new venture creation may yield important new insights (see Baron, 2008). To date, a few such studies have been

performed (e.g., Foo, Uy, & Baron, 2009), but much additional research is necessary to fully explicate the role of affect in the entrepreneurial process. A fourth avenue for future research involves the question, What are the effects, on entrepreneurs, of failure? It is well known that most new ventures fail and that many highly successful entrepreneurs experience failure in their initial efforts to found new companies. Yet, most research in the field of entrepreneurship focuses on successful entrepreneurs. In a sense, this creates an important survivor issue because typically, the entrepreneurs available for study are ones who have survived the winnowingout process imposed by markets and economic factors. This current state of affairs leaves important questions such as the following largely unanswered: What do entrepreneurs learn from initial failure experiences? What distinguishes those who move on to achieve subsequent success from those who do not? How do entrepreneurs differ in their reactions to failure? Efforts to address these issues may prove highly enlightening. A fifth topic deserving of future research attention concerns the role of expertise or expert performance in entrepreneurship. Research on this topic suggests that across many different fields (sports, music, chess, science, professional writing), expert performance does not derive primarily from special talents or growing experience. Rather, it rests strongly on participation in specific kinds of effortful, highly focused practice. In research on expert performance, this is known as deliberate practice (Ericsson, 2006), and it has been shown to play a key role in the generation of exceptional performance in many different fields. This principle raises an intriguing question: Do entrepreneurs—especially highly successful repeat entrepreneurs—also engage in such practice? If they do, what specific skills do they practice and improve? Research on expert performance suggests that engaging in deliberate practice yields many cognitive benefits, ranging from increased stores of domain-relevant information and improved cognitive frameworks for organizing such information to actual enhancements in basic memory systems (e.g., greatly improved access to information stored in long-term memory). Several researchers have suggested that applying the principles of expert 265

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

performance to entrepreneurship may prove highly informative (e.g., Baron & Henry, in press; Mitchell et al., 2007), but, to date, little (if any) research has directly investigated this possibility. A sixth area for future investigations involves the relationship between entrepreneurship and leadership (Vecchio, 2003). (See also chap. 7, this volume.) In many ways, entrepreneurs act as leaders, exerting a powerful influence on the new ventures they found and on the individuals who join these organizations (Vecchio, 2003). To the extent this is true, the principles and findings of research on leadership (e.g., Avolio, 2006; Yukl, 2002) may be applicable to key aspects of entrepreneurship. At the very least, this body of knowledge would add muchneeded methodological and theoretical sophistication to ongoing efforts to understand the factors that influence both entrepreneurs’ behavior and the success of their new ventures. For example, research on leadership has long recognized the importance of a multilevel theoretical perspective, the central role of moderating and mediating processes, and the importance of obtaining multiple dependent measures (e.g., Antonakis & Autio, 2006). Implementation of these approaches would, we believe, contribute significantly to progress in the field of entrepreneurship. Yet another topic well worthy of future investigation is entrepreneurial teams: How do they form and what factors determine their subsequent effectiveness? A considerable body of evidence exists regarding the nature of top management teams, and extending this body of knowledge, as well as general knowledge about the composition and effectiveness of teams (Mannix & Neale, 2005), to entrepreneurship would directly address another key aspect of new venture formation. Seventh, although many investigations have been performed to examine the role of entrepreneurs’ personal characteristics, most of these studies have focused on various aspects of personality (for a review, see Rauch & Frese, 2005, as cited in Rauch & Frese, 2007). Expanding the scope of this research so as to include variables not previously considered— for instance, entrepreneurs’ resistance to or tolerance for intense levels of stress (Kobasa, 1982), their stable levels of affect, and their work-related attitudes and values—would be informative. Given the 266

challenges involved in launching and operating a new venture, these and related factors may well play a key role in the entire process. Yet one more fertile area of future research should be briefly mentioned: cross-cultural investigations of the entrepreneurial process. Clearly, entrepreneurship occurs against a backdrop of cultural influences and factors. This essential fact has been noted in some previous entrepreneurship research (e.g., Tang et al., 2007) but has not yet been systematically examined. Investigating the complex interplay between cultural factors and basic aspects of entrepreneurship (e.g., opportunity recognition, acquisition of essential resources) is certainly a crucial task for the field of entrepreneurship. Finally, it is important that all of these issues and topics be investigated with appropriate research methods. In the past, there has been a dearth of longitudinal studies in entrepreneurship, despite repeated calls for such investigations (e.g., Baron & Shane, 2008). Clearly, such research is essential if we are ever to fully understand the complex, ever-changing processes through which new ventures are conceived, launched, and operated. Further, it is important to note that in the past, cross-sectional research in entrepreneurship has been interpreted as providing evidence for causality or process, when such inferences are actually not justified. For instance, in one recent study, now in press in a leading entrepreneurship journal, entrepreneurs with up to 30 years of experience were compared with students completing their master’s of business administration degrees who had never started a new venture with respect to the type of reasoning or logic they preferred to use in reaching decisions. Results indicated that the experienced entrepreneurs tended to prefer effectual logic to standard causal logic. Although such differences between the comparison groups clearly existed, it is very risky to attribute them, as the author did (see Baron, 2009), to differences in degree of entrepreneurial experience. The research used was clearly what Cook, Campbell, and Peracchio (1990) would describe as a posttest design with nonequivalent groups (in this case, entrepreneurs and master’s of business administration students). As a result, it is inappropriate to conclude that entrepreneurial experience causes a shift toward effectual logic. This, and similar mis-

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

matches of research methods and inferences, must be avoided if entrepreneurship, as a field, is to achieve the progress is seeks. Putting issues of research methodology aside, it is clear that each of the suggestions for future research offered herein relates to a central principle that, in closing, we wish to emphasize: Entrepreneurs are truly the active element in new venture creation; without them and their energetic, overt actions, there would simply be no new ventures (Baron, 2007a; Shane et al., 2003). Given this basic fact, we believe that entrepreneurs should be at, or near, the center of ongoing entrepreneurship research. To many I/O psychologists, this probably sounds like a virtual truism: How can we hope to understand new venture creation without knowing something about the people who actually undertake this task? However, in the past, many researchers who entered the field of entrepreneurship from disciplines other than psychology (e.g., economics, strategy, sociology) tended to view entrepreneurs as constituting a very small—and perhaps relatively unimportant— part of the total picture. In their view, economic factors, technological advances, government policies, and ever-shifting demographics were crucial, and the actions, motives, or skills of specific individuals were of little consequence. Although classical economic theory strongly points to such conclusions, it fails to take note of the fact that, ultimately, specific opportunities are recognized and developed by specific persons. In fact, contrary to what researchers adopting an economic perspective assume, viable opportunities sometimes exist for years or decades until the right person comes along to develop them. For instance, Chester Carlson developed the technology for a practical copy machine (the ancestor of all modern copiers) in 1938, but his technology was not commercially developed until the late 1940s—more than 10 years after it was available. Such instances, which are far from rare, suggest that market forces alone are not responsible for entrepreneurial activity. Rather, individuals play a key role, perhaps because their past experience, training, and personal interests provide them with the cognitive frameworks necessary for noticing and acting upon potentially valuable opportunities. To offer just one example,

consider a new device now in use in millions of kitchens throughout the world and by the expert chefs on every cooking show shown on TV: the Microplane. This new kind of grater handles everything from oranges and lemons to hard cheese in an amazingly effective manner. It is interesting to note that the basic principles on which this “new” device is based existed for many decades—woodworkers had long used similar devices to attain a smooth finish. However, the idea of applying these principles in a food grater did not emerge until the wife of an experienced woodworker (Lorraine Lee) picked up a tool her husband Leonard had brought home from their hardware store and tried it on a lemon. The results were so remarkable that they quickly moved the tool from the “sanding” category in their catalog of tools to the “kitchen” category; the Microplane is now available in many thousands of stores worldwide. In view of this and many other instances, we believe that entrepreneurs are indeed crucial to the entire process and should be at, or very near, the center of ongoing entrepreneurship research. Only when individuals who possess an array of skills, motives, interests, and cognitive frameworks recognize opportunities and then choose to act to develop them does anything concrete occur (McMullen & Shepherd, 2006). Certainly, we recognize that entrepreneurship occurs against a backdrop of environmental and economic factors and that these must be taken carefully into account to grasp the nature of the entire process. However, that in no way detracts from or minimizes the role of individual entrepreneurs or the importance of obtaining systematic knowledge of their motives, skills, perceptions, attitudes, and cognitions. Only by adopting this broader focus, we contend, will we ever fully understand why and how new ventures are launched and the factors that shape the extent to which they grow and prosper (or wither and die). In this respect, and in many others, the theories, methods, and perspectives of I/O psychology can, and do, constitute an invaluable source of insight for entrepreneurship researchers. The flow of benefits, however, is definitely not entirely one way. Entrepreneurship, in turn, offers I/O psychology several potential opportunities (no pun intended!), which can readily be pursued by I/O 267

Copyright American Psychological Association. Not for further distribution.

Baron and Henry

researchers. First, the new ventures entrepreneurs found provide a novel context for testing and refining some of I/O psychology’s basic knowledge (e.g., Baron, in press). In contrast to existing organizations, newly launched ones do not generally have a wellestablished structure or culture. What better environment in which to investigate how these important features emerge? Second, entrepreneurs are, in several respects, a unique population. Investigating entrepreneurs’ behavior in a wide range of contexts, therefore, can help extend the generality of previous findings concerning many important variables and populations. Third, entrepreneurs operate in relatively unusual contexts—ones that offer few clear guidelines concerning how to proceed and move toward key goals. In fact, as has often been noted, entrepreneurs face a situation in which they “make it up as they go along.” Investigating the ways in which they cope with extremely high levels of uncertainty and unpredictability may well provide important insights into the nature of motivation, decision making, and other basic processes—insights not readily obtainable in other, more well-established environments. Finally, it is intriguing to note, once again, that in a sense, all existing organizations, including even the largest and most successful, were at one time new ventures. Thus, when we investigate various aspects of individual and group behavior in existing organizations or consider the complex organizational processes that occur within them, we are, in a very real sense, focusing on a select sample of new ventures: ones that, for reasons still to be determined, survived and prospered. Gaining insights into why and how these particular organizations flourished in the uncertain and perilous environments they faced may greatly broaden our understanding of work settings and the key factors operating within them. In sum, closer dialogue between entrepreneurship and I/O psychology may truly enrich both fields. We close, therefore, with the hope that that this chapter, by encouraging such two-way discourse, contributes to enhanced progress in both fields.

References Adler, P., & Kwon, S. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27, 17–40. doi:10.2307/4134367 268

Aidis, R., & van Praag, M. (2007). Illegal entrepreneurship experience: Does it make a difference for business performance and motivation? Journal of Business Venturing, 22, 283–310. doi:10.1016/j.jbusvent.2006. 02.002 Aldrich, H. E., & Kim, P. H. (2007). Small worlds, infinite possibilities? How social networks affect entrepreneurial formation and search. Strategic Entrepreneurship Journal, 1, 147–165. doi:10.1002/sej.8 Aldrich, H. E., & Ruef, M. (2006). Organizations evolving (2nd ed.). Thousand Oaks, CA: Sage. Amabile, T. M. (1996). Creativity in context. Boulder, CO: Westview Press. Antonakis, J., & Autio, E. (2007). Entrepreneurship and leadership. In J. R. Baum, M. Frese, & R. A. Baron (Eds.), The psychology of entrepreneurship (pp. 189–208). Mahwah, NJ: Erlbaum. Ardichvili, A., Cardozo, R., & Ray, S. (2003). A theory of entrepreneurial opportunity identification and development. Journal of Business Venturing, 18, 105–123. doi:10.1016/S0883-9026(01)00068-4 Ariely, D. (2008). Predictably irrational: The hidden forces that shape our decisions. New York: Harper Collins. Astebro, T., Jeffrey, S. A., & Adomdza, G. K. (2008). Entrepreneurial perseverance: The role of personality and cognition in determining commercialization efforts and commercialization success. Manuscript submitted for publication. Avolio, B. J. (2006). Authentic leadership theory and practice: Origins, effects, and development. Stamford, CT: Elsevier, JAI Press. Baddeley, A. (1997). Human memory. New York: Psychology Press. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman. Baron, R. A. (2000). Thinking about “what might have been.” Journal of Business Venturing, 15, 79–91. doi:10.1016/S0883-9026(98)00024-X Baron, R. A. (2002). OB and entrepreneurship: The reciprocal benefits of closer conceptual links. In B. M. Staw & R. Kramer (Eds.), Research in organization behavior (pp. 225–269). Greenwich, CT: JAI Press. Baron, R. A. (2004). The cognitive perspective: A valuable tool for answering entrepreneurship’s basic “why?” questions. Journal of Business Venturing, 19, 221–240. Baron, R. A. (2006). Opportunity recognition as pattern recognition: How entrepreneurs “connect the dots” to identify new business opportunities. The Academy of Management Perspectives, 20, 104–119.

Entrepreneurship

Baron, R. A. (2007a). Behavioral and cognitive factors in entrepreneurship: Entrepreneurs as the active element in new venture creation. Strategic Entrepreneurship Journal, 1, 167–182. doi:10.1002/sej.12 Baron, R. A. (2007b). Entrepreneurship: A process perspective. In R. Baum, M. Frese, & R. A. Baron (Eds.), The psychology of entrepreneurship (pp. 19–40). Mahwah, NJ: Erlbaum.

Copyright American Psychological Association. Not for further distribution.

Baron, R. A. (2008). The role of affect in the entrepreneurial process. Academy of Management Review, 33, 328–340. Baron, R. A. (2009). “Effectual versus predictive logic in entrepreneurial decision making: Differences between experts and novices”: Does experience in starting new ventures change the way entrepreneurs think? Perhaps, but for now, “caution” is essential. Journal of Business Venturing, 24, 310–315. Baron, R. A. (in press). Job design and entrepreneurship: Why closer connections = mutual gains. Journal of Organizational Behavior. Baron, R. A., & Ensley, M. D. (2006). Opportunity recognition as the detection of meaningful patterns: Evidence from comparisons of novice and experienced entrepreneurs. Management Science, 52, 1331–1344. doi:10.1287/mnsc.1060.0538

Baron, R. A., & Tang, J. (in press). Positive affect, creativity, and innovation in new ventures: A moderated mediation model. Journal of Business Venturing. Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A metaanalysis. Personnel Psychology, 44, 1–26. doi:10.1111/ j.1744-6570.1991.tb00688.x Baum, J. R., Frese, M., & Baron, R. A. (Eds.). (2007). The psychology of entrepreneurship. Mahwah, NJ: Erlbaum. Baum, J. R., & Locke, E. A. (2004). The relationship of entrepreneurial traits, skill, and motivation to subsequent venture growth. Journal of Applied Psychology, 89, 587–598. doi:10.1037/ 0021-9010.89.4.587 Baum, J. R., Locke, E. A., & Smith, K. (2001). A multidimensional model of venture growth. Academy of Management Journal, 44, 292–303. doi:10.2307/ 3069456 Begley, T. M., & Boyd, D. R. (1987). Psychological characteristics associated with performance in entrepreneurial firms and smaller businesses. Journal of Business Venturing, 2, 79–93. doi:10.1016/ 0883-9026(87)90020-6 Blanchflower, D., & Oswald, A. (1998). What makes an entrepreneur? Journal of Labor Journal Economics, 16, 26–60.

Baron, R. A., & Henry, R. A. (in press). How entrepreneurs acquire the capacity to excel: Insights from basic research on expert performance. Strategic Entrepreneurship Journal.

Brockner, J., Higgins, E. T., & Low, M. B. (2004) Regulatory focus theory and the entrepreneurial process. Journal of Business Venturing, 19, 203–220.

Baron, R. A., & Markman, G. D. (2000). Beyond social capital: The role of social skills in entrepreneurs’ success. The Academy of Management Executive, 14, 1–15.

Busenitz, L. W., & Barney, J. B. (1997). Differences between entrepreneurs and managers in large organizations: Biases and heuristics in strategic decisionmaking. Journal of Business Venturing, 12, 9–30. doi:10.1016/S0883-9026(96)00003-1

Baron, R. A., & Markman, G. D. (2005). Toward a process view of entrepreneurship: The changing impact of individual level variables across phases of new venture development. In M. A. Rahim, R. T. Golembiewski, & K. D. Mackenzie (Eds.), Current topics in management (Vol. 9., pp. 45–64). New Brunswick, NJ: Transaction. Baron, R. A., Markman, G. D., & Hirsa, A. (2001). Perceptions of women and men as entrepreneurs: Evidence for differential effects of attributional augmenting. Journal of Applied Psychology, 86, 923–929. doi:10.1037/0021-9010.86.5.923 Baron, R. A., & Shane, S. A. (2008). Entrepreneurship: A process perspective (2nd ed.). Cincinnati, OH: Thomson South-Western. Baron, R. A., & Tang, J. (2009). Entrepreneurs’ social skills and new venture performance: Mediating mechanisms and moderating effects of industry. Journal of Management, 35, 282–306. doi:10.1177/ 0149206307312513

Cardon, M. S., Wincent, J., Singh, J., & Drnovsek, M. (2009). The nature and experience of entrepreneurial passion. Academy of Management Review, 34, 511–532. Chandler, G. N., & Jansen, E. (1992). The founder’s selfassessed competence and venture performance. Journal of Business Venturing, 7, 223–236. doi:10.1016/ 0883-9026(92)90028-P Choi, Y. R., Lévesque, M., & Shepherd, D. A. (2008). When should entrepreneurs expedite or delay opportunity exploitation? Journal of Business Venturing, 23, 333–355. doi:10.1016/j.jbusvent.2006.11.001 Ciavarella, M. A., Bucholtz, A. K., Riordan, C. M., Gatewood, R. D., & Stokes, G. S. (2004). The Big Five and venture success: Is there a linkage? Journal of Business Venturing, 19, 465–483. doi:10.1016/ j.jbusvent.2003.03.001 Collins, C. J., Hanges, P. J., & Locke, E. A. (2004). The relationships of achievement motivation to entrepreneurial behavior: A meta-analysis. Human 269

Baron and Henry

Performance, 17, 95–117. doi:10.1207/ S15327043HUP1701_5 Cook, T. D., Campbell, D. T., & Peracchio, L. (1990). Quasi experimentation. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 1, pp. 492–576). Palo Alto, CA: Consulting Psychologists Press.

Copyright American Psychological Association. Not for further distribution.

Cooper, A. C., Woo, C. Y., & Dunkelberg, W. C. (1988). Entrepreneurs’ perceived chances for success. Journal of Business Venturing, 3, 97–108. doi:10.1016/ 0883-9026(88)90020-1 Cropanzano, R., Goldman, B. M., & Folger, R. (2003). Deontic justice: The role of moral principles in workplace fairness. Journal of Organizational Behavior, 24, 1019–1024. doi:10.1002/job.228 Cummings, T. G. (Ed.). (2007). Handbook of organizational development. London: Sage. Davidsson, P., & Honig, B. (2003). The role of social and human capital among nascent entrepreneurs. Journal of Business Venturing, 18, 301–331. doi:10.1016/S0883-9026(02)00097-6 de Meza, D., & Southey, C. (1996). The borrower’s curse: Optimism, finance and entrepreneurship. The Economic Journal, 106, 375–386. doi:10.2307/2235253 Diener, E., & Seligman, M. E. P. (2002). Very happy people. Psychological Science, 13, 81–84. doi:10.1111/1467-9280.00415 Downs, A. C., & Lyons, P. M. (1991). Natural observations of the links between attractiveness and legal judgments. Personality and Social Psychology Bulletin, 17, 541–547. doi:10.1177/0146167291175009 Eddleston, I. A., & Powell, G. N. (2008). The role of gender identify in explaining sex differences in business owners’ career satisfier preferences. Journal of Business Venturing, 23, 244–256. doi:10.1016/ j.jbusvent.2006.11.002

preneurs’ affect and venture effort. Journal of Applied Psychology, 94, 1086–1094. doi:10.1037/a0015599 Forgas, J. P. (1995). Mood and judgment: The affect infusion model (AIM). Psychological Bulletin, 117, 39–66. doi:10.1037/0033-2909.117.1.39 Forgas, J. P. (Ed.). (2000). Feeling and thinking: Affective influences on social cognition. New York: Cambridge University Press. Forgas, J. P., & George, J. M. (2001). Affective influences on judgments, decision making and behavior in organizations: An information processing perspective. Organizational Behavior and Human Decision Processes, 86, 3–34. doi:10.1006/obhd.2001.2971 Fraser, S., & Greene, F. J. (2006). The effects of experience on entrepreneurial optimism and uncertainty. Economica, 73, 169–192. doi:10.1111/j.1468-0335. 2006.00488.x Frone, M. R. (2003). Work–family balance. In J. C. Quick & L. E. Tetrick (Eds.), Handbook of occupational health psychology (pp. 143–162). Washington, DC: American Psychological Association. doi:10.1037/10474-007 Gaglio, C. M., & Katz, J. (2001). The psychological basis of opportunity identification: Entrepreneurial alertness. Small Business Economics, 16, 95–111. doi:10.1023/ A:1011132102464 Gartner, W. B. (1989). Some suggestions for research on entrepreneurial traits and characteristics. Entrepreneurship Theory and Practice, 14, 27–37. Gartner, W. B., Bird, B. J., & Starr, J. A. (1991). Acting “as if”: Differentiating entrepreneurial from organizational behavior. Entrepreneurship Theory and Practice, 16, 13–31. Grant, A. M., & Ashford, S. J. (2008). The dynamics of proactivity at work. Research in Organizational Behavior, 28, 3–34. doi:10.1016/j.riob.2008.04.002

Ericsson, K. A. (2006). The influence of experience and deliberate practice on the development of superior expert performance. In K. A. Ericsson, N. Charness, P. J. Feltovich, & R. R. Hoffman (Eds.), The Cambridge handbook of expertise and expert performance (pp. 683–703). New York: Cambridge University Press.

Greenhaus, J. H., & Beutell, N. J. (1985). Sources of conflict between work and family roles. Academy of Management Review, 10, 76–88. doi:10.2307/258214

Ferris, G. R., Davidson, S. L., & Perrewé, P. L. (2005). Political skill at work: Impact on work effectiveness. Mountain View, CA: Davies-Black.

Hayward, M. L. A., & Hambrick, D. C. (1997). Explaining the premium paid for large acquisitions: Evidence of CEO hubris. Administrative Science Quarterly, 42, 103–127. doi:10.2307/2393810

Ferris, G. R., Witt, L. A., & Hochwarter, W. A. (2001). Interaction of social skills and general mental ability on job performance and salary. Journal of Applied Psychology, 86, 1075–1082. doi:10.1037/00219010.86.6.1075 Foo, M. D., Uy, M. A., & Baron, R. A. (2009). How do feelings influence effort? An empirical study of entre270

Hahn, U., & Chater, N. (1997). Concepts and similarity. In K. Lamberts & D. Shanks (Eds.), Knowledge concepts and categories (pp. 43–92). Cambridge, MA: MIT Press.

Hayward, M. L. A., Shepherd, D. A., & Griffin, D. (2006). A hubris theory of entrepreneurship. Management Science, 52, 160–172. doi:10.1287/mnsc.1050.0483 Herron, L., & Sapienza, H. J. (1992). The entrepreneur and the initiation of new venture launch activities. Entrepreneurship Theory and Practice, 16, 49–55.

Entrepreneurship

Higgins, E. T. (1998). Promotion and prevention: Regulatory focus as a motivational principle. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 30, pp. 1–46). New York, NY: Academic Press. Hitt, M. A., Beamish, P. W., Jackson, S. E., & Mathieu, J. E. (2007). Building theoretical and empirical bridges across level: Multilevel research in management. Academy of Management Journal, 50, 1385–1399.

Copyright American Psychological Association. Not for further distribution.

Hmieleski, K., & Baron, R. A. (2008). A contextual study of entrepreneurial self-efficacy. Strategic Entrepreneurship Journal, 2, 57–72. doi:10.1002/sej.42 Hmieleski, K., & Baron, R. A. (2009). Entrepreneurs’ optimism and new venture performance: A social cognitive perspective. Academy of Management Journal, 52, 473–488. Hochwarter, W. A., Ferris, G. R., Gavin, M. B., Perrewé, P. L., Hall, A. T., & Frink, D. D. (2007). Political skill as a neutralizer of felt accountability–job tension effects on job performance ratings: A longitudinal investigation. Organizational Behavior and Human Decision Processes, 102, 226–239. doi:10.1016/j.obhdp.2006.09.003 Hochwarter, W. A., Witt, L. A., Treadway, D. C., & Ferris, G. R. (2006). The interaction of social skill and organizational support on job performance. Journal of Applied Psychology, 91, 482–489. doi:10.1037/0021-9010.91.2.482 Ireland, R. D., Hitt, M. A., & Sirmon, D. G. (2003). A model of strategic entrepreneurship: The construct and its dimensions. Journal of Management, 29, 963–989. Isen, A. M. (1999). Positive affect. In T. Dalgleish & M. J. Power (Eds.), Handbook of cognition and emotion (pp. 521–539). Chichester, England: Wiley. Kaish, S., & Gilad, B. (1991). Characteristics of opportunities search or entrepreneurs versus executives: Sources, interests, general alertness. Journal of Business Venturing, 6, 45–61. Katz, J. E., & Shepherd, D. (Eds.). (2003). Cognitive approaches to entrepreneurship research. New York: JAI Press. Kirzner, I. M. (1979). Perception opportunity, and profit. Chicago: University of Chicago Press. Kirzner, I. M. (1985). Discovery and the capitalist process. Chicago: University of Chicago Press. Knight, E. H. (1921). Risk, uncertainty, and profit. New York: Kelly and Millman. Kobasa, S. C. (1982). The hardy personality: Toward a social psychology of stress and health. In J. Suls & G. Sanders (Eds.), The social psychology of health and illness (pp. 3–32). Hillsdale, NJ: Erlbaum.

Kozlowski, S. W. J., & Ilgen, D. R. (2006). Enhancing the effectiveness of work teams. Psychological Science in the Public Interest, 7, 77–124. Krueger, N. J., & Brazeal, D. H. (1994). Entrepreneurial potential and potential entrepreneurs. Entrepreneurship Theory and Practice, 19, 91–104. Lévesque, M., Shepherd, D. A., & Douglas, D. A. (2002). Employment or self-employment: A dynamic utilitymaximizing model. Journal of Business Venturing, 17, 189–210. doi:10.1016/S0883-9026(00)00063-X Lewicki, R. J., Saunders, D. M., & Barry, M. (2005). Negotiation. New York: McGraw-Hill/Irwin. Lichtenstein, B., Dooley, K. J., & Lumpkin, G. T. (2006). Measuring emergence in the dynamics of new venture creation. Journal of Business Venturing, 21, 153–175. doi:10.1016/j.jbusvent.2005.04.002 Locke, E. A. (2000). The prime movers: Traits of the great wealth creators. New York: AMACOM. Locke, E. A., & Baum, J. R. (2007). Entrepreneurial motivation. In J. R. Baum, M. Frese, & R. A. Baron (Eds.), The psychology of entrepreneurship (pp. 9–112). Mahwah, NJ: Erlbaum. Locke, E. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation: A 35-year odyssey. American Psychologist, 57, 705–717. doi:10.1037/0003-066X.57.9.705 Lumpkin, G. T. (2007). Intrapreneurship and innovation. In J. R. Baum, M. Frese, & R. A. Baron (Eds.), The psychology of entrepreneurship (pp. 237–264). Mahwah, NJ: Erlbaum. Lyubomirsky, S., King, L., & Diener, E. (2005). The benefits of frequent positive affect: Does happiness lead to success? Psychological Bulletin, 131, 803–855. doi:10.1037/0033-2909.131.6.803 Mannix, E., & Neale, M. A. (2005). What differences make a difference? The promise and reality of diverse teams in organizations. Psychological Science in the Public Interest, 6, 31–55. doi:10.1111/j.15291006.2005.00022.x Markman, G. D., Balkin, D. B., & Baron, R. A. (2002). Inventors and new venture formation: The effects of general self-efficacy and regretful thinking. Entrepreneurship Theory and Practice, 27, 149–165. doi:10.1111/1540-8520.00004 Markman, G. D., Baron, R. A., & Balkin, D. B. (2005). Are perseverance and self-efficacy costless? Assessing entrepreneurs’ regretful thinking. Journal of Organizational Behavior, 26, 1–19. doi:10.1002/job.305 Matlin, M. W. (2002). Cognition (5th ed.). Fort Worth, TX: Harcourt College. Mayer, K. B., & Goldstein, S. (1964). Manual workers and small businessmen. In A. B. Shostak & W. Bomgerg (Eds.), Blue-collar world: Studies of the 271

Baron and Henry

American worker (pp. 537–549). Englewood Cliffs, NJ: Prentice Hall. McClelland, D. (1961). The achieving society. New York: Free Press. McMullen, J. S., & Shepherd, D. A. (2006). Entrepreneurial action and the role of uncertainty in the theory of the entrepreneur. Academy of Management Review, 31, 132–152.

Copyright American Psychological Association. Not for further distribution.

Miner, J. B., & Raju, N. S. (2004). When science divests itself of its conservative stance: The case of risk propensity differences between entrepreneurs and managers. Journal of Applied Psychology, 89, 3–13. doi:10.1037/0021-9010.89.1.3 Mitchell, R. K., Busenitz, L., Bird, B., Gaglio, C. M., McMullen, J. S., Morse, E. A., & Smith, J. B. (2007). The central question in entrepreneurial cognition research. Entrepreneurship Theory and Practice, 31, 1–27. Molden, D. C., & Higgins, E. T. (2005). Motivated thinking. In K. J. Holyoak & R. G. Morrison (Eds.), The Cambridge handbook of thinking and reasoning (pp. 295–317). London: Cambridge University Press. Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational advantage. Academy of Management Review, 23, 242–266. doi:10.2307/259373 Naylor, J. C., Pritchard, R. D., & Ilgen, D. R. (1980). A theory of behavior in organizations. New York: Academic Press. Nosofsky, R. M., & Palmeri, T. J. (1998). A rule-plusexception model for classifying objects in continuousdimension spaces. Psychonomic Bulletin & Review, 5, 345–369. Ozgen, E., & Baron, R. A. (2007). Social sources of information in opportunity recognition: Effects of mentors, industry networks, and professional forums. Journal of Business Venturing, 22, 174–192. doi:10.1016/j.jbusvent.2005.12.001 Parker, S. K., Williams, H., & Turner, N. (2006). Modeling the antecedents of proactive behavior at work. Journal of Applied Psychology, 91, 636–652. doi:10.1037/0021-9010.91.3.636 Patzelt, H., Shepherd, D. A., Deeds, D., & Bradley, S. W. (2008). Financial slack and venture managers’ decisions to seek a new alliance. Journal of Business Venturing, 23, 465–481. doi:10.1016/j.jbusvent.2007. 06.001 Portes, A. (1998). Social capital. Annual Review of Sociology, 24, 1–24. doi:10.1146/annurev.soc.24.1.1 Putnam, F. (2000). Bowling alone: The collapse and revival of American community. New York: Simon & Schuster. Rauch, A., & Frese, M. (2007). Born to be an entrepreneur? In J. R. Baum, M. Frese, & R. A. Baron (Eds.), 272

The psychology of entrepreneurship (pp. 41–66). Mahwah, NJ: Erlbaum. Rauch, A., Wiklund, J., Frese, M., & Lumpkin, G. T. (2005). Entrepreneurial orientation and business performance: Cumulative empirical evidence. Manuscript submitted for publication. Riggio, R. E. (1986). Assessment of basic social skills. Journal of Personality and Social Psychology, 51, 649–660. doi:10.1037/0022-3514.51.3.649 Riggio, R. E., & Throckmorton, B. (1988). The relative effects of verbal and nonverbal behavior, appearance, and social skills on valuations made in hiring interviews. Journal of Applied Psychology, 18, 331–348. doi:10.1111/j.1559-1816.1988.tb00020.x Robbins, T. L., & DeNisi, A. S. (1994). A closer look at interpersonal affect as a distinct influence on cognitive processing in performance evaluations. Journal of Applied Psychology, 79, 341–353. doi:10.1037/00219010.79.3.341 Roese, N. J., & Olson, J. M. (1997). Counterfactual thinking: The intersection of affect and function. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 29, pp. 1–59). New York: Academic Press. doi:10.1016/S0065-2601(08)60015-5 Schjoedt, L., & Shaver, K. G. (2004, June). Does the potential for increased work and life satisfaction affect the decision to pursue an entrepreneurial career? An examination of the PSED data. Paper presented at the first annual Clemson/Kauffman Symposium on the PSED, Clemson, SC. Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: Detection, search, and attention. Psychological Review, 84, 1–66. doi:10.1037/0033-295X.84.1.1 Schumpeter, J. (1911). The theory of economic development. Cambridge, MA: Harvard University Press. Seibert, S. E., Kraimer, M. L., & Liden, R. C. (in press). A social capital theory of career success. Academy of Management Journal. Shane, S. (2000). Prior knowledge and the discovery of entrepreneurial opportunities. Organization Science, 11, 448–469. doi:10.1287/orsc.11.4.448.14602 Shane, S. (2003). The individual–opportunity nexus approach to entrepreneurship. Aldershot, England: Edward Elgar Smith. Shane, S. (2008). The illusions of entrepreneurship: The costly myths that entrepreneurs, investors, and policy makers live by. New London, CT: Yale University Press. Shane, S., Locke, E. A., & Collins, C. (2003). Entrepreneurial motivation. Human Resource Management Review, 13, 257–279. doi:10.1016/S1053-4822(03) 00017-2

Copyright American Psychological Association. Not for further distribution.

Entrepreneurship

Shane, S., & Venkataraman, S. (2000). The promise of entrepreneurship as a field of research. Academy of Management Review, 25, 217–226. doi:10.2307/ 259271

Tang, R. L., Walls, J., & Frese, M. (2007). Cross-cultural entrepreneurship: The case of China. In J. R. Baum, M. Frese, & R. A. Baron (Eds.), The psychology of entrepreneurship (pp. 265–286). Mahwah, NJ: Erlbaum.

Shaver, K., & Scott, L. (1991). Person, process, choice: The psychology of new venture creation. Entrepreneurship Theory and Practice, 16, 23–45.

Vecchio, R. P. (2003). Entrepreneurship and leadership: Common trends and common threads. Human Resource Management Review, 13, 303–327.

Simmel, G. (1995). Conflict and the web of group affiliations (W. K. Bendix, Trans.). Glencoe, IL: Free Pres.

Vroom, V. (1964). Work and motivation. New York: McGraw-Hill.

Simon, M., Houghton, S. M., & Aquino, K. (2000). Cognitive biases, risk perception, and venture formation: How individual decide to start companies. Journal of Business Venturing, 15, 113–134. doi:10.1016/S0883-9026(98)00003-2

Wayne, S. J., Liden, R. C., Graf, I. K., & Ferris, G. R. (1997). The role of upward influence tactics in human resource decisions. Personnel Psychology, 50, 979–1006. doi:10.1111/j.1744-6570.1997. tb01491.x

Small Business Association. (2001). New and small firm creation. Washington, DC: Government Printing Office.

Yukl, G. A. (2002). Leadership in organizations (5th ed.). Upper Saddle River, NJ: Prentice Hall.

Sternberg, R. J. (2004). Successful intelligence as a basis for entrepreneurship. Journal of Business Venturing, 19, 189–201. doi:10.1016/S0883-9026(03)00006-5

Zahra, S. A., & Hayton, J. C. (2008). The effect of international venturing on firm performance: The moderating influence of absorptive capacity. Journal of Business Venturing, 23, 195–220.

Stewart, W. H., & Roth, P. L. (2001). Risk propensity differences between entrepreneurs and managers: A meta-analytic review. Journal of Applied Psychology, 86, 145–153. doi:10.1037/0021-9010.86.1.145 Suchman, M. C. (1995). Managing legitimacy: Strategic and institutional approaches. Academy of Management Review, 20, 571–610. doi:10.2307/258788 Swets, J. A. (1992). The science of choosing the right decision threshold in high-stakes diagnostics. American Psychologist, 47, 522–532. doi:10.1037/0003-066X. 47.4.522

Zhao, H., & Seibert, S. E. (2006). The Big Five personality dimensions and entrepreneurial status: A meta-analytical review. Journal of Applied Psychology, 91, 259–271. doi:10.1037/0021-9010. 91.2.259 Zhao, H., Seibert, S. E., & Hills, G. E. (2005). The mediating role of self-efficacy in the development of entrepreneurial intentions. Journal of Applied Psychology, 90, 1265–1272. doi:10.1037/0021-9010. 90.6.1265

273

CHAPTER 9

Copyright American Psychological Association. Not for further distribution.

DEEPENING OUR UNDERSTANDING OF CREATIVITY IN THE WORKPLACE: A REVIEW OF DIFFERENT APPROACHES TO CREATIVITY RESEARCH Jing Zhou and Christina E. Shalley

At the societal level, creativity is essential for economic growth and social progress (Florida, 2004; Schumpeter, 1939). At the individual, team, and organization levels, it has also been argued that creativity is a key enabler and contributor to performance, entrepreneurship, growth, and competitiveness (Amabile, 1996; Oldham & Cummings, 1996; Shalley, 1991; Woodman, Sawyer, & Griffin, 1993; Zhou, 1998; Zhou & Shalley, 2008a). More important, this theoretical notion has started to receive encouraging, although at this stage still largely suggestive, support from empirical studies (Gilson, 2008). In the past 10 years, there have been a number of comprehensive reviews concerned with creativity in organizations. Therefore, in this review, our goal is to categorize previous theorizing and research on creativity into three broad approaches that capture three types of psychological processes—motivational, cognitive, and affective—rather than to provide an exhaustive list of variables investigated to date and details of previous studies’ designs and findings. It is our hope that by organizing previous theory and research into these three broad categories that represent different aspects of psychological processes, we can assist researchers to delve deeper into an understanding of what factors promote or inhibit creativity, how they influence creativity, and, above all, why these effects occur. Toward this goal, we only review representative studies in each of the three conceptual categories and briefly mention some other studies so that interested readers can follow up on them if

they wish. Selection criteria of studies included are (a) studies conducted in the workplace and (b) studies conducted in a controlled environment, such as behavioral laboratories, with the variables investigated having clear implications for creativity in the workplace. We refer interested readers to several comprehensive review articles, the majority of which have been published in the past few years: Anderson, De Dreu, and Nijstad (2004); Mumford and Gustafson (1988); Shalley, Zhou, and Oldham (2004); and Zhou and Shalley (2003), as well as a recently published volume devoted to covering theorizing and research on creativity in the workplace, Handbook of Organizational Creativity edited by Zhou and Shalley (2008b). BACKGROUND In this section, we present a commonly accepted definition of creativity and discuss its relationship with the concept of innovation. In addition, webriefly review the major research methods used in organizational behavior and organizational psychology for studying creativity.

Creativity Defined Creativity refers to the production of new and useful ideas concerning products, services, processes, and procedures (e.g., Amabile, 1996; Oldham & Cummings, 1996; Shalley, 1991; Woodman et al., 1993; Zhou, 1998). With regard to workplace creativity, both novelty and usefulness are necessary

We thank Sheldon Zedeck for his many helpful comments on earlier versions of this chapter.

http://dx.doi.org/10.1037/12169-009 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

275

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

conditions for something to be regarded as creative. If either one of them is absent, an idea, for example, would not be judged as creative. Therefore, an idea could be very unique and novel, but if it is not useful or feasible or does not have the potential to create value, it would not be considered creative. For example, if a NASA engineer working in the Mars program were to propose that in designing a spaceship that would be used to send astronauts to Mars, NASA should adapt the design of the spaceship used to send astronauts to the moon, it would be considered a useful idea. However, if the engineers were to propose that astronauts should be riding bicycles to Mars, this would not be considered a useful idea (although it is novel and unique), and indeed, it is a bizarre idea! In this regard, creativity research in organizational behavior and organizational psychology is different from the definition of creativity used by researchers in other fields. Creativity can be exerted by individual employees or a team of employees working together. Employees holding virtually all kinds of jobs, in all functional areas and at all levels of the organization, have the potential to be creative at work (Amabile, 1996; Oldham & Cummings, 1996; Shalley, Gilson, & Blum, 2000, Woodman et al., 1993; Zhou, 1998), although there are individual differences in terms of the magnitude of their potential. Also, the level of creativity realized can vary from something that is novel yet somewhat incremental to what may already be known to exist to something that is a radically new and different idea, product, or process. A related concept is innovation. The primary difference between definitions of creativity and innovation is that whereas creativity emphasizes the production of new and useful ideas by individuals and teams, innovation emphasizes the implementation of new ideas or practices in a unit or throughout an organization. For example, the Six Sigma program was the result of creativity at Motorola because it was created there. However, it was an innovation at General Electric because the original idea or program was created elsewhere (i.e., at Motorola), but General Electric implemented it throughout the company. As such, innovation can involve ideas invented outside of the focal organization, whereas creativity must involve the employees of an organization. In this sense, orga276

nizations that can effectively promote and use their employees’ creativity are said to do a better job of fully leveraging their employees’ capabilities and possess greater competitive advantage.

Research Setting, Design, and Measurement of Creativity Researchers have primarily used laboratory experiments, quantitative field studies, and qualitative field studies to investigate creativity. Each of these is discussed in the following sections. Laboratory research and creativity measurement. One unique feature of experimental research on creativity is that the experimental tasks need to be complex and open ended rather than simple and algorithmic or with demonstrably right or wrong answers (Amabile, 1996; McGraw, 1978; Zhou & Shalley, 2003). Some examples of tasks that have been used in more than three laboratory studies that may have direct or indirect implications for creativity in the workplace include (a) Amabile’s (1996) artistic, problem-solving, and verbal tasks; (b) Torrance’s (1974) unusual uses task; and (c) problem-solving tasks specifically targeting commonly seen problems in the workplace, such as Shalley’s (1991) memo task. Many studies using tasks in the first and last categories have followed Amabile’s consensual assessment technique to evaluate the extent to which research participants’ outputs are creative, whereas studies using tasks in the second category have used different criteria, which we discuss shortly. Amabile and colleagues have used three types of tasks in their creativity research program: artistic tasks (e.g., making collages and paintings), problemsolving tasks (e.g., generating ideas for products), and verbal tasks (e.g., writing poems and stories; Amabile, 1996). These tasks had been used with schoolchildren, artists, college students, and MBA students. For example, Amabile (1979) conducted an experiment in which participants were asked to work on a collage task. The experimenter gave each participant a piece of cardboard, a bottle of glue, and colored papers and asked the participant to make collages. The creativity of the collages was then measured by following Amabile’s (1979) consensual assessment technique, which is premised on an

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

operational definition of creativity—the extent to which an outcome resulting from working on a task is judged as creative by a panel of judges. The judges need to be capable of and appropriate for judging the particular task, and their ratings need to be reliable. More specifically, a panel of judges, who were artists familiar with making collages, independently rated the collages produced by the participants on several dimensions related to creativity, such as novelty of the idea, novelty of materials use, and complexity. A composite creativity score was then created by combining the normalized ratings on these dimensions, and an interjudge reliability index was calculated. Because the index showed satisfactory interjudge reliability (e.g., greater than .70), ratings by this panel of judges were averaged to create a creativity measure for each research participant. In Torrance’s (1947) unusual uses task, research participants are asked to generate unusual uses for objects such as a brick, newspaper, tire, and hanger. Derived from the notion that divergent thinking and creativity are closely related (Guilford, 1956), researchers typically obtain a set of indicators to measure the participants’ creativity on the unusual uses task. These indicators are the same indicators used in ideation tasks and group brainstorming research (e.g., Paulus, 2008). They may include fluency (i.e., total number of nonredundant uses generated for the objects), flexibility (i.e., how many different categories of ideas are generated), elaboration (i.e., whether the ideas are well developed), and originality (i.e., the extent to which the ideas are unusual or statistically infrequent). Research participants’ responses are evaluated on all or a subset of these dimensions by a panel of judges. For each of these dimensions, the judges’ ratings are averaged if their reliability scores are satisfactory (e.g., greater than .70). Note that although this task allows for measures focusing on divergent thinking, verbal fluency is essential for demonstrating high levels of divergent thinking and creativity on this task. As such, this task implies that verbal fluency is a critical component of creativity, which is a rather narrow definition of creativity (Amabile, 1996). In addition, one might argue that the unusual uses task does not resemble the most commonly seen creativity in the workplace. Finally, although creativity is defined as

the generation of new and useful ideas, some of the earlier laboratory studies measured creativity in terms of verbal fluency (e.g., Torrance, 1974) and complexity (e.g., Amabile, 1979). The definition and operationalization of creativity need to be matched. We recommend that in designing future studies, researchers provide a clear rationale that justifies their measuring indicators of creativity other than those that closely reflect the definition of creativity (i.e., novelty, usefulness, and overall creativity). However, some researchers prefer to use problemsolving tasks that closely represent commonly seen creativity in organizations, such as solving management problems (e.g., Shalley, 1991) and designing new products (e.g., Pearsall, Ellis, & Evans, 2008). For example, Shalley (1991) designed a memo task in which a series of human resource management– related problems are presented. Research participants are asked to play the role of the human resource management director and produce creative solutions to those problems. Following Amabile’s (1996) consensual assessment technique, a panel of appropriate judges (e.g., graduate students who have had significant management experience) is asked to rate the extent to which each idea or solution generated by each research participant is creative. If the judges’ ratings are reliable (e.g., greater than .70), the ratings are averaged across memos and judges to create an overall measure of creativity. Sometimes, in keeping with the definition of creativity, researchers also ask judges to rate the extent to which the judges’ ratings on participants’ solutions to Shalley’s (1991) memo problems are (a) useful and (b) novel. The researchers calculate interjudge reliability of the usefulness ratings and the novelty ratings, respectively. If the reliability scores are greater than .70, they then take the average ratings across memos and across judges to create an overall usefulness score and an overall novelty score. Finally, because by definition creativity needs to be both novel and useful, the researchers multiply the usefulness and the novelty scores to obtain a creativity measure. This approach was used in Zhou and Oldham (2001). Instead of calculating interjudge reliabilities alone, researchers sometimes also use a more rigorous approach by calculating both interjudge reliability and interjudge agreement. This is because interjudge 277

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

reliability reflects the extent to which the judges’ ratings are consistent or parallel, and interjudge agreement indicates the degree to which the judges assign the same ratings to the research participants’ responses (cf. Zhou & Shalley, 2003). Calculating both indices may reveal a more comprehensive picture of the compatibility of the judges’ ratings than calculating any one of these indices alone. An illustration of using this approach can be found in Zhou (1998). Laboratory studies are commonly used when researchers investigate effects of contextual factors on creativity or effects of the interaction between contextual factors and individual differences. To illustrate the laboratory design, let us take a close look at a study conducted by Zhou (1998). To save space and highlight the main issues concerning the laboratory experimental design, we focus on one main hypothesis tested in Zhou’s study rather than reviewing all hypotheses and results in their complexity and comprehensiveness. Using the motivational approach to creativity, Zhou hypothesized that feedback valence and feedback style would interact to affect creativity in such as way that individuals would exhibit the greatest creativity when they received positive feedback delivered in an informational style and would demonstrate the least creativity when they received negative feedback delivered in a controlling style. Feedback valence is defined as the positive or negative outcome resulting from a comparison between an individual’s creative output and normative or situational criteria. Feedback style is defined as the manner in which feedback is delivered: informational or controlling. In a laboratory experiment using a 2 × 2 factorial design, Zhou (1998) manipulated feedback valence at two levels (positive and negative) and feedback style also at two levels (informational and controlling). Research participants were randomly assigned to the four experimental conditions (positive feedback delivered in an informational style, positive feedback delivered in a controlling style, negative feedback delivered in an informational style, and negative feedback delivered in a controlling style). They worked on the aforementioned memo task designed by Shalley (1991). Using the consensual assessment technique developed by Amabile (1996), 278

Zhou asked three judges to rate the extent to which the solutions generated by the participants to solve the problems presented in the memos were creative. The reliability of the judges’ ratings was .71 (Jones, Johnson, Butler, & Main, 1983; Shrout & Fleiss, 1979; Tinsley & Weiss, 1975), indicating that the judges’ ratings were consistent and parallel. The agreement of the judges’ ratings was also satisfactory, χ2(1, N = 210) = 14.33, p < .05 (Lawlis & Lu, 1972; Tinsley & Weiss, 1975). Thus, the interjudge reliability and agreement results showed that it was appropriate to average the judges’ ratings to create an overall measure of creativity. Results obtained from a regression analysis using creativity measured after the participants had received their feedback (the main manipulation) as the dependent variable, and independent variables including feedback valence and style, supported the main hypothesis, stated earlier (ΔR2 = .01, p < .05). As is seen later in this chapter, our review of the three approaches to creativity research (motivational, cognitive, and affective) suggests that laboratory studies have been used in much of the earlier (e.g., before 2000) research guided by the motivational approach. Although laboratory studies continue to be conducted, studies have more recently been conducted in field settings using correlational design. In addition, laboratory studies have been used in many studies influenced by the cognitive approach. Finally, although studies using the affective approach have used both experimental and correlational designs, it is interesting to note that with two exceptions (i.e., Amabile, Barsade, Mueller, & Staw, 2005; Madjar, Oldham, & Pratt, 2002), most studies demonstrating positive affect’s relationship to creativity were conducted in the laboratory, whereas most of the studies showing facilitative effects of negative affect or joint effects of positive and negative affect (with the exception of Fong, 2006) have been conducted in field settings. Certainly, whether one should use an experimental design in the behavioral laboratory, a correlational or a longitudinal design in organizations depends on the priorities and focus of the research, and it is a trade-off between being able to manipulate the variables in a clean fashion and establish causality and considerations of external validity (Zhou & Shalley,

Understanding Creativity in the Workplace

Copyright American Psychological Association. Not for further distribution.

2003). As a field of inquiry in organizational behavior and organizational psychology, the creativity research field is still young, with many interesting questions to be answered and much knowledge to be discovered (Shalley & Zhou, 2008). Consequently, we believe that a variety of research designs and methods, including experimental studies in the lab, field correlational or longitudinal studies, and even field experimental studies are all valuable. Quantitative field studies. An increasing number of studies have been conducted in work organizations. These field studies have tended to be cross-sectional, correlational studies in that they typically use the survey method to collect data from employees and in some cases also supervisors. Creativity is usually the dependent variable in these field studies. It is typically measured by asking the focal employees’ supervisors to fill out a survey that contains a creativity scale. To illustrate a typical quantitative field study design, let us take a close look at a study conducted by Oldham and Cummings (1996). Because our purpose here is to illustrate the field study design, instead of reviewing the full detail of the investigation we again focus on one of their main hypotheses. Using the motivational approach to creativity research, Oldham and Cummings (1996) hypothesized that creative personalities, job complexity, and supervision would interact such that when employees have more creative personalities, hold complex jobs, and receive supportive and noncontrolling supervision, they would exhibit the highest levels of creativity. The researchers asked employees to fill out questionnaires measuring their creative personalities, the complexity of their jobs, and their supervisors’ support and control. These employees’ supervisors filled out separate questionnaires in which they rated their employees’ creativity. Regression results lent support to the hypothesis stated earlier (ΔR2 = .05, p < .05). In field studies, the most commonly used scales for measuring creativity include Oldham and Cummings’s (1996) 3-item scale (all items are listed in the Appendix in Oldham & Cummings, 1996); Scott and Bruce’s (1994) 6-item scale (the entire scale is listed in the Appendix in Scott & Bruce, 1994); Tierney, Farmer, and Graen’s (1999) 9-item

scale (all items are listed in the Appendix in Tierney et al., 1999); and Zhou and George’s (2001) 13-item scale (all items are listed in the Appendix in Zhou & George, 2001). These scales have usually shown satisfactory validity and reliability. For example, Oldham and Cummings reported a Cronbach’s alpha of .90 on their scale, Scott and Bruce reported that the Cronbach’s alpha for their scale was .89, Tierney et al. obtained a Cronbach’s alpha of .95 for their scale, and Zhou and George obtained a Cronbach’s alpha of .96 on their scale. Nonetheless, focused studies on scale development and comparison among the scales have been rare (Zhou & Shalley, 2003). When research samples are drawn from scientists and engineers working in the research and development function in organizations and when such data are available, archival data (e.g., number of patents obtained) have also been collected from participating companies’ records. Similar to studies on other behaviors or behavioral outcomes in organizations, the variance (R2) in creativity explained by many creativity studies has ranged from the low to mid-teens and in general has not exceeded .20. One of the emergent issues in field research on creativity is whether measures of research participants’ own personal assessment of engaging in creative activities would add value to creativity research. Some researchers have argued that the answer is affirmative because creativity is a process (Drazin, Glynn, & Kazanjian, 1999) and the focal individuals themselves are the first to be aware of their engagement in creative activities, whereas others such as supervisors and coworkers are likely to only notice and observe the individuals’ creativity at later stages, when creative outcomes have been produced. As such, to fully understand the creative process, we also need to document how individuals assess and report their own creativity (Hocevar, 1981; Hocevar & Bachelor, 1989). Toward this end, several recent studies asked the focal employees to report their own creativity at work (Carmeli & Schaubroeck, 2007; Kark & Carmeli, 2008; Shalley, Gilson, & Blum, 2009; Zhou, Shin, & Cannella, 2008). Compared with experimental studies conducted in the lab, field studies conducted in work organizations with actual employee samples have the potential to be 279

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

more generalizable and have greater external validity. However, studies conducted in field settings are by no means necessarily superior to experimental studies conducted in the lab. For example, because of constraints related to access, feasibility, and resources, most field studies have been cross-sectional. Usually, employees are invited to fill out questionnaires that contain measures of independent variables, and around the same time their supervisors are invited to fill out different questionnaires that contain ratings of their employees’ creativity (see Zhou & Shalley, 2003, for more detailed descriptions of field study design commonly used in research on creativity). Because of this cross-sectional design, many field studies cannot unequivocally determine direction of causality. For these reasons, researchers have recognized the need to conduct more longitudinal studies and field experiments. However, as Zhou and Shalley (2003) discussed, these more rigorous designs have been rare (see Amabile & Conti, 1999, and Amabile et al., 2005, for exceptions) because of resources and accessibility constraints. Qualitative field studies. Although the vast majority of field studies documented in the literature have been quantitative studies, there have also been a few qualitative field studies, including studies conducted by Hargadon and Sutton (1997) and Hargadon and Bechky (2006). To illustrate, let us take a close look at a qualitative field study conducted at a design firm and reported in Hargadon and Sutton (1997). On the basis of their on-site observations, interviews, and archival data, these researchers developed a grounded theoretical model in which they described and explained how the design firm used its networks and organizational memory to design creative products. Taking advantage of its network positions resulting from serving clients in 40 industries and its organizational memory systems and routines, the firm acted as a technology broker by learning and transferring technology and knowledge from one place to another place where such technology or knowledge was considered novel and useful and by recombining known technology, knowledge, and materials in new and useful ways. Because the creativity research field is relatively young (Shalley & Zhou, 2008) and because well280

executed qualitative studies could generate framebreaking and testable insights (Eisenhardt, 1989), more qualitative studies should be conducted to complement quantitative studies. Note that in the laboratory experiments and quantitative field studies we have reviewed, the measurement of creativity is consistent with the definition of creativity as an outcome, not as individual differences or individuals’ potential to be creative. For researchers who wish to study creativity as individuals’ potential rather than as actual outcome or behavioral responses, there are several self-report scales for measuring the potential to be creative. For example, Gough (1979) developed the Creative Personality Scale, in which he used a set of adjectives to differentiate creative from noncreative personalities. As another example, the Openness to Experience factor in the five-factor personality model (Costa & McCrae, 1992) may also be used to measure individuals’ creative potential. (See Vol. 2, chap. 5, this handbook.) We want to caution, however, that results concerning the correlations between each of the five factors (i.e., Openness to Experience, Conscientiousness, Agreeableness, Neuroticism, and Extraversion) in the five-factor personality model and creativity have been mixed. On one hand, in studies comparing artists with nonartists or scientists with nonscientists, among the five broad personality factors (not just one or two of their subcomponents), Openness to Experience was the factor that was most consistently related to creativity (Feist, 1998, 1999). Individuals high on Openness to Experience are those who are broad minded, curious, and untraditional. They are thought to be more flexible in absorbing information, combining new and unrelated ideas, and having a higher need to seek new experiences and perspectives. On the other hand, in studies on everyday employees (e.g., not artists or scientists) in work organizations, Openness to Experience was not directly related to creativity (e.g., George & Zhou, 2001). Interestingly, Feist (1998) showed that when comparing artists with nonartists, artists showed lower levels of conscientiousness than nonartists. In contrast, when comparing scientists with nonscientists, scientists showed higher levels of conscientiousness than nonscientists. In addition, compared with nonscientists,

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

scientists had higher confidence, which is a subcomponent of Extraversion. Some researchers have looked at individuals’ preferred problem-solving style as a predictor of how creative they would be (e.g., Jabri, 1991; Kirton, 1976, 1994). For example, Kirton’s (1976) adaption– innovation theory posited that individuals have a preferred style of problem solving. He proposed a bipolar continuum of cognitive styles with adaptors and innovators located at opposite ends. Adaptors prefer to problem solve using known procedures and paradigms, whereas innovators prefer to take risks and violate traditional ways of doing things to develop unique solutions. As such, this scale provides a measure of individuals’ problem-solving behaviors that are considered more or less creative, with the assumption that those who are more creative will be more likely to contribute to innovations. Cross-level research design. In a recent discussion on future directions of research on creativity in the workplace, Zhou and Shalley (2008b) suggested that cross-level research is one of the most promising future research directions. For example, integrating previous theories and research concerning achievement orientation, trait activation, and team learning, Hirst, van Knippenberg, and Zhou (2009) developed and tested a cross-level model theorizing how and why (a) individual differences in goal orientation affected individuals’ creativity and (b) goal orientation at the individual level of analysis and team learning behavior at the team level of analysis interacted to affect individuals’ creativity. Their results showed (a) learning goal orientation was positively related to creativity; (b) learning goal orientation interacted with team learning behavior so that when team learning behavior was high, learning goal had a cubic relation with creativity in such a way that learning orientation had a stronger, positive relation with creativity at moderate levels than at lower or higher levels; and (c) approach goal orientation interacted with team learning behavior so that when team learning behavior was high, approach goal was positively related to creativity. The full cross-level interaction model explained 10% of the variance in creativity. In sum, researchers have used laboratory studies, quantitative field studies, and qualitative field studies

to investigate creativity, with each method having its strengths and weaknesses (see Zhou & Shalley, 2003, for a fuller discussion on this issue). As is the case for most research topics in organizational behavior and organizational psychology, the use of multiple methods would be beneficial. We now turn to a review of previous theorizing and empirical findings concerning creativity in the workplace by organizing creativity research into three broad approaches that represent three general types of psychological processes: motivational, cognitive, and affective approaches. MOTIVATIONAL APPROACH Amabile’s (1996) componential theory of creativity highlights the role of motivation in enhancing or reducing individuals’ creativity. (See also Vol. 3, chap. 3, this handbook.) According to this theory, for individuals to exhibit high levels of creativity, three components must be present: (a) The individuals should possess domain-relevant knowledge and skills; (b) they need to have creativity-relevant skills and strategies (these are a more narrow set of skills than is discussed later in this chapter under creative cognition); and (c) they need to be intrinsically motivated to work on the task. The third component, intrinsic motivation, is defined as the type of motivation resulting from individuals’ interest and involvement in, curiosity about, satisfaction with, or positive challenge from the task itself (Amabile, 1996). Intrinsic motivation is said to be essential for creativity because without it, no matter how much knowledge or skills one possesses in a given field and no matter how skillful one is in thinking outside of the box and coming up with creative ideas, if one is not intrinsically motivated by the task, one simply will not engage and persist in creative activities. Thus, the componential theory sets the stage for investigating employee creativity taking a motivational approach, emphasizing the value of intrinsic motivation. Among the three approaches used to investigate workplace creativity, the motivational approach has attracted the most research attention in organizational behavior and has resulted in an impressive body of knowledge (Shalley et al., 2004; Zhou & Shalley, 2003). Most studies that have taken a 281

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

motivational approach to an understanding of creativity did not directly examine the relation between motivation and creativity (but see Shalley & PerrySmith, 2001, and Shin & Zhou, 2003, for exceptions). Instead, they relied on the motivational approach to identify contextual factors that were theorized to either boost or restrict intrinsic motivation, which, in turn, facilitates or inhibits creativity. Cognitive evaluation theory is especially useful here. According to cognitive evaluation theory (Deci & Ryan, 1980, 1985), whether a contextual factor boosts or diminishes individuals’ intrinsic motivation depends on whether this factor is informational or controlling. On one hand, when a contextual factor is informational, individuals are likely to feel competent and self-determining, and consequently their intrinsic motivation is likely to be high. On the other hand, when a contextual factor is controlling, individuals are likely to perceive that they are being pressured or constrained by external forces rather than being self-determining. Consequently, their intrinsic motivation is likely to be low. Thus, essentially, a motivational approach to creativity posits that contextual factors that are informational will enhance creativity, and contextual factors that are controlling will restrict creativity. In addition to examining effects of contextual factors on creativity, some previous studies that took a motivational approach have also taken an interactional approach to an understanding of creativity by investigating effects of interactions between contextual factors and personal factors on creativity. One of the most comprehensive conceptual works that has articulated the interactional approach to creativity is by Woodman et al. (1993). Woodman et al. (1993) argued that both contextual factors and individual differences factors affect employees’ creativity. They emphasized that instead of treating contextual factors and individual differences factors separately, to fully understand and predict creativity in the workplace researchers need to focus on the interactions between the two factors. The individual differences factors they reviewed include cognitive abilities or style identified by Carrol (1985, as cited in Woodman et al., 1993), such as associative fluency and figural fluency; personality traits identified by Barron and Harrington (1981), 282

such as broad interests, attraction to complexity, and high energy; intrinsic motivation as emphasized by Amabile (1996); knowledge such as domain-relevant knowledge theorized by Amabile (1996); and positive or negative effects of previous experiences discussed by B. S. Stein (1989). For consistency with the literature and clarity of presentation, we follow previous work (e.g., Shalley et al., 2004) and use the term personal factors to represent characteristics and attributes of individual employees—including their dispositions (e.g., personality traits), abilities (e.g., cognitive abilities), and knowledge—and the term contextual factors to represent situational factors (e.g., leadership and management practices, relationship with supervisors, relationship with coworkers) that are present in individual employees’ work environment.

Contextual Factors and Creativity Guided by the previously reviewed motivational approach to creativity, and often taking an interactional approach, researchers have investigated relations between various contextual factors and creativity, as well as effects of interactions between these contextual factors and personal factors on creativity. In the paragraphs to follow, we review some representative studies. The three scales mentioned earlier in this chapter, the Oldham and Cummings (1996) scale, the Tierney et al. (1999) scale, and the Zhou and George (2001) scale, are the most widely used scales for measuring creativity in research concerning contextual factors and creativity. Of importance, the studies regarding contextual factors and creativity mentioned in this section are concerned with creativity, not routine performance. This fact is evident in the theoretical arguments made in these studies—independent variables were chosen, and the relations among these variables were hypothesized on the basis of creativity theories. Indeed, empirical evidence has suggested that these independent variables differentially affected creativity and routine performance. For example, Oldham and Cummings (1996) demonstrated that the main and interactive effects of their independent variables (creativity-relevant personal characteristics, job complexity, noncontrolling supervision, supportive supervision) were different for their creativity measures and for their measure of routine performance.

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

Leadership or supervisory behaviors. One of the most salient contextual factors influencing employees’ creativity is leadership or behaviors exhibited by the employees’ supervisors (Amabile & Conti, 1999; Amabile, Conti, Coon, Lazenby, & Herron, 1996; Amabile & Gryskiewicz, 1989; Amabile, Schatzel, Moneta, & Kramer, 2004; Andrews & Farris, 1967; Frese, Teng, & Wijnen, 1999; George & Zhou, 2001; Oldham & Cummings, 1996; Shalley & Gilson, 2004; Shin & Zhou, 2003; Stahl & Koser, 1978; Tierney & Farmer, 2002, 2004; see also chap. 7, this volume). The motivational approach to creativity suggests that when leadership or supervisory behaviors are informational, the employees’ intrinsic motivation is likely to be maintained or enhanced, and consequently, they are likely to exhibit high levels of creativity at work. In contrast, when supervisors’ behaviors are controlling, their employees’ intrinsic motivation tends to be diminished, and subsequently the employees are unlikely to exhibit high levels of creativity. For example, using hierarchical regression analysis, Shin and Zhou (2003) found positive relations between transformational leadership and creativity (ΔR2 = .05, p < .01). In addition, employees’ intrinsic motivation partially explained these positive relations. However, several studies showed that managers’ controlling behaviors were negatively related to their employees’ creativity, presumably via reducing the employees’ intrinsic motivation (Stahl & Koser, 1978; George & Zhou, 2001 [β = −.28, p < .01]; Zhou, 2003 [Study 1, ΔR2 = .20, p < .01; Study 2, ΔR2 = .04, p < .01]). Coworker behaviors or influences. The motivational approach suggests that when employees are surrounded by coworkers whose behaviors are informational, the focal employees’ intrinsic motivation will be boosted, along with their creativity. In contrast, when their coworkers exhibit controlling behaviors, the employees’ intrinsic motivation and creativity will be lower. There have not been a large number of studies that have directly tested these ideas, and results from prior research have been mixed: Whereas some studies have yielded results consistent with these theoretical predictions (e.g., Amabile & Gryskiewicz, 1989; Cummings & Oldham, 1997; Madjar, Oldham, & Pratt, 2002;

Zhou, 2003; Zhou & George, 2001), other studies have found nonsignificant results (e.g., George & Zhou, 2001; Van Dyne, Jehn, & Cummings, 2002), and still other studies have found results that seemed to be contradictory to arguments based on the motivational approach (e.g., Shalley & Oldham, 1997). It is possible that in studies in which no results or contradicting results were found, individuals did not have highly interdependent working relationships with their coworkers, thereby rendering coworker influences relatively nonsalient or unimportant. It is also possible that these studies did not measure individual differences and examine the interaction effects of individual differences and coworker influences. Future research is needed to investigate these possibilities. Goals. The motivational approach would suggest that production goals would be controlling. Specifically, production goals could serve as an external constraint that pressures individuals to meet these goals, resulting in reduced intrinsic motivation and creativity. However, Shalley (1991) argued that depending on the type of goal assigned (i.e., production goal vs. creativity goal), a goal could have a positive effect on an individual’s intrinsic motivation and creativity. Specifically, she proposed that a creativity goal can direct one’s attention and help provide a standard so that an individual can judge his or her own performance. A creativity goal is a stated standard that an individual’s output should be creative or that an individual should attempt to engage in creative activities (e.g., playing with ideas, being flexible in their thoughts, widely scanning their environment). Therefore, the assignment of a creativity goal can help individuals to understand what is expected of them, so rather than constraining their intrinsic motivation and creativity, do-your-best and difficult creativity goals would inform participants of exactly what they should be trying to do and have a positive effect on their intrinsic motivation and subsequent creativity. A few studies have examined this issue, and they have found that it is the type of goals that are assigned that determine whether they are perceived as informational or controlling (Carson & Carson, 1993; Shalley, 1995). For example, Shalley (1991) 283

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

found that when either a do-your-best or a difficult productivity goal was assigned, this resulted in lower creativity. Conversely, in this same study, when a do-your-best or difficult creativity goal was also assigned, this resulted in higher creativity. Shalley (1991) reasoned that the creativity goal helps to motivate and focus individuals on the task at hand while disregarding any felt pressure when they also had a productivity goal. Finally, Gilson and Shalley (2004) examined customer service technician teams’ engagement in creativity processes at work. They found that the more creative teams were higher on shared goals for task accomplishment. These researchers argued that high levels of agreement among team members about what is important for their job can increase motivation, efficiency, and effectiveness. Feedback and evaluation. Receiving and providing feedback and evaluation are a fact of life in organizations. A series of studies have examined effects of the expectation of evaluation and the type of feedback actually given (Amabile, Goldfarb, & Brackfield, 1990; Shalley, 1995; Shalley & Perry-Smith, 2001; Zhou, 1998, 2008; Zhou & Oldham, 2001). Following the motivational approach to creativity, informational evaluation or feedback should boost intrinsic motivation and creative performance, whereas a more judgmental or controlling evaluation should be detrimental for both individuals’ intrinsic motivation and their creativity. Research in this area in general has not measured intrinsic motivation, but the results have supported this view, with studies finding that controlling evaluations or feedback restricts creativity (Amabile, 1979; Amabile et al., 1990; Bartis, Szymanski, & Harkins, 1988; Cheek & Stahl, 1986; Szymanski & Harkins, 1992; Zhou, 1998), whereas informational evaluations or feedback seems to facilitate creativity (Shalley, 1995; Zhou, 1998; Zhou & Oldham, 2001). One study (Shalley & Perry-Smith, 2001) directly manipulated the informational and controlling nature of the evaluation expected to be received and actually included a measure of intrinsic motivation. Specifically, Shalley and Perry-Smith (2001) found that the creativity of individuals who anticipated a judgmental evaluation was significantly lower than 284

that of those expecting a developmental evaluation in which experts would evaluate their work and provide suggestions for alternative approaches to consider in the future. However, they found no significant mediation for intrinsic motivation in the relation between expected evaluation and creativity. Shalley and Perry-Smith (2001) suggested that one possible reason for this is that high intrinsic motivation may be important for creativity but that to have a significant effect on creativity, it might need to exist along with other intervening variables. They proposed that future research should further examine this possibility as well as be open to considering other potential mediators. Job complexity. Another contextual factor influencing employees’ creativity is the nature of the employees’ job. Motivationally, job complexity should facilitate creativity via maintaining or enhancing the job holders’ intrinsic motivation. Job complexity refers to the extent to which jobs are significant and identifiable, provide the employees with autonomy and feedback, and provide the employees with opportunities to use a variety of skills (Hackman & Oldham, 1980). Especially for employees who welcome challenges and strive to grow on the job (cf. Hackman & Oldham, 1980), high levels of job complexity are likely to lead to high levels of intrinsic motivation, which, in turn, leads to high levels of creativity. Although in most studies documented in the literature, intrinsic motivation was not measured and hence the mediating role of intrinsic motivation was not directly tested, the pattern of relationships between job complexity and creativity was consistent with the theoretical prediction that job complexity was positively related to creativity (Amabile & Gryskiewicz, 1989; Farmer, Tierney, & Kung-McIntyre, 2003; Hatcher, Ross, & Collins, 1989; Oldham & Cummings, 1996; Shalley et al., 2009; Tierney & Farmer, 2002, 2004). The pattern of results held regardless of how job complexity (e.g., self-reported vs. objective job complexity measures) or creativity (e.g., supervisory ratings of creativity, self-ratings of creativity, or the number of creative ideas that employees submitted to their organization’s suggestion system) was measured or operationalized.

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

Rewards. The effects of rewards on creativity have been rather mixed (Shalley et al., 2004). A direct application of the motivational approach to predicting the effects of rewards on creativity would suggest that when rewards are informational, they maintain or boost individuals’ intrinsic motivation and subsequently their creativity. In contrast, when rewards are controlling, they diminish individuals’ intrinsic motivation and subsequent creativity. However, whereas some prior studies have found detrimental effects of rewards on creativity (e.g., Amabile, Hennessey, & Grossman, 1986; Kruglanski, Friedman, & Zeevi, 1971), others have found facilitative effects (Eisenberger & Armeli, 1997; Eisenberger & Rhoades, 2001). Thus, it appears that the effect of rewards on creativity is more complex than previously thought. Whether rewards are interpreted by the focal employee as informational or controlling may be contingent on a host of other factors. Indeed, more recent studies have shown that rewards interact with other personal and contextual factors to affect creativity. For example, Baer, Oldham, and Cummings (2003) found that rewards interacted with cognitive style and job complexity to be related to creativity such that rewards were positively related to creativity when employees had an adaptive cognitive style and worked on simple jobs and were negatively related to creativity when employees had an adaptive style and worked on complex jobs or when employees had an innovative style (see Kirton, 1976, 1994, for a complete set of definitions and measurement of adaptive vs. innovative styles) and worked on simple jobs (overall model R2 = .21, p < .01). George and Zhou (2002) found that rewards interacted with bad mood and clarity of feelings to be related to creativity in such a way that bad mood was positively related to employees’ creativity when rewards and clarity of feelings were both high (ΔR2 = .13, p < .01). It would be worthwhile for future research to continue investigating under what conditions individuals will interpret rewards as informational versus controlling and how these interpretations affected their intrinsic motivation and subsequently their creativity. In summary, because creativity is still a relatively young field of inquiry, continuing research using the

motivational approach at the individual level of analysis is still likely to yield valuable insights into what factors facilitate or inhibit creativity and why these effects occur. Furthermore, research focused on the team level or cross-levels of analysis is likely to produce even greater insights, particularly because creativity research at the team level and cross-levels of analysis will inform researchers and practitioners of the unique antecedents and processes of creativity at those levels because individuals are embedded in the team and social contexts and because research at those levels of analysis has been rare. For a more detailed conceptual treatment of various multilevel models of creativity and ideas for future research using these models than what limited space allows us to do here, see Zhou and Shalley (2008b). In addition to continuing multilevel (e.g., individual-, team-, and cross-level) research, future research focusing on the effects of motivation on creativity is also likely to benefit from conceptual advancement on whether intrinsic motivation includes different elements. For example, it may be productive to conceptually separate intrinsic motivation into multiple elements (e.g., enjoyment, mastery, curiosity, interest) and investigate how each of these elements influences the creative process (Gedo, 1997; Mumford, 2003). Finally, although the motivational approach has already guided a relatively large body of empirical research on creativity, theoretical advancement in this area is still vibrant. For example, Unsworth (2001) developed a conceptual framework that attempts to unpack creativity into different types on the basis of the drivers for creative engagement (external vs. internal) and the type of problems (open vs. closed). Also, research has indicated that employees can accurately identify when creative activity is required by their job (Shalley et al., 2000) and that job-required creativity is an important proximal determinant of employee creativity (Shalley, 2008; Unsworth, Wall, & Carter, 2005). Relatedly, Gilson and Shalley (2004) hypothesized and found that the more teams believed their jobs required creativity, the more frequently the team members reported engaging in creative processes. Although many creativity studies have focused on jobs that require creativity, such as research and development 285

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

(e.g., Perry-Smith, 2006; Scott & Bruce, 1994), other studies have looked at jobs that did not necessarily require or expect creativity (e.g., Madjar et al., 2002). For example, Shalley et al. (2009) studied the drivers of creative performance for a broad range of jobs that varied in their level of complexity and need for creativity. Theorizing and research along these lines are very important and promising because they will help researchers answer questions for which we currently do not have answers or do not have satisfactory answers. These questions include: What motivates employees to find creative ideas during routine task performance that does not require creativity? What role does serendipity play in creativity? When does creativity happen when it is required, and when does it happen spontaneously? COGNITIVE APPROACH Cognition has played a central role in the study of creativity. To produce creative outcomes, it is commonly accepted that individuals need to engage in certain creativity-relevant skills and processes that have been termed creative cognition (Finke, Ward, & Smith, 1992; Smith, Ward, & Finke, 1995). Creative cognition as a construct covers a broad range of cognitive processes that can facilitate creativity. Although both Amabile’s (1996) and Woodman et al.’s (1993) models refer to cognitive skills or creativity-relevant skills as being important for creativity, the ones they refer to are included within the construct of creative cognition, with many more also elaborated on. For example, of the nine cognitive skills and abilities that Amabile mentions, most of them would be subsumed under divergent thinking, which is an important part of the cognitive activities needed for creativity but does not represent all the necessary activities. Convergent thinking, for example, is also very important for creativity. Therefore, we use the concept of creative cognition to discuss the cognitive view because it is most comprehensive in its coverage. Creative cognition deals with the fundamental cognitive operations that can help produce creative thoughts. It begins with the premise that all individuals have the capacity to be creative. According to this approach, much of the observed differences in creativity can be explained by differences in indi286

viduals’ use and intensity of application of certain cognitive processes or combinations of processes, the capacity of memory systems, and the flexibility of stored cognitive structures (Ward, Smith, & Finke, 1999). In attempting to produce creative work, individuals need to search within and across categories of knowledge for diverse information that can be used to creatively explore problems, link ideas from multiple sources, and seek out new ways of working on a task. Sometimes individuals need to recognize the relevance of old or known information to new problems and combine concepts to generate more complex ones. As such, creative cognition involves generating a number of ideas about problems or work processes, combining ideas, evaluating them as to their merit, and selecting those that need further consideration and elaboration. Conceptual expansion is one strategy that helps generate novel ideas by using different categories of knowledge to generate ideas and making remote associations between seemingly unrelated ideas (Leung, Maddux, Galinsky, & Chiu, 2008; Shalley & Perry-Smith, 2008; Ward et al., 1999). A number of conceptual pieces have discussed the cognitive processes needed for creativity to occur (e.g., Amabile, 1996; Csikszentmihalyi, 1988, 1996; Drazin et al., 1999; Ford, 1996; Woodman et al., 1993). For example, Campbell’s (1960) evolutionary model of creativity proposed that creativity requires extensive trial and error and hard work. He argued that individuals have to generate multiple solutions to difficult problems and to do this they need to use a wide variety of approaches. As such, his model stressed the importance of variation in terms of ideas and the selective retention of promising ideas while culling out less desirable ones. The process of selective retention requires individuals to use what they know, in terms of their background, knowledge, and skills, to determine what ideas merit further consideration and which should be discarded. Building on Campbell’s (1960) model, Simonton’s (1999) evolutionary theory of creative thinking proposed a process of variation and selective retention. Simonton proposed that variation primarily contributes to idea novelty, whereas the process of

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

selection primarily contributes to idea usefulness. His theory was focused more on variation than on selective retention because novelty is what distinguishes creative ideas from more routine ones. According to Simonton, although ideas initially occur within one individual’s mind, as the individual tests ideas against relevant criteria for novelty and appropriateness or usefulness, they are shared with others. Additional variation and selection of ideas then occur as others have input as well. Empirical research on creative cognition has focused on examining the specific cognitive operations or processes that people need to engage in on a variety of tasks that require some level of creativity. This work has been experimental in nature, with research examining the effect of certain subsets of creative cognitive processes (e.g., Basudur, Graen, & Green, 1982; Reiter-Palmon & Illies, 2004; Runco & Chand, 1995). For example, Mumford and his colleagues (e.g., Dailey & Mumford, 2006; ReiterPalmon, Mumford, Boes, & Runco, 1997; Vincent, Decker, & Mumford, 2002) have conducted empirical studies to examine the cognitive skills needed to be able to find or construct problems, generate alternatives, and evaluate ideas. As an example, research on problem identification and construction has indicated that this is a critical process for creative problem solving. Specifically, some studies have found that problem construction is related to solution originality and quality and that the ability to construct problem solving effectively explains creativity over and above the contribution of intelligence (Okuda, Runco, & Berger, 1991; Reiter-Palmon, Mumford, & Threlfall, 1998; Smilansky, 1984). Recently, Yuan and Zhou (2008) found that when individuals engaged in variation, those who expected external evaluation generated fewer numbers of ideas. However, during selective retention, individuals who expected external evaluation performed better in improving idea appropriateness, and those who expected evaluation only during selective retention produced the most creative ideas. The effect sizes in their study ranged from .22 to .40.

1996; Koestler, 1964; Parnes, Noller, & Biondi, 1977; M. I. Stein, 1967). For example, Wallas’s (1926) classic model described four stages of creative thinking: preparation (e.g., examination of the problem and goals for addressing it), incubation (e.g., suspending conscious work on the problem but unconsciously working on it), illumination (e.g., the solution presents itself—the aha effect), and verification (e.g., use of logic and knowledge to evaluate the idea and refine it so that it is an appropriate solution). Hogarth (1980) proposed four stages: preparation, production, evaluation, and implementation. Although there are slight variations in each of the different models of the stages of the creative thought process, they all include some identification of a problem or opportunity, gathering information, generating ideas, and evaluating ideas. As such, these models characterize creativity as involving an iterative process that can include reflection and action, experimenting, seeking feedback, and searching for new ways to do things. When individuals engage in the cognitive processes needed for creativity (i.e., creative cognition), this can help move an idea through the stages of the creative thought process, such as from generation to implementation (Ford, 1996; Shalley & Perry-Smith, 2008). Regarding the specific stage of incubation, there has been some work that has tried to show its benefits and explore the theoretical reasons for how incubation works (e.g., Jett & George, 2003; Olton & Johnson, 1976; Segal, 2004). For example, Madjar and Shalley (2008) proposed that individuals’ focus of attention and experience of cognitive exhaustion may explain how multiple tasks, multiple goals, and discretion may effect incubation and subsequent creativity. Also, Elsbach and Hargadon (2006) suggested that breaks provide a time for nontaxing or mindless work in which the individual can think of non–task-related thoughts. They proposed that professionals’ creativity could be enhanced by designing workdays to include hours of cognitively challenging, high-pressure work interspersed with some hours of mindless work.

Stages of the Creative Thought Process

Personal Characteristics That Can Affect Creative Cognition

A great deal has been written about the different stages of the creative thought process (e.g., Amabile,

Some researchers have suggested that differences in how people approach problem solving can have an 287

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

effect on whether they engage in the cognitive processes necessary for creative performance (e.g., Jabri, 1991; Kirton, 1994). Some empirical studies (e.g., Tierney et al., 1999) have looked at the relationship between individuals’ preferred cognitive style and creative outcomes and found that individuals with a more innovative style tend to be more creative than those with a more adaptive style. Furthermore, some studies have looked at the effect of the interaction between cognitive style and contextual conditions on employees’ creative behavior (e.g., Baer et al., 2003; Miron, Erez, & Naveh, 2004; Scott & Bruce, 1994). For example, Baer et al. (2003) found a positive relation between extrinsic rewards and creativity for employees with an adaptive cognitive style who worked on relatively simple jobs. In this study, they also found a weak relationship between rewards and creativity for employees with an innovative cognitive style who worked on complex jobs and a negative relation for those in the adaptive style–complex job and innovative style–simple job conditions. There is also research that has examined the importance of what individuals focus on when they are attempting to solve problems or generate creative ideas. For example, Friedman and Forster (2001) found that when individuals focused on potential gains versus losses, this increased the accessibility of unconventional ideas and led them to generate more creative ideas. Also, Kray, Galinsky, and Wong (2006) found that problem solving using counterfactual thinking (i.e., structuring thought around salient relationships and associations) helps on creative association tasks but harms performance on creative idea-generation tasks. Their findings are consistent with Peterson and Nemeth’s (1996) premise that cognitive styles can have a differential influence on problem-solving effectiveness depending on what is required by the task. Also, a number of studies have examined personality factors that are related to an individual’s overall potential to engage in cognitive processes that lead to creativity (e.g., Barron & Harrington, 1981; Gough, 1979). For example, individuals who have more creative personalities tend to approach problems with broad interests and have the capacity to recognize divergent opinions. They are also self288

confident, persistent in developing new ideas, and tolerant of the ambiguity that arises from competing viewpoints. Some studies have empirically supported the association between creative personality and creativity (e.g., Oldham & Cummings, 1996; Zhou & Oldham, 2001). Some research has examined how individuals’ view of themselves or others’ view of them can translate to their being more creative. For example, Tierney and Farmer (2002, 2004) developed the construct of creative self-efficacy, which is the extent to which employees believe they have the ability to produce creative work, and found that creative self-efficacy was positively associated with creativity. Also, Farmer et al. (2003) examined the relationship between creativity and creative role identity, which is whether an individual views himor herself as a creative person. They found that the highest level of creativity occurred when employees had a strong creative role identity and perceived that their organization valued creative work. Carmeli and Schaubroeck (2007) found that the perceived expectations of leaders, customers, and family were positively associated with employees’ self-expectations for creativity, and these self-expectations were associated with creative involvement at work. Relatedly, Thatcher and Greer (2008) found that when others (i.e., team members) know the relative importance of an individual’s identity, this positively relates to the individual’s creativity. They argued that when individuals feel known and understood, they are more likely to freely share ideas, leading them to be more creative. Finally, Janssen and Huang (2007) found that individuals’ differentiation from others in their team and organization may be a critical driver for generating new ideas for organizational change.

Creativity as Planned Action Versus Unconscious Thought Process Research on whether individuals’ intention to be creative actually leads to creative outcomes has also emerged. In particular, Choi (2004) directly applied the theory of planned action (e.g., Ajzen, 1991) to creativity research in arguing that individuals’ subjective intention to be creative determines their actual creative performance. He found that individu-

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

als’ creative intention was positively related to their actual creativity. Whereas research linking individuals’ intention to be creative assumes that creativity is the result of explicit and conscious decisions (e.g., Choi, 2004), other studies have focused on investigating the relations between unconscious thought processes and creativity (e.g., Dijksterhuis & Meurs, 2006; Zhong, Dijksterhuis, & Galinsky, 2008). For example, Dijksterhuis and Meurs (2006) maintained that whereas conscious thought tends to be convergent, unconscious thought tends to be divergent. This is because conscious thought may restrict individuals to a limited set of attributes that they consider to be relevant and at the same time takes them away from attributes that are seemingly less relevant and less obvious but are more likely to lead to novel responses. Alternatively, unconscious thought leads to less focused and more associative processes that allow individuals to access materials that are less obvious and less accessible. In other words, unconscious thought allows for greater divergent thinking. To the extent that divergent thinking leads to creativity (Guilford, 1956, 1967), it is likely that unconscious thought processes result in greater creativity than do conscious thought processes. In a series of experiments (Dijksterhuis & Meurs, 2006), research participants worked on experimental tasks in which they were asked to generate items according to three experimental conditions (i.e., generating items immediately after receiving instructions, generating items after a few minutes of conscious thought, and generating items after a few minutes of distraction during which time unconscious thought was theorized to take place). In the unconscious thought condition, participants were found to produce more associative and original responses than the participants in the other two conditions.

Social Side of Creativity Recently, some work has started to explore different ways in which interactions with others can affect individuals’ creative cognition. We briefly discuss this research. Social networks. Perry-Smith and Shalley (2003) focused on the social side of creativity by highlighting the importance of others to individuals generat-

ing ideas. They proposed that when individuals connect with others, build networks, and interact with those not in their own field, this can allow them to make novel combinations of ideas and ultimately b e more creative. In a recent empirical piece, PerrySmith (2006) examined the role of several social network parameters on research scientists’ creative performance. She found that weak ties seemed to facilitate the generation of alternatives and encourage autonomous thinking. Results of this study highlighted the importance of understanding tie strength and the interactive effects of position within the broader social environment and outside ties for creativity. Finally, Zhou, Shin, Brass, Choi, and Zhang (in press) examined the influence of social networks and a personal value (i.e., conformity value) on employees’ creativity. They found a curvilinear relationship between the number of weak ties and creativity. Employees’ conformity value was found to moderate this curvilinear relationship, such that when conformity was low, employees exhibited greater creativity at intermediate levels of number of weak ties than when conformity was high. Presence of creative role models. The presence of creative role models who can facilitate the observer’s acquisition of creativity-relevant skills and strategies has been examined in two studies. For example, using social cognitive theory (Bandura, 1986) to develop a social learning perspective, Shalley and Perry-Smith (2001) found that observing creative role models allows individuals to acquire the strategies and approaches that can enable them to be more creative in their work. Zhou (2003) found that the presence of creative coworkers positively affected individuals’ creativity when supervisors’ close monitoring was low (ΔR2 = .16, p < .01, in Study 1, and ΔR2 = .04, p < .01, in Study 2) or developmental feedback was high (ΔR2 = .04, p < .01). Multicultural experiences. Leung et al. (2008) proposed that exposure to multiple cultures in and of itself can enhance creativity. They argued that multicultural experiences can facilitate creativity in a number of ways. First, multicultural experiences can broaden the range of accessible ideas and concepts 289

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

used in problem solving. Second, by having multicultural experiences, individuals may be better equipped to realize that the same form can have different functions and implications. Third, multicultural experiences can weaken and lead to questioning of established associations while increasing the individual’s willingness to seek out ideas from diverse outlets. Finally, multicultural experiences can build cognitive complexity. Thus, multicultural experiences can introduce individuals to a range of behavioral and cognitive scripts that enable the conceptual expansion process. In a series of studies, Leung et al. found that extensiveness of multicultural experiences is positively related to the use of creativity supporting cognitive processes (i.e., creative cognition) and creative performance. Furthermore, they found that the benefits for creativity of having multicultural experiences may depend on the extent to which an individual is open to the foreign culture. Identity integration. Cheng, Sanchez-Burks, and Lee (2008) examined the effects of identity integration on creativity. When individuals perceived that two social identities they have are compatible, they had higher levels of identity integration and demonstrated greater creativity. More specifically, the researchers showed that (a) Asian Americans who had higher identity integration demonstrated greater creativity on a task in which they were asked to develop new dishes when both Asian and American ingredients were available to them and (b) female engineers who had higher identity integration were more creative on a product design task when female users were the target user group of the product. Presumably, individuals who have high levels of identity integration have access to multiple knowledge domains, which provides a relatively large variety of raw materials for creative idea production.

Cognition at the Team Level Although cognition is generally an individual phenomenon, it is entirely possible that teams can engage in the cognitive processes needed to be creative. Despite the recognized importance of understanding team cognition as it relates to team creativity (Kurtzberg & Amabile, 2001), there has been very little work in this area. One notable exception is 290

Shalley and Perry-Smith’s (2008) development of the construct of team creative cognition. In their multilevel conceptual article, they introduced the construct of team creative cognition, a shared framework for how to approach problems creatively, and emphasized how it develops and evolves from one team member to others. In particular, they proposed that the emergence of team creative cognition comes from particular ties outside of the team and the infusion within the team is determined by the member’s centrality in the team’s sociocognitive network. Some researchers have examined teams’ overall engagement in creative processes at work (e.g., Gilson, Mathieu, Shalley, & Ruddy, 2005; Gilson & Shalley, 2004; Kazanjian, Drazin, & Glynn, 2000). Also, studies have examined how group processes and team climate can help facilitate engagement in cognitive processes needed for creativity (e.g., Amabile et al., 1996; Pirola-Merlo & Mann, 2004). For example, Leenders, van Engelen, and Kratzer (2003) studied new product development teams and found that a moderate frequency of communication was best for creativity because members could share their ideas and have a constructive dialogue while not being overloaded by the amount of information exchanged and still having the cognitive capacity to focus on the value of the information. As another example, Taggar (2002) examined the interaction between team members’ individual disposition to be creative (e.g., cognitive ability and personality dimensions) and team creativity-relevant processes (e.g., effective communication, addressing conflict) on the creativity of the teams’ products. The highest creativity was found in teams that had creative members and high levels of creativity-relevant processes. There is also a large body of work on group brainstorming (e.g., Goncalo & Staw, 2006; Paulus, 2000; Paulus & Yang, 2000). Group brainstorming research focuses on ideational creativity, which involves the generation of novel ideas related to a particular problem (e.g., find new uses for an existing object, how to market a new product). Traditionally, this research has been conducted in the laboratory with groups given the rules for brainstorming (e.g., generate a large number of ideas, do not criticize ideas; Osborn, 1953). Creativity is expected to arise from the statis-

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

tical advantage of having a large sample of ideas and the combining of ideas to form new ones. One potential benefit of group brainstorming is that individuals are exposed to diverse and potentially new categories of knowledge that can be applied to the problem. Also, within categories of knowledge, ideas generated can stimulate others to think of related ideas that are new (Paulus, 2008). Although this group idea-exchange process is primarily a cognitive activity, group social interactions can inhibit its effectiveness. For example, motivation losses such as loafing or free riding can occur, as well as the coordination losses that can lead to production blocking. A stream of research that has fascinated researchers for quite a number of years now is concerned with the relation between team diversity and creativity. Presumably, when team members have diverse backgrounds such as differences in age, gender, race, and educational specialization, the teams would possess a diverse set of information and benefit from the members’ differing perspectives and viewpoints, which, in turn, would lead to higher levels of team creativity. However, although not specifically focusing on team creativity as the outcome variable, a number of reviews of empirical studies have suggested that the relations between team diversity and team outcomes are more complex than previously thought (Jackson, Joshi, & Erhardt, 2003; Milliken & Martins, 1996; Williams & O’Reilly, 1998). Indeed, two recent studies that focused on team diversity and creativity did not find any main effects of team diversity on creativity (e.g., Pearsall et al., 2008; Shin & Zhou, 2007). Rather, whether team diversity facilitates or inhibits team creativity is likely to depend on context. For example, Pearsall et al. (2008) examined effects of gender diversity on team creativity. They found that only when gender faultlines (i.e., attributes existing in teams that potentially divide them into subgroups; Lau & Murnighan, 2005) were activated did gender diversity negatively affect team creativity (ΔR2 = .04, p < .05). Pearsall et al. further demonstrated that the moderated, negative relation between gender diversity and team creativity was partially mediated by emotional conflict.

As another example, Shin and Zhou (2007) examined the relation between employees’ educational specialization heterogeneity and team creativity. Shin and Zhou proposed that transformational leadership is especially useful in creating a unified team identity, energizing teams, and helping teams to fully take advantage of the diverse information and knowledge they possess to work together and come up with new and useful ideas. Using data from 75 research and development teams from 44 companies, they found that when transformational leadership was high, the teams’ educational specialization heterogeneity was greater and the teams’ creativity was higher (ΔR2 = .06, p < .05).

Cross-Level Hargadon and Bechky (2006) proposed a model of collective creativity in which the focus of creative problem solving can shift from individuals to interactions of a collective. They discussed how four different types of social interactions (i.e., help seeking, help giving, reflective reframing, and reinforcing) can lead to collective creativity by influencing comprehension of problems and generation of solutions that leverage the past experiences of this collection of individuals. As such, this is a cross-level model that describes how individuals’ creative cognition is influenced by the collective and at the same time may contribute to other individuals’ creative cognition in this collective. In sum, cognition has been central to the study of creativity, and much work has used this approach. However, further research and theorizing on the important role of cognition in facilitating creative performance is warranted. In particular, more work is needed in the new and emerging area of how social identity and interactions influence individual and team creative cognition. Furthermore, as the Hargadon and Bechky (2006) model attests, more work at the team level and cross-levels of analysis would help to further extend our knowledge of creativity using a cognitive approach. AFFECTIVE APPROACH The third theoretical approach that has guided creativity research is the affective approach. There are three related concepts used by researchers in this 291

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

area: affect, emotion, and mood (George & Brief, 1992). Moods refer to generalized affective states that are temporary and fluctuating and do not have a clear target. Compared with mood states, emotion is stronger in magnitude and clearer in terms of a target or an exogenous factor causing it. Affect is a general term that typically includes both mood states and emotion. Earlier research tended to categorize affect, mood, and emotion into two dimensions: positive and negative. Consistent with this research tradition in the affect literature, initial sociopsychological research using the affective approach to understanding creativity conceptualized affect as having positive and negative dimensions and focused on the role of positive affect.

Positive Affect and Creativity One of the leading researchers in this area is Isen, and much of her research concerning affect, divergent thinking, and creativity has used laboratory experiments in which positive affect was usually induced by showing the research participants a short film clip or by giving them small bags of candy. After showing the film clip for a few minutes, for example, the research participants were then asked to work on word association or insight tasks. In those experiments, Isen and her associates found that research participants in the positive affect condition produced more unique responses on unusual word association tasks (Isen, Johnson, Mertz, & Robinson, 1985) or were more likely to solve the functional fixedness problem presented in Duncker’s (1945) candle task (Isen, Daubman, & Nowicki, 1987). Although the candle task has a single demonstrably correct solution and it is thus straightforward to measure performance on this task, word associations were rated as unique or unusual if they were produced by a very small percentage of research participants (e.g., equal to or fewer than 2.5% of the participants). Although in Isen and colleagues’ (Isen, Daubman, & Nowicki, 1987; Isen, Johnson, Mertz, & Robinson, 1985) experiments, positive affect was usually induced rather than directly measured, Amabile et al. (2005) conducted a diary study in which they directly measured positive affect by using selfreported items, such as being happy or satisfied with 292

the team, which tapped the pleasantness dimension of affect (see Russell, 1980). They found that positive affect was an antecedent of creativity. As Davis (2009) concluded in a meta-analytic review that focused on understanding the relationship between positive mood and creativity, the studies he selected to examine demonstrated that this relationship is context dependent. Notably, he found that compared with negative mood, positive mood fostered ideation (d = 0.24), which by emphasizing fluency, flexibility, and cognitive variation mainly contributes to the novelty aspect, instead of the usefulness aspect, of creativity. Because workplace creativity is defined as ideas that need to be both novel and useful, the extent to which what he concluded can be translated into an understanding of the relation between positive mood and workplace creativity remains to be investigated. Indeed, Davis also found that compared with negative mood, positive mood showed little advantage on a problem-solving task (d = 0.02; the effect size did not significantly differ from zero).

Negative Affect and Creativity However, organizational life is not all positive and pleasant, and it is simply not realistic to expect organizational members to always be in a positive mood state. In fact, as convincingly argued by Anderson et al. (2004), there are numerous phenomena with negative connotations in organizational life, ranging from individual-level phenomena such as threats to job security and job dissatisfaction to group-level phenomena such as conflict to organizational-level phenomena such as budget deficiencies, shrinking market share, and pressures to restructure organizational processes. These negative events suggest that current ways of doing things no longer meet internal or external challenges. Therefore, creativity may be particularly needed at those negative moments. Although our knowledge concerning the relations between negative events and creativity is still quite limited, findings from a few initial studies are very interesting. In particular, Zhou and George (2001) formulated and tested a voice perspective of creativity that theorized conditions under which job dissatisfaction led to creativity. According to this perspective, under

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

certain circumstances (e.g., continuance commitment and coworker useful feedback), dissatisfied employees may engage in creative activities as an expression of voice. Empirical results supported their theoretical predictions (ΔR2s ranged from .03 [p < .05] to .08 [p < .01]). This study broke new conceptual grounds not only because it was the first published study that theorized and tested a voice perspective concerning job dissatisfaction and creativity but also because it challenged one of the fundamental assumptions in the field of organizational behavior and organizational psychology: that job satisfaction is always desirable and, conversely, that job dissatisfaction is always detrimental for organizations. Job dissatisfaction has both affective and cognitive underpinnings. So the next steps along this line of research would include studies that focus on each of these components separately. To begin, George and Zhou (2002) investigated when negative mood may be functional for creativity in the workplace. According to the mood-as-input model, individuals’ mood states provide them with information (e.g., Schwarz & Clore, 2003), and the significance and consequences of the information depend on the context (Martin & Stoner, 1996). Adapting this model to creativity research, George and Zhou theorized that under certain conditions, negative moods might foster creativity and positive moods might inhibit it. This is because employees’ work environment or context provides them with cues concerning their ongoing creative behaviors. These cues are valuable to them because when they are engaged in creative activities at work, they often have little objective information and have to decide for themselves when they have tried hard enough to come up with a new and improved procedure or put forth enough effort to come up with a new and better way of completing tasks. Consistent with these theoretical arguments, results showed that negative moods were positively related to creativity when perceived recognition and rewards for creativity and clarity of feelings (a meta-mood process) were high (ΔR2 = .13, p < .01). Additionally, results showed that under the same conditions, positive moods were negatively related to creativity (ΔR2 = .09, p < .05). Relatedly, De Dreu, Baas, and Nijstad (2008) demonstrated that negative mood did enhance cre-

ativity, but only when it was activating (e.g., angry). And, consistent with George and Zhou’s (2002) argument that under certain conditions, individuals experiencing negative mood are likely to discover an unsatisfying state of affairs, put forth great effort in trying to come up new and better ways of doing things, and be persistent until they come up with truly new and useful ideas, De Dreu et al. found that negative mood enhanced creativity by increasing persistence. In addition, they showed that positive mood also enhanced creativity, again only when it was activating (e.g., happy). Moreover, different from the mechanism through which negative mood facilitated creativity, positive mood enhanced creativity by boosting cognitive flexibility. In a meta-analysis concerning relations between moods and creativity, Baas, De Dreu, and Nijstad (2008) found that compared with mood-neutral control participants, positive moods were more related to creativity (r = .15). However, there were no significant differences between negative moods and mood-neutral control participants (r = −.03) or between negative moods and positive moods. Overall, they concluded that “the mood–creativity link is better understood as a function of various aspects of specific moods than simply in terms of hedonic tone or level of activation” (Baas et al., 2008, p. 795). Thus, instead of prolonging the debate on whether positive mood or negative mood enhances creativity, and whether positive mood enhances creativity to a greater extent than negative mood or vice versa, it would be more productive for future research to at least take a two-dimensional view of mood: hedonic tone (positive vs. negative moods) and activation (activating vs. deactivating). It would also be more productive to examine specific aspects of moods. Finally, the aforementioned results involving negative mood also suggest that future research needs to investigate effects of moods in organizational contexts and use samples from employees working in various functional areas in organizations. Much of the research conducted in the behavioral laboratory, especially studies focusing on effects of positive affect on creativity, has differed from characteristics of field settings in a few ways. First, the lab research used insight tasks such as Duncker’s 293

Copyright American Psychological Association. Not for further distribution.

Zhou and Shalley

(1945) candle problem or remote association tasks (e.g., Mednick, 1962) in which there is one demonstrably correct answer (Baas et al., 2008). In contrast, in organizations creative ideas and solutions often do not have a single demonstrably correct answer. Second, in the lab the research participants are often required to work on a task assigned by the experimenter. In contrast, in an organizational context, the first step of creativity often involves the employees’ discovering problems on their own and deciding for themselves to find solutions to these problems. Third, in the lab the research participants are often given a fixed amount of time, which is frequently less than 1 hour in duration, to work on the experimental task. In contrast, in organizations the employees, especially those who do not work in the research and development function, have to decide for themselves how long they want to work on solving a problem creatively. Fourth, whereas in the lab creativity measures place heavy emphasis on the novelty dimension, in organizations the usefulness dimension is also essential. In addition to examining effects of specific moods in specific contexts on specific dimensions of creativity, some researchers have also started to investigate effects of variables being defined as a combination of activation and physiological well-being on selfreported involvement in creative activities. For example, Kark and Carmeli (2008) found that individuals’ vitality was positively related to their selfreported involvement in creativity such that those who experienced high levels of vitality also selfreported high levels of involvement in creative activities. On the basis of Ryan and Bernstein (2004), they defined vitality as a mix of affective and physiological “aliveness” in that a high level of vitality is characterized by high levels of felt energy, vigor, and feelings of physiologically being capable and functioning fully.

Joint Effects of Positive Affect and Negative Affect Because positive, activating moods may facilitate cognitive flexibility and negative, activating moods may boost perseverance (De Dreu et al., 2008) and because both cognitive flexibility and perseverance are needed in coming up with creative ideas in the 294

workplace, it is important for research to examine the joint effects of positive and negative affect. In particular, on the basis of the mood-asinformation theoretical framework and its related research in psychology (e.g., Fiedler, 1988; Kaufmann, 2003; Martin & Stoner, 1996; Schwarz, 2002; Schwarz & Clore, 2003), George and Zhou (2007) advanced a “dual-tuning perspective” in arguing that both positive and negative moods may be functional for creativity in the workplace. Negative moods alert employees to problems, cause the employees to focus on the current situation rather than their preexisting assumptions, and motivate them to exert high levels of effort to make improvements (George & Zhou, 2002; Kaufmann, 2003; Martin & Stoner, 1996; Schwarz, 2002). Positive moods allow employees to be playful with ideas and willing to take risks and explore novel ways of doing things and facilitate divergent thinking. Moreover, and consistent with the person– context interaction approach to creativity (Woodman et al., 1993), the joint effects of positive and negative moods only manifest themselves in a supportive context provided by supervisors. Indeed, George and Zhou (2007) found that positive mood, negative mood, and supportive contexts interacted to affect creativity in such a way that when positive mood and supportive contexts were both high, negative mood had the strongest positive relation with creativity (ΔR2s ranged from .03 [p < .05] to .07 [p < .01]). Fong (2006) examined effects of emotional ambivalence, which is defined as simultaneously experiencing positive and negative emotions, on creativity. On the basis of the theoretical arguments developed in the affect literature that emotion provides information, Fong reasoned that because emotional ambivalence is commonly viewed as an unusual event, experiencing it increases individuals’ likelihood of making connections to other unusual associations among stimuli in the current environment. To the extent that such “remote” associations or divergent thinking facilitates creativity, individuals’ creativity will be enhanced after they experience emotional ambivalence. Fong conducted an experimental study in which she compared research participants’ creative performance on the remote

Copyright American Psychological Association. Not for further distribution.

Understanding Creativity in the Workplace

For example, using the motivational approach, research is needed to identify the full range of contextual factors that independently or jointly (with individual differences or other contextual factors) enhance or constrain creativity. Research is also needed to clarify the mixed results documented in the literature on the effects of some of the contextual factors (e.g., coworker influences) on creativity. Moreover, research is needed to directly measure and test whether motivation, especially intrinsic motivation, mediates the relationship between contextual factors and creativity. Furthermore, future research using any of the three approaches is needed to examine creativity at the team level and cross-levels of analysis. Finally, we believe that synthesizing and integrating these approaches presents yet another promising research direction. For example, future research may investigate the temporal sequencing of motivation, cognition, and affect in influencing creativity. As another example, research using the cognitive approach or the affective approach could benefit from results obtained by using the motivational approach, such as investigating whether the contextual factors that have been demonstrated to foster or restrict creativity by using the motivational approach can also influence creative cognition and affective states that are said to be linked to creativity. Last but not least, future research is needed to integrate the cognitive and the affective approaches. One way to do this would be to map specific aspects of mood states with the cognitive stages of the creative process. For example, are specific aspects of positive mood states especially useful at the idea-generation stage of creative cognition?

association task (Mednick, 1962). The participants worked in one of four experimental conditions: emotional ambivalent, positive emotion (i.e., happiness), negative emotion (i.e., sadness), and emotional neutral. Results showed that compared with the other three experimental conditions, the participants in the emotional ambivalent condition exhibited the highest level of creativity on the remote association task. Interestingly enough, neither positive emotion nor negative emotion had any main effect on creativity. In summary, in addition to the possibilities discussed earlier, future research using the affective approach may also identify contextual variables that are especially effective in influencing affect and examine whether these variables facilitate or inhibit creativity by influencing employees’ affect. For example, Madjar et al. (2002) conducted a study in which affect was shown as a mediator linking work and nonwork support received by employees from sources inside or outside their work organizations and those employees’ creativity in the workplace. Future research using the affective approach may also benefit from including the construct of emotional intelligence (George & Zhou, 2002; Zhou & George, 2003). Previous theory and research have suggested that emotional intelligence may influence creativity in two ways. First, George and Zhou (2002) showed that one dimension of emotional intelligence, clarity of feelings, was one of the two conditions under which negative moods were positively related to creativity. Second, in a conceptual article, Zhou and George (2003) theorized that leaders who have high levels of emotional intelligence are more likely to support their employees’ creativity than are leaders whose emotional intelligence is relatively low.

References

CONCLUDING REMARKS

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Performance, 50, 179–211.

As our review shows, research taking a motivational, cognitive, or affective approach has been conducted in parallel fashion. Although much progress has been made in understanding creativity by using these approaches, we contend that research continuing in any of the three traditions is necessary and likely to be both productive and informative.

Amabile, T. M. (1979). Effects of external evaluation on artistic creativity. Journal of Personality and Social Psychology, 37, 221–233. doi:10.1037/0022-3514. 37.2.221 Amabile, T. M. (1996). Creativity in context. Boulder, CO: Westview Press. Amabile, T. M., Barsade, S. G., Mueller, J. S., & Staw, B. M. (2005). Affect and creativity at work. Administrative 295

Zhou and Shalley

Science Quarterly, 50, 367–403. doi:10.2189/ asqu.2005.50.3.367 Amabile, T. M., & Conti, H. (1999). Changes in the work environment for creativity during downsizing. Academy of Management Journal, 42, 630–640. doi:10.2307/256984 Amabile, T. M., Conti, R., Coon, H., Lazenby, J., & Herron, M. (1996). Assessing the work environment for creativity. Academy of Management Journal, 39, 1154–1184. doi:10.2307/256995

Copyright American Psychological Association. Not for further distribution.

Amabile, T. M., Goldfarb, P., & Brackfield, S. C. (1990). Social influences on creativity: Evaluation, coaction, and surveillance. Creativity Research Journal, 3, 6–21. Amabile, T. M., & Gryskiewicz, N. D. (1989). The creative environment scales: Work environment inventory. Creativity Research Journal, 2, 231–252. Amabile, T. M., Hennessey, B. A., & Grossman, B. S. (1986). Social influences on creativity: The effects of contracted-for reward. Journal of Personality and Social Psychology, 50, 14–23. doi:10.1037/ 0022-3514.50.1.14 Amabile, T. M., Schatzel, E. A., Moneta, G. B., & Kramer, S. J. (2004). Leader behaviors and the work environment for creativity: Perceived leader support. Leadership Quarterly, 15, 5–32. doi:10.1016/j.leaqua.2003.12.003 Anderson, N., De Dreu, C. K. W., & Nijstad, B. A. (2004). The routinization of innovation research: A constructively critical review of the state-of-the-science. Journal of Organizational Behavior, 25, 147–173. doi:10.1002/job.236 Andrews, F. M., & Farris, F. (1967). Supervisory practices and innovation in scientific teams. Personnel Psychology, 20, 497–515. doi:10.1111/j.17446570.1967.tb02446.x Baas, M., De Dreu, C. K. W., & Nijstad, B. A. (2008). A meta-analysis of 25 years of mood–creativity research: Hedonic tone, activation, or regulatory focus? Psychological Bulletin, 134, 779–806.doi:10. 1037/a0012815 Baer, M., Oldham, G. R., & Cummings, A. (2003). Rewarding creativity: When does it really matter? Leadership Quarterly, 14, 569–586. doi:10.1016/ S1048-9843(03)00052-3 Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall. Barron, F., & Harrington, D. M. (1981). Creativity, intelligence, and personality. Annual Review of Psychology, 32, 439–476. doi:10.1146/annurev.ps.32.020181. 002255 Bartis, S., Szymanski, K., & Harkins, S. G. (1988). Evaluation and performance: A two-edged knife. 296

Personality and Social Psychology Bulletin, 14, 242–251. doi:10.1177/0146167288142003 Basudur, M. S., Graen, G. B., & Green, S. G. (1982). Training in creative problem solving: Effects on ideation and problem solving in an applied research organization. Organizational Behavior and Human Decision Processes, 30, 41–70. Campbell, D. T. (1960). Blind variation and selective retention in creative thought as to other knowledge processes. Psychological Review, 67, 380–400. doi:10.1037/h0040373 Carmeli, R., & Schaubroeck, J. (2007). The influence of leaders’ and other referents’ normative expectations on individual involvement in creative work. Leadership Quarterly, 18, 35–48. doi:10.1016/ j.leaqua.2006.11.001 Carson, P. P., & Carson, K. D. (1993). Managing creativity enhancement through goal setting and feedback. Journal of Creative Behavior, 27, 36–45. Cheek, J. M., & Stahl, S. S. (1986). Shyness and verbal creativity. Journal of Research in Personality, 20, 51–61. doi:10.1016/0092-6566(86)90108-X Cheng, C. Y., Sanchez-Burks, J., & Lee, F. (2008). Connecting the dots within: Creative performance and identity integration. Psychological Science, 19, 1178–1184. doi:10.1111/j.1467-9280.2008.02220.x Choi, J. N. (2004). Individual and contextual predictors of creative performance: The mediating role of psychological processes. Creativity Research Journal, 16, 187–199. doi:10.1207/s15326934crj1602&3_4 Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO PI-R) and NEO FiveFactor Inventory (NEO-FFI) professional manual. Odessa, FL: Psychological Assessment Resources. Csikszentmihalyi, M. (1988). Society, culture, and person: A systems view of creativity. In R. J. Sternberg (Ed.), The nature of creativity (pp. 325–330). Cambridge, England: Cambridge University Press. Csikszentmihalyi, M. (1996). Creativity. New York: HarperCollins. Cummings, A., & Oldham, G. R. (1997). Enhancing creativity: Managing work contexts for the high potential employee. California Management Review, 40, 22–38. Dailey, L., & Mumford, M. D. (2006). Evaluative aspects of creative thought: Errors in appraising the implications of new ideas. Creativity Research Journal, 18, 385–390. doi:10.1207/s15326934crj1803_11 Davis, M. A. (2009). Understanding the relationship between mood and creativity: A meta-analysis. Organizational Behavior and Human Decision Processes, 108, 25–38. doi:10.1016/j.obhdp. 2008.04.001

Understanding Creativity in the Workplace

Deci, E. L., & Ryan, R. M. (1980). The empirical exploration of intrinsic motivational processes. In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 39–80). New York: Academic Press. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. New York: Plenum Press.

Copyright American Psychological Association. Not for further distribution.

De Dreu, C. K. W., Baas, M., & Nijstad, B. A. (2008). Hedonic tone and activation level in the mood– creativity link: Toward a dual pathway to creativity model. Journal of Personality and Social Psychology, 94, 739–756. doi:10.1037/0022-3514.94.5.739 Dijksterhuis, A., & Meurs, T. (2006). Where creativity resides: The generative power of unconscious thought. Consciousness and Cognition, 15, 135–146. doi:10.1016/ j.concog.2005.04.007 Drazin, R., Glynn, M., & Kazanjian, R. (1999). Multilevel theorizing about creativity in organizations: A sensemaking perspective. Academy of Management Review, 24, 286–307. doi:10.2307/259083 Duncker, K. (1945). On problem-solving. Psychological Monographs, 58(5). Eisenberger, R., & Armeli, S. (1997). Can salient reward increase creative performance without reducing intrinsic creative interest? Journal of Personality and Social Psychology, 72, 652–663. doi:10.1037/ 0022-3514.72.3.652 Eisenberger, R., & Rhoades, L. (2001). Incremental effects of rewards on creativity. Journal of Personality and Social Psychology, 81, 728–741. doi:10.1037/ 0022-3514.81.4.728 Eisenhardt, K. M. (1989). Building theories from case study research. Academy of Management Review, 14, 532–550. doi:10.2307/258557 Elsbach, K. D., & Hargadon, A. B. (2006). Enhancing creativity through “mindless” work: A framework of workday design. Organization Science, 17, 470–483. doi:10.1287/orsc.1060.0193 Farmer, S. M., Tierney, P., & Kung-McIntyre, K. (2003). Employee creativity in Taiwan: An application of role identity theory. Academy of Management Journal, 46, 618–630. Feist, G. J. (1998). A meta-analysis of personality in scientific and artistic creativity. Personality and Social Psychology Review, 2, 290–309. doi:10.1207/ s15327957pspr0204_5 Feist, G. J. (1999). Influence of personality on artistic and scientific creativity. In R. Sternberg (Ed.), Handbook of creativity (pp. 273–296). Cambridge, England: Cambridge University Press. Fiedler, K. (1988). Emotional mood, cognitive style, and behavior regulation. In K. Fiedler & J. Forgas (Eds.), Affect, cognition and social behavior (pp. 101–119). Toronto: Hogrefe.

Finke, R. A., Ward, T. B., & Smith, S. M. (1992). Creative cognition: Theory, research, and applications. Cambridge, MA: MIT Press. Florida, R. (2004). America’s looming creativity crisis. Harvard Business Review, 10, 122–124. Fong, C. T. (2006). The effects of emotional ambivalence on creativity. Academy of Management Journal, 49, 1016–1030. Ford, C. M. (1996). A theory of individual creative action in multiple social domains. Academy of Management Review, 21, 1112–1142. doi:10.2307/259166 Frese, M., Teng, E., & Wijnen, C. J. (1999). Helping to improve suggestion systems: Predictors of making suggestions in companies. Journal of Organizational Behavior, 20, 1139–1155. doi:10.1002/(SICI)10991379(199912)20:73.0.CO;2-I Friedman, R. S., & Forster, J. (2001). The effects of promotion and prevention cues on creativity. Journal of Personality and Social Psychology, 81, 1001–1013. doi:10.1037/0022-3514.81.6.1001 Gedo, J. E. (1997). Psychoanalytic theories of creativity. In M. A. Runco (Ed.), Creativity research handbook (Vol. 1, pp. 29–40). Cresskill, NJ: Hampton. George, J. M., & Brief, A. P. (1992). Feeling good–doing good: A conceptual analysis of the mood at work– organizational spontaneity relationship. Psychological Bulletin, 112, 310–329. doi:10.1037/0033-2909. 112.2.310 George, J. M., & Zhou, J. (2001). When openness to experience and conscientiousness are related to creative behavior: An interactional approach. Journal of Applied Psychology, 86, 513–524. doi:10.1037/0021-9010.86.3.513 George, J. M., & Zhou, J. (2002). Understanding when bad moods foster creativity and good ones don’t: The role of context and clarity of feelings. Journal of Applied Psychology, 87, 687–697. doi:10.1037/ 0021-9010.87.4.687 George, J. M., & Zhou, J. (2007). Dual tuning in a supportive context: Joint contributions of positive mood, negative mood, and supervisory behaviors to employee creativity. Academy of Management Journal, 50, 605–622. Gilson, L. L. (2008). Why be creative: A review of the practical outcomes associated with creativity at the individual, group, and organizational levels. In J. Zhou & C. E. Shalley (Eds.), Handbook of organizational creativity (pp. 303–322). New York: Erlbaum. Gilson, L. L., Mathieu, K. E., Shalley, C. E., & Ruddy, T. M. (2005). Creativity and standardization: Complementary or conflicting drivers of team effectiveness? Academy of Management Journal, 48, 521–531. 297

Zhou and Shalley

Gilson, L. L., & Shalley, C. E. (2004). A little creativity goes a long way: An examination of teams’ engagement in creative processes. Journal of Management, 30, 453–470. doi:10.1016/j.jm.2003.07.001 Goncalo, J. A., & Staw, B. M. (2006). Individualismcollectivism and group creativity. Organizational Behavior and Human Decision Processes, 100, 96–109. doi:10.1016/j.obhdp.2005.11.003 Gough, H. G. (1979). A creativity scale for the Adjective Check List. Journal of Personality and Social Psychology, 37, 1398–1405. doi:10.1037/0022-3514.37.8.1398

Copyright American Psychological Association. Not for further distribution.

Guilford, J. P. (1956). The structure of intellect. Psychological Bulletin, 53, 267–293. doi:10.1037/ h0040755 Guilford, J. P. (1967). The nature of human intelligence. New York: McGraw-Hill. Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, MA: Addison-Wesley. Hargadon, A. B., & Bechky, B. A. (2006). When collections of creatives become creative collectives: A field study of problem solving at work. Organization Science, 17, 484–500. doi:10.1287/orsc.1060.0200 Hargadon, A., & Sutton, R. I. (1997). Technology brokering and innovation in a product development firm. Administrative Science Quarterly, 42, 716–749. doi:10.2307/2393655 Hatcher, L., Ross, T. L., & Collins, D. (1989). Prosocial behavior, job complexity, and suggestion contribution under gainsharing plans. Journal of Applied Behavioral Science, 25, 231–248. doi:10.1177/ 0021886389253002 Hirst, G., van Knippenberg, D., & Zhou, J. (2009). A cross-level perspective on employee creativity: Goal orientation, team learning behavior, and individual creativity. Academy of Management Journal, 52, 280–293. Hocevar, D. (1981). Measurement of creativity: Review and critique. Journal of Personality Assessment, 45, 450–464. doi:10.1207/s15327752jpa4505_1 Hocevar, D., & Bachelor, P. (1989). A taxonomy and critique of measurements used in the study of creativity. In J. A. Glover, R. R. Ronning, & C. R. Reynolds (Eds.), Handbook of creativity (pp. 53–75). New York: Plenum Press. Hogarth, R. (1980). Judgment and choice. Chichester, England: Wiley. Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). Positive affect facilitates creative problem solving. Journal of Personality and Social Psychology, 52, 1122–1131. doi:10.1037/0022-3514.52.6.1122 Isen, A. M., Johnson, M. M. S., Mertz, E., & Robinson, G. R. (1985). The influence of positive affect on the unusualness of word associations. Journal of 298

Personality and Social Psychology, 48, 1413–1426. doi:10.1037/0022-3514.48.6.1413 Jabri, M. (1991). The development of conceptually independent subscales in the measurement of modes of problem solving. Educational and Psychological Measurement, 51, 975–983. doi:10.1177/ 001316449105100417 Jackson, S. E., Joshi, A., & Erhardt, N. L. (2003). Recent research on team and organizational diversity: SWOT analysis and Implications. Journal of Management, 29, 801–830. Janssen, O., & Huang, X. (2007). Us and me: Team identification and individual differentiation as complementary drivers of team members’ citizenship and creative behaviors. Journal of Management, 34, 69–88. doi:10.1177/0149206307309263 Jett, Q. R., & George, J. M. (2003). Work interrupted: A closer look at the role of interruptions in organizational life. Academy of Management Review, 28, 494–507. Jones, A. P., Johnson, L. A., Butler, M. C., & Main, D. S. (1983). Apples and oranges: An empirical comparison of commonly used indices of interrater agreement. Academy of Management Journal, 26, 507–519. doi:10.2307/256262 Kark, R., & Carmeli, A. (2008). Alive and creating: The mediating role of vitality and aliveness in the relationship between psychological safety and creative work involvement. Journal of Organizational Behavior, 30, 785–804. doi:10.1002/job.571 Kaufmann, G. (2003). The effect of mood on creativity in the innovation process. In L. V. Shavinina (Ed.), The international handbook on innovation (pp. 191–203). Oxford, England: Elsevier Science. Kazanjian, R. K., Drazin, R., & Glynn, M. A. (2000). Creativity and technological learning: The roles of organization architecture and crisis in large-scale projects. Journal of Engineering and Technology Management, 17, 273–298. doi:10.1016/S09234748(00)00026-6 Kirton, M. J. (1976). Adaptors and innovators: A description and measure. Journal of Applied Psychology, 61, 622–629. doi:10.1037/0021-9010.61.5.622 Kirton, M. J. (1994). Adaptors and innovators: Styles of creativity and problem solving (2nd ed.). New York: Routledge. Koestler, A. (1964). The act of creation. New York: Macmillan. Kray, L. J., Galinsky, A. D., & Wong, E. M. (2006). Thinking within the box: The relational processing style elicited by counterfactual mind-sets. Journal of Personality and Social Psychology, 91, 33–48. doi:10.1037/0022-3514.91.1.33

Understanding Creativity in the Workplace

Kruglanski, A. W., Friedman, I., & Zeevi, G. (1971). The effects of extrinsic incentive on some qualitative aspects of task performance. Journal of Personality, 39, 606–617. doi:10.1111/j.1467-6494.1971. tb00066.x Kurtzberg, T. R., & Amabile, T. M. (2001). From Guilford to creative synergy: Opening the black box of teamlevel creativity. Creativity Research Journal, 13, 285–294. doi:10.1207/S15326934CRJ1334_06

Copyright American Psychological Association. Not for further distribution.

Lau, D. C., & Murnighan, J. K. (2005). Interactions within groups and subgroups: The effects of demographic faultlines. Academy of Management Journal, 48, 645–659. Lawlis, G. F., & Lu, E. (1972). Judgment of counseling process: Reliability, agreement, and error. Psychological Bulletin, 78, 17–20. doi:10.1037/h0032935 Leenders, R. T. A. J., van Engelen, J. M. L., & Kratzer, J. (2003). Virtuality, communication, and new product team creativity: A social network perspective. Journal of Engineering and Technology Management, 20, 69–92. doi:10.1016/S0923-4748(03)00005-5 Leung, A. K., Maddux, W., Galinsky, A. D., & Chiu, C. Y. (2008). Multicultural experience enhances creativity: The when and how. American Psychologist, 63, 169–181. doi:10.1037/0003-066X.63.3.169 Madjar, N., Oldham, G. R., & Pratt, M. G. (2002). There’s no place like home? The contributions of work and nonwork creativity support to employees’ creative performance. Academy of Management Journal, 45, 757–767. doi:10.2307/3069309 Madjar, N., & Shalley, C. E. (2008). Multiple tasks’ and multiple goals’ effect on creativity: Forced incubation or just a distraction? Journal of Management, 34, 786–805. doi:10.1177/0149206308318611 Martin, L. L., & Stoner, P. (1996). Mood as input: What we think about how we feel determines how we think. In L. L. Martin & A. Tesser (Eds.), Striving and feeling: Interactions among goals, affect, and selfregulation (pp. 279–301). Mahwah, NJ: Erlbaum. McGraw, K. (1978). The detrimental effects of reward on performance: A literature review and a prediction model. In M. Lepper & D. Greene (Eds.), The hidden costs of reward (pp. 33–60). Hillsdale, NJ: Erlbaum. Mednick, S. A. (1962). The associative basis of the creative process. Psychological Review, 69, 220–232. doi:10.1037/h0048850 Milliken, F. J., & Martins, L. L. (1996). Searching for common threads: Understanding the multiple effects of diversity in organizational groups. Academy of Management Review, 21, 402–433. doi:10.2307/ 258667 Miron, E., Erez, M., & Naveh, E. (2004). Do personal characteristics and cultural values that promote innovation, quality, and efficiency compete or com-

plement each other? Journal of Organizational Behavior, 25, 175–199. doi:10.1002/job.237 Mumford, M. D. (2003). Where have we been, where are we going? Taking stock in creativity research. Creativity Research Journal, 15, 107–120. doi:10.1207/ S15326934CRJ152&3_01 Mumford, M. D., & Gustafson, S. B. (1988). Creativity syndrome: Integration, application, and innovation. Psychological Bulletin, 103, 27–43. doi:10.1037/00332909.103.1.27 Okuda, S. M., Runco, M. A., & Berger, D. W. (1991). Creativity and the finding and solving of real-world problems. Journal of Psychoeducational Assessment, 9, 45–53. Oldham, G. R., & Cummings, A. (1996). Employee creativity: Personal and contextual factors at work. Academy of Management Journal, 39, 607–634. doi:10.2307/256657 Olton, R. M., & Johnson, D. M. (1976). Mechanisms of incubation in creative problem solving. The American Journal of Psychology, 89, 617–630. doi:10.2307/ 1421461 Osborn, A. F. (1953). Applied imagination. New York: Scribner’s. Parnes, S. J., Noller, R. B., & Biondi, A. M. (1977). Guide to creative action. New York: Scribner’s. Paulus, P. B. (2000). Groups, teams, and creativity: The creative potential of idea-generating groups. Applied Psychology, 49, 237–262. doi:10.1111/14640597.00013 Paulus, P. B. (2008). Fostering creativity in groups and teams. In J. Zhou & C. E. Shalley (Eds.), Handbook of organizational creativity (pp. 165–188). New York: Erlbaum. Paulus, P. B., & Yang, H. (2000). Idea generation in groups: A basis for creativity in organizations. Organizational Behavior and Human Decision Processes, 82, 76–87. doi:10.1006/obhd.2000.2888 Pearsall, M. J., Ellis, A. P. J., & Evans, J. M. (2008). Unlocking the effects of gender faultiness on team creativity: Is activation the key? Journal of Applied Psychology, 93, 225–234.doi:10.1037/0021-9010. 93.1.225 Perry-Smith, J. E. (2006). Social yet creative: The role of social relationships in facilitating individual creativity. Academy of Management Journal, 49, 85–101. Perry-Smith, J. E., & Shalley, C. E. (2003). The social side of creativity: A static and dynamic social network perspective. Academy of Management Review, 28, 89–106. Peterson, R. S., & Nemeth, C. J. (1996). Focus versus flexibility: majority and minority influence can both improve performance. Personality and Social 299

Zhou and Shalley

Copyright American Psychological Association. Not for further distribution.

Psychology Bulletin, 22, 14–23. doi:10.1177/ 0146167296221002

Academy of Management Journal, 38, 483–503. doi:10.2307/256689

Pirola-Merlo, A., & Mann, L. (2004). The relationship between individual creativity and team creativity: Aggregating across people and time. Journal of Organizational Behavior, 25, 235–257. doi:10.1002/ job.240

Shalley, C. E. (2008). Creating roles: What managers can do to establish expectations for creative performance. In J. Zhou & C. E. Shalley (Eds.), Handbook of organizational creativity (pp. 147–164). Hillsdale, NJ: Erlbaum.

Reiter-Palmon, R., & Illies, J. J. (2004). Leadership and creativity: Understanding leadership from a creative problem-solving perspective. Leadership Quarterly, 15, 55–77. doi:10.1016/j.leaqua.2003.12.005

Shalley, C. E., & Gilson, L. L. (2004). What leaders need to know: A review of social and contextual factors that can foster or hinder creativity. Leadership Quarterly, 15, 33–53. doi:10.1016/j.leaqua.2003.12.004

Reiter-Palmon, R., Mumford, M. D., Boes, J. O., & Runco, M. A. (1997). Problem construction and creativity: The role of ability, cue consistency, and active processing. Creativity Research Journal, 10, 9–23. doi:10.1207/s15326934crj1001_2

Shalley, C. E., Gilson, L. L., & Blum, T. C. (2000). Matching creativity requirements and the work environment: Effects of satisfaction and intention to leave. Academy of Management Journal, 43, 215–223. doi:10.2307/1556378

Reiter-Palmon, R., Mumford, M. D., & Threlfall, K. V. (1998). Solving everyday problems creatively: The role of problem construction and personality type. Creativity Research Journal, 11, 187–197. doi:10.1207/ s15326934crj1103_1

Shalley, C. E., Gilson, L. L., & Blum, T. C. (2009). Interactive effects of growth need strength, work context, and job complexity on self-reported creative performance. Academy of Management Journal, 52, 489–505.

Runco, M. A., & Chand, I. (1995). Cognition and creativity. Educational Psychology Review, 7, 243–267. doi:10.1007/BF02213373 Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39, 1161–1178. doi:10.1037/h0077714 Ryan, R. M., & Bernstein, J. H. (2004). Vitality. In C. Peterson & M. E. P. Seligman (Eds.), Character strengths and virtues: A handbook and classification (pp. 273–290). New York: Oxford University Press. Schumpeter, J. (1939). Business cycles: A theoretical, historical, and statistical analysis of the capitalist process. New York: McGraw-Hill. Schwarz, N. (2002). Situated cognition and the wisdom of feelings: Cognitive tuning. In L. Feldman Barrett & P. Salovey (Eds.), The wisdom in feelings (pp. 144–166). New York: Guilford Press. Schwarz, N., & Clore, G. L. (2003). Mood as information. Psychological Inquiry, 14, 296–303. doi:10.1207/ S15327965PLI1403&4_20

Shalley, C. E., & Oldham, G. R. (1997). Competition and creative performance: Effects of competitor presence and visibility. Creativity Research Journal, 10, 337–345. doi:10.1207/s15326934crj1004_5 Shalley, C. E., & Perry-Smith, J. E. (2001). Effects of social-psychological factors on creative performance: The role of informational and controlling expected evaluation and modeling experience. Organizational Behavior and Human Decision Processes, 84, 1–22. doi:10.1006/obhd.2000.2918 Shalley, C. E., & Perry-Smith, J. E. (2008). The emergence of team creative cognition: The role of diverse outside ties, sociocognitive network centrality, and team evolution. Strategic Entrepreneurship Journal, 2, 23–41. doi:10.1002/sej.40 Shalley, C. E., & Zhou, J. (2008). Organizational creativity research: A historical review. In J. Zhou & C. E. Shalley (Eds.), Handbook of organizational creativity (pp. 3–32). Hillsdale, NJ: Erlbaum.

Scott, S. G., & Bruce, R. A. (1994). Determinants of innovative behavior: A path model of individual innovation in the workplace. Academy of Management Journal, 37, 580–607. doi:10.2307/256701

Shalley, C. E., Zhou, J., & Oldham, G. R. (2004). The effects of personal and contextual characteristics on creativity: Where should we go from here? Journal of Management, 30, 933–958. doi:10.1016/ j.jm.2004.06.007

Segal, E. (2004). Incubation in insight in problem solving. Creativity Research Journal, 16, 141–148. doi:10.1207/ s15326934crj1601_13

Shin, S., & Zhou, J. (2003). Transformational leadership, conservation, and creativity: Evidence from Korea. Academy of Management Journal, 46, 703–714.

Shalley, C. E. (1991). Effects of productivity goals, creativity goals, and personal discretion on individual creativity. Journal of Applied Psychology, 76, 179–185. doi:10.1037/0021-9010.76.2.179

Shin, S. J., & Zhou, J. (2007). When is educational specialization heterogeneity related to creativity in research and development teams? Transformational leadership as a moderator. Journal of Applied Psychology, 92, 1709–1721. doi:10.1037/00219010.92.6.1709

Shalley, C. E. (1995). Effects of coaction, expected evaluation, and goal setting on creativity and productivity. 300

Understanding Creativity in the Workplace

Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420–428. doi:10.1037/0033-2909.86.2.420

Unsworth, K. (2001). Unpacking creativity. Academy of Management Review, 26, 289–297. doi:10.2307/ 259123

Simonton, D. K. (1999). Creativity as blind variation and selective retention: Is the creative process Darwinian? Psychological Inquiry, 10, 309–328.

Unsworth, K. L., Wall, T. B., & Carter, A. (2005). Creative requirements: A neglected construct in the study of employee creativity? Group & Organization Management, 30, 541–560. doi:10.1177/ 1059601104267607

Copyright American Psychological Association. Not for further distribution.

Smilansky, J. (1984). Problem solving and the quality of invention: An empirical investigation. Journal of Educational Psychology, 76, 377–386. doi:10.1037/ 0022-0663.76.3.377 Smith, S. M., Ward, T. B., & Finke, R. A. (Eds.). (1995). The creative cognition approach. Cambridge, MA: MIT Press.

Van Dyne, L., Jehn, K. A., & Cummings, A. (2002). Differential effects of strain on two forms of work performance: Individual employee sales and creativity. Journal of Organizational Behavior, 23, 57–74. doi:10.1002/job.127

Stahl, M. J., & Koser, M. C. (1978). Weighted productivity in R & D: Some associated individual and organizational variables. IEEE Transactions on Engineering Management, EM-25, 20–24.

Vincent, A. S., Decker, B. P., & Mumford, M. D. (2002). Divergent thinking, intelligence, and expertise: A test of alternative models. Creativity Research Journal, 14, 163–178. doi:10.1207/S15326934CRJ1402_4

Stein, B. S. (1989). Memory and creativity. In J. A. Glover, R. R. Ronning, & C. R. Reynolds (Eds.), Handbook of creativity (pp. 163–176). New York: Plenum Press.

Wallas, G. (1926). The art of thought. London: Cape.

Stein, M. I. (1967). Creativity and culture. In R. Mooney & T. Razik (Eds.), Explorations in creativity (pp. 109–119). New York: Harper.

Ward, T. B., Smith, S. M., & Finke, R. A. (1999). Creative cognition. J. Sternberg (Ed.), Handbook of creativity (pp. 189–212). New York: Cambridge University Press.

Szymanski, K., & Harkins, S. G. (1992). Self-evaluation and creativity. Personality and Social Psychology Bulletin, 18, 259–265. doi:10.1177/0146167292183001

Williams, K. Y., & O’Reilly, C. A. (1998). Demography and diversity in organizations: A review of 40 years of research. Research in Organizational Behavior, 20, 77–140.

Taggar, S. (2002). Individual creativity and group ability to utilize individual creative resources: a multilevel model. Academy of Management Journal, 45, 315–330. doi:10.2307/3069349

Woodman, R. W., Sawyer, J. E., & Griffin, R. W. (1993). Toward a theory of organizational creativity. Academy of Management Review, 18, 293–321. doi:10.2307/ 258761

Thatcher, S. M. B., & Greer, L. L. (2008). Does it really matter if you recognize who I am? The implications of identity comprehension for individuals in work teams. Journal of Management, 34, 5–24.

Yuan, F., & Zhou, J. (2008). Differential effects of expected external evaluation on different parts of the creative idea production process and on final product creativity. Creativity Research Journal, 20, 391–403. doi:10.1080/10400410802391827

Tierney, P., & Farmer, S. M. (2002). Creative self-efficacy: Potential antecedents and relationship to creative performance. Academy of Management Journal, 45, 1137–1148. doi:10.2307/3069429 Tierney, P., & Farmer, S. M. (2004). The Pygmalion process and employee creativity. Journal of Management, 30, 413–432. doi:10.1016/j.jm.2002. 12.001 Tierney, P., Farmer, S. M., & Graen, G. B. (1999). An examination of leadership and employee creativity: The relevance of traits and relationships. Personnel Psychology, 52, 591–620. doi:10.1111/j.1744-6570. 1999.tb00173.x Tinsley, H. E. A., & Weiss, D. J. (1975). Interrater reliability and agreement of subjective judgments. Journal of Counseling Psychology, 22, 358–376. doi:10.1037/ h0076640 Torrance, E. P. (1974). Torrance tests of creative thinking: Directions manual and scoring guide. Englewood Cliffs, NJ: Prentice Hall.

Zhong, C. B., Dijksterhuis, A., & Galinsky, A. D. (2008). The merit of unconscious thought in creativity. Psychological Science, 19, 912–918. doi:10.1111/ j.1467-9280.2008.02176.x Zhou, J. (1998). Feedback valence, feedback style, task autonomy, and achievement orientation: Interactive effects on creative performance. Journal of Applied Psychology, 83, 261–276. doi:10.1037/0021-9010. 83.2.261 Zhou, J. (2003). When the presence of creative coworkers is related to creativity: Role of supervisor close monitoring, developmental feedback, and creative personality. Journal of Applied Psychology, 88, 413–422. doi:10.1037/00219010.88.3.413 Zhou, J. (2008). Promoting creativity through feedback. In J. Zhou & C. E. Shalley (Eds.), Handbook of organizational creativity (pp. 125–146). Hillsdale, NJ: Erlbaum. 301

Zhou and Shalley

Zhou, J., & George, J. M. (2001). When job dissatisfaction leads to creativity: Encouraging the expression of voice. Academy of Management Journal, 44, 682–696. doi:10.2307/3069410

Zhou, J., & Shalley, C. E. (2008a). Expanding the scope and impact of organizational creativity research. In J. Zhou & C. E. Shalley (Eds.), Handbook of organizational creativity. Hillsdale, NJ: Erlbaum.

Zhou, J., & George, J. M. (2003). Awakening employee creativity: The role of leader emotional intelligence. Leadership Quarterly, 14, 545–568. doi:10.1016/ S1048-9843(03)00051-1

Zhou, J., & Shalley, C. E. (Eds.). (2008b). Handbook of organizational creativity. New York: Erlbaum.

Copyright American Psychological Association. Not for further distribution.

Zhou, J., & Oldham, G. R. (2001). Enhancing creative performance: Effects of expected developmental assessment strategies and creative personality. Journal of Creative Behavior, 35, 151–167. Zhou, J., & Shalley, C. E. (2003). Research on employee creativity: A critical review and directions for future research. In J. Martocchio (Ed.), Research in personnel and human resource management (pp. 165–217). Oxford, England: Elsevier.

302

Zhou, J., Shin, S. J., Brass, D. J., Choi, J., & Zhang, Z. (in press). Social networks, personal values, and creativity: Evidence for curvilinear and interaction effects. Journal of Applied Psychology. Zhou, J., Shin, S. J., & Cannella, A. (2008). Employee self-perceived creativity after mergers and acquisitions: Interactive effects of threat–opportunity perception, access to resources, and support for creativity. Journal of Applied Behavioral Science, 44, 397–421. doi:10.1177/ 0021886308328010

CHAPTER 10

PERFORMANCE MEASUREMENT AT WORK: A MULTILEVEL PERSPECTIVE

Copyright American Psychological Association. Not for further distribution.

Jessica L. Wildman, Wendy L. Bedwell, Eduardo Salas, and Kimberly A. Smith-Jentsch

For decades, one of the primary goals of organizational research has been the improvement and management of organizational performance. Inherent to the goal of improving performance is the concept of performance measurement (PM). PM is the mechanism that allows managers and researchers to gain an understanding of individual, team, and overall organizational performance. Without the ability to accurately measure a construct such as performance, it is impossible to truly understand, control, or improve it. As Sink and Tuttle (1989) asserted, one cannot manage what one cannot measure. Ultimately, the effective training and management of employees, teams, and organizations in any context is contingent on the quality of PM. Accordingly, much effort has been devoted over the past several decades to exploring theories, methods, and practices associated with PM (e.g., Bititci, Turner, & Begemann, 2000; Campbell, McCloy, Oppler, & Sager, 1993; Folan & Browne, 2005; Gershoni & Rudy, 1981; Kendall & Salas, 2004; Pun & White, 2005). The PM literature can generally be categorized into three distinct perspectives: individual-level PM, teamlevel PM, and organizational-level PM. Very little research has simultaneously examined multiple levels. This is problematic given that actual performance in organizations takes place at all three levels simultaneously, and perhaps more important, all three levels of performance are intertwined. Teams are becoming the predominant method for achieving organizational goals. These teams are made up of individual employees, who actually engage in behaviors that lead to per-

formance. Thus, there is a need to integrate these three streams of PM research into one comprehensive understanding of PM and its implications. To address this need, this chapter presents a multilevel perspective on the field of PM. First, we discuss the criterion problem, which represents a broad issue underscoring the importance of PM. Next, we briefly describe five critical considerations when choosing or designing any PM system. Then, after the core underlying issues are clear, we dive into PM as described from the individual, team, and organizational perspectives. This includes the general definition of performance, key theories, and common measurement strategies used in each stream of literature. Once each perspective is discussed separately, we discuss a multilevel approach to PM. The chapter concludes with a review of current trends requiring future research and some concluding remarks. (See also Vol. 2, chap. 9, this handbook.) THE CRITERION PROBLEM The development and measurement of appropriate performance criteria is of importance to both researchers and managers alike, as they are both focused on influencing performance. The practical significance of measurement on the basis of sound criteria has long been accepted (e.g., Scott, 1917); however, rigorous research on the “necessary conceptual, taxonomic, and methodological prerequisites for the pursuit of understanding criteria” ( J. T. Austin & Villanova, 1992, p. 836) did not become a prominent area of concern until the early 1990s. Researchers

http://dx.doi.org/10.1037/12169-010 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

303

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

have noted the necessity of well-developed criteria for measuring individual- and team-level performance as well as evaluation of organizational programs and training initiatives (Schmitt & Klimoski, 1991). Performance criteria are initially conceptual in nature and are thus defined on the basis of subjective statements of what is considered successful performance. Simply stated, performance criteria represent whatever aspects of performance a certain set of stakeholders have identified as critical. Thus, the selected dimensions of any given criterion measure are based largely on the defined conceptual criteria (Nagle, 1953; Toops, 1944). For example, if a set of stakeholders are conceptually interested in assessing the productivity of a professor, this could be assessed using measures of effectiveness such as number of publications, number of graduate students sponsored, and number of conference presentations. The important aspect of selecting criteria measures is to make sure these measures are rationally linked to the conceptual criteria and are sufficiently covering the criterion space. In other words, do the measures of performance effectiveness include all of the things that stakeholders deem important to performance in a particular job? One of the most troubling issues in performance research has been the lack of focus on the conscious choice and development of criteria measures. Unfortunately, organizations and researchers often select criteria on the basis of availability or how easy the criteria are to collect. This is problematic because the choice of a performance measure influences how well selected predictors can actually forecast future performance. The choice of outcome criteria is just as important as the choice of predictors, if not more so. The best selection test or training system in the world could be developed; however, without sound criteria to serve as a measure of effectiveness, it is difficult to provide evidence for its validity. If the performance measure chosen as the criterion is not conceptually related to the outcome of interest, or if the measure is contaminated, the selection process may be excellent or the training may be well designed; however, there will never be a demonstrated connection to performance. Therefore, there is a need for an increased focus on developing sound performance criteria and sys304

tematically linking those criteria to other constructs of interest. Additionally, performance criteria should be measured without the influence of halo and other sources of error to capture true performance. These issues fall under the term criterion problem (e.g., Flanagan, 1956; Smith, 1976). Essentially, this refers to problems associated with developing and measuring the multidimensional nature of performance criteria given the constraints of the measurement purpose and situational factors ( J. T. Austin & Villanova, 1992). For example, Viswesvaran, Schmidt, and Ones (2005) found that less than 10% of the variance in a set of job performance ratings or rankings could be attributed to valid performance-related information. The rest was attributed to such things as halo error, rater leniency error, and random error, among others, demonstrating that these problems are important to consider when measuring performance. These errors can be divided into three common categories: distributional errors, illusory halo, and other types of errors. Distributional errors relate to incorrect representations of performance distributions across employees being evaluated (Borman, 1991). These errors can occur in both the rating means (e.g., severity or leniency) and variance (e.g., range restriction and central tendency). If a rater provides ratings that are lower (severity) or higher (leniency) than actually warranted by the performance because of inaccurate norms, then ratings will be erroneously deflated (severity) or inflated (leniency). If a rater fails to sufficiently differentiate between two or more ratees on the same dimension, then restriction of range has occurred. This is similar to the error of central tendency; however, with central tendency, ratings tend to be clustered around the midpoint of any given scale (Tsui & Barry, 1986). The second category of errors is illusory halo, which results in correlations between ratings of two different dimensions being higher (or lower) than the correlation between the actual behaviors reflecting those dimensions. Essentially, raters are either overestimating (higher correlations) or underestimating (lower correlations) the relationship between dimensions (Borman, 1991; Fisicaro, 1988). The final category of other errors includes such perceptual errors as the similar-to-me error and

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

the first-impression error. Similar-to-me error occurs when the rater projects his or her own personal characteristics onto the employee (Latham, Wexley, & Pursell, 1975). If the rater is heavily influenced by early experiences with the ratee, then first-impression error has occurred (Latham et al., 1975). This can cause biased ratings that are lower or higher than the actual performance warrants, depending on whether first impressions are negative or positive. There are additional issues associated with the development of high-quality performance criteria. Because criteria essentially focus on the results or outcomes of performance, there are several steps required to directly link criteria measurement to associated predictors (J. T. Austin & Villanova, 1992). This is problematic in that other variables, such as situational factors, may constrain this translation from predictors to behaviors to results (Binning & Barrett, 1989). It is important to clearly define the constructs within this context. Borman (1991), in his seminal chapter on job performance, defined behavior (what people do), performance (individual contributions toward organizational goals), and effectiveness (outcomes such as promotion rate or salary level). Campbell et al. (1993) suggested that performance is the actual behavior and therefore measuring the behavior constitutes measuring performance. Regardless of the adopted definition, it is fairly easy to measure behaviors, as they are generally observable and can easily be recorded. It is also a relatively simple process to measure results using quantity, quality, or customer satisfaction. The difficulty lies in (a) tying specific behaviors to specific results in the context of performance and (b) measuring the cognitive aspects associated with behaviors. Others have suggested that criteria dimensions are also problematic because they are context sensitive (Bailey, 1983). Given this assertion, measures appropriate for use in one situation would be inappropriate within a different context. Also, as noted previously, the selected dimensions of a criterion construct are based on how the conceptual criteria are defined. An additional issue contributing to the criterion problem is the lack of description often provided as to why certain dimensions were selected and other seemingly important dimensions were ignored (J. T. Austin & Villanova, 1992).

SETTING THE STAGE: BASIC CONSIDERATIONS IN PERFORMANCE MEASUREMENT This section outlines five critical issues to consider when choosing or designing a PM system: (a) the purpose of the measurement, (b) the content of the measurement, (c) the timing of measurement, (d) the fidelity of the measurement setting, and (e) the technique or tools used for measurement. We refer to these issues simply as the why, what, when, where, and how of PM (see Figure 10.1). These considerations are important to keep in mind when examining existing PM strategies, because each strategy presents advantages and disadvantages regarding these considerations. Given that the primary focus of this chapter is on the methods for measuring performance at the individual, team, and organizational levels, the consideration of how (i.e., how to measure performance) is divided into these three perspectives for discussion and represents a large portion of the content in this chapter.

Why: Purpose of Measurement The first critical issue to consider when choosing or designing a PM system is the purpose for the measurement, as the purpose will drive the entire measurement process (Salas, Burke, & Fowlkes, 2006). The purpose determines whether multiple criteria measures (e.g., Bartram, 2005) or a single composite criterion measure (e.g., Viswesvaran et al., 2005) is used. There are numerous uses for PM data, ranging from basic research to a variety of applied purposes such as training development and strategic planning. The most common purposes for PM include research, feedback development, training development, performance evaluation, and organization planning. Multiple measures are appropriate if the purpose is to diagnose performance issues, as this allows for a more accurate picture of areas needing improvement and aids in planning for training and employee development. Composite measures, on the other hand, are better for comparing across units who may not do the same type of work. This is the basis of the Productivity Measurement and Enhancement System (ProMES), a PM system that is discussed later in the chapter. 305

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

FIGURE 10.1. Critical considerations in performance measurement.

Basic research. Accurate PM is absolutely critical to all research endeavors. As stated by Tannenbaum (2006), “measurement lies at the heart of scientific study” (p. 297). Without the ability to accurately and reliably measure performance and other constructs of interest to researchers, it would be impossible to gain any scientific knowledge. Brannick and Prince (1997) also pointed out that “measurement is central to the evaluation and elaboration of theories” (p. 5). Theories would not be validated, or basic relationships tested, without proper measurement. Measurement is the most basic ingredient in any research, for any purpose. Feedback development. There is a large base of literature connecting feedback to improved performance both for individuals and groups (e.g., Pritchard, Youngcourt, Philo, McMonagle, & David, 2007). PM plays a critical role in the development of feedback. Specifically, performance must be measured to assess how an individual or team is performing, including what they are doing right, what they are doing wrong, and where improvements in performance can be made. These performance data can then be used to develop focused feedback, centered on identified strengths, weaknesses, and areas for improvement. Therefore, accurate and thorough PM is the first step in any feedback system. By accurately measuring and describing the performance of an individual, feedback can be used as specific instructions for performance improvement. 306

Training development and evaluation. The use of PM data for feedback development is quite similar to the use of PM data for training development. Another intervention designed to improve performance, training, aims to develop an individual’s or team’s knowledge or skills by providing information and opportunities for practice. It is important that training systems are designed to address specific deficiencies in employee performance, as they can often be costly and time consuming both to design and implement. PM data are a necessary first step in the development of training. These data are used to identify deficiencies and pinpoint knowledge, skills, and abilities in need of improvement. PM also plays a role in the assessment of the training system effectiveness. Specifically, performance must be measured at the conclusion of the training program and linked to relevant outcomes to assess whether the training is imparting the desired knowledge or skills (i.e., whether there was learning) and ultimately contributing to the performance of the employees and organization as a whole (i.e., whether there was training transfer). Training is a cycle of providing instruction, assessing learning and outcomes, and adjusting instruction on the basis of that assessment. PM facilitates this cycle. In addition to remedial efforts, training can also be used to provide new knowledge, skills, or attitudes. For example, nearly 88% of organizations with revenues exceeding $10 billion have executive development programs aimed at providing executives with

Performance Measurement at Work

Copyright American Psychological Association. Not for further distribution.

skills or knowledge that are not necessarily directly related to their current positions (Czarnowsky, 2008). PM can aid in the development of these programs as well. By measuring the performance of top executives, an organization can establish criteria of what a successful executive looks like by focusing on the strengths of each individual that positively contribute to personal and organizational effectiveness. These criteria can then be used to develop training targeting future executive development efforts.

such as these would be made blindly. Finally, organizations are interested in overall measures to inform decisions regarding human resources (e.g., recruitment, training).

What: Content of Measurement

Performance evaluations. PM data can also be more simply used for evaluation purposes. Employee evaluations are yearly or quarterly assessments used to determine the comparative success of individuals within an organization. Data from these evaluations, usually in the form of subjective ratings performed by supervisors, can then be used to determine various human resources decisions such as promotions, salary changes, or bonuses. PM data for evaluation purposes often serve as the justification behind these types of decisions that must be made in all organizations. Given that performance evaluation data can often impact individual employees in salient and life-changing ways (e.g., firing, promotion), it is absolutely critical that performance data used for this purpose are accurate and nonbiased. Team-level PM can also be used for evaluative purposes, similarly to individual-level data.

Once the purpose of the measurement has been identified, it is necessary to determine what corresponding content should be captured. Depending on the reason behind the PM, and how performance is being defined, there are numerous behavioral aspects that could be measured. For example, if the purpose of the measurement is to develop taskwork training for pilots, then measuring task-related performance would likely be the best choice of content. However, if the purpose is to look at how aircrews function together as a cohesive unit, teamworkrelated behaviors should be the focus of measurement. As is described later in the chapter, there are many different types of performance that can be measured that focus on task performance, interpersonal performance, or the outcomes of performance. Which type of performance is measured should be decided on carefully to best match the purpose of the PM system. This consideration relates heavily to the criterion problem and the importance of choosing performance measures that represent the conceptual criteria of interest.

Organizational planning. Up to this point, every purpose for PM discussed has focused on the individual or team level. However, PM is also critical and necessary at the overall organizational level as well. All organizational-level decision making and planning relies on accurate measurement of performance at the individual, team, and organizational level. For example, if an organization puts a new policy or program in place, they will undoubtedly need to measure performance at some point after implementation to assess whether that program is working as intended and to decide whether the program should be modified, expanded, or eliminated (Tannenbaum, 2006). Assessing big picture performance also allows for an organization to keep track of their organizational health, which can lead to high-level decisions such as mergers or acquisitions. Without PM at the organizational level, decisions

Criteria: A deeper look. The conceptual criterion can be described as a verbal statement of the important outcomes related to a particular problem (Borman, 1991). Conceptual criteria are abstract statements of what is important to the stakeholder and represent the starting point that drives the development of performance measures. Essentially, conceptual criteria are the gold standard of what a highly successful employee, team, or organization would look like if performing at the highest level. Consequently, conceptual criteria are very subjective in nature. Subject matter experts (SMEs) can provide insight, but the bottom line is that the conceptual criteria should conceptually relate back to the organizational mission and goals. Measures of effectiveness (outcomes) are developed on the basis of the conceptual criteria. These should be developed rationally to ensure they map onto the conceptual crite307

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

ria. Measures of effectiveness can be considered the operational definition of the conceptual criteria, with an evaluative component. Muchinsky (2009) provided an example of conceptual criteria by using the example of a successful college student. He suggested that one important dimension is intellectual growth, noting that highly successful college students experience more intellectual growth than unsuccessful, or less successful, colleagues. Additionally, he pointed to emotional growth as a second dimension, positing that a college education should allow successful students to clarify values and beliefs, aiding in their development and stability. A final dimension to be considered might be citizenship, whereby successful college students desire to engage in civic activities and positively contribute to their surrounding community. Muchinksy suggested that these three factors are the defining criteria, or conceptual criteria, of what constitutes a successful college student. Yet, these are theoretical; therefore, the challenge is to convert these theoretical ideas of desired behaviors into something that can be quantifiably measured. Performance measures: General characteristics. Generally, performance measures should provide information regarding products, services, or the tasks that individuals or teams complete to produce those products or services. Performance measures are essentially tools that let decision makers see how well individuals, groups, teams, or organizations are doing. Additionally, they provide insight into whether goals are being met, whether customers are satisfied, whether processes are indeed working as desired, and where improvements are needed. Performance measures can also be multidimensional. There are numerous examples of this dimensionality. For example, number of accidents or injuries per million hours worked is one indicator of a company’s safety program. However, the cost of injuries provides additional information regarding safety program effectiveness. This type of measure provides more detailed information than just the first example, which is a single dimensional measure. Essentially, whatever is measured must be expressed in measurement units that are meaningful given the entire purpose of measurement. 308

When: Timing of Measurement Another consideration focuses on when the construct will be measured. Performance is dynamic and changes over time. Processes that are happening at the beginning of a performance cycle may change or even be replaced with different processes at the end of the performance cycle. Therefore, the point during performance at which a construct is measured and the amount of times it is measured (i.e., once or repeated measures) may have a significant impact on what information is captured. For example, many performance measures can be considered “lagging measures” in that they capture performance outcomes long after the behavior that led to those outcomes occurred. End-of-the-year performance reviews and archival data are two good examples of this type of measurement. These sources are practical and useful measures of past performance, and they can be used to link specific performance behaviors to more distal organizational outcomes such as financial success. However, they may provide an inaccurate, or outdated, understanding of current performance, especially if the performance in question is likely to change quickly or often over time (i.e., is cyclical in nature). Measuring performance throughout a performance period is advantageous because it provides a real-time understanding of what behaviors are actually occurring that lead to the performance outcome. Several of the measures commonly used in the team literature, such as event-based measurement and communication analysis, take this approach. However, the limitation with midperformance measurement is that usually it is a more intrusive method of measurement. If individuals are aware they are being observed or evaluated, there may be an issue with eliciting maximum versus typical performance, which is discussed in more detail in the following section. One advantage of a repeated measures design for PM is the ability to determine the magnitude of any gains in performance. For example, assume performance is being measured to assess the effectiveness of a training program or some other intervention. A pre- and postmeasurement approach allows for a comparison of levels prior to training and levels posttraining. It is also a useful method for capturing the dynamic aspect of performance. Specifically, dif-

Performance Measurement at Work

ferent performance processes may become more or less critical at different points during a performance cycle, and by measuring performance repeatedly, these changes in process can be captured. However, there are also limitations with this design. Without a control group, it is difficult to conclude with certainty that any noted improvements in performance were specifically due to the intervention and not some outside influence.

Copyright American Psychological Association. Not for further distribution.

Where: Fidelity of the Measurement Setting Another issue to consider when measuring performance is how characteristics of the setting will impact the process of PM. This issue pertains mostly to observational or rating-type measures, as knowledge tests and financial data are generally independent of the setting. Observational methods, however, are directly influenced by the realism of the measurement setting. In both laboratory and on-the-job settings, the level of fidelity influences the process of PM. Hays and Singer (1989) defined fidelity as “the similarity between the . . . situation and the operational situation which is simulated” (p. 50). In the case of PM, this refers to how closely the measurement setting replicates the actual performance situation it is intended to represent. Fidelity can be further defined in terms of two dimensions: (a) the physical characteristics of the measurement environment (i.e., the look and feel of the equipment and environment) and (b) the functional characteristics of the measurement environment (i.e., the functional aspects of the task and equipment). Depending on the purpose and nature of the performance measures, different levels of fidelity will be more or less appropriate. For example, if the measurement is intended to capture day-to-day performance of employees on the job, and the job in question is highly dependent on various changes that occur in a fast-paced dynamic environment, a strictly controlled laboratory setting with a low level of fidelity may result in misleading findings. Imagine trying to measure the performance of a team of emergency medical technicians (EMTs) while they are responding to a severe vehicle collision. In this situation, PM may be more accurate if gathered from the natural job environment or in a high-fidelity laboratory setting (i.e.,

simulation) designed to closely mimic the complex dynamic environment faced by the EMTs. If the simulated environment does not accurately represent the potentially complex environmental factors that are inherent in situations commonly faced by EMTs (e.g., quickly changing medical status of victims, severe weather, vehicle fires or explosions), the measures may not capture performance that is indicative of day-to-day actions. On the other end of the spectrum, some tasks can be appropriately measured using lower fidelity situations. For example, an assembly line worker most likely could perform his or her task in an artificially contrived task simulation in relatively the same manner as he or she performs in the actual work environment. Overall, the level of fidelity of the setting, nature of the task, and purpose of the measurement must be considered in tandem when choosing the setting in which to conduct PM. Another important issue to consider when examining the fidelity of a measurement environment is the problem of maximum versus typical performance. Maximum performance can be defined as the highest level of performance possible to achieve under optimal conditions, whereas typical performance is the average performance on a day-to-day basis (Mangos & Arnold, 2008). Sackett, Zedeck, and Fogli (1988) provided an example to illustrate the differences between the two constructs, using grocery store register clerks. Typical performance was operationalized as the average number of items scanned less the number of voids per shift, whereas maximum performance was operationalized as the speed and accuracy of scanning items averaged across several timed observations. They found that the measures were not statistically related, which suggests that typical and maximum performance are distinct constructs. Additionally, research has shown that each construct has different antecedents. For example, intelligence is more highly related to maximum performance and personality is more predictive of typical performance (Dubois, Sackett, Zedeck, & Fogli, 1993). There are cultural implications to maximum and typical performance as well. Dubois et al. (1993) found that Caucasians outperformed African Americans on typical performance; however, the differences were minimal when looking at maximum performance. 309

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

Certain environmental cues can elicit maximum performance conditions, which can be problematic if the measurement is intended to capture typical, day-to-day performance. When individuals are acutely aware that they are being observed and evaluated, they will likely try to perform to the best of their ability (Sacket et al., 1988). Consequently, maximum performance may be unintentionally elicited and measured. This same phenomenon can occur when observers are present in the natural work environment and the individuals being observed are aware of their presence. This is a significant problem, as research has shown that both maximum and typical performance are predicted by different variables (Campbell et al., 1993; Lim & Ployhart, 2004). Therefore, if the goal of measurement is to represent typical performance, the knowledge of being observed may trigger maximum performance instead and may distort findings regarding the relationships between performance criteria and predictor variables. Fidelity is often associated with a trade-off in terms of the level of experimental control in a measurement setting. For example, although measurement in the operational work environment is as realistic (i.e., high fidelity levels) as possible, this usually makes it more difficult to isolate and identify the causes underlying performance because so many uncontrolled variables are freely influencing performance. In other words, because the experimenter does not design and control the setting of the measurement, there is the potential for any number of environmental factors to influence or, more important, confound, results. Therefore, it is often more difficult to assess the effectiveness of training or other performance interventions in the field because effects can be hidden by various outside influences. However, results found while measuring performance in the operational setting will most often be more externally valid than results found in more artificial or lower fidelity settings (i.e., lab setting), because data are collected directly in the environment to which they are intended to generalize.

How: Measurement Techniques One final consideration when choosing or designing a PM system is the technique or approach for mea310

surement. Rather than first presenting the theoretical background for each level, and then the measurement techniques separately, we group them together within the realm of individual, team, and organizational perspectives to ensure that each measurement technique is considered within the appropriate theoretical context. Our hope is that this delineation will provide insight into the most commonly used measurement techniques at the different levels in addition to providing the necessary context for why consideration of multilevels with regard to measurement warrants attention. Therefore, in the next sections, we discuss each level (individual, team, and organizational) and describe common techniques frequently used for PM at that level. It is absolutely critical to note that the measurement strategies discussed in each perspective are in no way used exclusively within that perspective. Many of the measurement tools mentioned are clearly applicable to, and consequently have been used across, multiple levels of PM. Additionally, the list of measurement strategies we provide is by no means exhaustive. However, as our primary goal is to compare and integrate three distinct streams of literature, we discuss each selected measurement strategy within the theoretical perspective in which it is discussed or most frequently used. We bring all three perspectives together at the conclusion of the chapter. PERFORMANCE MEASUREMENT FROM THE INDIVIDUAL PERSPECTIVE In the following section, we discuss PM approaches focused on capturing individual-level phenomenon. First, we define individual performance. Second, we describe several key theories of individual performance that have been developed over the years. Finally, we summarize and describe the most common performance measurement approaches used at the individual level.

Defining Individual Performance A majority of the PM literature has been devoted to measurement at the individual level. The most basic resource in an organization is the individual employee, and therefore the first place to start man-

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

aging performance is at the individual level. For as long as organizational science has existed, the term performance has been misused and vaguely defined. Campbell et al. (1993) attempted to rectify this misuse of the term by defining performance as synonymous with behavior. In this view, performance must be the actions of the individual in question. Job performance, specifically, “includes only those actions or behaviors that are relevant to the organization’s goals and that can be scaled in terms of each individual’s proficiency” (Campbell et al., 1993, p. 40). Job performance at the individual level, simply stated, is what employees are hired to do. Therefore, performance is not the outcome or the consequence of behavior; it is the behavior itself (Campbell et al., 1993). This distinction defines the difference between performance, effectiveness, and productivity. Performance is the actions taken by the individual, effectiveness is the “evaluation of the results of performance” and productivity is “the ratio of effectiveness to the cost of achieving that level of effectiveness” (Campbell et al., 1993, p. 41). Each of these terms is an independent construct. One can measure performance without evaluating that performance or without comparing that evaluation with cost. However, when trying to use PM data for any practical purpose, evaluating that performance is a critical step.

Theories of Individual Performance Several theories of individual performance have been developed looking at various aspects of performance such as job performance, organizational citizenship behavior, contextual performance, adaptive performance, integrated work role performance, and counterproductive work behavior. Each of these theories is described in more detail in this section. Job performance behaviors. Along with their definition of performance as synonymous with behavior, Campbell et al. (1993) also broke job performance down into eight major behavioral components: (a) job-specific task proficiency, (b) non–job-specific task proficiency, (c) written and oral communication task proficiency, (d) demonstrating effort, (e) maintaining personal discipline, (f ) facilitating peer and team performance, (g) supervision or leadership, and (h) management or administration. Job-

specific task proficiency reflects the “degree to which the individual can perform the core substantive or technical tasks that are central to the job” (p. 46). Non–job-specific task proficiency is the degree to which the individual can perform tasks in the workplace that are not specific to a particular job (i.e., teamwork skills). Written and oral communication task proficiency is the proficiency with which a job incumbent can write or speak. Demonstrating effort is a reflection of the consistency, frequency, and willingness of an individual to demonstrate effort. Maintaining personal discipline involves the extent to which negative behaviors (i.e., alcohol and substance abuse, excessive absenteeism) are avoided at work. Facilitating peer and team performance is the extent to which an individual supports their peers. Supervision or leadership is the degree to which an individual engages in behaviors directed at influencing the performance of subordinates. Last, management or administration includes performance behaviors directed at management tasks such as articulating goals or monitoring progress. One critical contribution of the Campbell et al. theory of performance is the conceptualization of performance as a multidimensional construct. By breaking job performance down into multiple components, they acknowledged that performance is not just one behavior that can be captured by one simple measure. Organizational citizenship behavior. One limitation of the Campbell et al. (1993) model of job performance is that it focuses solely on task performance as defined by the job description. It does not account for behaviors that are not technically part of the job yet contribute to job performance. In response to this gap in the literature, several new performance concepts were developed, such as citizenship behavior (e.g., Borman et al., 2001). In a review of organizational citizenship behavior (OCB), Podsakoff, MacKenzie, Paine, and Bachrach (2000) summarized the literature into seven core types of citizenship behaviors: (a) helping behaviors, (b) sportsmanship, (c) organizational loyalty, (d) organizational compliance, (e) individual initiative, (f) civic virtue, and (g) self-development. (See also Vol. 2, chap. 10, this handbook.) Helping behaviors include helping others with work-related problems as well as actively preventing 311

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

problems for others. Sportsmanship includes behaviors such as not complaining even when inconvenienced and generally maintaining a positive attitude in the face of difficulty. Organizational loyalty is composed of behaviors such as protecting, endorsing, and defending the organization and its objectives. Organizational compliance is another organizationally focused type of citizenship behavior that includes internalized and accepting the organization’s rules, regulations, and procedures even when not being directly observed. Individual initiative OCBs are behaviors in which the individual goes above and beyond minimum requirements in task-related situations. Civic virtue describes a high level commitment to the organization as a whole. This commitment is displayed through behaviors such as attending voluntary meetings and monitoring environmental changes that could impact the organization. Finally, selfdevelopment refers to voluntary behaviors aimed at improving one’s own knowledge, skills, and abilities. Overall, OCBs pose an interesting dilemma for PM in that by definition they are not explicit requirements of the job, and thus including them as part of a formal review or performance evaluation may be unethical. Simply stated, it may be inappropriate to evaluate an employee in terms of behaviors that are not explicitly stated as part of their job role, especially if the evaluation is then used as the basis for pay and promotion decisions. If OCBs are included in formal performance reviews, this in essence makes them part of the job description, and therefore they are no longer extrarole. If an organization chooses to include OCBs as part of their official job descriptions, then this approach is appropriate. However, the defining feature of OCBs is that they are performed without being required (i.e., extrarole behaviors; Organ, 1997); therefore, formally measuring them for evaluative purposes could potentially change the nature of these behaviors. In fact, there is an ongoing debate in the literature regarding whether measuring and evaluating OCBs changes the fundamental nature of the behaviors. Contextual performance. The concept of contextual performance is very similar to citizenship behavior in that they both describe on-the-job behavior that is not directly recognized as part of the 312

job (i.e., it is not a job requirement) yet still contributes to job effectiveness. Contextual performance is behavior that contributes to organizational effectiveness through its impact on the psychological, social, and organizational context (Motowidlo, 2003). Borman et al. (2001) presented a refined model of contextual performance that categorizes behaviors as personal support, organizational support, and conscientious initiative. Personal support includes behaviors such as helping others with tasks and showing courtesy and tact when interacting with others. Organizational support includes actions such as defending and promoting the organization. Conscientious initiative focuses on behaviors such as devoting extra effort to the job or taking advantage of opportunities for self-development. There is a noticeable amount of overlap between the conceptualizations of contextual behavior and OCB as described previously. The same issues regarding measurement of OCBs applies to measuring contextual behavior. Given that contextual performance is composed of behaviors that are not formally recognized as part of the job, it is unethical to evaluate individuals on the basis of contextual performance without their explicit knowledge. Accordingly, if contextual performance is required as part of a job, it is by definition no longer contextual. Adaptive performance. The Campbell et al. (1993) model of job performance also does not account for work behaviors that contribute to effectiveness in dynamic, complex, uncertain, and interdependent settings (Griffin, Neal, & Parker, 2007). Pulakos, Arad, Donovan, and Plamondon (2000) developed a theoretically and empirically based model of performance focused on the concept of adaptivity, with eight dimensions of adaptive performance. This model is intended to assess how well individuals adjust or adapt to new conditions or unexpected job requirements. The eight dimensions of adaptive performance are (a) handling emergencies or crisis situations; (b) handling work stress; (c) solving problems creatively; (d) dealing with uncertain and unpredictable work situations; (e) learning work tasks, technologies, and procedures; (f ) demonstrating interpersonal adaptability; (g) demonstrating cultural adaptability; and (h) demonstrating physically oriented adaptability.

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

The first dimension, handling emergencies or crisis situations, involves reacting appropriately in lifethreatening or dangerous situations. The dimension of handling work stress includes remaining calm when faced with difficulties and effectively managing frustration. Solving problems creatively refers to behaviors such as finding innovative ideas to complex problems and considering a wide range of possibilities when solving a problem. Dealing with uncertain and unpredictable work situations is similar to handling emergency and crisis situations and handling work stress in that it involved reacting appropriately to a cue, but this dimension focuses on changing plans, goals, and strategies in response to unexpected events or situations rather than just remaining calm. Learning work tasks, technologies, and procedures is the most task-relevant dimension of adaptive performance and includes keeping up to date with changing procedures and technology necessary for the job. The dimension of demonstrating interpersonal adaptability includes being openminded and considerate when dealing with other people and maintaining effective relationships. Demonstrating cultural adaptability specifically focuses on interacting with people from other cultures and adjusting behavior to make these interactions effective. Finally, demonstrating physically

oriented adaptability refers to adjusting to physical environmental conditions such as temperature or training to become more physically proficient. Integrated work role performance. Bringing together several of the previous understandings of work performance, Griffin et al. (2007) recently developed an integrated model of work role performance (see Table 10.1). They proposed that context plays a major role in the behaviors that will be viewed as valuable performance in an organization. Specifically, they proposed that uncertainty in the environment influences to what extent roles can be formalized and that interdependence with the environment influences how embedded work roles are in the larger system. Their model attempts to address the difficulty of capturing the total set of performance dimensions in a job by cross-classifying the three levels at which work behaviors can contribute to effectiveness (individual, team, and organization) with the three different forms of work behavior (proficiency, adaptivity, and proactivity). This crossclassification resulted in nine subdimensions of work role performance: (a) individual task proficiency, (b) individual task adaptivity, (c) individual task proactivity, (d) team member proficiency, (e) team member adaptivity, (f ) team member proactivity,

TABLE 10.1 Model of Positive Work Role Behaviors

Individual work role behaviors

Proficiency: Fulfills the prescribed or predictable requirements of the role

Adaptivity: Copes with, responds to, and supports change

Proactivity: Initiates change, is self-starting and future directed

Individual task behaviors: Behavior contributes to individual effectiveness

Individual task proficiency

Individual task adaptivity

Individual task proactivity

Team member behaviors: Behavior contributes to team effectiveness rather than individual effectiveness

Team member proficiency

Team member adaptivity

Team member proactivity

Organization member behaviors: Behavior contributes to organization effectiveness rather than individual or team effectiveness

Organization member proficiency

Organization member adaptivity

Organization member proactivity

Note. From “A New Model of Work Role Performance: Positive Behavior in Uncertain and Interdependent Contexts,” by M. A. Griffin, A. Neal, and S. K. Parker, 2007, Academy of Management Journal, 50, p. 330. Copyright 2007 by the Academy of Management. Adapted with permission. 313

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

(g) organization member proficiency, (h) organization member adaptivity, and (i) organization member proactivity. Individual task proficiency represents the formal task performance that contributes to individual effectiveness. Team member proficiency moves one level beyond individual task proficiency and includes taskrelated behaviors that an individual engages in that contribute to team effectiveness. This includes behaviors such as helping other team members or monitoring the work of other team members. Organization member proficiency describes task-related behaviors that contribute to organizational effectiveness, such as defending the organization’s reputation. This pattern of behavior remains consistent throughout the rest of the model. Specifically, individual task adaptivity, team member adaptivity, and organization member adaptivity all refer to behaviors such as appropriately responding to changes in the environment that contribute to individual, team, and organizational effectiveness. Similarity, individual task proactivity, team member proactivity, and organization member proactivity represent the three levels of self-starting, futuredirected behavior. This model of performance is more robust than previously mentioned models in that it integrates the broad concepts of role behavior, adaptive behavior, and proactive behavior and applies these concepts across multiple levels of analysis; however, it is still lacking in comprehensiveness (e.g., counterproductive work behaviors [CWBs]). Counterproductive work behavior. Thus far, all theories of individual performance have focused on the positive behaviors job incumbents can engage in. However, humans are also capable of negative, or dysfunctional, work behaviors, and research has labeled this CWB (Sackett, 2002). CWB refers to any type of intentional employee behavior that is contrary to the organization’s interests. CWBs include various deviant acts such as theft, destruction of property, drug abuse, and poor attendance. Some have argued that CWB is not a distinct construct but is rather a representation of the negative end of the citizenship behavior continuum. Recently, however, Sackett, Berry, Wiemann, and Laczo (2006) empirically supported that CWB is a separate and distinct construct from OCB. 314

Strategies for Measuring Individual Performance The most common approaches for measuring individual performance include performance appraisals, multiple-source ratings, objective measures, job knowledge tests, and work sample tests. It is important to note that these measurement strategies have not been used exclusively for capturing individual performance but are most often seen in the individual realm. Further detail is provided for each strategy in this section. Performance appraisals. Performance appraisals (PAs) are one of the most commonly used methods of PM in organizations. Traditionally, the term performance appraisal referred to a process involving a supervisor completing an annual report on an employee’s performance and discussing it with the employee in an interview (Fletcher, 2001). In a traditional PA system, the supervisor prepares a written evaluation of the employee on the basis of information gathered from coworkers, customers, and any pertinent documentation regarding the employee (Aldakhilallah & Parente, 2002). Then the supervisor schedules a meeting with the employee to review his or her job description, his or her performance against this description, and the organization’s goals. The supervisor also addressees the employee’s career progress and identifies opportunities for further development and improvement. This information is forwarded to higher management to be used in promotion and salary decisions. PA systems are often used to make promotional decisions based on past performance, to identify skill deficiencies and need for training, to make salary decisions, and to provide employees with feedback. As with most PM approaches, PAs have received both praise and criticism. Some voice the merits of PA systems, including feedback, goal setting, career management, objective assessment, and legal protection (Nickols, 2007). Also, some have claimed quite simply, “having a traditional PA system is better than having no system at all” (Aldakhilallah & Parente, 2002, p. 44), although this claim should be qualified to include only accurate and effective PA systems. It is likely that a bad PA system could actually be worse than no system in that it might force employees to

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

focus on the “wrong” behaviors—ones that are not really important to effective performance. Another advantage that has been mentioned previously relates to salary—that it is a motivational tool to improve employee performance, and therefore, salary decisions are often tied to the PA system (Rynes, Gerhart, & Parks, 2005). Finally, the outcomes of PAs provide the necessary justifications for decisions such as termination, promotion, transfer, or a change in salary. The more supported vein of thought, however, posits that supervisor-based PA systems are costly in terms of both time and money, and usually only result in negative emotions with very little demonstrable value (Nickols, 2007). Because the system is almost entirely controlled by the supervisor, there is a high chance that the employee’s appraisal will be based on the supervisors’ opinion alone. Depending on the ethical nature and personality of the supervisor in question, this could make PAs unfair or inaccurate. Data from an Internet survey revealed several other issues associated with PAs (Nickols, 2007). Often, PAs are accompanied by periods of reduced productivity; heightened negative emotional states such as anxiety, depression, and stress; and lowered morale and motivation. When they are linked to short-term rewards or consequences, they also tend to foster a focus on short-term goals at the sacrifice of long-term goals, which can result in negative consequences in the long run. Additionally, people are often hesitant to convey negative feedback, and therefore supervisors may engage in avoidance, delay, or distortion of PA feedback. Benedict and Levine (1988) found that, in particular, female raters may delay appraisals, delay scheduling feedback sessions, and more positively distort their ratings, especially when rating low performers. Because of these inherent issues, much of the research on PA has focused on making more objective and accurate ratings (Fletcher, 2001). Multiple-source ratings. Multiple-source ratings, also known as “360-degree feedback,” have been defined as “evaluations gathered about a target participant from two or more rating sources, including self, supervisor, peers, direct reports, internal customers, external customers, vendors, or suppliers” (Dalessio, 1998, p. 278). This PM method extends the PA concept by retaining subjective evaluations

as the main form of measurement, but this time including ratings from multiple sources, rather than from the supervisor alone. 360-degree feedback systems were originally developed for purely developmental purposes, with no intention of evaluative use. Very quickly, however, organizations began integrating this method into their PA systems, making it evaluative rather than only developmental. The use of 360-degree feedback at first seems like a logical choice for evaluative purposes (Waldman, Atwater, & Antonioni, 1998). This measurement system, unlike a traditional PA system, provides performance feedback from not only the supervisor but also from subordinates, peers and coworkers, clients (if applicable), and the self, reducing the chances for bias or unfair evaluations (Waldman et al., 1998). Specifically, by having peers rate the performance of other workers at their level, they are in a position to have a deep understanding of the job requirements and conditions, and therefore should provide more accurate ratings. They also generally have more opportunities to observe and monitor the work of the ratee because they work directly with them. Supervisors often are too far removed to have this level of understanding. On this basis, peer-report ratings seem to be a good alternative to self-report ratings because the tendency for inflating performance is reduced. Additionally, the traditional PA system flows only downward from supervisors to subordinates; supervisors and higher management do not receive any feedback or evaluation. Having feedback flowing in all directions allows for the development of management and leadership, and subordinates and customers are in a good position to evaluate managerial performance (Morgeson, Mumford, & Campion, 2005). 360-degree feedback is also intended to be given anonymously, which theoretically should result in more honest (and therefore accurate) feedback (Ghorpade, 2000). A review of the literature by Morgeson et al. (2005) delineated several other advantages of 360-degree feedback, such as an increase in information and formal feedback between employees, an increase in management learning, encouragement of goal setting and skill development, a change in corporate culture, and improved managerial effectiveness. 315

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

However, 360-degree feedback has limitations as a performance evaluation approach. Waldman et al. (1998) suggested that the bidirectional nature of the system could encourage employees to deliberately try to sabotage the system, with supervisors striking deals with their subordinates to give high ratings in exchange for high ratings. They also contended that this type of sabotage is much less likely if feedback is purely developmental, as there is no immediate or direct tangible outcome associated with good or bad evaluations. Also, peer-report ratings come with problems of their own. Specifically, Viswevaran et al. (2005) found that peer ratings had more halo error than supervisor ratings. Other issues arise regarding the improper implementation of 360-degree feedback systems. Often, organizations implementing the process do not clearly define the mission and scope of the program beforehand or do not provide clear rules for information sharing. Consequently, feedback can end up unrelated to actual employee performance, and employees may be unable to use the results to develop goals or plans (Ghorpade, 2000). Toegel and Conger (2003) argued for the separation of developmental and evaluative tools because of the contradictions and competing goals between using 360-degree feedback as developmental versus an evaluative technique. Objective measures. Another type of measure often used to assess individual performance is objective data. Specifically, indices such as absences, production rates, sales, or number of disciplinary cases can be used as a measure of an individual’s performance (Borman, 1991). Objective measures are appealing to many organizations because they are easy to gather and interpret, and are not as vulnerable to rater error or subjectivity as subjective measures. However, there are disadvantages to objective measures of individual effectiveness as well. To begin, the quality of an objective measure depends on what the stakeholder in question considers important. If the measurement system is designed to reduce absenteeism, then measuring absences is an appropriate choice. However, if the measurement system is intended to improve customer satisfaction, individual absences would be a very deficient measure. Another important factor to 316

consider with regard to objective measures is the notion of controllability. Controllability of measures can be conceptualized as the extent to which individuals (or teams) control the indicator of performance by varying the amount of effort they allocate to the measured tasks. Controllability—both the actual level and the perception of control—can impact motivation, performance, and ultimately organizational effectiveness. For example, if a measure of performance is bed utilization in a hospital, doctors and nurses have very little control over this, as it is largely determined by the nature of the illnesses that affect the patients. The only way to affect this measure is for patients to remain longer than needed to “use a bed.” This is clearly not an effective practice. Therefore, it is critical that objective measures of individual performance are used only when they appropriately represent the criteria of interest to the stakeholder and they are under the control of the individual being measured. Job knowledge tests. Job knowledge tests are usually written or computer-based tests that assess the extent of an individual’s knowledge regarding the content and procedures necessary for the job (Borman, 1991). Prior to developing a job knowledge test, a thorough job analysis should be conducted to gain a clear understanding of the knowledge, skills, and abilities necessary to perform the job. From this information, a set of items can be developed. As in any other written test, items can take many forms, such as multiple choice, true–false, or essay. Job knowledge tests are inherently best suited for positions that require high levels of declarative and procedural knowledge. It is important to note that job knowledge tests assess only the extent to which an individual can recall the appropriate information or procedure for a job but not his or her skill in applying that knowledge or performing that procedure. Job knowledge tests can therefore be appropriate performance measures when combined with other indicators (i.e., work-sample tests; see below) or for jobs that are highly dependent on declarative knowledge (e.g., tour guides). Work-sample tests. Work-sample tests are the practical, organizationally based equivalent of a laboratory-based measurement system (Cascio & Phillips, 1979). Because hands-on work-sample tests

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

require employees to engage in a simulated version of their normal taskwork specifically for measurement purposes, they are best suited for jobs with tasks that can be easily replicated in artificial settings. When broadly defined to include assessment centers, this type of measurement is suited for many different jobs. Example jobs that commonly use work-sample tests include baggage screeners, assembly line workers, aircraft pilots, or business executives. These jobs include tasks that can be easily simulated or tasks with behaviors that can be easily recorded and scored. Jobs that would not be as wellsuited for work-sample tests are more complex or ambiguously defined strategic planning or research positions, as these jobs require tasks that occur over a much longer period of time and tend to include aspects of performance that cannot be outwardly observed as scored during a short, simulated session.

Summary Several of the most commonly used measures of individual performance were described, including PAs, multiple-source ratings, objective measures, job knowledge tests, and work-sample tests (see Table 10.2 for a summary). These measures capture a wide range of performance behaviors and outcomes from a variety of perspectives. However, to reiterate,

the measures discussed do not represent an exhaustive list of the techniques for assessing individual performance. They represent only a sampling of the most commonly used and studies techniques. PERFORMANCE MEASUREMENT FROM THE TEAM PERSPECTIVE The science of teams has developed rapidly in recent years, and developments in the measurement of team performance have increased as well. Before we describe the measurement strategies most commonly used to capture team performance, we define team performance and provide a summary of the key theories regarding team performance. (See also chap. 19, this volume.)

Defining Team Performance “Teams do not behave, individuals do” (Zalesny, Salas, & Prince, 1995, p. 99). This assertion creates a challenge for teams researchers in that it assumes teams do not engage in measurable behaviors. However, it can be argued that although teams do not behave, the behavioral interactions between team members and the behavioral processes the team engages in as a whole can be measured at the team level or aggregated to the team level from the individ-

TABLE 10.2 Individual-Level Measurement Techniques Technique

Description

References

Performance appraisals

Traditionally, a measurement process involving a supervisor completing a written evaluation of an employee’s performance

Aldakhilallah and Parente (2002); Arvey and Murphy (1998); Nickols (2007)

Multiple-source ratings

Evaluations about a target participant collected from two or more source ratings such as self, supervisors, peers, customers; also referred to as “360-degree feedback”

Beehr et al. (2001); Ghorpade (2000); Morgeson et al. (2005); Toegel and Conger (2003)

Objective measures

Objective indices such as absences, production rates, sales, or number of disciplinary cases

Borman (1991); Ghalayini and Noble (1996); Pandey (2005); Paranjape et al. (2006)

Job knowledge tests

Written or computer-based tests that assess the extent of an individual’s knowledge regarding the content and procedures necessary for the job

Osborn and Campbell (1976)

Work-sample tests

Hands-on simulated versions of everyday taskwork used for measurement purposes

Cascio and Phillips (1979)

317

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

ual level. Various models of team performance have identified critical behaviors that facilitate team performance, such as communication, coordination, mutual performance monitoring, and backup behavior (e.g., Marks, Mathieu, & Zaccaro, 2001; Salas, Sims, & Burke, 2005). Volumes have been devoted to the development of behavioral, team-level measurement systems (e.g., Brannick, Salas, & Prince, 1997). When measuring team performance, it is critical to first start with a definition. Salas, Stagl, Burke, and Goodwin (2007) noted that team performance is often considered either a behavioral or cognitive act, and it is the resulting outputs that are considered performance outcomes. They suggested that team performance is multilevel, characterized by both taskwork (e.g., writing a paper) and teamwork (e.g., backup behavior) competencies exhibited by one or more team members, as well as team-level action (e.g., adaptation). This is not a new conceptualization. Early research has pointed to the multilevel nature of team performance as well (e.g., Kozlowski & Klein, 2000). Salas et al. (2007) postulated that team performance is a bottom-up emergent process, beginning with individuals and progressing toward teams. However, it is important not to ignore the top-down process inherent in team performance. Higher level (i.e., organizational) factors can directly impact, or have a moderating effect on, team performance. Additionally, it is important to acknowledge that process measures are distinct from, although related to, outcome measures. Process measures capture and “describe the strategies, steps, or procedures used to accomplish a task” (Smith-Jentsch, Cannon-Bowers, Tannenbaum, & Salas, 1998, p. 62). Outcome measures “evaluate the quantity or quality of the end result of those processes” (Smith-Jentsch et al., 1998, p. 62). Although outcome measures are ultimately what the researcher or organization is interested in predicting or managing, it is necessary to measure both process and outcome. This is due to the fact that outcome measures are influenced by much more than just individual or team performance. Outcomes can be affected by environmental factors, situational factors, or even luck. Cannon-Bowers and Salas (1997) suggested that team PM should capture both processes and outcomes at the individual and team levels. They posited 318

that outcome measures are not very diagnostic because they often do not indicate the underlying causes of that outcome. This is where process measures provide additional information by capturing the underlying behavioral mechanisms of performance. Therefore, it is critical to measure processes as well as outcomes for a robust and comprehensive understanding of team performance.

Theories of Team Performance Salas et al. (2007) conducted a thorough review of the literature to determine what is known regarding team performance. They found over 130 models of team performance or effectiveness that addressed at least three constructs believed to be relevant in the nomological network of team performance. The inclusion criteria prohibited inclusion of the thousands of additional “models” that only considered one or two constructs. They reviewed 11 of these models (e.g., Gersick, 1988; Hackman, 1987; Nieva, Fleishman, & Reick, 1978), noting that their selection in no way invalidated the significance of the remaining models. On the basis of this review, Salas et al. (2007) created an integrated model of team effectiveness. This model attempted to provide a comprehensive snapshot of the variables that impact team effectiveness. Below, we use a similar approach in summarizing the team performance literature. Space precludes us from thoroughly examining all potential models of team performance; therefore, we highlight just a selection of the plethora of team performance models. However, through the above discussion, we wish to illustrate that there is no universally accepted model of team performance to use as a departure point for measurement purposes. Team processes. Capturing performance at multiple levels is also a critical part of the framework presented by Cannon-Bowers and Salas (1997). Specifically, they believed measures of both individual- and team-level competencies are necessary for team PM. Teams are made up of individuals, and teamwork is composed of individual behaviors, so individual-level measurement is critical. However, individual-level measurement is not sufficient, as team performance is more than the

Performance Measurement at Work

Copyright American Psychological Association. Not for further distribution.

sum of its parts. There must also be team-level measurement to capture the processes and outcomes emerging from team interactions. Marks et al. (2001) recently delineated 10 core team processes that occur throughout the performance cycle of a team, such as mission analysis, coordination, and conflict management (see Table 10.3). Completing the circle, the outcomes of team performance are also multidimensional. Prior research has examined performance outcomes such as SME ratings of operational readiness in experienced military battalions (Lim & Klein, 2006), performance of

undergraduate students in a low-fidelity flight simulation (Mathieu, Heffner, Goodwin, Salas, & CannonBowers, 2000), and safety and efficiency in air-traffic controllers (Smith-Jentsch, Mathieu, & Kraiger, 2005). The definition, and measurement, of team performance varies widely across different populations, settings, and purposes. With so many interrelated inputs, processes, and outcomes, it is impossible for one measurement tool to capture every aspect of team performance simultaneously. Team performance is also a dynamic phenomenon, which means that static measurement systems

TABLE 10.3 Team Processes Process

Definition Transition processes

Mission analysis formulation and planning

Interpreting and evaluating the team’s mission, including identifying its main tasks as well as the operative environmental conditions and team resources available for mission execution

Goal specification

Identifying and prioritizing goals and subgoals for mission accomplishment

Strategy formulation

Developing alternative courses of action for mission accomplishment Action processes

Monitoring progress toward goals

Tracking task and progress toward mission accomplishment, interpreting system information in terms of what needs to be accomplished for goal attainment, and transmitting progress to team members

Systems monitoring

Tracking team resources and environmental conditions as they relate to mission accomplishment, which involves (a) internal systems monitoring (tracking team resources such as personnel, equipment, and other information that is generated and contained within the team) and (b) environmental monitoring (tracking the environmental conditions relevant to the team)

Team monitoring and backup behavior

Assisting team members to perform their tasks; assistance may occur by (a) providing a teammate verbal feedback or coaching, (b) helping a teammate behaviorally in carrying out actions, or (c) assuming and completing a task for a teammate

Coordination

The process of orchestrating the sequence and timing of interdependent actions Interpersonal processes

Conflict management

Preemptive conflict management involves establishing conditions to prevent, control, or guide team conflict before it occurs; reactive conflict management involves working through tasks and interpersonal disagreements among team members

Motivation and confidence building

Generating and preserving a sense of collective confidence, motivation, and task-based cohesion with regard to mission accomplishment

Affect management

Regulating member emotions during mission accomplishment, including (but not limited to) social cohesion, frustration, and excitement

Note. From “A Temporally Based Framework and Taxonomy of Team Processes,” by M. A. Marks, J. E. Mathieu, and S. J. Zaccaro, 2001, Academy of Management Review, 26, p. 363. Copyright 2001 by the Academy of Management. Adapted with permission. 319

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

capture only a cross-section rather than the entire performance episode as it progresses over time. Marks et al. (2001) specifically focused on this temporal aspect of team performance when developing the previously mentioned 10 core teamwork processes. They posited that different team processes are critical during different phases of task execution. Specifically, teams cycle through action and transition phases, and depending on which phase the team is currently in, the processes occurring will differ. Therefore, measuring team performance at one point in time may paint a completely different picture than measuring it at a different point in time or even at multiple points in time. Researchers have long been working to overcome the measurement challenges inherent in the complex, multidimensional, dynamic nature of team performance (e.g., Salas, Priest, & Burke, 2005). Guzzo and Dickson (1996) reviewed the literature on team performance and noted empirical support for several important team process and outcome variables. Research has found positive associations between cohesiveness and performance (Evans & Dion, 1991; Guzzo & Shea, 1992). For example, Evans and Dion (1991) found that 18% of variance in performance was accounted for by cohesion after correcting for measurement error. Campion, Medsker, and Higgs (1993) looked at several variables, including composition. They found a positive relationship between team size and effectiveness (r = .23, p < .05) and a composite measure of composition including heterogeneity, flexibility, size, and preference for group work (r = .21, p < .05). Although Campion et al. found no significant effect, or a negative one (r = −.05, ns), on heterogeneity and performance, Watson, Kumar, and Michaelsen (1993) found time to be influential, specifically that heterogeneous teams (operationalized by cultural diversity) who worked together long enough overcame the performance deficits relative to homogeneous teams. Others have found positive relationships between performance and (a) familiarity (e.g., Goodman & Leyden, 1991; finding increases from 1.8% to 11% in productivity), (b) leader expectations (e.g., Eden, 1990; ω2 = .19 and .17 for performance operationalized as theoretical specialty and practical specialty, respectively), (c) leader mood (e.g., George & Bettenhausen, 1990; 320

r = .43, p < .01, for prosocial behavior, which was significantly correlated with sales performance), (d) motivation as defined by team efficacy and group potency (e.g., Gully, Incalcaterra, Joshi, & Beaubien, 2002; ρ = .41 and .37, respectively), (e) self-efficacy (e.g., Earley, 1994; effort and self-efficacy accounted for a .66 change in R2 over demographic variables alone, with all variables together accounting for 69% of the variance in performance), (f) team goals (e.g., Weingart & Weldon, 1991; r = .45, p < .05) and (g) quality of feedback (e.g., Pritchard, Harrell, DiazGranados, & Guzman, 2008; r = .45, p < .01). The variability in these findings suggests the presence of moderating variables. Considering the negative effects that Campion et al. found, these moderating variables could change not only the magnitude of the relationships but also the direction. This is important to note for measurement. Because team performance is complex, if relevant impacting variables are not measured, it may lead to incorrect conclusions regarding team performance. Salas, Sims, and Burke (2005) took a different approach to team performance with the advancement of a theory called the “Big Five in Teamwork.” Their goal was to present a parsimonious framework, highlighting the essence of teamwork. They theorized that teamwork is composed of five core processes: (a) team leadership, (b) team orientation, (c) mutual performance monitoring, (d) backup behavior, and (e) adaptability. They noted the importance of three additional variables of interest: (a) shared mental models (SMMs), (b) closed-loop communication, and (c) mutual trust. This model incorporates not only core team skill-based competencies but also important affective competencies essential for effective team performance. Current research is focusing on the development of performance measures based on this model (e.g., Wiese et al., 2006). Team adaptability. Burke, Stagl, Salas, Pierce, and Kendall (2006) proposed that a critical skill of any team, and by extension, a critical skill to be measured, is the ability of a team to adapt. Team adaptation is defined as a change in team performance resulting from an identified cue or cue pattern, leading to effective and efficient outcomes for the entire team (Burke et al., 2006). Team adaptability is cru-

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

cial to organizations, as teams are called on to make decisions and solve problems in complex, dynamic environments with increasing frequency. Adaptation of team performance processes is of great importance to those working to understand effective team performance and how teams should alter in response to rapidly changing task demands. Burke et al. (2006) developed a cross-level mixed determinants model that integrates several perspectives of organizational theory. This multidisciplinary, multilevel, and multiphasic model is one of the most comprehensive conceptualizations of team adaptation available in the literature. The adaptive cycle comprises four process-oriented phases: (a) situation assessment, (b) plan formulation, (c) plan execution, and (d) team learning, as well as emergent cognitive states that function as both proximal outcomes and inputs throughout the cycle. Operating on the assumption that different team processes are critical at different phases of the team cycle (Marks et al., 2001), Burke et al. posited that plan formulation consists of transition processes and that plan execution consists of action processes, with interpersonal processes occurring in both. Team cognition. The study of individual mental models quickly made the leap to the team level, resulting in the concept of SMMs. SMMs are measured by capturing the amount of “sharedness,” or overlap, between a set of mental models, usually through an aggregation technique. Other team-level cognition constructs that can be measured as a dimension of team performance include transactive

memory systems, team situation awareness, and metacognition (e.g., J. R. Austin, 2003; Hinsz, 2004; Prince, Ellis, Brannick, & Salas, 2007; see Table 10.4). Kraiger and Wenzel (1997) suggested that SMMs can be measured on the basis of three key components: knowledge, behaviors, and attitudes, including perceptions, reactions, and structures. They provided numerous methodologies for measuring these three components of SMM, such as card sorts, structural assessments, and attitude perception surveys. They provided several hypotheses with regard to framework of antecedents, outcomes, and components of SMMs. Kraiger and Wenzel also noted another important issue with regard to measuring SMMs: measure weighting. They postulated that what measures to use and how to weight them are very situation specific yet cautioned that any selected measures should be sensitive to factors such as organizational culture.

Strategies for Measuring Team Performance There is no one universally accepted measure of team performance. Guzzo and Dickson (1996) defined team performance effectiveness on the basis of earlier work by Hackman (1987) and Sundstrom, De Meuse, and Futrell (1990). They suggested that team performance effectiveness is characterized by (a) team outputs (e.g., quantity or quality, customer satisfaction), (b) consequences for members, or (c) an increase in ability of a team to effectively perform at a later time. The approaches described below tend to fall under one or more of these three areas.

TABLE 10.4 Components of Team Cognition Component

Definition

Source

Team mental models

Team-level stable mental representations, including key knowledge about undertaking team tasks related to both teamwork and taskwork

Rico et al. (2008, p. 167)

Transactive memory system

A cognitive system teams use to encode, store, and retrieve information

Lewis (2003)

Metacognition

What group members know about the way groups process information

Hinsz (2004, p. 35)

Team situation awareness

The mental representation associated with a dynamic understanding of the current situation that is developed by team members moment by moment

Rico et al. (2008, p. 167)

321

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

Observational approaches. Observational methods of PM, at first glance, are enticing because they offer the objectivity of an outside observer. However, evidence has shown that techniques that use human observers to rate performance have very low interrater reliability (Fowlkes, Lane, Salas, Franz, & Oser, 1994). This could be because of several reasons, such as inadequate rater training, leniency effect, or halo effect. Because of this reliability issue, some observational methods have been developed that focus on capturing more objective data such as frequency or existence of specific performance dimensions rather than subjective ratings of behavior as positive or negative. Examples of both rating-based and checklist-based observational methods are described in the following sections. In both categories, measurement requires a third party observer to capture performance. Behavioral observation scales. The behavioral observation scale (BOS) approach to PM involves the use of observers who provide subjective ratings of the frequency of team performance. In this method, an observer physically watching a team uses a Likerttype scale to rate the amount of times that the team engages in a certain specified process. For example, an observer may be recording the frequency of information exchange in a team. Behaviors representative of information exchange would be rated on a scale from 1 to 5, representing intervals between none and always. Specifically, the item may ask “How often does this team share information about the environment with each other?” The observer would then rate the behavior as happening never, seldom, sometimes, frequently, or always on a 1-to-5 scale. The distinguishing characteristic of this measurement method is that it assesses the typical behavior of a team over time. However, this characteristic also makes ratings more susceptible to recency effects. Over time, observers may ignore earlier behavior in favor of the more recently observed behaviors. For example, a team may begin their task by sharing information 50 times an hour but may slow down to 20 times an hour as they become more familiar with the task and each other. Because the most recent behavior viewed by the observer was a subjectively “low” 20 utterances an hour, they may rate this team as “seldom” communicating, when on average they 322

were actually communicating 35 times an hour. Another problem with the BOS method is the subjectivity of the rating scales. Specifically, frequency is a very subjective concept. Some people may consider the same objective amount of communication (e.g., 50 utterances) as never communicating, whereas others may consider that as always communicating. Finally, there is a temporal aspect that bears discussion. Consider the information exchange example above. Perhaps in the beginning, teams should have been communicating 50 times, but as team members become more familiar with each other, they should only communicate 20 times. If the measurement scale is not adapted to reflect optimal levels of performance, teams may inadvertently receive low ratings (i.e., 20 times per hour as low), when in reality that amount of information exchange is quite appropriate for effective performance. Behaviorally anchored rating scales. The behaviorally anchored rating scales (BARS; Smith & Kendall, 1963) method is very similar to the BOS approach in that it requires an outside expert to observe, classify, and rate behavior (Kendall & Salas, 2004). Originally, this measurement technique was created to evaluate individuals; however, BARS has been readily adopted for use within the team performance arena. Just as in BOS methods, the observer rates the occurrence of team behavior on a numerical scale. However, rather than simply rating the frequency of behavior within the team, behavior is rated in regard to quality. There are specific examples of high-quality and low-quality behavior attached, or anchored, to each rating point in the scale. For example, the scale may range from 1 to 5, with 1 being poor behavior and 5 being excellent behavior. The behavioral anchors are usually generated by SMEs who provide performance episodes that represent both exceptional and unacceptable performance in the specified situation. These behavioral examples, in the form of short written descriptions, are intended to facilitate more accurate ratings by observers by making sure that similar behaviors are rated as the same number across raters. The point of this measurement technique, as used with teams, is to capture the relative frequency of a behavior as well as providing context as to where that frequency lies on a good to bad continuum. Essentially, just noting that a behavior

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

occurred 20 times does not provide evaluative information as to what level of quality that frequency represents. Therefore, BARS provide more usable information than BOS in most cases. There are some issues inherent in the BARS method, however. Because behavioral anchors focus on specific types of behaviors, observers tend to watch for those types of behaviors and rate performance on the basis of them regardless of the possibility that the team is using an equally successful, but different, behavior and regardless of the overall performance (Kendall & Salas, 2004). This issue actually relates back to the issue of criterion dimensionality, meaning several different approaches can be used to successfully complete a given task. Therefore, BARS are best suited for PM in work situations that requires very specific behavioral responses and in which criterion dimensionality is not a likely possibility. Event-based performance measurement. Eventbased measurement techniques, often used in training exercises, are “event-based” because they involve systematically scripting events into a relevant exercise, task, or scenario to trigger specific behavioral responses from the team in question (Fowlkes et al., 1994). In this measurement approach, similarly to BOS and BARS, specific behaviors are identified by SMEs as critical, and these behaviors are recorded as they are observed. It is the high level of control over the appearance of relevant behaviors that makes this a unique measurement technique (Salas, Burke, Fowlkes, & Priest, 2003). The intention of scripting scenarios and identifying behaviors a priori is to increase the level of reliability in measurement. Additionally, event-based measurement is distinct from BOS and BARS techniques because behaviors are not necessarily rated on the basis of quality. One specific method for event-based team PM is known as the targeted acceptable responses to generated events or tasks (TARGETs) methodology (Fowlkes et al., 1994). Targeted acceptable responses to generated events or tasks. In the TARGETs method of PM, a checklist of specific, observable behaviors is generated on the basis of the purpose of the observer and the task characteristics of the team being observed. For example, TARGETs for a helicopter aircrew may

include statements such as “Pilots question unsafe navigation procedure” or “Pilots acknowledge communications” (Fowlkes et al., 1994, p. 52). These are specific, easily observable behaviors that an observer could capture during team task execution. Therefore, a trained observer could simply watch the team in action and check off boxes as each behavior occurs. One important characteristic of the TARGETs method is that the events within the scenario being measured are controlled to elicit the behaviors of interest, because routine scenarios generally will not result in a wide enough range of observable behaviors. This makes the TARGETs methodology of PM, or any other event-based approach, much better suited for laboratory-based research than for field or practical use. Additionally, it may be noted that this particular technique (among others used within team PM) can be equally applied to individual PM. It is important to remember that team performance is frequently measured at the individual level and then aggregated to the team level. Therefore, it is important to use measures that fully capture individual performance yet can account for the complexities of team performance. Team dimensional training. Another form of event-based measurement, specifically designed for training teamwork skills, is team dimensional training (TDT; Smith-Jentsch et al., 1997). This training system focuses on measuring and improving four core teamwork behaviors: information exchange, initiative or leadership, supporting behavior, and communication. As in all event-based approaches, a training scenario is carefully designed to provide ample opportunities for the trainees to engage in the four teamwork behaviors. As the trainees go through the scenario, an instructor observes and records examples of strong and weak execution of the teamwork behaviors using a coding sheet. The coding sheet includes a column with the times when each scripted event should occur and a column where the instructor writes a description of how team performs in reaction to those events. Communication analysis. Another common method for examining teamwork performance has been through the analysis of communication transcripts (Salas et al., 2003). The analysis of communi323

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

cation data generally takes two forms: content analysis and flow analysis. Content analysis focuses on analyzing the linguistic content of the communication data, such as the topics of discussions, the frequency of specific words, or the frequency of questions asked. Typically in content analysis, communication utterances are categorized into groups that represent behavioral constructs such as information exchange or backup behavior, and the frequency of those utterances represent the amount of that behavior occurring. Flow analysis differs from content analysis in that it ignores the meaning or content of the communications and focuses only on the pattern of the team’s interactions. For example, flow analysis might investigate whether there is a consistent pattern in the types of communications occurring, such as responses occurring immediately after questions. This is a unique method for during-performance PM in that it requires no participation during the team task except for the presence of audio or video recording equipment. The actual evaluation of performance is performed post hoc. Archival measurement such as communication analysis is appealing because the source of the data is permanent, easy to access, and objective in nature. The problem is that measures capture performance well after it has occurred and therefore may draw an outdated picture. Additionally, analysis of communication transcripts requires the prior development of a coding or classification scheme to guide coders in interpreting communication data in the context of teamwork. As this technique requires outside raters to qualitatively assess behavior, it is critical that the raters are properly trained in the classification scheme. Classification schemes for communication analysis can take many forms. Communication transcripts could be coded for the frequency of communication, the content of communication, the pattern communication, or a combination thereof. For example, the number of request for information could be counted, as well as whether these requests for information were regarding the taskwork of others or the environment. These communication instances could be considered as representations of team and systems monitoring, which are two of the critical team processes identified by Marks et al. (2001). 324

Automated measurement. Automated measurement techniques are one of the more recently discussed methods in team performance literature. This is not a set of specific measurement tools but is rather a specific strategy for implementing various performance metrics. Specifically, automated computer systems can be used to continuously monitor team processes such as communication utterances or body movements (Kendall & Salas, 2004). The recorded behaviors are compared against an expert standard, and the system can provide automated feedback to the team. Automated measures are less obtrusive than other measures (Salas, Priest, & Burke, 2005), but they can be used to measure only overt behaviors and are often quite expensive to implement and maintain.

Summary The team-level measurement techniques discussed in this section included observational methods such as BOS and BARS, event-based methods such as TARGETS and TDT, as well as communication analysis and automated measurement (see Table 10.5). There are clearly some parallels between team-level and individual-level measurement in that BOS and BARS are both rating methods just as PAs or multiple-source feedback are. However, the team PM literature tends to be heavily focused on the use of measurement for feedback and training purposes, and places a heavy emphasis on the measurement of performance at multiple levels. As we move on to the organization PM literature, the emphasis shifts from multiple levels of analysis to multiple simultaneous dimensions of performance. This point is more fully developed throughout the remainder of this chapter. PERFORMANCE MEASUREMENT FROM THE ORGANIZATIONAL PERSPECTIVE The third level of performance measurement focuses on the processes and outcomes of the organization as a whole. In the following sections, we define organizational performance, summarize the key theories of organizational performance, and present the most common measurement strategies for capturing organizational-level phenomena.

Performance Measurement at Work

TABLE 10.5 Team-Level Measurement Techniques

Copyright American Psychological Association. Not for further distribution.

Technique

Description

Source

Behavioral observation scales

A subjective evaluation of team performance completed by an uninvolved third-party observer, usually based on a Likert-scale

Dominick et al. (1997); Kendall and Salas (2004); Taggar and Brown (2001)

Behaviorally anchored rating scales

A subjective evaluation of team performance completed by an uninvolved third-party observer that includes illustrative examples of good and bad behavior on which to anchor the ratings

Kendall and Salas (2004); Motowidlo and Borman (1977)

Targeted acceptable responses to generated events or tasks

A measurement system that involves scripting a scenario to elicit behaviors of interest and then completing a checklist of those behaviors

Dwyer et al. (1997); Fowlkes et al. (1994)

Team dimensional training

A form of event-based measurement that focuses on measuring and providing feedback regarding information exchange, initiative or leadership, supporting behavior, and communication

Smith-Jentsch et al. (1997, 2008)

Communication analysis

The analysis of communication data or transcripts for indicators of performance behaviors

Bowers et al. (1998); Dong (2005); Landauer et al. (1998); Muniz et al. (1996)

Automated measurement

Any measurement system that uses automated computer systems to continuously monitor and record team processes such as communication or body movements

Kendall and Salas (2004); Salas, Priest, and Burke (2005)

Defining Organizational Performance The larger system under which organizational teams are embedded provides the context for team performance (Guzzo & Dickson, 1996). Researchers have long argued that there has been a lack of emphasis on tying team performance to the overall organizational performance (Levine & Moreland, 1990). McGrath (1991) argued that teams are partially nested and loosely coupled to the broader organization. This refers to the fact that individuals are often part of more than one team and that teams can be part of more than one organization. Team performance, thus, impacts organizational performance. The literature pertaining to organizational PM is incredibly vast and ranges from theories and reviews of performance (e.g., Folan & Browne, 2005; Herman & Renz, 2008) to case studies and descriptions of specialized PM systems developed for individual companies (e.g., Bhasin, 2007; Khan & Wibisono, 2008). A comprehensive review of this literature is far beyond the scope of this chapter; thus we focus on several of the more commonly studied theories and strategies for organizational PM, as well as some of the newest approaches.

Traditionally, scholars of organizational performance have focused on measuring financial outcomes (Ghalayini & Noble, 1996). Financial outcomes such as return on investment, productivity, or sales per employee provide a concrete, easily accessible measure of overall organizational performance. Yet in terms of assessing the entire organizational performance domain, these measures may be deficient. Therefore, the focus on organizational-level PM has shifted from financial outcomes to more integrated measures of performance that include financial performance along with other dimensions of performance such as customer service and organizational learning. Some define organizational performance from a systems management approach, and therefore collect measures of both internal and external performance information (Jensen & Sage, 2000). Only very recently has organizational performance literature begun to conceptualize performance as a process and explore the measurement of processes in relation to strategic goals and company policy (Nenadal, 2008).

Theories of Organizational Performance The measure of organizational-level performance used depends on the model of organizational effectiveness 325

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

serving as the theoretical basis. Ahmed (1999) reviewed several of the most widely studied models of organizational effectiveness, including the goal model, system model, internal process model, human relations model, and the political approach. These models differ in their emphasis and prioritization of different dimensions of performance. The goal model (Georgopoulos & Tannenbaum, 1971) defines organizational effectiveness in terms of an organization’s achievements of its stated official goals. This is one of the earliest and most dominant models of effectiveness, given it directly attempts to align PM with the organizational strategy. The system model of effectiveness views organizations as open systems that work in a close relationship with the external environment. In this model, effectiveness is seen as the organization’s ability to survive and adapt in a changing environment. The internal process model conceptualizes effectiveness as the extent to which the internal business processes of an organization are smooth, orderly, continuous, predictable, and with minimal conflict. The human relations approach has a person-centered focus and maintains employee needs (i.e., employee satisfaction, self-efficacy) as the most important aspect of organizational effectiveness. The political approach is a very unusual model of organizational effectiveness, and it uses criteria such as “responsiveness, accountability, representativeness, and adherence to democratic values” (Ahmed, 1999, p. 544). Clearly, each of these theories focuses on very different aspects of organizational performance and would require very different methods of PM. Total quality management. From a business perspective, total quality management (TQM) represents a management philosophy of satisfying the customer’s requirements continually, at a low cost, by involving everyone’s daily commitment (Kanji, 1990). TQM defines quality as a process that must be managed. The objective of TQM is to train all levels of an organization to accept a new set of rules, methods, and lifestyles that focus on continuous improvement (Chung, Tien, Hsieh, & Tsai, 2008). There are four stages to the TQM process: (a) identifying and collecting information about the areas that need improvement in the organization, (b) making 326

sure that management understands and accepts the TQM philosophy, (c) identifying and resolving issues by involving all of management and supervision in a proper scheme of training and communication, and (d) starting new initiatives with new targets and spreading the improvement process to all aspects of the organization including supplier and customer links. TQM has a definite multilevel perspective in that it advocates a philosophy of continuous PM and improvement in all systems within a business. A recent study examined the value of TQM in actual organizations and found that 15 organizations that used TQM had above-average financial ability (Chung et al., 2008). Knowledge management. The knowledge management (KM) approach to PM budded out of the organizational management paradigm shift from focusing on managing physical goods to focusing on managing intangible assets (Nielsen, 2005). In this management philosophy, knowledge is viewed as a critical resource in an organization, and management of this resource can lead to competitive advantage. Of course, to manage knowledge in an organization, it must be measured. The literature dealing with KM can generally be grouped into content and process perspectives (Nielsen, 2005). The content perspective looks at the “what” of knowledge in organizations, specifically examining the categorization and transferability of different types of knowledge in organizations (Nielsen, 2005). Organizational knowledge can be categorized into explicit and tacit knowledge. Explicit knowledge is the knowledge that is exchanged using formal, systematic language and includes things such as explicit facts (Nielsen, 2005). Often, this knowledge can be found in instantiated formats such as books, databases, and computer programs (Small & Sage, 2005/2006). Tacit knowledge, on the other hand, is more intuitive, nonverbalized knowledge that is not directly articulated but still exists (Nielsen, 2005). This type of knowledge is hard to put into words and usually is rooted in contextual experiences (Small & Sage, 2005/2006). Other content-based KM research has looked at the holistic concept of organizational knowledge and the distinctions between data, information, and knowledge (Small & Sage, 2005/2006).

Performance Measurement at Work

Copyright American Psychological Association. Not for further distribution.

The process perspective of KM focuses on the “how” of organizational knowledge and therefore looks at the different stages of information processing in which knowledge is “embodied, embedded, embrained, encultured, and encoded” within members of the organization (Nielsen, 2005, p. 4). One goal of process perspective KM research is to uncover how the processes for accumulating and internalizing knowledge can be enhanced or improved (Nielsen, 2005).

Strategies for Measuring Organizational Performance The approaches used to measure organizational performance have changed over the years. Earlier measures included simpler outcomes, such as financial data and balanced scorecards (BSCs). More complex, integrated measures have emerged, such as Six Sigma and ProMES. Each of these approaches in described in more detail in this section. Financial data. As mentioned previously, traditional organizational PM literature focused on financial outcome measures. Financial measures are easy to obtain, easy to interpret, address the bottom line, and often are already recorded in organizational reports, thereby removing the need for a separate PM system. This makes them an enticing option for organizations looking to measure their overall levels of performance. Productivity, earlier defined as the ratio of effectiveness to the cost of achieving that level of effectiveness, has been one of the most widely used indicators of financial performance in organizations (Ghalayini & Noble, 1996). Although financial performance measures appear to be a simple, easy-to-use option for measuring organizational performance, the enticing simplicity of financial data brings with it many inherent problems. One of the problems attributed with financial measures such as return on investment is the potential for managers to engage in short-sighted decision making to maximize the short-term return, and consequently, sacrificing the long-term well-being of the firm (Pandey, 2005), a potential contributor to the current economic crisis. Along the same lines, some have argued that financial measures “tell the story of past events” and therefore are inadequate

for information age companies moving at a fast pace (Paranjape, Rossiter, & Pantano, 2006, p. 6). Additionally, they may not be the best choice of organizational performance indicators for nonprofit organizations because profit is not the primary focus of those groups. Ghalayini and Noble (1996) outlined several general limitations of traditional financial measures. One limitation is the missing link between measuring performance in financial terms and the improvement efforts needed to improve that financial performance. It is often difficult to diagnose the causes of performance problems at the individual or team level if all one has is organizational-level financial data. It is this problem of translation between the process of performance and its outcomes that makes financial performance measures less than ideal. Additionally, traditional financial measures do not take into account the strategy or goals of the organization as an entity, unless those goals are purely financial to begin with. For example, an organization may have high-level qualitative goals such as becoming well known in a particular target population or developing a reputation for timely and accurate service. The attainment of these goals cannot be measured using financial data. This is problematic, especially if the goals are central to the organization’s identity. This is where the measurement of individualand team-level behaviors, cognitions, and attitudes can improve a PM system. Balanced scorecard. One of the first PM tools created in response to the limitations of traditional financial measures is known as the BSC. The BSC is “a system of combining financial and nonfinancial measures of performance in one single scorecard” (Pandey, 2005, p. 51). BSC is not a strategy but is rather a management tool focusing on the financial and nonfinancial goals of an organization (Pandey, 2005). It was developed in response to the realization that quite often, nonfinancial goals are the drivers of business success, and therefore purely financial measures of performance are not sufficient alone. This integrated PM system complements financial measures with critical nonfinancial perspectives (Paranjape et al., 2006). A recent estimate 327

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

states that over 60% of organizations use a scorecard (Kaplan & Norton, 2005). BSC suggests the inclusion of measures of four perspectives—(a) financial, (b) customer, (c) internal business processes, and (d) learning and growth— although it does not have to be restricted to these four dimensions of performance (Pandey, 2005). All four of these perspectives are necessary to get an accurate picture of organizational performance. One of the unique aspects of this measurement approach is that BSC implies a complex causality between the four dimensions, with learning and growth influencing the internal business processes, and the processes have an impact on the financial outcomes, either directly or through the customer perspective (Dror, 2008). Financial measures such as growth, profit margin, return on investment, and so on, are backwardlooking indicators but are necessary to see whether process improvements are translating into financial success. Customer-focused measures include customer satisfaction, customer retention, market share, and customer profitability. For example, poor customer satisfaction is a leading indicator of future performance decline (Pandey, 2005). The internal process perspective looks at the quality of the business processes within the organization, and the key objectives are process improvement and suppliers’ relations. Finally, the learning and growth perspective focuses on innovation, creativity, and capability. Measures of learning and growth include employee satisfaction, employee retention, and employee productivity. In the BSC approach, each of the four perspectives includes (a) objectives to attain (e.g., higher customer satisfaction, higher return on investment), (b) measures of those objectives (e.g., financial data, customer survey data), (c) target values of those measures, and (d) initiatives needed to achieve those targets. The success of BSC depends on the clear identification of nonfinancial and financial variables, their accurate measurement, and linking performance to rewards and penalties (Pandey, 2005). As literature on BSC has grown, there have been two distinct and conflicting viewpoints emerging. One division of the literature advocates the successes of BSC. The main advantage of the BSC system is the ability to translate an organization’s vision and cen328

tral mission into tangible, achievable objectives and measures (Kanji, 2002; Pandey, 2005). BSC is unique from other measurement systems in that it contains both the outcome measures (i.e., financial measures) and the performance drivers of those outcomes (i.e., customer, innovation, internal process; Kanji, 2002). Kanji (2002) delineated several additional strengths of the BSC approach. While focusing on multiple dimensions of performance to get a more holistic view of organizational effectiveness, it simultaneously limits the number of measures so as to avoid information overload. It is also a relatively flexible PM approach and can be individualized for each organization. Finally, it places strong emphasis on customers and the market, which are often ignored by traditional measures. Alternatively, literature also criticizes the approach and calls attention to the lack of scientific evidence linking BSC to improved organizational performance (Paranjape et al., 2006). Other criticisms of BSC claim that it lacks a mechanism for building and maintaining relevance of measures once they are defined (Dror, 2008). In response, Kaplan and Norton (2004) introduced the concept of a strategy map, which provides such a mechanism for connecting strategic objectives with each other. Kanji (2002) described several other weaknesses of BSC. It was designed as a conceptual model and, therefore, is hard to convert into measurement. It is not a comprehensive system approach and leaves out important stakeholders in addition to the role of individual employees and suppliers. The effectiveness of the strategy map has yet to be empirically investigated. Six Sigma. Six Sigma is another performance management technique that has gained increasing popularity. Technically, the name Six Sigma can be translated into “3.4 defects per million opportunities” (Banuelas & Antony, 2003). This technique was developed at Motorola as a response to a large loss in productivity due to a lack of quality (Raisinghani, Ette, Pierce, Cannon, & Daripaly, 2005). Although the application of Six Sigma is spreading into increasingly different business sectors, historically, Six Sigma has been predominantly manufacturingbased (McAdam, Hazlett, & Henderson, 2005). This method of performance management “relies on

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

planned change, team-based collaboration, a focus on performance improvement, a systems perspective, and reliance on the scientific method and statistical methodologies” (Jeffery, 2005, p. 21). Six Sigma is a very process-focused method and therefore attempts to reduce errors and mistakes by experimenting with processes until more stable and robust processes are achieved. When Six Sigma was first developed, it was primarily seen as a statistical method for tracking and reducing variability in processes (McAdam et al., 2005). However, as the popularity and use of Six Sigma have increased, the approach has grown into a much more complex operations improvement and problem-solving methodology. One of the most interesting aspects of the Six Sigma approach is the change in organizational culture that is often associated with its use. Many organizations using Six Sigma view it not as a statistical method for reducing error, but as a complete business philosophy requiring high levels of organizational and management support. This buy-in is crucial to the success of the strategy. The Six Sigma method can be implemented using two different approaches—a continuous improvement design (reactive strategy) or a design–redesign approach known as Design for Six Sigma (DFSS; Banuelas & Antony, 2003). The continuous improvement design follows a sequence of five phases (a) define, (b) measure, (c) analyze, (d) improve, and (e) control (DMAIC). In the define phase, the problem to be addressed is identified and defined, and the critical stakeholders are identified. In the measure phase, the measurement capability of the organization is assured, current levels of performance are identified (most often using existing corporate performance measures), and goals for improvement are then set. In the analyze phase, the causes of problems and the key variables that may be linked to defects are identified. In the improve phase, experimentation is used to quantify the impact of variables on the process, and the process is subsequently improved. Last, in the control phase, continuous monitoring and adjustments are used to maintain the improved process (Goh & Xie, 2004). This improvement approach assumes the initial process is essentially correct and thus only needs to be adjusted for optimal performance (Banuelas & Antony, 2003).

Raisinghani et al. (2005) described five general steps in the DMAIC approach. These include (a) determining the appropriateness of the existing PM equipment through analysis, (b) looking for process deviations that require improvement, (c) using the “design of experiments” technique, (d) conducting a failure mode and effects analysis, and (e) taking one final measure of quality to ensure that Six Sigma levels have been achieved. DFSS follows a slightly different sequence of phases: (a) define, (b) measure, (c) analyze, (d) design, and (e) verify (DMADV). It differs only in that rather than improving an existing process, and controlling that change, a brand new process is designed and then verified. DFSS is a more proactive approach in that it involves designing processes capable of reaching Six Sigma levels, effectively preventing problems before they arise rather than after (Banuelas & Antony, 2003). This approach does not assume that the initial process is correct; instead, it aims to replace the process with a better, more effectively designed process. Therefore, both efficiency and effectiveness of the process can be improved using the DMADV approach. As with the implementation of any PM system, the context surrounding the Six Sigma approach is critical. McAdam et al. (2005) delineated a list of key steps in the Six Sigma processes on the basis of a review of the literature. The organization must gain senior manager support and involvement to create the desired deep level. A team of specialists must be trained to lead the process, requiring an initial investment of time devoted to training and developing a set of experts in the process. Much of the limited empirical evaluation of Six Sigma focuses on case studies (e.g., Antony, Kumar, & Madu, 2005). Productivity Measurement and Enhancement System. ProMES is an intervention designed to improve performance by actively measuring it and using those measures to provide detailed performance-based feedback (Pritchard, 1990; Pritchard et al., 2008; Pritchard, Holling, Lammers, & Clark, 2002). A typical ProMES intervention involves a design team (composed of people responsible for doing the work) who identifies the main objectives of the unit and develops quantitative 329

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

measures of these objectives. These measures are designed to be valid measures of all the important aspects of work required from the unit. The system is then reviewed and approved by higher management. On the basis of the measures, a regularly occurring feedback report is provided to the work units to guide desired improvements. The conceptual framework underlying ProMES is the NPI theory (Naylor, Pritchard, & Ilgen, 1980). NPI theory is a refinement of expectancy theory (Arvey, 1972; Campbell & Pritchard, 1976; Vroom, 1964) with a novel approach to motivation. Motivation within NPI is considered a process of allocating energy and time to various activities (Pritchard et al., 2002). The NPI motivational process involves acts (individual behavior), focusing on both direction (why certain acts are chosen over others) and amplitude (the amount of effort devoted to completion of the act). These acts lead to the generation of products (individual outputs). These products are then evaluated, which results in evaluations. These evaluations come from supervisors, peers, self, subordinates, family, and so on. Outcomes are tied to those evaluations and can either be intrinsic (i.e., increased satisfaction with one’s work) or extrinsic (i.e., pay raise, promotion). Outcomes are motivating because of their tie to needs satisfaction, which are very individual desires regarding everything from status to health and more. Acts, products, evaluations, outcomes, and needs satisfaction are combined into motivational force, which refers to an individual belief that changes in devoted time and energy (effort) toward different acts (tasks) will lead to changes in the amount of needs that are satisfied (Pritchard et al., 2002). A key characteristic of NPI is contingencies. Between each of the motivational components are component links (contingencies), signifying the interconnectivity among components (Pritchard et al., 2002). For example, actions generate products, thus the creation of products is contingent on the amount of effort directed toward acts or behaviors that directly relate to generating the product. Each contingency is based on individually perceived relationships. Again looking to the actions to products link, acts-to-products contingencies depict the relationship between the amount of effort or time devoted 330

toward an act and the expected amount of generated product. ProMES consists of seven steps (Pritchard, 1990; Pritchard et al., 2002). For a more detailed review, see Pritchard (1990). The first step is creating a design team of seven to eight people who are ultimately responsible for creating the new system. On the basis of the notion of participation in decision making, the team is composed of employees who do the work under evaluation. The second step focuses on the identification of overall unit objectives, developed through group discussion to consensus. The objectives should describe the purpose of the work group. All dimensions of the work must be reflected in the overall objectives. The third step involves identifying indicators or quantifiable measures of effectiveness to determine how well the objectives are being met. The objectives and indicators are presented to upper management for approval, and any disagreements are discussed to consensus. The fourth step moves toward defining contingencies. Contingencies are basically a function of the amount of the indicator as compared with its value to the organization. Indicators are charts, plotting effectiveness on the y-axis and the objective on the x-axis. The fifth step is to design the feedback system. Data are collected from each indicator, and indicator effectiveness scores are calculated. These individual scores are aggregated into an overall effectiveness score. A feedback report is then prepared with these data and is compared with historical data to determine improvements and well as areas that require further attention. The sixth step is the actual feedback session between the supervisor and employee, who discuss causes for improvements and any noted decreases, and devise a plan for continued improvement or changes. The seventh step involves monitoring the project overtime. If it is determined after several feedback sessions that some aspect of the measurement system needs to be changed, then the entire process should start again to come to agreement on how it should be fixed. If there are significant changes in the work or policy changes, then the system should also be reviewed. This system is highly effective. Not only does it lead to large productivity improvements in many different types of settings, these effects have also been

Performance Measurement at Work

Copyright American Psychological Association. Not for further distribution.

shown to last over time (Pritchard et al., 2008). The mean weighted effect size is 1.44, which is equivalent to the 93rd percentile of performance under baseline. Pritchard et al. (2008) also investigated a number of potential moderators. They found that the degree to which the study followed the ProMES methodology, the quality of the feedback given, whether changes were made to the new system, the degree of interdependence of the work group, and the degree of centralization of the organizations moderated the effectiveness of a ProMES intervention.

Summary Financial data, BSC, Six Sigma, and ProMES represent some of the most widely studied and discussed approaches for measuring organizational performance (see Table 10.6). As the science of organizational PM has grown, there has been a definite push toward more integrated, multidimensional PM systems. Financial data alone are no longer considered an appropriate gauge of organizational effectiveness. There has also been a push toward PM systems that are more diagnostic of the causes underlying performance. We contend that taking a multilevel approach to PM is one of the most effective ways to diagnose organizational performance. This approach is described in detail in the following section.

MULTILEVEL PERFORMANCE MEASUREMENT: THE LINKAGES BETWEEN LEVELS We contend that PM should serve three fundamental purposes: describing, evaluating, and diagnosing. The most basic purpose for PM is the accurate description of performance. Simply stated, PM is “the process of quantifying action” (Neely, Gregory, & Platts, 1995, p. 80). Often, what are referred to as PM systems in modern organizations are truly performance evaluation systems: that is, they are systems used to determine whether a particular measure of performance is satisfactory. However, the first basic step in PM is the accurate description of performance without any judgment. Once performance has been described, the next step is to evaluate that described action. As previously mentioned, PM can occur with or without a subsequent evaluation of that performance. However, the link between these two systems is a critical one if PM data are to serve any practical purpose. If the interested party cannot gauge whether the performance level captured by a PM system is “good” or “bad,” then they are unable to manage and improve that performance. However, the reverse is also true—if a well-designed evaluation system is applied to an inaccurate or unreliable measure, the product will

TABLE 10.6 Organizational-Level Measurement Techniques Technique

Description

Sources

Financial data

Outcomes measures focused on financial success such as return on investment or productivity

Ghalayini and Noble, (1996); Pandey (2005); Paranjape et al. (2006)

Balanced scorecard

A performance measurement system that combines financial and nonfinancial measures into a single scorecard for organizational effectiveness

Ahn (2005); Kanji (2002); Kaplan and Norton (1996); Norreklit (2000); Pandey (2005); Paranjape et al. (2006); Schwartz (2005)

Six Sigma

A performance measurement system that focuses on improving process design and reducing process error and waste

Antony (2006); Banuelas and Antony (2003); Goeke and Offodile (2005); Goh and Xie (2004); Raisinghani et al. (2005)

Productivity Measurement and Enhancement System

A performance intervention designed to improve performance by carefully measuring it and providing detailed feedback

Pritchard et al. (1989, 2002, 2008)

331

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

again be meaningless. Therefore, it is imperative that organizations have both sound PM strategies as well as sound evaluation strategies for interpreting performance data. The third step after describing and evaluating performance is to diagnose the causes of effective and ineffective performance. This step is perhaps the most important in the PM process because without an understanding of the causes underlying performance, the development of feedback and training becomes difficult and suboptimal, if not impossible. Identifying deficiencies in performance is not very useful unless one can also identify the underlying causal mechanisms to change to rectify those deficiencies. Diagnosing the underlying causes of effective and ineffective performance is the only way PM can be used to manage and improve performance. This concept of linking specific performance outcomes to specific causal mechanisms to achieve effective PM has been called to our attention by other organizational researchers (e.g., Ittner & Larcker, 2003). Therefore, we suggest that for any PM system to provide the most accurate description, evaluation, and diagnosis of performance, it should take an integrative, multilevel approach to measurement (see Figure 10.2). This position has been also been advocated quite strongly in the team PM literature (e.g., Salas et al., 2003). Individuals are the units that make up teams, and therefore team PM should include both team-level and individual-level indices. Following the same pattern, organizations are made up of teams, and teams are made up of individuals, so organizational PM should include all three levels of measurement if it is intended to capture the most beneficial and applicable information. The team PM literature focuses heavily on the distinctions between process and outcome, and how measurement of both is necessary if one is to accurately diagnose the underlying mechanisms influencing performance outcomes. Applying this concept to the organizational-, individual- and team-level processes are often the underlying causal mechanisms behind organizational outcomes. Therefore, for organizational PM to be diagnostic, it must capture and connect performance at all active levels of analysis. 332

FIGURE 10.2. Linkages between levels of performance.

Although the paradigm shift in organizational PM from financial measures to more integrated multidimensional measurement systems illustrates a much needed change, organizational performance measures still tend to focus heavily on tangible outcomes at a very high level. Therefore, there is still some level of disconnect between effective performance at the organizational level and the processes that enable that performance at the individual and team levels. This disconnect makes it difficult to translate organizational-level measures into actionable goals for improvement. For example, a corporation may measure profit per unit production to assess overall organizational-level performance and find they are making $5 in profit for each unit produced, and decide they would ultimately like an increase of $2 in profit. Yet, how does this translate into terms of improvement? One can not tell employees to work “$2 harder.” How much extra effort is necessary for two more dollars of profit? What changes in the production process are necessary? What factors influence level of profit? This is why organizational-level PM is important, but not sufficient, for improved organizational performance. To truly diagnose and improve organizational performance, multiple levels of performance must be captured to link overall financial outcomes to underlying individual- and team-level processes. Effective PM systems should focus on making con-

Copyright American Psychological Association. Not for further distribution.

Performance Measurement at Work

nections between the different levels of performance and should take advantage of the various techniques available from each perspective. This will allow the organization to gain an understanding of what individual-level behaviors and processes impact team processes and team outcomes, what team-level processes impact organizational outcomes, and so on. By doing this, changes in organizational-level financial performance can be stimulated by managing performance at the employee and team levels. Ultimately, a multilevel understanding of performance will enable the organization to pinpoint the causes behind performance problems and effectively address them. To illustrate the potential use of multilevel PM, a hypothetical example is described. Assume an organization is planning to implement a new PM system throughout all levels of its personnel. Perhaps the company employs factory workers that work individually to make products on an assembly line, as well as new product development teams that create new ideas for the products, among many other positions. For the sake of parsimony, we discuss only these two specific job positions. Now assume the organization chooses to use an automated measure of performance for the assembly line workers after carefully considering the purpose, content, timing, and setting desired. This automated system records the number of units produced, number of errors made, and several other measures of productivity. This choice of measurement allows for continuous on-the-job, real-time capturing of data. The organization then decides, after careful consideration, that the new product development will be individually rated using 360-degree feedback from themselves, their teammates, and the team leader on several dimensions including innovation, cooperation, and communication. The team will also be rated as a unit on the same dimensions. These subjective rating measures are better suited for measurement of highly creative and innovative jobs. Finally, organizational performance will be measured using overall customer satisfaction surveys as well as archival data reporting the annual profit of the company. Now, because the organization collected PM data from all three levels of analysis, the organization is capable of linking outcomes to the underlying causal

processes. Specifically, perhaps analysis of the data suggests that higher ratings on communication and innovation at the team level within the new product development teams are related to higher levels of customer satisfaction, whereas the number of units produced by assembly line workers is quite unrelated to customer satisfaction scores. This information would allow the organization to pinpoint the process that needs remediation (i.e., communication and innovation within teams) if an increase in customer satisfaction is their top priority. Similarly, they may find that the number of errors made on the assembly line is highly related to overall profit, and therefore an increase in profit would more likely occur because of a reduction in these errors. The main point of this example is that multilevel PM leads to a richer understanding of performance and the links between different processes and different outcomes, and therefore enables the organization to more effectively and efficiently manage their performance. The connections between different levels of analysis occur not only in a bottom-up direction but also from the top down. Because individuals and teams operate within the particular context of the organization, factors from the organizational level may impact both team and individual functioning (Salas et al., 2003). For example, a lack of resources provided by the organization or a set of organizational procedures may limit the ability of a team or individual to perform in a certain way. Thus it is critical to measure contextual variables at the organizational level to take those into consideration when trying to assess individual or team performance. EMERGING ISSUES FOR FUTURE RESEARCH IN PERFORMANCE MEASUREMENT An astounding amount of research has examined PM at the individual, team, and organizational levels. So much so that only a fraction of that research was touched on in this chapter. However, despite the abundance of research in the field of PM, there is always room for growth and further exploration. In particular, as the world changes and technologies advance, several areas have become paramount to the understanding of performance at any level: dis333

Wildman et al.

tribution, diversity, dysfunction, and culture. We posit that these concepts represent the future of PM research.

Copyright American Psychological Association. Not for further distribution.

Performance Measurement in Distributed Environments Globalization has impacted the way organizations work. Corporate offices are distributed throughout the world. Teams are distributed throughout various offices. Yet, another important phenomenon is also occurring—an increase in the number of workers who are working outside of the traditional office environment. A survey of senior-level executives worldwide indicated that two thirds of the global workforce was involved in distributed work arrangements (AT&T, 2004). In the United States, telecommuting, the most common form of distributed work, has increased exponentially over the past several years. In 1997, 11.6 million full-time employees worked remotely at least 1 day per month (WorldatWork, 2006). This number more than quadrupled to 45 million by 2006, the majority of whom are women, ages 35–44 (Gajendran & Harrison, 2007; Roitz, Allenby, & Atkyns, 2002). The increased prevalence creates a measurement issue—how does one measure the performance of employees when one never observes them actually “performing”? Many theories have put forth the notion that technology-mediated interaction leads to decrements (e.g., Media Richness Theory; Daft & Lengel, 1986). This is seemingly evidence that perhaps distributed performance should be measured differently than performance that is directly observed, to account for these issues. However, others speculate that good measurement is good measurement regardless of the setting and that, for example, what constitutes good PM for teams that are collocated is the same for PM for teams that are distributed— specifically, that similar principles should apply to the development of a system for collocated or distributed employees (Blackburn, Furst, & Rosen, 2003). Blackburn et al. (2003) suggested that the what, how, why and when of PM are closely related in either face-to-face or distributed environments. They outlined several characteristics that should apply to either situation (e.g., performance measures should always be tied to desired organizational goals, the 334

purpose of the measure should be clear). However, there are some noted differences. In face-to-face environments, performance ratings are based not only on the outcomes but also on the perceived effort on the part of the employee. However, this is not possible when measuring virtual performance and thus should be taken into consideration. Additionally, it is important to consider the multiple ways at which to arrive at the same outcome (criterion dimensionality). These issues and many more need to be systematically addressed by future research. Researchers need to understand the impact of distribution on performance ratings; the value added by distributed team or organizational members; and if performance should be measured differently, then how? Some initial work has been done in this area (e.g., Hofmann, Klar, Mohr, Quick, & Siegle, 1994); however, more research is needed, given the changing nature of technology and the increased prevalence of this type of work. The implications of this research extend beyond teleworkers to any employees who are working away from their home office (e.g., expatriates).

Performance Measurement in Diverse Settings Today’s workforce is more diverse than ever before. This poses a unique challenge for designing processoriented PM systems that evaluate fairly across demographic groups. Specifically, criterion dimensionality issues can play a big role in diverse settings. As mentioned previously, criterion dimensionality is the idea that two people on the same job may be equally effective but engage in very different behaviors to reach that level of effectiveness (Borman, 1991). This can become a problem if there are gender or ethnicity differences in the strategies used to complete a task and the measurement system assumes only one correct strategy. For example, men and women may differ systematically on leadership styles because of their differences (i.e., women tend to exhibit more democratic or participative styles whereas men tend to portray autocratic or directive styles; Eagly & Johnson, 1990). If a measure of performance assumes one style of leadership to be more effective than another without actually validating this effect, the measurement system could lead to unfair test bias and adverse impact issues. Therefore, it is critical that researchers

Performance Measurement at Work

examine issues of demographic difference in process and how this influences subsequent effectiveness. If findings indeed suggest that multiple strategies can lead to the same level of overall effectiveness, this dimensionality should be reflected in the measurement system.

Copyright American Psychological Association. Not for further distribution.

Performance Measurement and Dysfunction The study of performance has been pervasive in the industrial and organizational psychology literature for decades. Yet only recently has the field begun to consider unintended consequences of measuring performance. As noted by Grizzle (2002), “we expect that measuring efficiency leads to greater efficiency and measuring outcomes leads to better outcomes, but we don’t always get the results we expect” (p. 363). She noted several examples such as distorting financial data as an unintended consequence of measuring quarterly earnings per share. Many companies, such as Enron, resorted to an incomplete disclosure of financial data in an attempt to manipulate this measure. In terms of process measures, she noted that measuring quantity can lead to a diminished focus on quality. The measure may be indicating positive findings (i.e., high levels of quantity); however, the actual product may be of very poor quality. Others have noted the unintended consequences of measuring performance on the employees. For example, performance monitoring in call centers has been shown to cause employees distress, which impacts their well-being and satisfaction (Holman, Chissick, & Totterdell, 2002). More research is needed that focuses on these unintended consequences: What are they? Under what circumstances do they occur? How much and in what ways do they affect performance measures? Finally, what types of interventions alleviate the potentially harmful effects of these unintended consequences of PM?

Performance Measurement and Culture One emerging issue in PM research is the impact of cultural differences on PM. As the globalization of business increases, organizations are beginning to measure performance across national and cultural

boundaries like never before. Cultural background influences the qualities and traits people value, and therefore changes the way people rate performance dimensions. For example, Nonaka and Takeuchi (1995) argued that Westerners and Japanese view knowledge differently, which could influence how KM is viewed in multinational organizations. For example, Japanese view knowledge as being primarily tacit, whereas Western cultures tend to focus on explicit knowledge. Therefore, in a PM system that focuses on capturing levels of explicit knowledge, there may be cultural differences between Western and Eastern employees. These cultural differences in how people perceive performance dimensions could indicate that an employee being rated by a manager from one culture may receive a completely different rating from a manager of another culture. The challenge in this situation is either selecting dimensions of performance that are comparable across cultures or building a common understanding of the performance dimensions already being measured. Otherwise, ratings will not be equivalent. For example, Gillespie (2005) compared 360-degree feedback ratings across Great Britain, Hong Kong, Japan, and the United States and found that there were differences between countries in responses to the survey. Specifically, Gillespie found that respondents from different countries had different understandings of the constructs included on the survey and how these constructs related to each other. Therefore, they responded to the survey differently, meaning that the results of 360-degree feedback across cultures were not comparable. This calls into question the appropriateness of using 360feedback in multicultural contexts because the ratings given by members of different cultures may not be compatible for aggregation. This is a very important issue for multinational corporations to consider, and further research is needed to explore this issue and possible ways to account for these differences. CONCLUDING REMARKS The purpose of this chapter was to highlight PM issues as they relate to various levels (individuals, teams, and organizations). Given the scope of the literature on PM developed over the past several 335

Copyright American Psychological Association. Not for further distribution.

Wildman et al.

decades, we have presented a brief overview of this topic. By noting issues relevant to individuals, teams, and organizations as well as presenting several measurement strategies common to each level, we hope to synthesize existing literature on this topic in a systematic manner. Although the literature has taught researchers a great deal about PM, there is still much the field does not know. Future research suggestions are designed to spur further investigation in areas that are currently underdeveloped in terms of both scientific and practical understanding. It is clear that progress has been made with regard to improved measurement techniques, the consideration of multiple levels of performance, and a more concerted effort to consider conceptual criteria. Yet for a more broad understanding of PM, more research is still required.

References Ahmed, S. (1999). The emerging measure of effectiveness for human resource management: An exploratory study with performance appraisal. Journal of Management Development, 18, 543–556. Ahn, H. (2005). Insights from research: How to individualise your balanced scorecard. Measuring Business Excellence, 9, 5–12. Aldakhilallah, K. A., & Parente, D. H. (2002). Redesigning a square peg: Total quality management performance appraisals. Total Quality Management, 13, 39–51. Antony, J. (2006). Six Sigma for service processes. Business Process Management Journal, 12, 234–248. Antony, J., Kumar, M., & Madu, C. (2005). Six Sigma in small- and medium-sized UK manufacturing enterprises: Some empirical observations. International Journal of Quality & Reliability Management, 22, 860–874. Arvey, R. D. (1972). Task performance as a function of perceived effort performance and performance reward contingencies. Organizational Behavior & Human Performance, 8, 423–433. Arvey, R. D., & Murphy, K. R. (1998). Performance evaluations in work settings. Annual Review of Psychology, 49, 141–168. AT&T. (2004). The remote working revolution. Retrieved August 21, 2008, from http://www.corp.att.com/ emea/docs/remote_working_2004.pdf Austin, J. R. (2003). Transactive memory in organizational groups: The effects of content, consensus, specialization, and accuracy, on group performance. Journal of Applied Psychology, 88, 866–878. Austin, J. T., & Villanova, P. (1992). The criterion problem: 1917–1992. Journal of Applied Psychology, 77, 836–874. 336

Bailey, C. T. (1983). The measurement of job performance. Aldershot, England: Gower Press. Banuelas, R., & Antony, J. (2003). Going from Six Sigma to Design for Six Sigma: An exploratory study using analytic hierarchy process. The TQM Magazine, 15, 334–344. Bartram, D. (2005). The great eight competencies: A criterion-centric approach to validation. Journal of Applied Psychology, 90, 1185–1203. Beehr, T. A., Ivanitskaya, L., Hansen, C. P., Erofeev, D., & Gudanowski, D. M. (2001). Evaluation of 360-degree feedback ratings: Relationships with each other and with performance and selection predictors. Journal of Organizational Behavior, 22, 775–788. Benedict, M. E., & Levine, E. L. (1988). Delay and distortion: Tacit influences on performance appraisal effectiveness. Journal of Applied Psychology, 73, 507–514. Bhasin, S. (2007). Lean and performance measurement. Journal of Manufacturing Technology Management, 19, 670–684. Binning, J., & Barrett, G. Y (1989). Validity of personnel decisions: A review of the inferential and evidential bases. Journal of Applied Psychology, 74, 478–494. Bititci, U. S., Turner, T., & Begemann, C. (2000). Dynamics of performance measurement systems. International Journal of Operations & Production Management, 20, 692–704. Blackburn, R. S., Furst, S. A., & Rosen, B. (2003). Building a winning virtual team. In C. B. Gibson & S. G. Cohen (Eds.), Virtual teams that work: Creating the conditions for virtual team effectiveness (pp. 95–120). San Francisco: Jossey-Bass. Borman, W. C. (1991). Job behavior, performance and effectiveness. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2., pp. 271–326). Palo Alto, CA: Consulting Psychologists Press. Borman, W. C., Buck, D. E., Hanson, M. A., Motowidlo, S. J., Starks, S., & Drasgow, F. (2001). An examination of the comparative reliability, validity, and accuracy of performance ratings made using computerized adaptive rating scales. Journal of Applied Psychology, 86, 965–973. Bowers, C. A., Jentsch, F., Salas, E., & Braun, C. C. (1998) Analyzing communication sequences for team training needs assessment. Human Factors, 40, 672–679. Brannick, M. T., & Prince, C. (1997). An overview of team performance measurement. In M. T. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods, and applications (pp. 3–16). Mahwah, NJ: Erlbaum. Brannick, M. T., Salas, E., & Prince, C. (Eds.). (1997). Team performance assessment and measurement: Theory, methods, and applications. Mahwah, NJ: Erlbaum.

Performance Measurement at Work

Burke, C. S., Stagl, K. C., Salas, E., Pierce, L., & Kendall, D. L. (2006). Understanding team adaptation: A conceptual analysis and model. Journal of Applied Psychology, 91, 1189–1207. Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. (1993). A theory of performance. In N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations (pp. 35–70). San Francisco, CA: Jossey-Bass.

Copyright American Psychological Association. Not for further distribution.

Campbell, J. P., & Pritchard, R. D. (1976). Motivation theory in industrial and organizational psychology. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 63–130). Chicago: Rand McNally. Campion, M. A., Medsker, G. J., & Higgs, A. C. (1993). Relations between work group characteristics and effectiveness: Implications for designing effective work groups. Personnel Psychology, 46, 823–850. Cannon-Bowers, J. A., & Salas E. (1997). A framework for developing team performance measures in training. In M. T. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods, and applications (pp. 45–62). Mahwah, NJ: Erlbaum. Cascio, W. F., & Phillips, N. F. (1979). Performance testing: A rose among thorns? Personnel Psychology, 32, 751–766. Chung, Y. C., Tien, S. W., Hsieh, C. H., & Tsai, C. H. (2008). A study of the business value of total quality management. Total Quality Management & Business Excellence, 19, 367–379 Czarnowsky, M. (2008, September). Executive development: An ASTD research report. Training and Development, 62, 44–45. Daft, R. L., & Lengel, R. H. (1986). Organizational information requirements, media richness and structural design. Management Science, 32, 554–571. Dalessio, A. T. (1998). Using multisource feedback for employee development and personnel decisions. In J. W. Smither (Ed.), Performance appraisal: State-ofthe-art in practice (pp. 278–330). San Francisco: Jossey-Bass. Dominick, P. G., Reilly, R. R., & McGourty, J. W. (1997). The effects of peer feedback on team member behavior. Group & Organization Management, 22, 508–520. Dong, A. (2005). The latent semantic approach to studying design team communication. Design Studies, 26, 445–461. Dror, S. (2008). The balanced scorecard versus quality award models as strategic frameworks. Total Quality Management & Business Excellence, 19, 583–593. Dubois, C. L. Z., Sackett, P. R., Zedeck, S., & Fogli, L. (1993). Further exploration of typical and maximum performance criteria: Definitional issues, prediction,

and White–Black differences. Journal of Applied Psychology, 78, 205–211. Dwyer, D. J., Fowlkes, J. E., Oser, R. L., Salas, E., & Lane, N. E. (1997). Team performance measurement in distributed environments: The TARGETs methodology. In M. T. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods, and applications (pp. 137–153). Mahwah, NJ: Erlbaum. Eagly, A. H., & Johnson, B. T. (1990). Gender and leadership style: A meta-analysis. Psychological Bulletin, 108, 233–256. Earley, P. (1994). Self or group? Cultural effects of training on self-efficacy and performance. Administrative Science Quarterly, 39, 89–117. Eden, D. (1990). Pygmalion without interpersonal contrast effects: Whole groups gain from raising manager expectations. Journal of Applied Psychology, 75, 394–398. Evans, C. R., & Dion, K. L. (1991). Group cohesion and group performance: A meta-analysis. Small Group Research, 22, 175–186. Fisicaro, S. A. (1988). A re-examination of the relation between halo error and accuracy. Journal of Applied Psychology, 73, 239–244. Flanagan, J. C. (1956). The evaluation of methods in applied psychology and the problem of criteria. Occupational Psychology, 30, 1–9. Fletcher, C. (2001). Performance appraisal and management: The developing research agenda. Journal of Occupational and Organizational Psychology, 74, 473–487. Folan, P., & Browne, J. (2005). A review of performance measurement: Towards performance management. Computers in Industry, 56, 663–680. Fowlkes, J. E., Lane, N. E., Salas, E., Franz, T., & Oser, R. (1994). Improving the Measurement of Team Performance: The TARGETs Methodology. Military Psychology, 6, 47–61. Gajendran, R. S., & Harrison, D. A. (2007). The good, the bad, and the unknown about telecommuting: Metaanalysis of psychological mediators and individual consequences. Journal of Applied Psychology, 92, 1524–1541. George, J. M., & Bettenhausen, K. (1990). Understanding prosocial behavior, sales performance, and turnover: A group-level analysis in a service context. Journal of Applied Psychology, 75, 698–709. Georgopoulos, B. S., & Tannenbaum, A. S. (1971). A study of organisational effectiveness. In J. Ghorpade (Ed.), Assessment of organizational effectiveness: Issues, analysis, and readings (pp. 177–188). Pacific Palisades, CA: Goodyear. 337

Wildman et al.

Gershoni, H., & Rudy, N. (1981). An analysis of the total variation of work measurement techniques. International Journal of Production Research, 19, 303–316.

Herman, R. D., & Renz, D. O. (2008). Advancing nonprofit organizational effectiveness research and theory: Nine theses. Nonprofit Management and Leadership, 18, 399–415.

Gersick, C. G. (1988). Time and transition in work teams: Toward a new model of group development. Academy of Management Journal, 31, 9–41.

Hinsz, V. B. (2004). Metacognition and mental models in groups: An illustration with metamemory of group recognition memory. In E. Salas & S. M. Fiore (Eds.), Team cognition: Understanding the factors that drive process and performance (pp. 33–58). Washington, DC: American Psychological Association.

Ghalayini, A. M., & Noble, J. S. (1996). The changing basis of performance measurement. International Journal of Operations & Production Management, 16, 63–80.

Copyright American Psychological Association. Not for further distribution.

Ghorpade, J. (2000). Managing five paradoxes of 360degree feedback. Academy of Management Executive, 14, 140–150.

Hofmann, R., Klar, R., Mohr, B., Quick, A., & Siegle, M. (1994). Distributed performance monitoring: Methods, tools, and applications. Parallel and Distributed Systems, 5, 585–598.

Gillespie, T. L. (2005). Internationalizing 360-degree feedback: Are subordinate ratings comparable? Journal of Business and Psychology, 19, 361–380.

Holman, D., Chissick, C., & Totterdell, P. (2002). The effects of performance monitoring on emotional labor and well-being in call centers. Motivation and Emotion, 26, 57–81.

Goeke, R. J., & Offodile, O. F. (2005). Forecasting management philosophy life cycles: A comparative study of Six Sigma and TQM. The Quality Management Journal, 12, 34–46.

Ittner, C. D., & Larcker, D. F. (2003, November). Coming up short on nonfinancial performance measurement. Harvard Business Review, 88–95.

Goh, T. N., & Xie, M. (2004). Improving on the Six Sigma paradigm. The TQM Magazine, 16, 235–240. Goodman, P. S., & Leyden, D. P. (1991). Familiarity and group productivity. Journal of Applied Psychology, 76, 578–586. Griffin, M. A., Neal, A., & Parker, S. K. (2007). A new model of work role performance: Positive behavior in uncertain and interdependent contexts. Academy of Management Journal, 50, 327–347 Grizzle, G. A. (2002). Performance measurement and dysfunction: The dark side of quantifying work. Public Performance and Management Review, 25, 363–369. Gully, S., Incalcaterra, K., Joshi, A., & Beaubien, J. (2002). A meta-analysis of team efficacy, potency, and performance: Interdependence and level of analysis as moderators of observed relationships. Journal of Applied Psychology, 87, 819–832. Guzzo, R. A., & Dickson, M. W. (1996). Teams in organizations: Recent research on performance and effectiveness. Annual Review of Psychology, 47, 307–338.

Jeffery, A. B. (2005). Integrating organization development and Six Sigma: Six Sigma as a process improvement intervention in action research. Organizational Development Journal, 23(4), 20–31. Jensen, A. J., & Sage, A. P. (2000). A system management approach for improvement of organizational performance measurement systems. Information, Knowledge, Systems Management, 2, 33–61. Kanji, G. K. (1990). Total quality management: The second industrial revolution. Total Quality Management, 1, 3–12. Kanji, G. K. (2002). Performance measurement system. Total Quality Management, 13, 715–728. Kaplan, R. S., & Norton, D. P. (1996, January–February). Using the balanced scorecard as a strategic management system. Harvard Business Review, 150–161. Kaplan, R. S., & Norton, D. P. (2004). The strategy map: Guide to aligning intangible assets. Strategy & Leadership, 32, 10–17. Kaplan, R. S., & Norton, D. P. (2005). The office of strategy management. Strategic Finance, 87, 8–60.

Guzzo, R. A., & Shea, G. P. (1992). Group performance and intergroup relations in organizations. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 3, 2nd ed., pp. 269–313). Palo Alto, CA: Consulting Psychologists Press.

Kendall, D. L., & Salas, E. (2004). Measuring team performance: Review of current methods and consideration of future needs. In J. W. Ness, V. Tepe, & D. Ritzer (Eds.), The science and simulation of human performance (Vol. 5, pp. 307–326). New York: Elsevier.

Hackman, J. R. (1987). The design of work teams. In J. Lorsch (Ed.), Handbook of organizational behavior (315–342). Englewood Cliffs, NJ: Prentice-Hall.

Khan, M. K., & Wibisono, D. (2008). A hybrid knowledge-based performance measurement system. Business Process Management, 14, 129–146.

Hays, R. T., & Singer, M. J. (1989). Simulation fidelity in training system design: Bridging the gap between reality and training. New York: Springer-Verlag.

Kozlowski, S. W. J., & Klein, K. (2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In

338

Performance Measurement at Work

K. J. Klein (Ed.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 3–90). San Francisco: Jossey-Bass. Kraiger, K., & Wenzel, L. H. (1997). Conceptual development and empirical evaluation of measures of shared mental models as indicators of team effectiveness. In M. T. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods, and applications (pp. 63–84). Mahwah, NJ: Erlbaum.

Copyright American Psychological Association. Not for further distribution.

Landauer, T. K., Foltz, P. W., & Darrel, L. (1998). An introduction to latent semantic analysis. Discourse Processes, 25, 259–284. Latham, G. P., Wexley, K. N., & Pursell, E. D. (1975). Training managers to minimize rating errors in the observation of behavior. Journal of Applied Psychology, 60, 550–555. Levine, J. M., & Moreland, R. L. (1990). Progress in small group research. Annual Review of Psychology, 41, 585–634. Lewis, K. (2003). Measuring transactive memory systems in the field: Scale development and validation. Journal of Applied Psychology, 88, 587–604. Lim, B. C., & Klein, K. J. (2006). Team mental models and team performance: A field study of the effects of team mental model similarity and accuracy. Journal of Organizational Behavior, 27, 403–418. Lim, B., & Ployhart, R. E. (2004). Transformational leadership: Relations to the five-factor model and team performance. Journal of Applied Psychology, 89, 610–621. Mangos, P. M., & Arnold, R. D. (2008). Enhancing military training through the application of maximum and typical performance measurement principles. Performance Improvement, 47, 29–35. Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based framework and taxonomy of team processes. Academy of Management Review, 26, 356–376. Mathieu, J. E., Heffner, T. S., Goodwin, G. F., Salas, E., & Cannon-Bowers, J. A. (2000). The influence of shared mental models on team processes and performance. Journal of Applied Psychology, 85, 273–283. McAdam, R., Hazlett, S.-A., & Henderson, J. (2005). A critical review of Six Sigma: Exploring the dichotomies. International Journal of Organizational Analysis, 13, 151–174. McGrath, J. E. (1991). Time, interaction, and performance (TIP): A theory of groups. Small Group Research, 22, 147–174. Morgeson, F. P., Mumford, T. V., & Campion, M. A. (2005). Coming full circle: Using research and practice to address 27 questions about 360-degree feedback programs. Consulting Psychology Journal: Practice and Research, 57, 196–209.

Motowidlo, S. J. (2003). Job performance. In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.), Handbook of psychology: Vol. 12. Industrial and organizational psychology (pp. 39–53). New York: Wiley. Motowidlo, S. J., & Borman, W. C. (1977). Behavioral anchoring scales for measuring morale in military units. Journal of Applied Psychology, 62, 177–183. Muchinsky, P. M. (2009). Psychology applied to work: An introduction to industrial and organizational psychology. Summerfield, NC: Hypergraphic Press. Muniz, E., Stout, R. J., & Salas, E. (1996, March). Communications as an indicator of team situation awareness. Paper presented at the 42nd Annual Meeting of the Southeastern Psychological Association, Norfolk, VA. Nagle, B. (1953). Criterion development. Personnel Psychology, 6, 211–289. Naylor, J. C., Pritchard, R. D., & Ilgen, D. R. (1980). A theory of behavior in organizations. New York: Academic Press. Nenadal, J. (2008). Process performance measurement in manufacturing organizations. International Journal of Productivity and Performance Management, 57, 460–467. Nickols, F. (2007). Performance appraisal: Weighed and found wanting in the balance. The Journal for Quality and Participation, 30, 13–16. Nielsen, B. B. (2005). Strategic knowledge management research: Tracing the co-evolution of strategic management and knowledge management perspectives. Competitiveness Review, 15, 1–13. Nieva, V., Fleishman, E. A., & Reick, A. (1978). Team dimensions: Their identity, their measurement, and their relationships (Contract No. DAHC19-78-C-0001). Washington DC: Response Analysis Corp. Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company. New York: Oxford University Press. Norreklit, H. (2000). The balance on the balanced scorecard: A critical analysis of some of its assumptions. Management Accounting Research, 11, 65–88. Organ, D. W. (1997). Organizational citizenship behavior: It’s construct clean-up time. Human Performance, 10, 85–97. Osborn, W. C., & Campbell, R. C. (1976, October). Developing skills qualifications tests. Paper presented at the annual convention of the Military Testing Association, Gulf State Park, AL. Pandey, I. M. (2005). Balanced scorecard: Myth and reality. The Journal for Decision Makers, 30, 51–66. Paranjape, B., Rossiter, M., & Pantano, V. (2006). Insights from the balanced scorecard performance measurement systems: Successes, failures, and future— A review. Measuring Business Excellence, 10, 4–14. 339

Wildman et al.

Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach, D. G. (2000). Organizational citizenship behaviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26, 513–563. Prince, C., Ellis, E., Brannick, M. T., & Salas, E. (2007). Measurement of team situation awareness in low experience level aviators. The International Journal of Aviation Psychology, 17, 41–57.

Copyright American Psychological Association. Not for further distribution.

Pritchard, R. D. (1990). Enhancing work motivation through productivity measurement and feedback. In U. Kleinbeck, H. Quast, H. Thierry, & H. Hacker (Eds.), Work motivation: Past, present, and future (pp. 119–132). Hillsdale, NJ: Erlbaum. Pritchard, R. D., Harrell, M. M., DiazGranados, D., & Guzman, M. J. (2008). The productivity measurement and enhancement system: A meta-analysis. Journal of Applied Psychology, 93, 540–567. Pritchard, R. D., Holling, H., Lammers, F., & Clark, B. D. (Eds.). (2002). Improving organizational performance with the productivity measurement and enhancement system: An international collaboration. Huntington, NY: NOVA Science Publishers. Pritchard, R. D., Jones, S. D., Roth, P. L., Stuebing, K. K., & Ekeberg, S. E. (1989). The evaluation of an integrated approach to measuring organizational productivity. Personnel Psychology, 42, 69–115. Pulakos, E. D., Arad, S., Donovan, M. A., & Plamondon, K. E. (2000). Adaptability in the workplace: Development of a taxonomy of adaptive performance. Journal of Applied Psychology, 85, 612–624. Pun, K. F., & White, A. S. (2005). A performance measurement paradigm for integrating strategy formulation: A review of systems and frameworks. International Journal of Management Reviews, 7, 49–71. Raisinghani, M. S., Ette, H., Pierce, R., Cannon, G., & Daripaly, P. (2005). Six Sigma: Concepts, tools, and applications. Industrial Management & Data Systems, 105, 491–505. Rico, R., Sánchez-Manzanares, M., Gil, F., & Gibson, C. (2008). Team implicit coordination processes: A team knowledge-based approach. The Academy of Management Review, 33, 163–184. Roitz, J., Allenby, B., & Atkyns, R. (2002). 2001/2002 employee survey results—Telework, business benefit, and the decentralized enterprise. Retrieved June 4, 2003, from http://www.att.com/telework/ article_library/survey_results_2002.html Rynes, S. L., Gerhart, B., & Parks, L. (2005). Personnel psychology: Performance evaluation and pay for performance. Annual Review of Psychology, 56, 571–600. Sackett, P. R. (2002). The structure of counterproductive work behaviors: Dimensionality and relationships 340

with facets of job performance. International Journal of Selection and Assessment, 10, 5–11. Sackett, P. R., Berry, C. M., Wiemann, S. A., & Laczo, R. M. (2006). Citizenship and counterproductive behavior: Clarifying relations between the two domains. Human Performance, 19, 441–464. Sackett, P. R., Zedeck, S., & Fogli, L. (1988). Relations between measures of typical and maximum job performance. Journal of Applied Psychology, 73, 482–486. Salas, E., Burke, C. S., & Fowlkes, J. E. (2006). Measuring team performance “in the wild”: Challenges and tips. In W. Bennett, C. Lance, & D. Woehr (Eds.), Performance measurement: Current perspectives and future challenges (pp. 245–272). Mahwah, NJ: Erlbaum. Salas, E., Burke, C. S., Fowlkes, J. E., & Priest, H. A. (2003). On measuring teamwork skills. In J. C. Thomas & M. Hersen (Eds.), Comprehensive handbook of psychological assessment (pp. 427–442). Indianapolis, IN: Wiley. Salas, E., Priest, H. A., & Burke, C. S. (2005). Teamwork and team performance measurement. In J. R. Wilson & N. Corlett (Eds.), Evaluation of human work (3rd ed., pp. 793–808). New York: Taylor & Francis. Salas, E., Sims, D. E., & Burke, C. S. (2005). Is there “big five” in teamwork? Small Group Research, 36, 555–599. Salas, E., Stagl, K. C., Burke, C. S., & Goodwin, G. F. (2007). Fostering team effectiveness in organizations: Toward an integrative theoretical framework of team performance. In J. W. Shuart, W. Spaulding, & J. Poland (Eds.), Nebraska Symposium on Motivation: Vol. 51. Modeling complex systems: Motivation, cognition and social processes (pp. 185–243) Lincoln: University of Nebraska Press. Scott, W. D. (1917). A fourth method of checking results in vocational selection. Journal of Applied Psychology, 1, 61–67. Schmitt, N. W., & Klimoski, R. J. (1991). Research methods in human resources management. Cincinnati, OH: Southwestern Publishing. Schwartz, J. (2005). The balanced scorecard versus total quality management: Which is better for your organization? Military Medicine, 170, 855–858. Sink, D. S., & Tuttle, T. C. (1989). Planning and measurement in your organization of the future. Norcross, GA: Industrial Engineering and Management Press. Small, C. T., & Sage, A. P. (2005/2006). Knowledge management and knowledge sharing: A review. Information Knowledge Systems Management, 5, 153–169. Smith, P. C. (1976). Behavior, results, and organizational effectiveness: The problem of criteria. In M. D. Dunnette (Ed.), Handbook of industrial and

Performance Measurement at Work

organizational psychology (pp. 745–775). Chicago: Rand McNally. Smith, P. C., & Kendall, L. M. (1963). Retranslation of expectations: An approach to the construction of unambiguous anchors for rating scales. Journal of Applied Psychology, 47, 149–155. Smith-Jentsch, K. A., Cannon-Bowers, J. A., Tannenbaum, S. I., & Salas, E. (2008). Guided team self-correction: Impacts on team mental models, processes, and effectiveness. Small Group Research, 39, 303–327.

Copyright American Psychological Association. Not for further distribution.

Smith-Jentsch, K. A., Mathieu, J. E., & Kraiger, K. (2005). Investigating linear and interactive effects of shared mental models on safety and efficiency in a field setting. Journal of Applied Psychology, 90, 523–535. Sundstrom, E., De Meuse, K. P., & Futrell, D. (1990). Work teams: Applications and effectiveness. American Psychologist, 45, 120–133. Tannenbaum, S. I. (2006). Applied performance measurement: Practical issues and challenges. In W. Bennett, C. E. Lance, & D. J. Woehr (Eds.), Performance measurement: Current perspectives and future challenges (pp. 297–318). Mahwah, NJ: Erlbaum. Toegel, G., & Conger, J. A. (2003). 360-degree assessment: Time for reinvention. Academy of Management Learning and Education, 2, 297–311. Toops, H. A. (1944). The criterion. Educational and Psychological Measurement, 4, 271–297. Tsui, A. S., & Barry, B. (1986). Interpersonal affect and rating errors. Academy of Management Journal, 29, 586–599. Viswesvaran, C., Schmidt, F. L., & Ones, D. S. (2005). Is there a general factor in ratings of job performance?

A meta-analytic framework for disentangling substantive and error influences. Journal of Applied Psychology, 90, 108–113. Vroom, V. H. (1964). Work and motivation. New York: Wiley. Waldman, D. A., Atwater, L. E., & Antonioni, D. (1998). Has 360-degree feedback gone amok? Academy of Management Executive, 12, 86–94. Watson, W. E., Kumar, K., & Michaelsen, L. K. (1993). Cultural diversity’s impact on interaction process and performance: Comparing homogeneous and diverse task groups. Academy of Management Journal, 36, 590–602. Weingart, L., & Weldon, E. (1991). Processes that mediate the relationship between a group goal and group member performance. Human Performance, 4(1), 33–54. Wiese, E. E., Merket, D., Stacy, W., Nelson-Walwanis, M., Freeman, J., & Aten, T. (2006, December). Enhancing distributed debriefs with performance measurement objects. Proceedings of the 2006 Interservice/Industry Training, Simulation, and Education Conference (pp. 1–11). Orlando, FL. WorldatWork. (2006). Telework trendlines for 2006. Retrieved August 21, 2008, from http://www.workingfromanywhere.org/news/Trendlines_2006.pdf Zalesny, M. D., Salas, E., & Prince, C. (1995). Conceptual and measurement issues in coordination: Implications for team behavior and performance. In G. R. Ferris (Ed.), Research in personnel and human resources management (Vol. 13, pp. 81–115). Greenwich, CT: JAI Press.

341

CHAPTER 11

STRATEGIC REWARD AND COMPENSATION PLANS

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

In recent years, the role of compensation plans has been expanding beyond some of the traditional conceptions that have historically driven psychological research. Questions such as “How many dimensions of pay satisfaction emerge from factor analysis?” and “Does satisfaction with pay influence job performance?” have been the focus. Increasingly, the term compensation has been used to convey a much broader and different set of activities than it once did. For instance, researchers are asking questions such as “Are employers offering employee benefits that fit well with their employees’ needs and preferences (to avoid dysfunctional turnover)?” and “Does the choice of compensation plan combined with particular training programs influence organizational financial performance?” Although compensation practice and research continue to follow traditions set forth by the disciplines that study these areas (i.e., primarily economics, psychology, sociology), researchers and practitioners alike have cast traditional rewards and compensation systems in a role that may facilitate or hinder organizational success. Following Cascio and Aguinis’s (2008) content analysis of industrial and organizational psychology research, practitioners largely make recommendations about the management or development of individuals in organizations, or they advise those who do. Such research is termed strategic rewards and compensation systems. The purpose of this chapter is threefold. First, a synthesis is presented of the changing landscape of compensation practice (placed into the context of the evolution of the human resource management

[HRM] function in organizations) and the signals that indicate a far more strategic role for compensation systems than it has ever previously enjoyed. Second, the psychological research on compensation is briefly characterized. Third, a selective, critical analysis is provided of the extant literature on strategic rewards and compensation that argues for and illustrates an expanded perspective for approaching the study of strategic compensation. Much of what is discussed in this chapter has been described using other terms (e.g., base pay, pay-for-performance, short-term incentives, long-term incentives, employee benefits). Past research has illuminated our understanding of the role of compensation at the individual level of analysis. (For an extensive review of important published research, see Gerhart & Rynes, 2003; Rynes & Gerhart, 2000.) For our purposes here, research conducted at the individual level of analysis is referred to as compensation and/or rewards and compensation. Research that argues for compensation having a role in influencing organizational outcomes such as market share, stock price, and so forth, is referred to as strategic rewards and compensation plans. COMPONENTS OF REWARD AND COMPENSATION SYSTEMS Compensation represents both the intangible and tangible rewards or returns employees receive for performing their jobs (Martocchio, 2009). Intangible compensation includes recognition, status, employment security, challenging work, and unplanned

http://dx.doi.org/10.1037/12169-011 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

343

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

learning opportunities. Tangible rewards include pay and employee benefits. Both intangible and tangible rewards have been the subject of strategic human resource management (SHRM) research (Toh, Morgeson, & Campion, 2008). Compensation includes both monetary and nonmonetary rewards. Monetary compensation programs reward employees according to their job performance levels or for learning job-related knowledge or skills (i.e., competencies). Monetary compensation comes in a variety of forms, including base pay and incentives. Base pay is adjusted based on seniority, merit pay increases (i.e., past job performance), and competency-based pay (i.e., learning and demonstrating competence in knowledge and skill sets required by the employer). Nonmonetary compensation includes a variety of employee benefits. Three fundamental roles characterize employee benefits programs: (a) protection, (b) paid time off, and (c) accommodation and enhancement. Protection programs provide family benefits, promote health, and guard against income loss caused by catastrophic factors such as unemployment, disability, or serious illnesses. Paid time-off policies compensate employees for a limited duration and specified purpose when they are not performing their primary work duties. Accommodation and enhancement benefits promote opportunities for employees and their families. A wide variety of programs exist, including day care facilities, stress management classes, flexible work schedules, and tuition reimbursement. Employee benefits derive from two broad sources: those required by law and those offered at the discretion of employers. Laws such as the Social Security Act of 1935 mandate a variety of programs designed to provide income to retired workers, monetary benefits to the beneficiaries of deceased workers, and medical protection for older Americans. Organizations may offer additional benefits on a discretionary basis. This source distinction is an important one. Generally speaking, legally required benefits are a constant across most organizations; employers have no control over the choice of whether to offer these benefits or how to design them. Discretionary benefits naturally provide employers with virtually total control over the choice to 344

offer or design health insurance and retirement plans within the limits of available resources and government regulations such as the Employee Retirement Income Security Act of 1974 and the Pension Protection Act of 2006. For example, the Employee Retirement Income Security Act, coordinated with the Internal Revenue Code (2000; i.e., the code for all taxation in the United States), provides employers with substantial tax savings when they design defined benefit pension plans that meet a variety of criteria. Two examples include vesting, the nonforfeitable right of employees to receive employer contributions to their retirement plan, and also the employer meeting minimum funding rules to ensure that there will be sufficient money to provide retired employees with lifelong pensions. Offering discretionary benefits can lead to competitive advantage by attracting and retaining the best-qualified individuals, particularly when employers offer a choice of benefits that fit the needs and preferences of the workforce, such as day care benefits for employees with young children. PUTTING “STRATEGIC” IN REWARD AND COMPENSATION PLANS Strategic rewards and compensation address compensation from a strategic management perspective. Strategy has been defined as “the determination of the basic long-term goals and objectives of an enterprise, and the adoption of courses of action and the allocation of resources necessary for carrying out these goals” (Chandler, 1962, p. 13). Ultimately, the strategic management perspective focuses on promoting the viability of companies in competitive market places for as long as potential market opportunities are evident. For example, the U.S. Postal Service is experiencing deficit spending in large part because its services (i.e., delivering written correspondence) are less relevant to the younger generation, who communicate through alternative media, including electronic social networks, e-mail, blogs, and instant messaging. In addition, increasingly, organizations advertise their products and services on Internet sites, relying less on direct mail marketing. Presently, compensation practice is one among many HRM practices that influence employment status. Compensation practices may be designed to

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

promote cooperation among team members, such as with the use of team incentive pay plans. Besides compensation, other HRM practices that influence the employment relationship include recruitment, selection, structuring the flow of work, performance appraisal, training, career development, employee separations, and, more recently, involvement in labor–management relations (Kochan, Katz, & McKersie, 1994). Within most companies, three kinds of HRM decisions take place that correspond to three areas of organizational functioning: strategic decisions, tactical decisions, and operational decisions. At the highest levels of organizational hierarchies, HRM executives make strategic decisions that are conveyed within the organization’s competitive strategy. A human resource strategy specifies the use of particular HRM practices in a manner that promotes the attainment of competitive strategy: compensation strategy and training strategy. The time horizon for strategic decisions is broadest, and it may span in excess of 3 years. Beneath the strategic level, somewhere in the middle range of a company’s hierarchy, are tactical decisions. HRM professionals make tactical decisions that specify policy or practice that supports the attainment of a company’s strategy. Developing compensation programs, recruitment plans, and methods to reduce the turnover of employees whose job performance is exemplary are just a few examples of HRM practices, the design of which reflect tactical decisions. The time horizon for tactical decisions is suited to meet the imperatives of competitive strategy. Finally, at the lowest levels, HRM professionals make operational decisions that consist of administering compensation programs, implementing performance appraisal procedures, and developing monitoring systems for phenomena such as employee absenteeism and turnover. Operational decisions occur daily. Today, not every organization practices HRM at each of these three levels. Some companies practice HRM only at the operational level, usually relying on previously relied-on methods and practices. In fact, the activities of the HRM field in its early stages were characteristically operational. Over the course of the previous century, the HRM function evolved from what is known as manpower planning and

personnel administration to a field that is increasingly becoming entrenched in tactical and strategic decisions (Kaufman, 2008). ON BECOMING STRATEGIC REWARDS AND COMPENSATION SYSTEMS A chronological review of the developments in compensation research and practice from the early 20th century through to the publication in this chapter in the year 2010 is essential for an understanding of the necessity of rewards and compensation systems as strategic partners. Six categories were selected to organize this review to highlight key developments. The five categories include (a) industrial revolution, (b) human relations, (c) World War II, (d) civil rights and women’s rights movements, (e) strategic human resource period, and (f) global financial crisis, which began in 2007.

Industrial Revolution Family agricultural farms and small family craft businesses were the bases for the U.S. economy before the 1900s. The turn of the 20th century marked the beginning of the industrial revolution in the United States. During the industrial revolution, the economy’s transition from agrarian and craft businesses to large-scale manufacturing or factory systems began. Increasingly, individuals were becoming employees of large factories instead of self-employed farmers or small business owners. Factory owners, also referred to as capitalists, sought profits. That is, the term profits refers to the money that capitalists enjoy after deducting the costs of doing business, such as employee wages. Highly efficient workforces were therefore an essential part to the capitalist’s profit motive. This shift from the agricultural sector to the industrial sector promoted the beginnings of the field of human resource management (Baron, Dobbin, & Jennings, 1986; Kaufman, 2008). The profit motive and the sheer size of factory employment gave rise to divisions of labor on the basis of differences in worker skill, effort, and responsibilities. The growth in the size of the workplace necessitated practices to guide such activities as hiring, training, setting wages, handling grievances, 345

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

and terminating employment for work rule violations or poor job performance. Factory owners sought out the expertise of mechanical engineers to promote efficient production systems and productive workers. The work of these engineers defined the scientific management movement. The goals of scientific management practices were to contribute to labor cost reductions by replacing inefficient production methods with efficient production methods. Factory owners used time-and-motion studies and job analysis to meet the labor cost reduction objective. The scientific management approach was based on a mechanistic view of employees, and it was assumed that employees were primarily motivated by economic reasons. The time-and-motion studies analyzed the time it took employees to complete their jobs. These studies literally focused on employees’ movements and the identification of the most efficient steps to complete jobs in the least amount of time (Person, 1929). At the time, employers used job analysis to classify the most efficient ways to perform jobs. Scientific management practices gave rise to individual incentive pay systems and the use of welfare practices. The most noteworthy incentive plans to emerge during this era were piecework plans. The term piecework plans refers to types of incentive pay plans that have several general features. That is, an employee’s compensation depends on the number of units she or he produces over a given period. Specifically, these plans reward employees on the basis of their individual hourly production measured against an objective output standard, which is determined by the pace at which manufacturing equipment operates. For each hour, workers receive piecework incentives for every item produced over the designated production standard. The early personnel (and compensation) function emphasized labor cost control and management control over labor. Many employers instituted so-called scientific management practices to control labor costs, as well as welfare practices to maintain control over labor. Welfare practices represent the forerunner of modern discretionary employee benefits practices.

Human Relations Movement The human relations movement of the 1920s and 1930s produced a great deal of academic literature 346

and prescriptive practices about effective worker management. The goal of this movement was to identify causes of and solutions for worker dissatisfaction, absenteeism, excessive turnover, and conflicts between workers and management over working conditions. Perhaps the best-known psychological research during this time was the Hawthorne studies, conducted between 1924 and 1927 (Mayo, 1933). The findings from these studies provided important insights into the role of tangible compensation in worker motivation. The Hawthorne studies were named after the location at which the research took place: the Western Electric Company Hawthorne Works. Company management and university researchers investigated the effects of hours, wages, rest periods, lighting conditions, and degree of supervision on worker productivity. The investigators hypothesized that differences in these factors would significantly influence worker productivity. The findings indicated that these environmental variables accounted for some, but not all, of the differences in worker effectiveness. Interpretations of the results of the Hawthorne studies have led to various conclusions about their implications for industrial and organizational psychology. Conclusions from a reanalysis of the original data from the Hawthorne studies seem reasonable: The experiments drew attention to small group processes, and the studies’ conclusions led to widespread acceptance of human relations as a primary factor in worker performance. Following dissemination of the findings, previously attempted and conceptually simpler mechanisms such as those of scientific management tended to be given less emphasis as determinants of work performance. These variables include the possible benefits of fatigue reduction, use of economic incentives, the exercise of discipline, and other aspects of managerial control. But it is precisely such factors to which we are directed by empirical analyses of the Hawthorne data. (Franke & Kaul, 1978, p. 638)

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

Harvard Business School professor Elton Mayo expressed an interest in the findings of the Hawthorne studies, which prompted him to conduct a series of additional experiments. These experiments were based on his strong beliefs that management does not hold the physical or social welfare of workers in high regard. In turn, he maintained that management’s lack of concern for employees leads to worker dissatisfaction and counterproductive behaviors, such as absenteeism. Thus, he embarked on a program of research to examine these influences. Work done during the human relations movement fundamentally challenged the value of scientific management’s sole focus on worker efficiency and gave way to the importance of workers’ psychological states. Four general conclusions were drawn from the human relations movement (Encyclopedia Britannica, 2007). First, the physical potential and mental potential of workers are useful predictors of job performance but not necessarily the only useful predictors. That is, social factors strongly influence job performance as well. Second, creating formal organizational charts does help organize workflow, but this kind of organization does not account for the social dynamics between individuals within the workplace. Third, work group norms have an important effect on worker productivity. Fourth, the workplace as a social system is made up of interdependent parts, and politics is a prevalent social dynamic in organizations, influencing why and how employees interact with each other. Ultimately, the human relations movement demonstrated that social and psychological factors increased worker productivity more than did changes in wages and work hours. These studies also challenged long-standing assumptions of economists that only wages and work hours mattered. The insights from this period provided support to the positive role intangible compensation could play on employee job performance. Arguably, this shift opened the door to adopting nonmonetary recognition programs and to the importance of intangible rewards and compensation.

World War II The federal government’s imposition of wage freezes during World War II gave rise to many present-day

discretionary benefits. They were referred to as welfare practices. Welfare practices were generous endeavors undertaken by some employers, motivated in part to minimize employees’ desire for union representation, to promote good management, and to enhance worker productivity. Welfare practices were “anything for the comfort and improvement, intellectual or social, of the employees, over and above wages paid, which is not a necessity of the industry nor required by law” (U.S. Bureau of Labor Statistics, 1919, p. 37). Companies’ welfare practices varied. For example, some employers offered facilities such as libraries and recreational areas; others offered financial assistance for education, home purchases, and home improvements. In addition, employer sponsorship of medical insurance coverage became common. The use of welfare practices created the need to administer them. Welfare secretaries served as an intermediary between the company and its employees, and they were essentially predecessors of HRM professionals. Employers withdrew costly offerings after the government ended the wage freeze. The withdrawal of these benefits created discontent among employees: Many viewed employer-sponsored benefits as an entitlement. For instance, employees strongly reacted to the withdrawal of medical insurance. Legal battles followed based on the claims of employees that health protection was a fundamental right. Health insurance benefits subsequently became a mandatory subject of collective bargaining in union settings.

Civil Rights and Women’s Rights Eras The 1960s and 1970s marked substantial strides toward more fair treatment of minorities and women in the workplace through the passing of various laws that promoted employment-related decisions according to qualifications to perform a job, or actual job performance, by prohibiting decisions about employment status on the basis of personal characteristics such as race or sex (see Vol. 2, chap. 15, this handbook). In the process, federal laws led to the bureaucratization of compensation practice. Personnel department administrators took the lead in developing and implementing employment practices that upheld the myriad federal employment laws. The personnel department also 347

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

maintained written records, creating documentation in the event of legal challenges to employment practices and, for most employers with federal contracts, affirmative action plans. In short, compensation professionals were largely administrators who managed within the strict confines of government regulation. In the 1960s, the U.S. government instituted major legislation aimed at protecting individual rights to fair treatment in the workplace. Most often, fair treatment means making employment-related decisions according to job performance; for example, awarding higher merit pay increases to the better performers. The Equal Pay Act of 1963, Title VII of the Civil Rights Act of 1964, and the Pregnancy Discrimination Act of 1974 are among the greatest achievements toward promoting minority and women’s rights in the workplace. The Equal Pay Act of 1963 is an amendment to the minimum wage provisions of the Fair Labor Standards Act of 1938. Congress enacted the Equal Pay Act to remedy a serious problem of employment discrimination in private industry: The wage structure of “many segments of American industry has been based on an ancient but outmoded belief that a man, because of his role in society, should be paid more than a woman even though his duties are the same” (S. Rep. No. 176, 1963). The Equal Pay Act is based on a simple principle: Men and women should receive equal pay for performing equal work. The definition of wages in the Equal Pay Act of 1963 also encompasses employee benefits. The Equal Employment Opportunity Commission (EEOC) defined wages to include all payments made to, or on behalf of, an employee as compensation for employment. Thus, employers must provide equal employee benefits to male and female employees who perform equal work, regardless of cost differences. Male and female employees must receive equal benefits for their beneficiaries as well. Stepping out of the 1960s momentarily, it had become more difficult to sue employers for pay discrimination under Title VII of the Civil Rights Act of 1964. Title VII imposes a statute of limitations period—typically 180 days—after which employees may file claims of illegal discrimination against employers. In 2007, the U.S. Supreme Court rendered a strict interpretation as to when the statute of limita348

tions period begins for women to sue their employers for discrimination in pay. In Ledbetter v. Goodyear Tire & Rubber Co. (2007), a female employee named Lilly Ledbetter sued the Goodyear Tire and Rubber Company after she learned that some male employees with the same job had been paid substantially more than she had been over a period of several years. Ledbetter claimed that the statute of limitations period began when each discriminatory pay decision was made and communicated to her. She argued that multiple pay decisions were made over the years each time Goodyear endorsed each paycheck, making each paycheck a separate act of illegal pay discrimination. The Supreme Court rejected Ledbetter’s allegation that each paycheck (i.e., following the initial paycheck when the pay disparity first existed) represented an intentionally discriminatory act by the employer. The Court argued that, instead, any act of discrimination occurred each time pay raise decisions were made. In Ledbetter’s case, the Court argued that any discriminatory decisions literally occurred years prior to Ledbetter raising her concerns with the EEOC. The later effects of past discrimination do not restart the clock for filing an EEOC charge, making Ledbetter’s claim an untimely one. Not all of the judges agreed with this ruling, noting that, given pay secrecy policies in most companies, many employees would have no idea within 180 days that they had received a lower raise than others. In addition, an initial disparity in pay may be small, leading a woman or minority group member to not make waves in order to try to succeed in their jobs. In 2009, the Ledbetter court decision was overturned with the passage of the Lilly Ledbetter Fair Pay Act, a key initiative in closing the pay gap between men and women. This act restored the law as it was prior to the narrowly decided Supreme Court Ledbetter decision in 2007 (Ledbetter v. Goodyear Tire & Rubber Co., 2007). That decision tossed aside long-standing prior law and made it much harder for women and other workers to pursue pay discrimination claims, stating that a pay discrimination charge must be filed within 180 days of the employer’s initial decision to pay an employee less. The Lilly Ledbetter Fair Pay Act restores prior law, providing that a pay discrimination charge must simply be filed within 180 days of a discriminatory paycheck.

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

The Paycheck Fairness Act of 2008 is a second key initiative in closing the pay gap between men and women. This bill strengthens the Equal Pay Act of 1963 by strengthening the remedies available to put sex-based pay discrimination on par with race-based pay discrimination. That is, employers are now required to justify unequal pay by showing that the pay disparity is not sex based but, rather, job related. In the 1960s and 1970s, notwithstanding the passage of the Equal Pay Act of 1963, women’s rights advocates essentially interpreted the pervasive pay gap between women and men as discriminatory. Concerns about the fairness of alternative job evaluations were raised (see Martocchio, 2009; Milkovich & Newman, 2008, for reviews of alternative job evaluation methods). Compensation professionals use job evaluation to systematically recognize differences in the relative worth among a set of jobs and establish pay differentials accordingly. Based on job content and the organization’s priorities, pay differentials are set for virtually all jobs within the company. A critical decision was whether a single job evaluation technique is broad enough to assess a diverse set of jobs. In particular, the decision is prompted by such questions as “Can we use the same compensable factors (following the Equal Pay Act of 1963, the universal compensable factors of skill, effort, responsibility, and working conditions) to evaluate a forklift operator’s job and the plant manager’s job?” If the answer is yes, then a single job evaluation technique is appropriate. If not, then more than one job evaluation approach should be used. It is not reasonable to expect that a single job evaluation technique, based on one set of compensable factors, can adequately assess diverse sets of jobs—operative, clerical, administrative, managerial, professional, technical, and executive. Clearly, a cardiac surgeon’s job is distinct from a day care worker’s job. Some job evaluation methods include reliance on market pay rates, pay incentives, individual rates, and collective bargaining. Many companies determine the value of jobs by paying relative to the average or median rate in the external labor market. These questions about job evaluation decisions and the women’s rights movement fueled the comparable worth debate. Comparable worth represents an ongoing debate in society regarding the pay dif-

ferentials between men and women who perform similar, but not identical, work where women earn substantially less than men. In fact, a sex-based pay gap has endured for over 150 years (Goldin, 1991). Although the pay gap has decreased in recent decades, it is still substantial. In recent years, the pay gap has been about 75%; that is, women earned about 75 cents for every dollar that men earned (U.S. Bureau of Census, 2009). Comparable worth advocates argue that women who perform work that is comparable with men’s work in terms of compensable factors should be compensated equally to men. This debate is fueled, in part, by the use of particular compensation practices. As noted, job evaluation procedures influence the comparable worth debate. Using market wage or “typical” wage rates from compensation surveys also contributes to this continuing issue. England (1992) provided a theoretical analysis and review of the research on the comparable worth debate. For the purposes of this discussion, only a brief summary of the key issues based on her work is presented. Since World War II, American women have joined the workforce in increasing numbers out of economic need, increased opportunities for jobs and higher wages, and career aspirations. Single or divorced mothers have had to support themselves financially, and many women entered the workforce during the 1970s, when economic recessions resulted in the temporary layoffs of their husbands. Besides economic need, the number of employment opportunities for women rose. This growth resulted from a restructuring of the economy that produced declining employment in both agriculture and manufacturing and increases in service industries such as food service, retail sales, and the health industry. Job segregation also influences comparable worth. Few jobs in the United States are substantially integrated by gender. In other words, men are the predominant incumbents for certain jobs, whereas women are the predominant incumbents for others. Presently, in American society, men dominate high-status occupations, such as medical doctors, lawyers, and college or university professors, and women dominate lower status occupations, such as housekeepers, clerical workers, nurses, and elementary school teachers, even though the gap 349

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

is becoming less pronounced in some fields. Even when women become doctors, engineers, lawyers, or university professors, pay differences still often exist (U.S. Bureau of Labor Statistics, 2009). Many comparable worth advocates believe that certain compensation practices, job evaluations, and compensation surveys represent the media through which alleged discrimination against women operates. Job evaluation per se is not discriminatory. In fact, many comparable worth advocates argue that the use of a single job evaluation system that reflects the complete range of female- and male-dominated jobs can establish the “true” worth of jobs on the basis of content (National Committee on Pay Equity, 1987). However, the use of multiple job evaluation systems to evaluate jobs may be discriminatory when one system is applied to male-dominated jobs and the other is applied to female-dominated jobs. Compensation surveys also have been implicated as culprits in pay disparity between men and women holding jobs of comparable worth. Specifically, comparable worth advocates (among others) have argued that market rates are biased against women. In Lemons v. City and County of Denver (1980), Lemons claimed that her job as a nurse, a female-dominated job, was illegally paid less than comparable jobs held predominantly by men. In fact, Lemons argued that nursing requires more education and skill than some male-dominated jobs and so should be paid at a higher rate despite market factors, the fact that the male-dominated jobs generally earn higher rates in the local labor market. She also argued that local labor markets were inherently biased against women and therefore should not be a legitimate basis for establishing pay rates. The court disagreed with her charge on the basis that Title VII of the Civil Rights Act of 1964 did not focus on equalizing market disparities. In summary, the courts treat the market as a reality that allows companies little room for discretion in terms of gender-based salary differentials. Specifically, the courts hold that no single company should be held accountable for any pay-related biases against women. They have chosen not to interfere with companies’ reliance on market pricing because doing so could potentially undermine companies’ ability to compete. Mandating that a company 350

increase some female employees’ pay would represent a substantial cost burden, hindering ability to compete.

Strategic Compensation Period Personnel administration was transformed from a purely administrative function to a competitive resource in many companies during the 1980s, as technology transformed the workplace and pressures from global competitors intensified. Since the early 1980s, researchers and practitioners have given thought to the notion that HRM practices could contribute to companies’ competitive advantage by motivating employees to excel, learn new knowledge and skills, and take on a sense of ownership in the company (Pfeffer, 1995). Competitive advantage describes a company’s success. Specifically, competitive advantage refers to a company’s ability to maintain market share and profitability over a sustained period of several years (Becker & Huselid, 2006). Employers began to recognize that employees are key resources necessary for a company’s success, particularly in changing business environments characterized by rapid technological change and intense business competition from foreign countries. Employers’ recognition that employees represent an important resource led to the view of employees as human resources. In line with this view, companies design human resource management practices to promote competitive advantage. For example, organizations are striving to lower the costs of employment to help maintain profitability in ways to promote competitive advantage. First, two noteworthy changes in the employee benefits field include a shift from defined benefit plans to defined contribution plans. Defined benefit pension plans obligate companies to make lifelong payments to retirees. Defined contribution plans, however, do not create long-term obligations because organizations simply set up tax-deferred savings plans for employees under Section 401(k) of the Internal Revenue Code (2000). Second, organizations are moving toward offering high-deductible health care plans. Typically, employer-sponsored plans require relatively low deductible amounts (i.e., the amount the insured pays for medical care before the insurance company

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

begins to pay benefits), but such plans have become too costly to employers as insurance companies charge substantially more for employer-sponsored plans, in large part because of runaway costs of health care. High-deductible plans shift greater responsibility to employees for their medical care costs and enable organizations to spend substantially less for such plans than for lower deductible plans. According to the U.S. Bureau of Labor Statistics, nearly 55% of workers employed in the private sector participated in at least one company-sponsored retirement plan from 1992 to 1993 (Costo, 2006). Since then, the participation rate has declined slightly to approximately 50% in 2006, the latest available data at the time of publication. However, there has been a noticeable decrease in participation rates for defined benefit plans over the past 15 years. From 1992 to 1993, 32% of private sector employees participated in defined contribution plans, and slightly fewer participated in defined benefit plans (Costo, 2006). In 2006, 42% participated in defined contribution plans, but only 20% participated in defined benefit plans. These trends in retirement plan participation have two important explanations (Costo, 2006). First, there has been a shift in the labor force toward different occupations and industries. Specifically, there has been a relative decline in employment among full-time workers, union workers, and workers in goods-producing businesses. The decline in full-time workers and the increase in part-time workers have led to fewer opportunities for participation in company-sponsored retirement plans. Quite simply, employers often employ part-time workers to save benefits costs. The decline in union affiliation (i.e., union members or just part of the bargaining unit) also contributes to the overall trends described earlier. In 2006, nearly 90% of employees affiliated with unions were eligible to participate in a retirement plan, whereas only about half of the nonunion workers were eligible. Labor unions represent workers in negotiations with management over terms of employment. The inclusion of lucrative retirement plans was among the top priorities in negotiations to maintain the support of middle-aged and older workers. Finally, among employment trends, the expansion of service industries relative to some-

what stable employment in the goods-producing sector helps to explain retirement plan participation. Fewer service-oriented workers have access to defined benefit plans (19% vs. 33%), though the percentage of workers with access to defined contribution plans is higher and similar in both industries (approximately 50%–60%). However, actual employee participation in defined contribution plans is drastically lower in service employers than in goods-producing companies. Wages in the service-producing companies tend to be lower than in goods-producing companies. The second reason for changes in moving from participation in defined benefit plans to defined contribution plans is that defined benefit plans are quite costly to employers compared with defined contribution plans. Companies struggle to adequately fund these plans to ensure that retirees receive entitled benefits for the remainder of their lives. Also, the Pension Benefit Guaranty Corporation (PBGC) serves as the insurer by taking over pension obligations for companies that terminate their defined benefit plans because of severe financial stress. Companies with defined benefit plans pay premiums to the PBGC to insure defined benefit plans in the event of severe financial distress. The Pension Protection Act of 2006 requires that companies at high risk of not meeting their pension obligations pay substantially more to insure defined benefit plans, adding to the substantial cost. Companies can choose from three broad classes of health insurance programs in the United States, including fee-for-service plans, managed care plans, and point-of-service plans, the latter of which combines features of fee-for-service and managed care plans. An emerging class of health insurance programs is based on consumer-driven health care, in which employees play a greater role in decisions about their health care, have better access to information with which to make informed decisions, and share more in the costs to receive employer-sponsored health insurance and substantially higher annual deductibles (i.e., the amount an individual pays for health care before the insurance benefits defray part or all of additional health care costs). It is also important to mention that health care in the United States is classified as a multiple-payer 351

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

system. In a multiple-payer system, more than one party is responsible for covering the cost of health care, including the government, employers, labor unions, employees, or individuals not currently employed (e.g., retirees, the unemployed, employees whose employer does not pay for health care coverage). As I discuss shortly, a variety of forces have contributed to the existence of a multiple-payer health care system in the United States. A multiple-payer system stands in contrast to a single-payer system in which the government regulates the health care system and uses taxpayer dollars to fund health care, as in Canada and some other countries. Single-payer systems are often referred to as universal health care systems because the government ensures that all of its citizens have access to quality health care regardless of their ability to pay. These approaches to health care coverage have been at the heart of political and social debate for years. The debate has taken center stage since President Barack Obama was elected in 2008. A variety of forces contributed to the current multiple-payer health care system in the United States, including tax incentives to employers that provide insurance to employees, and employers’ recognition that healthier workforces are likely to be more productive. However, the dramatically rising costs of health care services have made it cost-prohibitive for many employers—particularly small employers—to provide health insurance coverage because insurance companies continually raise their rates to maintain profits. In addition, the rampant rate of inflation in health care costs and insurance far outpaced the rise in employee wages and salaries, forcing most companies that could still afford to offer health insurance to employees to reduce the level of insurance coverage by increasing deductibles (to be discussed shortly) or to contribute more to receive group coverage. Effectively, changes such as these continue to put health care out of the reach of millions of individuals (i.e., adults and children). In fact, the number of uninsured Americans has risen noticeably during the past several years to nearly 50 million. As an aside, other societal forces have contributed to the rise in the number of uninsured Americans. Among them are the outsourcing of high-paying 352

jobs to other countries, the decline of high-paying unionized jobs such as in the automobile industry, and the rise of lower paying jobs typical of the expanding service sector of the economy. The confluence of these and other unmeasured forces makes it difficult to isolate the effects of a single source on the uninsured problem. Nevertheless, this problem of more and more people becoming uninsured has caught the attention of politicians, business executives, and the labor movement. These groups are calling for government intervention and collaborative efforts between the groups to reduce the number of uninsured. Extreme proposals advocate a single-payer health care system such as in Canada, based on the belief that access to quality health care is a basic right rather than a privilege available only to those who can afford to pay for it. For example, in early 2007, the executives from four major U.S. Corporations (Wal-Mart, Intel, AT&T, and Kelly Services) and the leadership from the Service Employee International Union held a national press conference to express the desperate need for a system that provides universal health care coverage for every person in the United States. This coalition, named Better Health Care Together, advocated several goals during the press conference. Specifically, by the year 2012, they hope to see “quality, affordable health insurance coverage” for every American and “businesses, governments, and individuals all . . . contribute to managing and financing a new American health care system” (Better Health Care Together, n.d.). Critics from other large labor unions, such as the United Food and Commercial Workers Union, have argued that large nonunion employers such as Wal-Mart and others that hope to limit union influence are simply attempting to improve their public image by expressing concern about the uninsured problem in America. In any event, the coalition did not address how to fund a universal health care system, that is, who becomes the single payer. At least for now, their objective is to raise awareness of the problem. President George W. Bush advanced a plan in early 2007 to place the burden of health insurance costs on individuals by offering tax breaks to everyone who purchases health insurance coverage.

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

Currently, the tax laws make employer-sponsored health insurance tax-free for workers and employers. That is, contributions to purchase health insurance are not included in an employer’s or employee’s taxable income. However, because employers pay the majority of the health insurance premium, typically, the tax deduction for employees is relatively small. Needless to say, debates about the issue of moving toward a single-payer system in the United States will continue for some time to come for a variety of reasons. Since his election, President Obama authorized the establishment of the White House Office of Health Reform. President Obama has pledged to provide all Americans access to affordable and high-quality health care. This office is charged with establishing policies, priorities, and objectives for the federal government’s comprehensive efforts to improve access to health care, the quality of such health care, and the sustainability of the health care system.

Global Financial Crisis Since 2007, fiscally irresponsible mortgage lending practices led to the collapse of a house of cards when millions of individuals began defaulting on their home mortgage payments. Residential values tumbled; consumer confidence dropped precipitously, leading to lower spending. Drastically lower property values left countless individuals with negative equity in their homes, cutting off their ability to obtain home equity lines of credit for spending or to sell their home for any profit. The resulting drop in consumer confidence and ability to spend caused company profits to tumble, precipitating wide-scale layoffs. The global financial crisis has led to changes in compensation and benefits practices in organizations. To control layoffs, many companies adopted either or both across-the-board pay cuts or furlough policies that require some employees to take days off without pay. Merit pay increase budgets ranged between 0% and 3%, representing the lowest payfor-performance increase budgets in more than 3 decades. Organizations also sought to reduce expenditures on benefits by terminating defined benefit pension plans or suspending matching contributions to employees’ 401(k) savings accounts.

REWARDS AND COMPENSATION RESEARCH: A SELECTIVE REVIEW AND ASSESSMENT The review in this section is organized into two parts. The first entails a review of the literature based on the two types of compensation, intangible and tangible, discussed earlier. Specifically, in part one, the discussion centers on intangible compensation and the particular elements of tangible compensation as well as the character of research prior to the emerging literature on SHRM. The second part is a review of established taxonomies of HRM activities and strategies, including tangible reward and compensation practices and key theoretical perspectives that have been suggested as relevant to the study of SHRM. The SHRM literature has been silent on intangibles such as nonmonetary recognition of employee accomplishments. Following this review, a synthesis is provided of the most critical research gaps that remain and that limit our understanding of strategic rewards and compensation in increasingly complex environments. A detailed review of specific psychological studies on compensation and reward systems can be found in Rynes and Gerhart (2000). Since the publication of their review, Cascio and Aguinis (2008) reported a dramatic decline in the number of scholarly articles on rewards systems published in the Journal of Applied Psychology and Personnel Psychology.

Compensation Research Prior to SHRM Research As a field of study, HRM is continuing to undergo a major transformation that began in the early 1980s. Prior to this period, almost all research in HRM could be grouped into the practices of selection, training, appraisal, and rewards, generally mirroring the established functions within the typical HRM department. Not surprisingly, the literature in each of those functions evolved in relative isolation from each other. As a result, whereas researchers in compensation focused intently on issues of job evaluation, setting pay ranges, and designing effective incentive practices, they did so with little consideration of the relationship between compensation system variables and those associated with selection, training, and performance appraisal. 353

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

Moreover, the traditional HRM research perspective was decidedly micro in nature; that is, it focused on specific HRM practices with virtually no consideration for how those practices might fit within the strategies and context of the organization. Compensation researchers conducted (and still conduct) research on issues such as pay satisfaction dimensionality, the relationship between pay satisfaction and job performance, procedural and distributive justice attributed to compensation practices, and the role of alternative compensation schemes in the attraction and retention of talent (Dulebohn & Werling, 2007; Gerhart & Rynes, 2003; Werner & Ward, 2004). In general, the focus is on individuals’ reactions to rewards and compensation practices. Over the years, multiple reviews and two volumes provided comprehensive reviews of the content of this literature (e.g., Dulebohn & Werling, 2007; Gerhart & Milkovich, 1992; Gerhart & Rynes, 2003; Milkovich, 1988; Werner & Ward, 2004). Altogether, this body of work described and critiqued rewards and individual compensation research from economic, psychological, and sociological perspectives. Because of these excellent reviews, the goal is not to reinvent the wheel by reviewing individual studies; the past reviews of the compensation literature have done an exemplary job of that. Although past research has given us a better sense of how individuals are likely to react to particular compensation practices, our understanding is limited. It is possible that differential reactions may be due to something other than the design of compensation practice. What is apparent from past psychological research on compensation is that researchers have given little consideration to the psychological meaning of money.

Psychological Research on the Meaning of Money As noted by Krueger (1986), “Money is probably the most emotionally meaningful object in contemporary life; only food and sex are its close competitors as common carriers of such strong and diverse feelings, significance, and strivings” (p. 3). Perhaps the most noteworthy gap in rewards and compensation research is a careful construct definition of the meaning of 354

money and theory development about individual differences in the meaning of money (Mitchell & Mickel, 1999). After all, it is the locus of tangible compensation practices, yet it is not well understood. Presumably, two individuals who make the same annual salary and receive identical pay increases based on objective performance measures (e.g., units produced per unit of time) should have identical pay satisfaction. Or should they? A brief review of the meaning of money is provided in this section. At a minimum, we know that money is a multidimensional construct, perhaps for most individuals in the United States. U.S. culture emphasizes instrumentality. Expectancy theory suggests that individuals will strive for high levels of performance when they believe that better performance leads to greater pay (Vroom, 1964; see also Vol. 3, chap. 2, this handbook.) In this context, money derives importance from what it can buy. Greater amounts of money enable people to buy more things. What things do people buy? We buy housing (or rent it), utilities, food, clothing, transportation, and entertainment. Money serves instrumental and symbolic meanings (Mitchell & Mickel, 1999; Tang, 1992). It is often associated with four of the most important symbolic attributes human strive for: (a) achievement and recognition, (b) status and respect, (c) freedom and control, and (d) power (Tang, 1992). Money can provide the luxury of time, autonomy, and freedom of choice (Parsons, 1967). It recognizes accomplishments (Kirkaldy & Furnham, 1993), and it can provide power and access to resources (Mitchell & Mickel, 1999). Cultural values are probably a key factor in defining a meaning of money construct (see Vol. 3, chap. 23, this handbook). Culture is defined as “the human-made part of the environment (Herskovits, 1955). It has both objective elements—tools, roads, appliances—and subjective elements—categories, associations, beliefs, attitudes, norms, roles, and values” (Triandis, 1994, p. 113). Cultural values shape work-related attitudes and behaviors (Triandis, 1994). Culture also influences the domain of normative behavior (i.e., behavior that is desirable versus condemned for members of the culture), defines roles for individuals in the social structure, and prescribes guiding principles and values in one’s life. As a result,

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

culture specifies how things in the environment, including an organization’s practices, policies, and procedures, are to be evaluated, and subsequent reactions to such procedures (Robert, Probst, Martocchio, Drasgow, & Lawler, 2000). A striking contrast exists between U.S. and Japanese culture. In U.S. businesses, strategic business decisions generally originate from top management. Japanese business leaders cultivate consensus on business decisions, or nemawashi. U.S. culture promotes a sense of individualism, which translates into high career mobility. Japanese culture promotes a sense of collectivism, which leads to heightened loyalty for employers. These cultural values are apparent in compensation systems. The predominant bases for pay in the United States are performance and knowledge, which represent equity (Heneman & Werner, 2005). In Japan, the predominant basis for pay is seniority, which represents equality. As a result, pay differences among the Japanese tend to be smaller than pay differences among U.S. employees. Another noteworthy cultural contrast exists between the United States and the People’s Republic of China (PRC). The differences between the U.S. market economy and the PRC’s centralized governmentcontrolled economy sets the stage for cultural clashes. For decades, the Chinese government owned and operated virtually all business organizations. The Communist Party places substantial emphasis on equal contributions to society, group welfare, and the concern for interpersonal relationships. In addition, the Communist Party calls for greater emotional dependence of Chinese citizens on their employers. Further, it expects employers to assume a broad responsibility for their members. Zhou and Martocchio (2001) demonstrated support for the difference in ideals when Chinese (employed in the PRC) and U.S. managers made decisions about monetary and nonmonetary compensation awards. These ideals are evident in the Chinese workplace and in compensation practices. Employers provide housing and modest wages for food and clothing. The Chinese receive health care under government-sponsored protection programs. Because of the communist ethic, the Chinese do not identify well with payfor-performance programs, though that is slowly

changing as a capitalist presence creates for-profit organizations. A third example involves the compensation packages for U.S. and Mexican managerial employees, who differ substantially. The most important elements of U.S. managers’ compensation are base pay and long-term incentives. Base pay and cash allowances represent the lion’s share of Mexican managerial employees’ compensation packages. In fact, the Mexican government mandates that employers award Christmas bonuses, profit sharing, and a minimum 20% vacation pay premium (i.e., employers must pay employees at least an additional 20% of the regular pay while on vacation). U.S. employers offer these allowances at their discretion, not by government mandate. The most noteworthy difference is Mexico’s acquired right law: Employees possess the right to benefit from compensation practices that have been in effect for at least 2 years. For example, if an employer instituted 40 paid vacation days per year, employees acquire the right to 40 paid vacation days per year every year if the company continued this practice for at least 2 consecutive years. Although U.S. companies have never operated under an acquired right law, U.S. employees viewed benefits as an entitlement. Nowadays, U.S. companies discourage that view because benefits represent a significant cost.

Rewards and Compensation in the Era of SHRM Research In this section, my objective is to assess how well the extant literature on SHRM in general and strategic rewards and compensation in particular address the changing reality facing organizations. Because a number of excellent review articles of the SHRM literature have already been published (e.g., Becker & Huselid, 2006; Wright & McMahan, 1992), the study of the role of SHRM practices in organizational performance, such as compensation systems, sets the stage for reviewing what we know about strategic rewards and compensations as well as the limitations of this body of research. Overall, there has been excellent work that has conceptualized the role of individual HRM practices in organizational level success. To date, empirical work has lagged behind conceptual 355

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

work, in large part due to challenges put forth by organizations to make reliable and valid data available to researchers. Where data are available, threats to internal validity, such as history (e.g., rapidly contracting economy) and maturation (e.g., mergers and acquisitions leading to changes in strategic direction, adding new product lines, dropping others), are alternative explanations to predictions about the influence of HRM practices on organizational performance. Indeed, as Becker and Huselid (2006) pointed out, those researchers generally have neglected clarifying the “black box” between HRM practices and organizational performance. Since 1980, we have seen an explosion of research and writing that has attempted to bridge the fields of competitive strategy and HRM. Many well-known strategy researchers are abandoning the deeply rooted notion that strategy is the province of top management and, therefore, outside the scope of functional areas such as HRM (Barney, 2001). This movement opens the door for researchers to study HRM-related strategic issues at multiple organizational levels. For instance, Mintzberg and Waters (1985) argued that strategy results from a combination of deliberate planning by senior executives and emergent decisions at different levels in the organization. Mintzberg and Waters also have challenged the entrenched view that top executives are responsible for strategic decisions and generally relegate other members of the organization to subordinate implementation roles. Many others have advanced the notion that one of the most significant trends for the 1990s and into the 21st century has been to incorporate contextual factors when examining the role of HRM practices in attaining competitive advantage (Schuler & Jackson, 1987; Wright & McMahan, 1992). Such contextual factors include business strategies, national culture, organizational culture and history, business life cycle, organizational size and structure, and economic and legal conditions. This notion conveys a deterministic view of HRM strategy, casting it as a dependent variable. Since then, many HRM scholars have recognized the need for linking HRM policies and practices to organizational strategy and now study HRM functions with organizational performance as the focal point (e.g., Huselid, Beatty, & Becker, 2005; Lepak & Snell, 1999; Wright & Snell, 1998). This 356

stream of research has characterized HRM strategy as an antecedent of organizational performance. When studying the relationship between HRM practices and organizational performance, researchers have focused their attention on the HRM architecture. HRM architecture is a metaphor many researchers have used to highlight the locus of value creation in SHRM, namely, the systems, practices, competencies, and employee performance behaviors that reflect the development and management of the organization’s human capital (Wright & Snell, 1998). Most of the attention has been on the configuration of HRM architecture. For example, researchers raised questions about whether single HRM practices had an independent effect on employees and, subsequently, organizational performance, or whether bundles of HRM practices were more effective (Becker & Huselid, 2006). Bundles were believed to possibly amplify individual ability and motivation over the influence of any single practice (Huselid, 1995; Ichniowski, Shaw, & Prennushi, 1997). Until recently, relatively few researchers have explicitly attempted to extend understanding of the relationship between HRM practices and organizational performance by integrating the possible mechanisms through which the HRM architecture influences organizational performance. For instance, Toh et al. (2008) found that organizations used one of five bundles of HRM practices based on organizational values and structure. Each bundle included combinations of different staffing, development, reward, and evaluation practices that were associated with an organization’s strategic objectives: cost minimizers, contingent motivators, competitive motivators, resource makers, and commitment maximizers. For example, contingent motivators, compared with other organizations, adopt contingent pay systems to a relatively greater extent to motivate their employees. Competitive motivators seek to increase employee motivation by paying employee market competitive wages and benefits. Toh et al. (2008) also sought to explain the choices of bundles based on Schneider’s (1987) attraction– selection–attrition (ASA) model (see Vol. 3, chap. 1, this handbook). Briefly, the ASA model maintains that through attraction, selection, and attrition, members of an organization tend to exhibit more

Strategic Reward and Compensation Plans

Copyright American Psychological Association. Not for further distribution.

characteristics similar to each other than to members of different organizations (Schneider, Smith, Taylor, & Fleenor, 1998). In addition, organizational founders establish “processes, structures, and culture [that] emerge to facilitate the achievement of these goals” (Schneider, Goldstein, & Smith, 1995, p. 749). From a strategic management perspective, MacDuffie (1995) reasoned that there is a pull toward a complementary relationship among an organization’s culture, structure, and practices. STRATEGIC REWARDS AND COMPENSATION SYSTEMS: THE CASE FOR A PROCESS THEORY PARADIGM As an alternative to treating HRM strategy as a dependent variable, Lengnick-Hall and Lengnick-Hall (1988) proposed a framework of SHRM as an interactive phenomenon. They levied criticisms against the SHRM literature. Among their criticisms, they noted that SHRM models emphasized strategy implementation over formulation. They maintained that the implication of treating human resources as a means precludes viewing human resources as a strategic capacity from which competitive choices should be derived. Moreover, they indicated that SHRM models cast organizational life cycles as single and uncontrollable catalysts of change. Consequently, they argued that organizational leaders may underestimate their potential for choice and influence. Lengnick-Hall and Lengnick-Hall’s (1988) model emphasizes reciprocal interdependence; that is, the choice of strategy is not a given. They further suggested that the HRM function should contribute directly to both strategy formulation and strategy implementation. Conditions that influence what types of questions should be asked are not the same contingencies that determine the answers to those questions. Both asking the right questions and making acceptable choices are necessary for sustained high performance.

Established Taxonomies and Typologies Clearly, the predominant categorization of HRM activity has been, and continues to be, based on the HRM functions of planning, staffing, rewards and compensation, appraisal and training, and develop-

ment (Becker & Huselid, 2006; Schuler & Jackson, 1987). More recently, authors have stressed the need for integrating HRM activities across the individual functions, and some have suggested a slightly broader list, for example, including labor relations (Cutcher-Gershenfeld, 1991), but the functional theme has generally prevailed. Despite widespread acceptance of the functional view as the convention for describing HRM activity, however, some alternative categorization schemes have appeared (Beer, Spector, Lawrence, Mills, & Walton, 1984; Wright & Snell, 1991). For example, Beer et al. (1984) suggested that HRM should be viewed as managing employee influence, employee flows, reward systems, and work systems rather than addressing individual HRM practices. Beer et al. argued that viewing the HRM function in this way provides a much more generalist approach and helps to overcome limitations of the functional view related to narrowness of perspective and unrecognized synergies among functions. Compensation has been explicitly used as an example of the advantages of a broadened perspective. For example, because compensation systems are ubiquitous in organizations, they may be thought of as the solution to problems in employee job performance and organizational effectiveness. Of course, in many cases, such problems may be more appropriately addressed via selection or training systems approaches. Incentive pay systems are well suited to aligning the interests of employees and employers. For example, gains sharing programs are most appropriate where companies’ objectives center on increasing organizational efficiency through cost reduction. The adoption of pay-forknowledge systems can provide employees with the essential knowledge and skills needed in the conversion of outmoded production systems into lean manufacturing systems. However, there is no place for compensation systems to make improvements when an organization selects individuals who do not have a sufficiently high enough level of general mental ability to understand complex instructions. In his groundbreaking and now well-known work on strategic “menus” associated with HRM subfunctions, Schuler (1986) identified nine strategic choices related to compensation and reward systems. 357

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

Specifically, those nine choices are (a) low base salaries versus high base salaries, (b) internal equity versus external equity, (c) few perks versus many perks, (d) standard fixed package versus flexible package, (e) low participation versus high participation, (f) no incentives versus many incentives, (g) short-term incentives versus long-term incentives, (h) no employment security versus high employment security, and (i) hierarchical versus high participation. Schuler noted that his intent was to communicate the variety of characteristics and methods that shape and give flavor to each HRM function. He further noted that his intent was also to provide a menu of choices for the HRM practitioner to select after a determination is made of what employee characteristics are necessary based on organizational strategy. Although Schuler’s model is extremely useful in helping to categorize and think about different compensation dimensions, it is primarily beneficial for thinking about the type of compensation plan design and implementation that falls into the domain of psychological research and is therefore limiting. That is, Schuler’s dimensions are most apt when thinking about compensation developed with known problems and the “gaps” between desired and current employee motivation can be reliably identified.

Emerging Theoretical Perspectives In response to criticisms regarding the lack of theoretical orientation in the field, authors recently have begun to invoke a number of different theories from the organizational literature (Becker & Huselid, 2006; Wright & McMahan, 1992). Many of the perspectives have their roots in contingency theory (Lawrence & Lorsch, 1967), which suggests that organizations whose internal features best match the demands of their environments will achieve the optimum adaptation (see chap. 5, this volume). As Schoonhoven (1981) noted, contingency theory is not really a theory at all in the conventional sense of theory as a well-developed set of interrelated propositions. It is more of an orienting strategy or metatheory, suggesting ways in which a phenomenon ought to be conceptualized, or as an approach to the phenomenon explained. As applied to compensation, the general contingency prediction would be that congruency or fit 358

among compensation strategies, organizational strategies, and organizational characteristics will have a beneficial effect on organizational performance. Specifically, compensation embodies a resource that facilitates organizations’ adaptation and responses to environmental demands such as aligning employees’ interests with the long-term success of the organization. Drawing from contingency theorists (e.g., Lawrence & Lorsch, 1967), the guiding premise is that compensation systems are essential resources for aligning organizations with the external environment. In particular, compensation represents an opportunity for organizations to incent employees that contribute, in aggregate, to an organizations’ competitiveness (Pfeffer, 1993). Resource-based perspective. Resource-based theory is currently receiving a significant amount of attention in the strategic management literature (e.g., Barney, 2001). According to the resource-based view of the organization, competitive advantage can only occur in situations of organizational resource heterogeneity and organizational resource immobility, and it is these assumptions that serve to differentiate the resource-based model from the traditional strategic management model. In order for an organization’s resource to provide sustained competitive advantage, four criteria must be attributable to the resource: (a) the resource must add positive value to the organization, (b) the resource must be unique or rare among current and potential competitors, (c) the resource must be imperfectly imitable, and (d) the resource cannot be substituted with another resource by competing organizations. The resource-based approach provides a framework for examining the pool of human resources that may be either able or unable to carry out a given strategy during the formulation phase of strategic management. Thus, unlike the behavioral perspective, the resource-based view assumes that strategies are not universally implementable but are contingent on having the human resource base necessary to implement them. Put another way, this approach assumes a reciprocal interdependence between an organization’s business strategy and its human resources capability (Lengnick-Hall & Lengnick-Hall, 1988). Since then, Lepak and Snell (1999, 2002)

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

have urged researchers to pay greater attention to theoretical development of differentiating the HRM architecture. Specifically, they have said that SHRM researchers have treated all employees as alike and essential to competitive advantage. Moreover, SHRM research focuses on “the extent to which a set of practices is used across employees of an organization as well as the consistency of these practices across the firm [organization]” (Lepak & Snell, 1999, p. 32). For example, take the relationship between Dell Computers and any one of its parts suppliers. Consistent with resource dependency theory, augmented employee skills, acquired through the systematic training curricula embedded within a skill-based pay program, may well be one basis for creating a dependency between organizations. An offering of highest quality components by a particular parts supply manufacturer may be a deciding factor for Dell in choosing one supplier over the other. To the extent that training helped equip employees with the needed skills to manufacture highest quality components, it can be said that the supplier’s payfor-knowledge compensation program contributed to Dell’s dependence. This imperative does not extend to the part of the architecture that is not directly related to producing highest quality components. Agency and institutional perspective. The agency perspective is based on the notion that because most modern for-profit organizations separate owners (in the hands of dispersed stockholders) and managers (hired to maximize shareholder wealth), an agency relationship exists. Jensen and Meckling (1976) defined the agency relationship as “a contract under which one or more persons engage another person (the agent) to perform some service on their behalf which involves delegating some decision-making authority to the agent” (p. 308). In any relationship in which the owner is not present, the agent is in a position to engage in self-serving behaviors and pursue short-run rather than long-term goals. Unconstrained discretion, potential conflict of interest, asymmetrical information between owners and managers, and the ability to shift risks into the future when present management is likely to have moved on to other organizations, have implications for managerial behaviors that are at odds with eco-

nomic models such as the resource-based approach. For example, management could actually invest in suboptimal projects at the expense of long-term profits. With respect to compensation, this may most often be manifested in the neglect of maintaining competitive base pay (i.e., minimizing short-term direct costs) at the expense of a stable as well as highly skilled and motivated workforce. A related noneconomic theory that has been suggested as an explanation for some SHRM practices is the institutional perspective (Meyer & Rowan, 1977; Scott, 1987). The basic tenet of institutionalism is that many structures, programs, and practices in organizations attain legitimacy through the social construction of reality. More specifically, advocates of the institutional perspective argue that (a) what looks on the surface to be rationally derived organizational structures and practices may only appear to be so and (b) structures may serve some functional goal, although they had not been designed for that particular purpose. Considerable anecdotal and case study data suggest that a large number of merit pay or seniority programs continue to exist not because of their contribution to resource enhancement or demonstrated link to business strategy but rather because of a host of institutional factors (Heneman & Werner, 2005). GAPS IN THE STRATEGIC REWARDS AND COMPENSATION LITERATURE: WHAT RESEARCHERS KNOW AND WHAT THEY DO NOT KNOW The review and critique of the extant literature reveal two particularly acute gaps with respect to our understanding of strategic rewards and compensation systems. First, although considerable progress is being made in some areas of SHRM, the compensation function seems particularly underdeveloped. In this regard, perhaps the most pressing need is to determine the domain of strategic rewards and compensation decisions and the processes that lead to these decisions. That is, strategic rewards and compensation decisions connote compensation decisions responsive to environmental opportunities and threats and that are linked to the overall longterm directions and purposes of the organization. 359

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

Questions such as “What are these threats and opportunities?” and “How do these threats and opportunities influence the formulation of strategic rewards and compensation decisions?” merit researchers’ attention. All compensation decisions are clearly not strategic. Deciding which compensation consultants to use or which human resource information system to purchase are not strategic decisions. Instead, choosing whether to load the mix of pay elements toward base pay versus incentives or whether to focus incentives on short-term performance goals or long-term organizational goals probably are strategic decisions. The work in this area is still particularly ill-defined. A richer conceptualization of the dimensionality of strategic rewards and compensation plans would help to promote more theoretical and comparative research. More specifically, the major research tasks still to be accomplished are (a) to identify the compensation decisions that are strategic; (b) to develop descriptions of these decisions; and (c) to determine whether compensation strategies affect workforce attitudes and behaviors, which in turn affect the implementation of competitive strategy. In this chapter, my focus is on understanding and describing the compensation decisions that are strategic. Understanding and describing these decisions is a necessary but not sufficient condition for attaining (c). Second, with a few notable exceptions, the SHRM literature is long on models, propositions, and new perspectives, but it still is woefully short on empirical research relative to the psychological research on rewards and compensation practices. (For example, Huselid’s [1995] study included a single item that was part of an employee skills and organizational structures scale.) The lack of empirical research is probably attributable to a number of factors, including lack of specification of strategic compensation dimensions and difficulty in getting even remotely comparable measurements across organizations, not to mention likely threats to internal validity, as discussed earlier. For example, I have witnessed many practitioners and researchers interchange one type of compensation component for another. Despite efforts to educate individuals about the proper use of compensation 360

components through professional certification, textbooks, or professional reference publications, one of the greatest obstacles to comparing compensation plan effectiveness is a lack of standard common terms, implementation methods, and evaluation measures across organizations. For example, a timely blog entry on the WorldatWork Web site (http://www.worldatwork.org) illustrates this concern (Kovac, 2008): Not surprised that the effect of sharply rising food and energy prices is having an impact on the workforce and challenging most businesses to respond in a sensitive way to help employees out; but surprised that the cost-of-living adjustment (COLA) terminology would still be alive, and apparently well. COLA is a by-product of union contract terminology where COLA adjustments are the primary way that wage increases have been determined under collective bargaining agreements. In the 1970s and early 1980s, doubledigit inflation was the norm and we typically managed two merit increases cycles a year. We did this to try and keep pace with wage inflation that was the result of an out-of-control inflation cycle that characterized much of the late 70s and early 80s. But even in this environment, we carefully avoided the phrase COLA to describe our practice. The primary criteria used to allocate merit increases, based on a substantial merit budget at least compared to today, was performance. Differentiating salaries and merit increases based on performance was at the core of our merit programs. COLAs, on the other hand, make no differentiation and are allocated across-the-board irrespective of merit. I would caution anyone, as I did in our webinar, from resorting either to the practice or the language of COLA as it is inconsistent with a pay-for-performance

Strategic Reward and Compensation Plans

Copyright American Psychological Association. Not for further distribution.

culture. Businesses are successful, or not, based on their ability to compete for customers based on the price, value, and quality of their products and services. Performance-based pay translates this business reality into an organization’s culture by requiring increases in pay to be earned. COLAs are inconsistent with this message and have no place in today’s environment. Much anecdotal evidence suggests a strong relationship between compensation decisions and organizational performance. More attention now needs to be paid to the type of research that will provide the necessary support or disconfirmation (Dyer, 1984). What do researchers know about strategic rewards and compensation decisions in light of the changing realities (i.e., increased environmental turbulence) facing decision makers in organizations? Compared with conditions of low environmental turbulence, we know relatively little about strategic compensation systems because the conventional approach to generating knowledge is not well suited to turbulent contexts in which organizations now operate. As a corollary, it is important to emphasize that our limited knowledge is not due to a lack of researchers’ creativity or that existing perspectives (i.e., the variety previously reviewed) are without merit. Only until recently have industrial and organizational psychologists frequently incorporated multiple levels of analysis in their research. Previously, the focus was on individuals without explicit recognition of broader environmental variables. Clearly, the work to date has provided us with a running start toward better understanding SHRM. Nevertheless, as organizational realities evolve, so must our approaches to effectively assessing SHRM phenomena. Gaps in our knowledge about strategic rewards and compensation decisions will be discussed by assessing the approaches to knowledge generation currently in use.

Rewards and Compensation Prior to the SHRM Research Era Many of the theoretical and empirical advances in this traditional compensation research on pay satis-

faction, justice perceptions, and individual job performance have followed the scientific process that Kane (1991) advocated: Science is properly conducted via the processes of conjecture, mapping, and evaluation. Conjecture consists of the positing of causes of phenomena or of the causes of relationships between phenomena. Mapping consists of specifying the correspondence between the abstract constructs posited in a theory and their observable manifestations envisioned in the notion of a nomological network (Cronbach & Meehl, 1955). Evaluation refers to the empirical assessment of predictions deduced from the conjectured theory. (p. 247) Inherent in this scientific process is deductive reasoning as a basis for formulating testable hypotheses, and an investment in research designs that maximize the degree of internal validity. Thus far, this approach to studying individuals’ reactions to compensation practices has yielded a rich body of theoretical and empirical knowledge. Deductive reasoning. The process of deduction (Dewey, 1933) begins with general laws about operations that must have some empirical content. These laws are combined with empirical observations to generate specific, testable hypotheses. A widely tested hypothesis in the compensation literature is that pay relative to others doing similar work in other companies will positively predict pay level satisfaction (Rynes & Gerhart, 2000). Equity theory would predict that a major influence on pay level satisfaction is comparisons of one’s pay relative to that of referent others (Scholl, Cooper, & McKenna, 1987). Internal validity. Internal validity deals with the conclusions drawn about whether there is a causal relationship between two variables, and if so, the direction of the relationship (Cook & Campbell, 1979). In compensation research, there appear to be careful assumptions about the nature of causal relationships between variables at both theoretical 361

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

and observed levels. According to J. S. Mill (as cited in Cook & Campbell), there are three criteria for inferring cause: (a) covariation between the presumed cause and effect, (b) temporal precedence of the cause, and (c) enough control to rule out alternative explanations. Much past research on individual responses to compensation practices has not suitably met Mill’s criteria. Much of that work was based on cross-sectional designs and included limited controls based on convenience, not theory (i.e., mainly, demographic variables that researchers easily include on survey forms). (See Gerhart and Rynes [2003] for a detailed review of individual studies and critique of methodological shortcomings.)

Summary On balance, I feel that researchers have generated a rich body of knowledge pertaining to individual reactions to compensation plans. The process of deduction lends itself well to gaining knowledge about employee attitudes and job performance because this stream of research is largely based on established theories of motivation (e.g., goal setting, equity theory). Researchers have done less well specifying the theoretical rationale that connects constructs (Cook & Campbell, 1979). For example, we have a respectable understanding about why and how perceptions of fairness follow the application of procedurally questionable pay determination processes, but less is known about other issues such as the psychological determinants of “just meaningful” pay increase amounts. The lack of understanding may center on the fundamental difference in interests between academic researchers and practitioners. Most often, academic researchers seek answers to why variable A predicts variable Z. Practitioners, however, tend to focus on which variables—A, B, or C—are stronger predictors of variable Z. The merits of the extant research notwithstanding, further compensation research can no longer focus narrowly on psychological phenomena. Indeed, it is more likely that the greatest contributions to our understanding of individual reactions to compensation practices will be generated by effectively assessing the context in which reactions occur. Researchers may consider the theoretical perspectives that were described earlier insofar as institutional and agency 362

factors probably affect how individuals will respond to compensation practices. STRATEGIC REWARDS AND COMPENSATION SYSTEMS The character of SHRM research since the mid-1980s has remained unchanged. Many of the models of SHRM were developed as generalizations of particular instances. This inductive mode of knowledge generation led some SHRM researchers to offer prescriptions about how to configure HRM strategy based on a set of contextual influences (Wright & McMahan, 1992). Moreover, another characteristic of this literature was to configure the content of these strategies (e.g., cost leadership vs. differentiation) so as to maximize the variance explained in organizational performance (Dyer, 1984). Although it is informative, inductive reasoning does not lend itself well to explaining the compensation strategies that address imminent business needs and unknown business development concerns under conditions of notable environmental turbulence. Similarly, as is discussed later, relying on the percentage of variance explained in organizational performance is also inappropriate. Inductive approaches to developing research questions are the predominant alternative to deduction. Research questions are suggested by behavior in specific examples or in findings from specific studies and then are used to make inferences about the general case. Inferences generated by induction do not follow logically or necessarily from the phenomenon to be explained (Cappelli, 1985). Also, there is only a weak logical basis to expect that past experience will be a reliable foundation on which to predict future events. For example, a number of strategic rewards and compensation research questions have been derived inductively (e.g., Schuler, 1986). The dearth of empirical assessments of these questions notwithstanding, such generalizations most likely do not hold across organizational contexts, largely due to the variability in environmental conditions over time. A problem with inductive research centers on the question, How does one know that events that occurred previously will be a reliable basis on which to explain or predict future events? According to the

Strategic Reward and Compensation Plans

realist perspective, science is governed by rigorous rules of logic, but science cannot be considered incontestable because these rules are socially constructed by the community of scientists (Cook & Campbell, 1979). Thus, the extraction of prescriptive statements from observations made across different times and organizations is problematic.

Liabilities of Prescription

Copyright American Psychological Association. Not for further distribution.

As noted by Mintzberg (1979), The orientation to as pure a form of description as possible has enabled us to raise doubts about a good deal of accepted wisdom [prescription]. . . . It is the literature of management that often emerges as naked, since much of what it says becomes transparent when held up to the scrutiny of descriptive research. (p. 583) Four specific shortcomings are associated with building a knowledge base solely on prescription. First, Brief and Dukerich (1991) noted that prescription tends to narrow the theorist’s vision. As we noted earlier, common to most SHRM research is the presumption that strategy determines the structure of HRM practices (e.g., Schuler & Jackson, 1987). For example, competitive strategies that require employee innovation and creativity could be well served by offering substantially higher merit pay increases to innovators (see chap 9, this volume). Unfortunately, the reality is that typical merit pay increase budgets provide for limited pay raise amounts that are of similar magnitude to increases in the cost of living (see Martocchio, 2009). Yet, a plausible alternative view is that the capabilities of organizations influence the choice of which competitive strategies to pursue (Lengnick-Hall & Lengnick-Hall, 1988). The inductive approach to studying SHRM and the prescriptions that ensue are likely to reinforce the status quo because induction is made post hoc with regard to strategy development and implementation as well as choice of HRM practices. It is not possible to observe from post hoc observations whether competitive strategy shapes HRM practices or whether interactive processes determine both competitive strategy and HRM practices.

A second limitation of relying on prescription for generating knowledge is a deemphasis of process knowledge (Brief & Dukerich, 1991). Becker and Huselid (2006) emphasized that much of the SHRM literature is based on content issues, instead of process issues, from which prescriptions for action are made. They argued that process research that focuses on the mechanisms and procedures used to identify, analyze, and ultimately decide on strategic issues lies at the heart of understanding. Dubin (1976) observed that practitioners who provide access to researchers are concerned with predicting outcomes (e.g., target profit margins, market share, customer satisfaction) and deem process knowledge as irrelevant. In fact, practitioners within the field tend to shape research agenda more than researchers influence practitioners’ agendas (Barley, Meyer, & Gash, 1988). Consequently, it appears that researchers’ attempts to acquire process knowledge will be difficult given this reality. A third shortcoming of prescription is a tendency to attempt to generalize to other contexts the conceptual conclusions that were derived inductively from a single case. Concepts based on a single instance may well not apply to other situations for two reasons. The first reason for restricted generalizability, obsolescence, is grounded in “transitory regularities, deriving from the existence of temporally restricted technological or institutional patterns” (Rescher, 1970, p. 156). In a related manner, Brief and Dukerich (1991) argued that although regularities do exist, the equivocal stability of such patterns renders probabilistic predictions infeasible. For example, until recently, it made good sense to study an individual’s decision to enroll in a 401(k) retirement savings plan. Since then, the Pension Protection Act of 2006 permits employers to automatically enroll new employees in the company’s 401(k) plan and allows employees to opt out at any time. Now, the focus on why employees choose to participate has been supplanted by the more relevant question pertaining to why some employees may choose to opt out of the plan. The second reason for limited generalizability, sampling, centers on the extent to which samples used in research are representative of the population of interest. Much of SHRM research involves detailed 363

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

examination of single organizations. Given the complexity of organizations due to variations in key characteristics such as size, life cycle, and industry sector, applying prescriptions based on a single organization to a heterogeneous set of organizations is inappropriate. For example, HRM architecture includes human capital. In competitive labor markets, organizations attempt to attract and retain the best individuals for employment partly by offering lucrative wage and benefits packages. However, systematic differences in compensation across industries make it difficult for some organizations to compete for the best talent. Interindustry differentials can be attributed to a number of factors, including the industry’s product market, the degree of capital intensity, and the profitability of the industry, unionization, and gender mix of the workforce (Osburn, 2000). Companies that operate in product markets where there is relatively little competition from other companies tend to pay higher wages because these companies exhibit substantial profits. This phenomenon can be attributed to such factors as higher barriers to entry into the product market and an insignificant influence of foreign competition. Government regulation and extremely expensive equipment represent entry barriers. The U.S. defense industry and the U.S. Postal Service have high entry barriers and no threats from foreign competitors. Capital intensity, the extent to which companies’ operations is based on the use of large-scale equipment, also explains pay differentials between industries. The amount of average pay varies with the degree of capital intensity. On average, capitalintensive industries (e.g., manufacturing) pay more than industries that are less capital intensive (e.g., service industries). Capital-intensive businesses require highly capable employees who have the aptitude to learn how to use complex technology. Service industries are not capital intensive, and most have the reputation of paying low wages. The operation of service industries depends almost exclusively on employees with relatively common skills rather than on employees with specialized skills to operate physical equipment such as casting machines or robotics. Unionization also influences compensation. Highly unionized industries tend to pay higher wages, on 364

average, than nonunion industries (and still do, notwithstanding noteworthy concessions by unionized employees in the commercial airlines and automobile manufacturing industries). At the same time, most highly unionized industries (e.g., manufacturing, construction, mining) are capital intensive, requiring employees with the aptitude to learn and use complex production technology. Following the earlier discussion of comparable worth, industries that employ substantially larger percentages of women than men tend to pay less. According to comparable worth proponents, women’s work is undervalued relative to men’s work in society. At the same time, most female-dominated industries appear in the service sector, such as in retail (e.g., store clerks) and education (e.g., primary school teachers). A fourth limitation of prescription is that the uncritical reader of the SHRM literature may take prescription as laws or definitive assessments of proposed action. Brief and Dukerich (1991) maintained that researchers should be self-critical and not allow “what might be” to be taken as “what will be.” As social scientists, we must recognize that our theories are incapable of supplying solutions to problems, and, at best, represent one of several inputs into the problem-solving process (Lindblom & Cohen, 1979). Financial, economic, and legal considerations are among some of the other inputs into the problem-solving process. For example, changes in financial accounting standards have prompted myriad organizations to eliminate the offer of employer-sponsored health care throughout retirement. Provisions in the Employee Retirement Income Security Act of 1974 enable financially troubled companies to terminate costly defined benefit pension plans.

Limitations of Emphasizing Variance Explained The main criteria for judging the merit of virtually all HRM research are (a) whether the observed relationships among the variables are statistically significant and (b) the amount of variance explained in the dependent variable by the independent variables. In this paradigm, more variance explained is considered to be better than less variance explained.

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

Without exception, all quantitative research on compensation has been based on the variance theory paradigm (Gerhart & Rynes, 2003). A variance theory is one that states (Mohr, 1982) that X is a necessary and sufficient condition for the outcome Y. In other words, Y is completely determined by X. By definition, compensation research lends itself well to variance theories. Compensation represents both the intangible and tangible rewards or returns employees receive for performing their jobs, and researchers have studied individual reactions or the influence of compensation practices on such outcomes as individual attitudes and job performance as well as on organizational performance (Gerhart & Rynes, 2003). Variance theories of compensation, then, suggest that the underlying mechanisms that drive outcomes, such as pay satisfaction, perceptions of justice, job performance, and organizational performance, are mechanistic (i.e., in conventional ordinary least squares regression, R2 has a maximum value of 1.0, and unexplained variance is determined by the equation 1 − R2). Thus, virtually all compensation researchers have constructed theories with the prime objective of maximizing the variance explained in the dependent variable by the independent variables, and random variance is considered error. Variance theories are best suited to studying issues of attitude change and individual job performance; however, they do not lend themselves well to the adoption of this paradigm for studying strategic rewards and compensation in turbulent environments. By nature, turbulent environments are antithetical to the orderliness of variance theories, representing stochastic shocks that are simply treated as error variance. For example, mergers and acquisitions, economic contraction, and lower priced products (of equal or better quality) are exogenous shocks that disrupt the expected positive influence of incentive pay systems on sales performance. Longitudinal studies may better enable to account for these shocks; however, longitudinal studies are not prevalent, perhaps due in part to the difficulty of ascertaining these data and the lack of control of rival alternative explanations. Besides the limitations of variance theories noted thus far, researchers should think carefully before judging the efficacy of their theories on the basis of R2.

Percent variance explained can be a misleading index of the influence of systematic factors on attitudinal and job performance criteria when small influences manifested at any instance (i.e., indexed by a relatively small R2) cumulate to produce a meaningful outcome over time (Abelson, 1985). Abelson used a baseball batting average analogy to illustrate that on any single batting attempt, a baseball batter may strike out, yet have an impressive batting average for the season. The point is that when a systematic factor such as batting skill operates, it may not be evident at any particular point in time because variables that are not part of the theory are likely to intervene and appear to undermine the expected. For instance, in the case of pay-for-performance, recessionary economic conditions often lead to reductions in an organization’s merit-pay increase budget. Such reductions stand to lower employee motivation for future performance in two ways. First, perceptions of underpayment inequity (Adams, 1965; Jacques, 1961) may intensify because budget reductions make it more challenging to recognize differences in employee performance. Exemplary performers may receive pay raises that are only slightly higher than average performers. Second, overall lower pay increases may fall below the threshold of justmeaningful pay differences (Zedeck & Smith, 1968), possibly exacerbating perceptions of inequity. CONCLUSIONS This review and critique suggests that the future is bright for all types of compensation research, subject to two considerations. First, researchers of psychological issues should include more the role of the meaning of money more prominently in the development of theoretical models, particularly in light of the global financial crisis, when organizations are reducing tangible compensation in various ways. Second, researchers interested in the strategic role of rewards and compensation should invest in learning the theoretical and empirical advances in the strategic management literature, just as psychologists have mastered the underpinning theories for employee motivation and behavior. For studying strategic rewards and compensation, researchers may consider adopting a shift in paradigm 365

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

from a variance theory approach to a process theory approach, adding qualitative research methodologies to their extant repertoire of quantitative research methodologies. Ultimately, understanding the strategic rewards and compensation phenomenon will come from getting closer to executive and managerial decision processes that underlie the unfolding of strategic rewards and compensation programs. Notwithstanding the focus on individuals or organizations, fruitful research will also likely result from understanding two salient contextual issues in employee benefits issues: generational diversity in the workforce, and the changing employment relationship. For a variety of reasons, including rising health care costs, employee benefits increasingly represent a larger proportion of total compensation expenditure and stand to become an important focus of compensation research in the future. A recent report suggests that employee benefits and reward systems are among the top 10 critical concerns facing organizations today (Center for Advanced Human Resource Studies [CAHRS], 2006). These findings are commonly noted elsewhere in surveys of practitioners that are sponsored by professional associations (e.g., the International Foundation of Employee Benefit Plans [http://www.ebri.org], the Society for Human Resource Management [http://www.shrm.org], and WorldatWork [http://www.worldatwork.org]). Various survey results suggest that although companies recognize the importance of their rewards and benefits programs for attracting talent, few companies have ascertained employee expectations regarding their benefits programs or have measured the return on investment from their benefit offerings.

Workforce Diversity Research on the effects of demographic characteristics on employee preferences, especially benefits preferences, has been scant. In addition, this research is at least a decade or more old and may not be reflective of the changed and changing workforce characteristics of today. Some current research has examined employee choice behaviors with regard to pension plans and found effects for organizational tenure (Dulebohn, Murray, & Sun, 2000), age (Gunderson & Luchak, 2001), and gender (Gunderson & Luchak, 2001). Research has also 366

found that preference for family-oriented benefits is influenced by age, gender, and number of children (Gunderson, Rozell, & Kellogg, 1995). Although early research provided only mixed support for the relationship between demographics and benefit preferences (Milkovich & Newman, 2008), there appears to be little current research examining the role of personal characteristics in determining preferences and needs for different kinds of benefits. A more extensive “rewards dialogue” between employers and employees is likely to make reward and benefits systems more responsive to employee needs. Also, companies hope that the set of employee benefits will motivate employees to perform their jobs as well as possible. This hope is based on a scientific assumption that employees have similar attitudes toward, and will be motivated similarly by, benefits systems. However, several demographic trends in the U.S. workforce call into question this assumption. Consider, for instance, the increasing median age of the working population, the growing dependency ratios—defined as the number of children and elderly per 100 working aged individuals—and the increasing life expectancy of present day workers (Dencker, Joshi, & Martocchio, 2007). These trends may influence employee attitudes regarding benefits and call for greater scrutiny of these systems and their ability to address the changing needs and expectations of employees. To motivate employees and elicit desired performance outcomes, companies need to take a more audience-driven approach to their benefits systems (i.e., employees form the audience). Such an approach would involve a greater scrutiny of the motivational potential of these systems in relation to the varying needs and expectations of employees (CAHRS, 2006). In the present business environment, a one-size-fits-all approach or flexible benefits programs under Internal Revenue Code, Section 125 (2008) may not be conducive to this audience-driven approach to employee benefits. Scientific models of employee motivation (e.g., goal setting theory, expectancy theory, equity theory) have been shown to be effective in explaining and predicting behavior where the context has been held constant. However, given the changing demographics of the U.S. workforce, the context has

Copyright American Psychological Association. Not for further distribution.

Strategic Reward and Compensation Plans

changed, and many of the assumptions underlying our reward and motivation theories may not be applicable in all organizational contexts. For instance, due in part to variation across generations in norms regarding the appropriate age of marriage, the age at which an employee becomes concerned with life insurance has likely increased over time. Thus, it is possible that employee benefits systems do not effectively motivate all employees to strive for or attain first-rate job performance. For example, older employees without dependent children are unlikely to set higher work goals, put forth greater effort, or feel that the benefits offerings are equitable because their employer provides various family-friendly programs. An alternative explanation is the length of the period between performance and reward (e.g., it may be decades before an employee receives the rewards of lifetime retirement income). In short, although the demographic composition of the U.S. labor force has undergone a major transformation in the 20th century, assumptions underlying employee benefits policies remain largely unchanged.

Changing Employment Contract The concept of psychological contracts holds that in exchange for their services, employees believe they will receive something from the organization, compensation being the most basic form of such an exchange (Rousseau, 1995; see also Vol. 3, chap. 5, this handbook.) Although a few employee benefits are legally mandated, organizations can exercise a lot of discretion in the range and type of benefits they offer. Psychological contracts are highly subjective beliefs about what the organization should provide in return for employment and these may not necessarily be enforceable (Rousseau, 1995). Psychological contracts implicitly establish terms of employment. This is in contrast to more explicit economic exchange agreements such as wage or salary levels. Thus, as an example, company policies might imply that an employee will be eligible for educational assistance after 5 years of continuous employment and satisfactory levels of performance. An employee who is interested in making use of this benefit would reciprocate by remaining with the company and working hard. Thus, the employee’s psychological contract with the company would

include the employee’s obligations (i.e., 5 years of hard work) and the employer’s obligations (i.e., educational assistance). Recall the discussion earlier in the chapter about the evolving nature of social exchange. Psychological contracts are a part of the social exchange process in the employment relationship. Psychological contracts result in employees holding a range or continuum of expectations from the employer, from pay and promotions to career development and family welfare. The continuum of expectations that employees hold of an employer can be seen as having two poles: transactional psychological contracts at one end and relational psychological contracts at the other. Toward the transactional end of the continuum, employees’ expectations of the employer are more economic and extrinsic in nature. Thus, employees’ expectations of high pay and promotions or career advancement in exchange for hard work would represent transactional types of expectations in the psychological contract. However, toward the relational end, employees’ expectations of the employer may be either economic or noneconomic but are also emotional, subjective, and intrinsic in nature. Thus, employees’ expectations of job security in exchange for loyalty to the employer would represent relational types of expectations in the psychological contract. Psychological contracts that are transactional in nature can be understood with an example of short-term employment. An independent contractor or consultant hired by an organization is more likely to have transactional expectations from the hiring organization. The independent consultant or contractor would expect the organization to provide good pay as well as the opportunity to build their marketability by adding the organization to their client portfolio. Once the project or assignment for which the independent contractor was hired is completed, the exchange relationship with the organization may end. Relational expectations can be understood by looking at the employee–employer relationship. Employees hired by a company or organization with the understanding of full-time employment are more likely to hold both transactional and relational expectations of their employer. For instance, not only will such employees expect pay, promotion, and career advancement in exchange for 367

Copyright American Psychological Association. Not for further distribution.

Joseph J. Martocchio

work efforts, but they will also expect job security, recognition, and support in exchange for commitment and loyalty to the employer. Employee benefits practices can be seen to fulfill both transactional and relational expectations of employees. Some employee benefits might fulfill more transactional expectations. For example, as suggested earlier, American employees might expect employers to provide health insurance in addition to wages. There might also be legally required benefits that the employer would be required to provide to employees. Thus, health insurance and other legally required benefits would form a part of employees’ transactional expectations from the employer. Employee benefits might also fulfill employees’ relational expectations. For instance, employee benefits such as the paid time off and accommodation and enhancement benefits (e.g., day care and transportation services) might help fulfill employees’ relational expectations. An employee might expect paid vacation time for completing a certain length of continuous employment or might start expecting family welfare practices to be provided to employees who have been with the organization for more than 1 year. In addition, some employee benefits might fulfill both transactional and relational expectations of employees. Retirement plans are a good example of employee benefits that fulfill both transactional and relational expectations. For instance, employees might not only expect employers to enroll them in a retirement plan when they are hired but also to make increasingly larger contributions as they stay committed and loyal to the company. This would increase their sense of security from the employment relationship. Similarly, educational assistance benefits aimed at rewarding continued employment as well as career development would help fulfill both transactional and relational expectations. The traditional employment relationship was characterized by an expectation of mutual long-term commitment between employers and employees, strong internal promotion and employee development practices, and interest in promoting equitable compensation. Times have changed. Nowadays, competitive pressures encourage employees and employers to share a mutual expectation that the employment 368

relationship is temporary. For example, many companies are adopting pension programs that by design discourage long-term employment. Specifically, there is a shift in many companies from the use of traditional defined benefit pension plans to defined contribution plans. Typically, companies contribute an amount equal to a fixed percentage of an employee’s annual pay to the defined benefit pension plan. Over time, an employee’s pay will rise as he or she receives permanent pay increases for earning promotions, better performance, or gaining new knowledge or skills. As a result, the actual dollar amount contributed to an employee’s retirement plan increases, though the percentage remains constant. Defined contribution plans are much like a savings account balance that is invested in a variety of ways chosen by employees. Employers are not required to make contributions to these plans, nor are they responsible for guaranteeing a minimum account balance or income from the plan at any time. These features make it easy for employees to move from one company to the next without any consequence for investment growth over time. The changing employment contract also has implications for the relevance of particular benefits to younger employees who are not likely to have expectations of long-term employment. For example, younger employees stand to view life insurance and health insurance benefits as less valuable because the likelihood of using such benefits is not as great as it is for older workers who, statistically speaking, are more likely to die or become seriously ill. Instead, younger workers are likely to consider those benefits addressing priorities fitting their life stage to possess greater value, for instance, day care assistance and tuition reimbursement programs. In conclusion, my goal in this chapter was to chronicle the changing landscape of compensation and the signals that indicate a far more strategic role for compensation than it has ever previously enjoyed. The major research tasks still to be accomplished are (a) to identify the compensation decisions that are strategic; (b) to develop descriptions of these decisions; and (c) to determine whether (and how) compensation strategies affect workforce behaviors, which, in turn, affect the implementation of competitive strategy.

Strategic Reward and Compensation Plans

References Abelson, R. P. (1985). A variance explanation: When a little is a lot. Psychological Bulletin, 97, 129–133. doi:10.1037/0033-2909.97.1.129 Adams, J. S. (1965). Injustice in social exchange. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 267–299). New York: Academic Press.

Copyright American Psychological Association. Not for further distribution.

Barley, S. R., Meyer, G. W., & Gash, D. C. (1988). Cultures of culture: Academics, practitioners, and the pragmatics of normative control. Administrative Science Quarterly, 33, 24–60. Barney, J. (2001). Is the resource-based “view” a useful perspective for strategic management research? Yes. Academy of Management Review, 26, 41–56. doi:10.2307/259393 Baron, J. N., Dobbin, F., & Jennings, P. D. (1986). War and peace: The evolution of modern personnel administration in U.S. industry. American Journal of Sociology, 92, 350–383. doi:10.1086/228504 Becker, B. E., & Huselid, M. A. (2006). Strategic human resources management: Where do we go from here? Journal of Management, 32, 898–925. doi:10.1177/ 0149206306293668 Beer, M., Spector, B. Lawrence, D. Q., Mills, D. Q., & Walton, R. E. (1984). Managing human assets: The groundbreaking Harvard Business School Program. New York: Free Press. Better health care together. (n.d.). Retrieved October 11, 2009, from http://www.betterhealthcaretogether.org/ ourvision Brief, A. P., & Dukerich, J. M. (1991). Theory in organizational behavior: Can it be useful? Research in Organizational Behavior, 13, 327–352. Cappelli, P. (1985). Theory construction in IR and some implications for research. Industrial Relations, 24, 90–112. doi:10.1111/j.1468-232X.1985.tb00982.x Cascio, W. F., & Aguinis, H. (2008). Research in industrial and organizational psychology from 1963 to 2007: Changes, choices, and trends. The Journal of Applied Psychology, 93, 1062–1081. doi:10.1037/0021-9010. 93.5.1062 Center for Advanced Human Resource Studies. (2006, July). CAHRS top ten list, July 2006, Center for Advanced Human Resource Management Research. Retrieved August 30, 2006, from http://www.ilr.cornell.edu/cahrs

Costo, S. L. (2006, February). Trends in retirement plan coverage over the last decade. Monthly Labor Review, 129(2), 58–64. Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281–302. doi:10.1037/h0040957 Cutcher-Gershenfeld, J. (1991). The impact on economic performance of a transformation in workplace relations. Industrial & Labor Relations Review, 44, 241–260. doi:10.2307/2524806 Dencker, J. C., Joshi, A., & Martocchio, J. J. (2007). Employee benefits as context for intergenerational conflict. Human Resource Management Review, 17, 208–220. doi:10.1016/j.hrmr.2007.04.002 Dewey, J. (1933). How we think. Boston: Heath. Dubin, R. (1976). Theory building in applied areas. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 17–39). Chicago: Rand McNally. Dulebohn, J. H., Murray, B., & Sun, M. (2000). Selection among employer-sponsored pension plans: the role of individual differences. Personnel Psychology, 53, 405–432. Dulebohn, J. H., & Werling, S. E. (2007). Compensation research past, present, and future. Human Resource Management Review, 17, 191–207. doi:10.1016/ j.hrmr.2007.03.002 Dyer, L. (1984). Studying human resource strategy: An approach and agenda. Industrial Relations, 23, 156–169. doi:10.1111/j.1468-232X.1984.tb00894.x Employee Retirement Income Security Act of 1974, Pub. L. No. 93–406, § 88 Stat. 829 (1974). Encyclopedia Britannica. (2007). Hawthorne research. Retrieved May 15, 2008, from Encyclopedia Britannica Online: http://search.eb.com/eb/article? eu=40460 England, P. (1992). Comparable worth: Theories and evidence. New York: Aldine de Gruyter. Equal Pay Act, 29 U.S.C.A. § 206(d) (1963). Fair Labor Standards Act of 1938, 29 U.S.C.A. § 201, et seq. (1938). Franke, R. H., & Kaul, J. D. (1978). The Hawthorne experiments: First statistical interpretation. American Sociological Review, 43, 623–643. doi:10.2307/2094540

Civil Rights Act of 1964 (Title VII), 78 Stat. 253 (1964).

Gerhart, B., & Milkovich, G. T. (1992). Employee compensation: Research and practice. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., pp. 29–41). Palo Alto, CA: Consulting Psychologists Press.

Cook, T. D., & Campbell, D. T. (1979). Quasiexperimentation: Design & analysis issues for field settings. Boston: Houghton Mifflin.

Gerhart, B., & Rynes, S. L. (2003). Compensation: Theory, evidence, and strategic implications. Thousand Oaks, CA: Sage.

Chandler, A. D., Jr. (1962). Strategy and structure. Cambridge: MIT Press.

369

Joseph J. Martocchio

Goldin, C. (1991). The role of World War II in the rise of women’s employment. The American Economic Review, 81, 741–747. Gunderson, D. E., Rozell, E. J., & Kellogg, C. E. (1995). Family supportive organizational benefits as influences on entry level job preferences: An empirical analysis using a policy capturing methodology. Benefits Quarterly, 11, 58–68. Gunderson, M., & Luchak, A. (2001). Employee preferences for pension plan features. Journal of Labor Research, 22, 795–808.

Copyright American Psychological Association. Not for further distribution.

Heneman, R. L., & Werner, J. M. (2005). Merit pay: Linking pay to performance in a changing World. Greenwich, CT: Information Age. Herskovits, M. J. (1955). Cultural anthropology. New York: Knopf. Huselid, M. A. (1995). The impact of human resource management practices on turnover, productivity, and corporate financial performance. Academy of Management Journal, 38, 635–672. doi:10.2307/256741 Huselid, M. A., Beatty, R. W., & Becker, B. E. (2005, December). “A players” or “A positions?” The strategic logic of workforce management. Harvard Business Review, 83(12), 110–117. Ichniowski, C., Shaw, K., & Prennushi, G. (1997). The effects of human resource management practices on productivity: A study of steel finishing lines. The American Economic Review, 87, 29–51. Internal Revenue Code § 61 (2000). Internal Revenue Code § 125 (2008). Jacques, E. (1961). Equitable payment. New York: Wiley. Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3, 305–360. doi:10.1016/0304-405X(76)90026-X

as a symbol and reality in psychotherapy and psychoanalysis (pp. 3–16). New York: Brunner/Mazel. Lawrence, P. R., & Lorsch, J. W. (1967). Organization and environment: Managing differentiation and integration. Boston: Graduate School of Business Administration, Harvard University. Ledbetter v. Goodyear Tire & Rubber Co., No. 05–1074 (U.S. Supreme Court, May 29, 2007). Lemons v. City and County of Denver, 620 F.2d 228 (1980). Lengnick-Hall, C. A., & Lengnick-Hall, M. L. (1988). Strategic human resources management: A review of the literature and proposed typology. Academy of Management Review, 13, 454–470. doi:10.2307/ 258092 Lepak, D. P., & Snell, S. A. (1999). The human resource architecture: Toward a theory of human capital allocation and development. Academy of Management Review, 24, 31–48. doi:10.2307/259035 Lepak, D. P., & Snell, S. A. (2002). Examining the human resource architecture: The relationships among human capital, employment, and human resource configurations. Journal of Management, 28, 517–543. Lily Ledbetter Fair Pay Act of 2009, Pub. L. No. 111–2 (2009). Lindblom, C. E., & Cohen, D. K. (1979). Usable knowledge: Social science and social problem solving. New Haven, CT: Yale University Press. MacDuffie, J. P. (1995). Human resource bundles and manufacturing performance: Organizational logic and flexible production systems in the world auto industry. Industrial & Labor Relations Review, 48, 197–221. doi:10.2307/2524483 Martocchio, J. J. (2009). Strategic compensation: A human resource management approach (5th ed.). Upper Saddle River, NJ: Pearson Education.

Kane, J. S. (1991). Toward a modernized model of science. Human Resource Management Review, 1, 245–251. doi:10.1016/1053-4822(91)90001-S

Mayo, E. (1933). The human problems of an industrial civilization. New York: MacMillan.

Kaufman, B. E. (2008). Managing the human factor: The early years of human resource management in American industry. Ithaca, NY: Cornell University Press.

Meyer, J. W., & Rowan, E. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83, 340–363. doi:10.1086/226550

Kirkcaldy, B., & Furnham, A. (1993). Predictors of beliefs about money. Psychological Reports, 73, 1079–1082. Kochan, T. R., Katz, H. C., & McKersie, R. B. (1994). The transformation of American industrial relations. Ithaca, NY: ILR Press. Kovac, J. C. (2008, June 27). Merit, COLA and the economy. Retrieved July 8, 2008, from http://www.worldatwork. org/waw/adimComment?id=26953 Krueger, D. W. (1986). Money, success, and success phobia. In D. W. Krueger (Ed.), The last taboo: Money 370

Milkovich, G. T. (1988). A strategic perspective on compensation management. In K. Rowland & G. Ferris (Eds.), Research in Personnel and Human Resources Management (3rd ed., pp. 263–288). Greenwich, CT: JAI Press. Milkovich, G. T., & Newman, J. M. (2008). Compensation (9th ed.). Boston: Irwin/McGraw-Hill. Mintzberg, H. (1979). An emerging strategy of “direct” research. Administrative Science Quarterly, 24, 582–589. doi:10.2307/2392364

Strategic Reward and Compensation Plans

Mintzberg, H., & Waters, J. A. (1985). Of strategies, deliberate and emergent. Strategic Management Journal, 6, 257–272. doi:10.1002/smj.4250060306 Mitchell, T. R., & Mickel, A. E. (1999). The meaning of money: An individual-difference perspective. Academy of Management Review, 24, 568–578. doi:10.2307/259143 Mohr, L. B. (1982). Explaining organizational behavior. San Francisco: Jossey-Bass.

Copyright American Psychological Association. Not for further distribution.

National Committee on Pay Equity. (1987). Job evaluation: A tool for pay equity. Washington, DC: Author. Osburn, J. (2000, February). Interindustry wage differentials: Patterns and possible sources. Monthly Labor Review, 123, 34–46. Parsons, T. (1967). Sociological theory and modem society. New York: Free Press. Paycheck Fairness Act, H.R. 1338, 110th Cong. § 2(2) (2008). Pension Protection Act of 2006, Pub. L. No. 109–280, § 120 Stat. 780 (2006). Person, H. S. (1929). The new attitude toward management. In H. S. Person (Ed.), Scientific management in American industry (pp. 65–75). New York: Harper & Brothers. Pfeffer, J. (1993). Competitive advantage through people: Unleashing the power of the work force. Boston: Harvard Business School Press. Pfeffer, J. (1995). Producing sustainable competitive advantage through the effective management of people. The Academy of Management Executive, 9, 55–69. Pregnancy Discrimination Act, 42 U.S.C.A. § 2000e et seq. (1974). Rescher, N. (1970). Scientific explanation. New York: The Free Press. Robert, C., Probst, T., Martocchio, J. J., Drasgow, F., & Lawler, J. J. (2000). Empowerment and continuous improvement in the United States, Mexico, Poland, and India. The Journal of Applied Psychology, 85, 643–658. Medline doi:10.1037/0021-9010.85.5.643 Rousseau, D. M. (1995). Psychological contracts in organizations: Understanding written and unwritten agreements. San Francisco: Sage. Rynes, S. L., & Gerhart, B. (2000). Compensation in organizations: Current research and practice. San Francisco: Jossey-Bass. Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437–453. doi:10.1111/j.1744-6570. 1987.tb00609.x Schneider, B., Goldstein, H. W., & Smith, D. B. (1995). The ASA framework: An update. Personnel Psychology, 48, 747–773. doi:10.1111/j.1744-6570.1995.tb01780.x

Schneider, B., Smith, D. B., Taylor, S., & Fleenor, J. (1998). Personality and organizations: A Test of the homogeneity of personality hypothesis. The Journal of Applied Psychology, 83, 462–470. doi:10.1037/0021-9010.83.3.462 Scholl, R., Cooper, E. A., & McKenna, J. (1987). Referent selection in determining equity perceptions: Differential effects on behavior and attitudinal outcomes. Personnel Psychology, 40, 113–124. doi:10.1111/j.1744-6570.1987.tb02380.x Schoonhoven, C. B. (1981). Problems with contingency theory: Testing assumptions hidden within the language of contingency theory. Administrative Science Quarterly, 26, 349–377. doi:10.2307/2392512 Schuler, R. S. (1986). Fostering and facilitating entrepreneurship in organizations: Implications for organization structure and human resource practices. Human Resource Management, 25, 607–629. doi:10.1002/hrm.3930250408 Schuler, R. S., & Jackson, S. E. (1987). Linking competitive strategies with human resource management practices. Academy of Management Executive, 1, 207–219. Scott, W. R. (1987). The adolescence of institutional theory. Administrative Science Quarterly, 32, 493–511. doi:10.2307/2392880 Social Security Act, 42 U.S.C.A. § 301 et seq. (1935). S. Rep. No. 176 at 1 (1963). Tang, T. L. P. (1992). The meaning of money revisited. Journal of Organizational Behavior, 13, 197–202. doi:10.1002/job.4030130209 Toh, S. M., Morgeson, F. P., & Campion, M. A. (2008). Human resource configurations: Investigating fit with the organizational context. The Journal of Applied Psychology, 93, 864–882. doi:10.1037/ 0021-9010.93.4.864 Triandis, H. C. (1994). Cross-cultural industrial and organizational psychology. In H. C. Triandis, M. Dunnette, & L. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 4, pp. 103–172). Palo Alto, CA: Consulting Psychologists Press. U.S. Bureau of the Census. (2009). Statistical abstracts of the United States: The national data book. Retrieved October 11, 2009, from http://www.census.gov U.S. Bureau of Labor Statistics. (1919). Welfare work for employees in industrial establishments in the United States. Bulletin, 250, 119–123. U.S. Bureau of Labor Statistics. (2009). National compensation survey. Retrieved October 11, 2009, from http://www.bls.gov/eci/ Vroom, V. (1964). Work and motivation. New York: Wiley. 371

Joseph J. Martocchio

Wright, P. M., & Snell, S. A. (1998). Toward a unifying framework for exploring fit and flexibility in strategic human resources management. Academy of Management Review, 23, 756–772. doi:10.2307/259061

Wright, P. M., & McMahan, G. C. (1992). Theoretical perspectives for strategic human resource management. Journal of Management, 18, 295–320. doi:10.1177/014920639201800205

Zedeck, S., & Smith, P. C. (1968). A psychophysical determination of equitable payment: A methodological study. The Journal of Applied Psychology, 52, 343–347. doi:10.1037/h0026245

Wright, P. M., & Snell, S. A. (1991). Toward an integrative view of strategic human resource management. Human Resource Management Review, 1, 203–225. doi:10.1016/1053-4822(91)90015-5

Zhou, J., & Martocchio, J. J. (2001). Chinese and American managers’ compensation award decisions: A comparative policy-capturing study. Personnel Psychology, 54, 115–145. doi:10.1111/j.1744-6570.2001.tb00088.x

Copyright American Psychological Association. Not for further distribution.

Werner, S., & Ward, S. (2004). Recent compensation research: An eclectic review. Human Resource Management Review, 14, 201–227. doi:10.1016/ j.hrmr.2004.05.003

372

CHAPTER 12

PERSPECTIVES ON ORGANIZATIONAL CLIMATE AND CULTURE

Copyright American Psychological Association. Not for further distribution.

Benjamin Schneider, Mark G. Ehrhart, and William H. Macey

The chapter is about two siblings. They share much in common but have typically failed to see the benefits that might accrue to them if they learned more about each other and the ways that each could complement its own strengths to be more effective. The siblings go by the names organizational climate and organizational culture. They emerged from different academic traditions and times, but their focus—in the psychological life of organizations—has produced perspectives with oddly great conceptual similarity and increasingly similar research tactics, similarities that some culture researchers frequently deny although there have been attempts at rapprochement (Denison, 1996; Reichers & Schneider, 1990; Schneider, 2000). In what follows, we first review the history and current status of research and theory on organizational climate, the older sibling. Organizational climate inherited much of its psychological parents’ heritage with a focus on psychological issues involving perception, affect, and attitudes. Organizational climate has been variously defined—more on this later—but for present purposes, climate concerns the policies, practices, and procedures as well as the behaviors that get rewarded, supported, and expected in a work setting and the meaning those imply for the setting’s members (Schneider & Reichers, 1983; Schneider, White, & Paul, 1998). The younger sibling, organizational culture, arrived later than climate, having inherited characteristics associated more with sociology and anthropology and thus a focus on the collective rather than the individual. Organizational culture has also had a

variety of definitions, as we see later, but for now we define culture as beliefs, ideologies, and values, and the ways these are transmitted through symbols, language, narratives (myths, stories), and practices (rituals and taboos) especially during socialization to the workplace (Trice & Beyer, 1993). We review this sibling as well from both research and conceptual vantage points. As the chapter unfolds, and especially in the section on organizational culture, we suggest approaches to research and theory that could yield some rapprochement for the siblings. We do this because it is so clear to us that if there were more “sharedness” (to coin a term) across the boundaries that have been created, there is much to be gained for an understanding of the ways climate and culture define organizations for the people who work in them and determine organizational effectiveness. ORGANIZATIONAL CLIMATE In this section, we present an overview of the early history of climate research as well as the conceptual and research methods issues that have attracted thinking and research. These latter issues are presented in two subsections. The first subsection outlines the emergence of the issues in some detail (levels of analysis issues, the focus of climate research, and the dimensions of climate), and the second subsection presents the more recent conceptual and research approaches that have resolved many of the issues with which earlier researchers struggled.

http://dx.doi.org/10.1037/12169-012 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

373

Schneider, Ehrhart, and Macey

Copyright American Psychological Association. Not for further distribution.

Early Conceptual Background to the Study of Organizational Climate: From Lewin, Lippitt, and White (1939) to Likert (1967) Lewin, Lippitt, and White (1939) studied the social climate of boys’ groups in the late 1930s. In this research, Lewin and his colleagues demonstrated that boys who worked under a more democratic (participative) leadership style produced as much as those who were taught under an authoritarian style. Of considerable interest was the finding that under democratic leadership conditions, the boys behaved more cooperatively with each other and more openly with the teacher (leader), experienced less stress, and appeared more positive about their experience (they smiled more and chatted with each other more). Lewin and his colleagues thought of these different leadership styles as tactics for creating a sense in the boys about how they should behave— cooperatively versus individually—and feel—positive versus negative. The leader did this without telling the boys how to behave or what to feel; the boys sensed it based on both the leader’s behavior and each other’s behavior. Lewin et al. called the confluence of the behaviors and attitudes that were a result of the leader behaviors a social climate as a way to capture the diverse patterns of social behavior that emerged. It is important for our purposes here to note that they did not call the result just a “climate” in the abstract, but a “social climate” as a way to capture the particular kind of climate that was created in the boys’ groups. It is also important to note that they did not measure social climate in any formal way (e.g., with a survey). While Lewin et al. (1939) were interested in how the young boys behaved and felt when confronted with a specific leadership style, it was the work of the founders of organizational psychology in the late 1950s and early 1960s that stimulated the application of the climate idea to business and industry. That is, until the writings of people such as Argyris (1957), McGregor (1960), Likert (1961), Schein (1965), and Katz and Kahn (1966), the closest industrial psychologists came to a climate idea was the issue of employee morale (Viteles, 1953). By this, we mean that industrial psychologists were generally unconcerned with the ways employees experienced the organizations in which they worked except for 374

the concepts of attitudes and morale. So, for example, Tiffin’s (1946) classic text Industrial Psychology had long sections on testing, vision problems in industry, fatigue, and accident prevention but only 15 pages in a concluding chapter to summarize all of the work on employee attitudes and morale. In contrast, Argyris (1957) wrote in his book Personality and Organization about how employees were infantilized by modern industry practices and reacted to this by behaving as children, just as management expected them to. Although he did not use the phrase, we might today speak of the “climate for infantilization.” McGregor (1960) proposed that “managerial cosmology”—the way managers believe employees are motivated—determines their behavior toward employees, which in turn affects employee behavior. Thus, in his famous Theory X and Theory Y he illuminated the idea that managerial behavior toward employees creates a “managerial climate” (his phrase) and that this climate determines employee behavior. It is, of course, interesting to note that both Argyris and McGregor were students of Lewin, so the concern for the psychological environments created by managers for employees is clear. At the University of Michigan, Lewin also had an influence on the work of Likert (1961) and Katz and Kahn (1966). Likert, in particular, conceptualized the important role of leadership as a determinant of employee experiences and organizational effectiveness. In many ways, Likert’s (1961) conceptualization was similar to that of McGregor’s: Reliance is not placed solely or fundamentally on the economic motive of buying a man’s time and using control and authority as the organizing and coordinating principle of the organization. On the contrary . . . [t]he full strength of all economic, ego, and other motives is generated and put to use. (pp. 99–100) Likert’s (1967) later book took his thinking a step further toward a more detailed specification of different systems of organizing vis-à-vis employees in pursuit of organizational effectiveness. In the 1961 book, Likert introduced four systems by which organizations might function, generally arrayed

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

along a continuum from completely autocratic to completely participative. In the 1967, book he labeled these System 1 (exploitative autocratic), System 2 (benevolent authoritative), System 3 (consultative), and System 4 (participative group); today we might call his System 1–System 4 typology different levels of a “climate for productivity through employee participation.” As a sidelight, it is interesting to note that in the 1967 volume is found one of the earliest, if not the earliest, formal statements of human asset accounting going as far as delineating, for example, how to value the human assets of a firm as a function of the investments made in them via recruiting, selection, and training and of the costs to replace present workers (see Miner, 2002, for an extended discussion of Likert’s theory). Thus Argyris, McGregor, and Likert were concerned with how humans were treated by organizations and how they responded to that treatment as a way to understand organizational effectiveness and not as a means for understanding human behavior in the abstract. Their goal was not for employees to feel good or to achieve states of self-actualization; instead, they were interested in understanding how employees experienced their work organizations when those organizations were effective, and through their research and theory, deduced that treating people as humans with human feelings was a key to organizational effectiveness. Likert had brought social psychologist Daniel Katz from Princeton to Michigan, and it was his background in social psychology that influenced the work Katz did with Robert Kahn (who obtained his degree and spent his entire career at Michigan) on The Social Psychology of Organizations (Katz & Kahn, 1966). This book, more than the earlier works by Argyris and McGregor and to a lesser extent Likert, emphasized the total social situation encountered by employees rather than a more focused leadership perspective. The importance of this emphasis was that organizations were viewed as total systems, not only in interaction with their subparts but also in interaction with the external world from which they drew resources and to which they provided output. It is useful to note that Katz and Kahn freely used alternative terms in describing the importance of social psychological concepts as vehicles for under-

standing the psychology of organizations. On page 50 of the 1978 edition of their book, Katz and Kahn (1978) used each of the following terms to capture the essence of the organization as a social psychological enterprise: norms, values, roles, climate, culture, subculture, collective feelings and beliefs, atmosphere, taboos, folkways, and mores. To our mind, this is a useful listing because it points to ways in which terminology from both climate research and culture research literatures might be simultaneously used to capture a broad range of related phenomena. It is critical to understand that at this point in the history of the psychological study of work organizations, the experience of workers and the role of those experiences in contributing to organizational effectiveness were just beginning to be understood. Schein (1965) is the author of probably the first book called Organizational Psychology, and it was his book that insightfully summarized much of the conceptual work that had been going on up to that time. Schein was and still is at the Massachusetts Institute of Technology, where Lewin had been prior to the University of Michigan and where McGregor was one of his colleagues. Schein (1965) wrote in this first statement of the field of organizational psychology: “The material covered in this book will reflect the general historical trend from an individualoriented industrial psychology toward a group- and systems-oriented organizational psychology” (p. 5). His statement was overly optimistic because the route to a more group- and systems-focused field of organizational psychology is still unfolding to this day and, as we see later in the chapter, even in the study of organizational climate took many decades to occur.

Summary By the late 1960s and early 1970s, a rich conceptual history of what came to be called organizational psychology existed. The focus of scholars in this field was on the human issues surrounding organizational effectiveness, especially on the roles of leadership and the larger social system in which people worked. While the theories had many psychological concepts in them—perception, motivation, attitudes—the focus was on the design of organizations that were effective through collective human attitudes and 375

Schneider, Ehrhart, and Macey

action and not on individual employees as the unit of theory or analysis.

Copyright American Psychological Association. Not for further distribution.

Early Conceptual and Research Dilemmas: Levels of Analysis, Dimensions, and Focus Working in this relatively new area of research, scholars naturally followed their training and their interests without much to draw on in the way of conceptual and methodological approaches. So, faced with this ambiguity, researchers did work at both the individual and the unit levels of analysis, on dimensions of climate that varied from study to study and that implicitly focused on outcomes of interest not tightly tied conceptually to the climate they were studying; that is, there was little focus in the measures developed and used against specific criteria of interest. As reviews of the research being published began to accumulate, the dilemmas surrounding these issues became clear. In this section we illuminate them, and in the next section we show how many of them were resolved through discussion and hard conceptual and methodological work. Levels of analysis issues. For the most part, the early survey research on organizational climate focused on individual responses to climate surveys or individual responses to experimentally created situations or conglomerations of objective organizational attributes (e.g., structure and size; Payne & Pugh, 1976), as well as more perception-based experiential indicators of the ways humans experienced their organizations (e.g., leader support; Forehand & Gilmer, 1964). In these cases, the organizational climate data were in turn related to the individual behavior of the respondents and not to organizational effectiveness. For example, in contrast to the Michigan researchers (e.g., Seashore, 1964; Tannenbaum, 1962), who collected survey data from individual employees and related those data in the aggregate to organizational performance, others who began to study organizational climate did not make the step early on to aggregate individual data and relate it to organizational effectiveness. For example, Schneider and Bartlett (1968, 1970) developed a climate survey and retained the individual as the unit of analysis for both the climate and the outcomes (life insurance agent sales) of interest. As another example, consider 376

the development of the Litwin and Stringer (1968) measure of climate, which focused on individual differences in responses to experimentally manipulated work settings as a function of those individuals’ need profiles. The surveys used in these early research efforts were surprisingly similar to contemporary surveys in that they asked respondents to report on generic environmental practices and procedures, not on how they personally felt about them. The items asked about many features of the environment with a focus on the social characteristics of that environment (including leadership, coworkers, and conflict). For example, in Schneider and Bartlett’s (1968, 1970) survey there was not an explicit theory of organizations guiding the effort but an attempt to capture a wide variety of what employees experienced generally surrounding the issue of the social environment created for them. Sample items included the following: (a) for the assessment of Managerial Support (one of their six dimensions), “Managers take an active interest in the progress of their agents”; (b) for the assessment of New Employee Concerns, “Agents receive sufficient field training prior to being left on their own”; and (c) for Intra-Agency Conflict, “There are definite agent cliques within the agency.” The response scale used concerned how characteristic the item was of their workplace: never, slightly, sometimes, considerably, always. It is clear from these examples that the environment and not personal responses to that environment was the target, but the outcome of interest remained individual performance. L. R. James and Jones (1974) were perhaps the early and most explicit commentators on the issue of levels of theory and levels of analysis, and they coined the term psychological climate to refer to studies in which the unit of data collection as well as the unit of analysis was the individual versus organizational climate when the data were collected from individuals and were aggregated to reflect an organizational attribute (climate). In that article, they also characterized psychological climate as an intervening process between the organization’s attributes and individual attitudes and behaviors. In his recent writings, L. R. James (in L. R. James

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

et al., 2008) took a perhaps more extreme vantage point on the centrality of individuals’ affective processing of environmental information than he did earlier. Thus the newer review proposed “that psychological climate should be strongly related to affective variables” (L. R. James et al., 2008, p. 13). The potential role of affect in climate perceptions led to Guion’s (1973) critique that if climate is in the heads of individuals, are not such perceptions the same as any other job attitude, like job satisfaction? The answer of course is no—it is not the same. Indeed, LaFollette and Sims (1975) and Schneider and Snyder (1975) showed that satisfaction was not the same as climate by demonstrating that classical dimensions of job satisfaction were at best moderately related to what were coming to be classical dimensions of organizational climate. This was apparently at least partially a function of the fact that climate items were relatively pure in their descriptions of external characteristics (like those presented earlier), whereas job satisfaction items were more evaluative and personal in their focus and/or confounded the descriptive with the evaluative. For example, the classic job satisfaction measure of Smith, Kendall, and Hulin (1969) asks respondents to evaluate their leaders by asking if they are “impolite,” “tactful,” and “annoying,” along with some more descriptive items like “Asks my advice,” “Tells me where I stand,” and “Leaves me on my own.” But note how even the descriptive items are personally referenced (me, my) rather than asking what the leader generally does. When the items are presented in the same survey with the same directions and studied at the individual level of analysis, they will, of course, yield significant relationships among them. Separating out the descriptive from the evaluative is difficult, but the conclusion that they are assessing the same thing would be erroneous. Other perspectives that came along in the 1980s (Ashforth, 1985; Glick, 1985) helped to clarify some of these levels of analysis and conceptual issues. Ashforth (1985) proposed that the question of the locus of climate research (individual perceptions, structure, process) is not as important as the fact that it is the shared perceptions and meaning attached to those perceptions that are important. Glick (1985) put the issue this way:

The conceptual morass that grew out of attempts to understand organizational climate was criticized by Guion (1973) who emphasized the unit of theory problem in organizational climate research. Was organizational climate to be conceptualized as an individual or an organizational attribute? (p. 601) For Glick, the unit of theory for organizational climate research was the organization (or subunit), not the individual. Glick succinctly argued that unless (a) climate survey items assessed organizational functioning, (b) the data were aggregated to the organizational level of analysis, and (c) the climate measurement was focused on important organizational outcomes, then climate research was no different from other individual-level attitudinal research. Consistent with Glick’s (1985) perspective, the issue of whether climate is different from other attitudes is most salient when psychological climate is assessed at the individual level of analysis. As L. R. James and Tetrick (1986) documented, psychological climate clearly correlates, perhaps reciprocally, with other more affectively loaded measures like job satisfaction at the individual level of analysis. While perhaps of some interest to perceptually and cognitively oriented psychologists, research on psychological climate reveals little about organizational functioning and organizational effectiveness, although some attention to psychological climate is still evident in the recent research literature (Carr, Schmidt, Ford, & DeShon, 2003; D’Amato & Zijlstra, 2008; L. R. James et al., 2008). For us that research, even when it has a more descriptive rather than evaluative focus, is not conceptually or methodologically useful for understanding organizational behavior and/or the effects of climate on organizational (or at a minimum, unit) effectiveness, what we view as the appropriate frame of reference for an organizational psychology (Schein, 1965). Our position is in agreement with Schneider (1975) and Schneider and Reichers (1983) that employee perceptions of their work and work world constitute, in the aggregate, the psychology of the organization and are as real as any other attribute of 377

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

the organization (e.g., structure, size). As a result, when climate survey items target how the organization functions (events, practices, and procedures and behaviors that get rewarded, supported and expected; Ostroff, Kinicki, & Tamkins, 2003), then they will typically have a significant level of consensus within organizations and a significant level of variance between organizations (Bliese, 2000; issues surrounding the aggregation of individual perceptions to higher levels of analysis are discussed later in this chapter). In the language of Kozlowski and Klein (2000), even though the level of measurement is the individual, the level of the construct is the organization. In other words, climate is a property of the unit—it is the organization’s climate, not the individual’s. Dimensionality issues. Strangely, the early dimensions of organizational climate did not emerge from the formative conceptual materials summarized earlier. Thus, in contrast to the theoretical work of people like Argyris (1957) and McGregor (1960) and the methodological advances of especially Likert (1961, 1967) and his coworkers at Michigan (e.g., Seashore, 1964), the focus of early empirical research on climate was on employee well-being and on individual outcomes rather than on organizational effectiveness. We think this happened implicitly, not explicitly, because most of the research was being done primarily by industrial psychologists trained in the individual differences and psychometric traditions that dominated that era. The focus on individual well-being and individual performance yielded surveys focused on those facets of the work environment related to well-being, especially social well-being, and the result was that the dimensions were relatively heavily oriented to personal experiences and were analyzed at the individual level of analysis. Hellriegel and Slocum (1974), in their early and comprehensive review of climate measures (see their Table 1, pages 264–269, for a listing of climate research to 1974), put it this way: “[T]here seems to be an over-emphasis, relatively speaking, on people-oriented scales. This may be partially a consequence of abstracting climate items from satisfaction scales” (p. 261). An early and relatively complete review of the dimensions emerging from these surveys was accom378

plished by Campbell, Dunnette, Lawler, and Weick (1970). Campbell et al.’s review was important because it focused on the climate experienced by managers in organizations and because it summarized the then-extant survey research and the dimensions that had emerged from that research. In many ways, this was the first time a review summarized such dimensions, and they reported that at the time there were four dimensions that emerged in survey research on climate: individual autonomy; degree of structure imposed on the situation; reward orientation; and consideration, warmth, and support. Hellriegel and Slocum (1974) thought that there were more dimensions at the time but acknowledged that the four identified by Campbell et al. were reasonable. A review that included much of the work being done simultaneously in Europe, especially in the United Kingdom, was that of Payne and Pugh (1976). While much of the early work on climate in the United States owes a debt of gratitude to the University of Michigan researchers, in Europe much of the early work was accomplished at Aston University in England under the influence of Derek Pugh (e.g., Pugh, 1966). That work began with the idea that it is critical to understand the importance of organizational structure if one is to understand the impact of organizations on the psychology of those organizations. Structure in this model is far more than size and levels of hierarchy; structure has to do with authority systems (like those noted by Likert, 1967, in his Systems 1–4), status systems, and the structuring of role activities. Climate in Payne and Pugh’s conceptualization concerned degree of risk taking, warmth, support, and innovativeness. One of the key findings from their review of structure– climate relationships was that more bureaucratic structures were not perceived to be cold, threatening, and/or low in cohesiveness. Another was that organizations with decentralized decision making were perceived to be warm and supportive and to encourage risk taking. Between 1968 and 1973, various combinations of Pugh and Payne along with Pheysey and Mansfield produced seven papers on these and related topics in Administrative Science Quarterly, two in Organizational Behavior and Human Performance (OBHP, now OBHDP), two in Human Relations, and a review in

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

the Psychological Bulletin (see Payne & Pugh, 1976). This publication record suggests that although relevant research was being conducted in the United Kingdom, U.S. publication outlets received much of the benefit. Furthermore, this early work does not seem to have produced a rich literature on the topic of organizational climate in Europe more generally, with the recent exception being research on service climate (to be discussed later) in Spain (e.g., GonzalezRoma, Peiro, & Tordera, 2002). Payne and Pugh relied somewhat on the classification of climate dimensions delineated by Campbell et al. (1970) and then expanded on the relationships of those dimensions with structural variables. L. A. James and James (1989), like Campbell et al. (1970), had 4 dimensions in their review: role stress and lack of harmony; job challenge and autonomy; leadership support and facilitation; and work group cooperation, friendliness, and warmth, with each of those subsuming between three and six more microissues. For example, leader support and facilitation included hierarchical influence, psychological influence, leader trust and support, leader interaction facilitation, and leader goal emphasis and facilitation. Similarly, Ostroff (1993) proposed 3 overall dimensions of climate (affective, cognitive, and instrumental) that encompass 12 more specific facets. As the number of possible dimensions of organizational climate expanded, researchers became concerned with what such dimensions might mean. M. J. Burke, Borucki, and Hurley (1992) concluded that rather than have the 4 higher order dimensions of L. A. James and James (1989), they could reduce the number of dimensions to 2: concern for employees and concern for customers. We return to the issue of customers in a moment, but for now it is useful to note that research by Patterson et al. (2005) began with 19 hypothesized dimensions of climate— reduced to “only” 17 as the research unfolded. On the focus of climate research. In his 1975 review of climate research, Schneider presented the thought that perhaps the wide range in the number of dimensions of climate that had been identified was a function of there being no focus for measurement. In essence, he raised the bandwidth problem: Without a focus for the outcomes of interest, you need a very

wide bandwidth to compensate for the lack of focus! He proposed that when climate measurement had a focus, like service quality or accidents or innovation, then dimensions of climate appropriate to those foci would follow more naturally and they would focus on organizational events, practices, and procedures relevant for the attainment of specific (e.g., safety, service, innovation) goals. Subsequent research on safety (Zohar, 1980, 2000) and service quality (Schneider, Parkington, & Buxton, 1980; Schneider, White, & Paul, 1998) suggested this was true. The research on service climate in fact revealed that aggregated employee perceptions at the bank branch level of analysis were significantly related to customer satisfaction. In M. J. Burke et al.’s (1992) work, the finding for service climate is one of the reasons why one of the two factors they proposed was concern for customers, with the other being a climate for employee well-being—the focus, if you will, of much of the earlier work on climate. As we review in more detail later, this approach to research—the “climate for what” or the “focused climate” or the “strategic climate” approach—has resulted in the demonstration of considerable validity for the climate construct, a characteristic not generally true of the more generic climate model (Ostroff, Kinicki, & Tamkins, 2003; Schneider, Bowen, Ehrhart, & Holcombe, 2000). Summary. The dilemmas confronted by early climate researchers were numerous, and it took many years of work by dedicated scholars to reach some consensus on issues. For example, L. R. James and Jones’s (1974) characterization of individual-level climate research as psychological climate was freeing because the distinction clarified the point that if one was claiming to do organizational climate thinking and research, then both the climate variable and the outcome should be at the unit level of analysis. The dimensions of climate in early research varied greatly in their number and focus, with most perhaps emphasizing employees’ personal experiences, especially their experiences with other people at work (leadership, coworkers, conflict, and so forth). Not surprisingly, studies of these dimensions against outcomes revealed that (a) when the outcome was satisfaction, relationships were statistically significant and at least moderate in level (e.g., .50; Schneider & 379

Schneider, Ehrhart, and Macey

Snyder, 1975); and (b) when the outcome was productivity, the outcomes were variable and/or not statistically significant (Hellriegel & Slocum, 1974; Pritchard & Karasick, 1973). A proposal by Schneider (1975), seconded strongly by Glick (1985), to have climate research focus on organizationally relevant outcomes had implications not only for the content of the dimensions to be assessed but also for the level of analysis of the research itself.

Copyright American Psychological Association. Not for further distribution.

Contemporary Climate Research: Resolving Some Dilemmas Ashkanasy, Wilderom, and Peterson (2000) wrote about the most vexing issue in climate research, data aggregation, in the opening chapter to their excellent Handbook of Organizational Culture and Climate: The field [i.e., research on organizational climate] seemed to be getting weary of endless technical haggling about whether attitudes and perceptions of individuals could be aggregated to represent something at the organizational level, what metric should represent agreement, what criterion should be used to justify aggregation, which particulars are most important in the increasingly overwhelming morass of organizational behavior variables generating inconsistent, weak, contingent relationships and so on. (p. 4) The point Ashkanasy et al. (2000) were making is that the doldrums in research resulting from the focus on statistical issues was one of the reasons for the rise of the organizational culture construct in both the world of academic research and application by practitioners. Organizational culture was always thought of as an organizational variable, studied by sociologists and anthropologists for whole organizations, and therefore did not require the statistical squabbling that characterized psychologists wanting to aggregate individual perceptions. Most climate researchers, and especially L. R. James, Demaree, and Wolf (1984) and Glick (1985), recognized that some uniformity was required in ways of estimating the outcomes of aggregation in terms of some index or indices of consensus. As Ashkanasy, Wilderom, and Peterson (2000) noted, 380

there was considerable confusion about this issue, and while more details are presented shortly, here let it be noted that both within-organization agreement (L. R. James et al., 1984) and between-organization differences (Glick, 1985) have served as frames of reference for resolution of the issue (Bliese, 2000). Once some agreement on the statistical issues of data aggregation had been achieved, climate research emerged from the doldrums brought on by that morass and, in many ways, has progressed considerably in the past 10 or so years. Interest in the topic has certainly increased. According to Kuenzi and Schminke (2009), there were over three times as many articles on organizational climate published in top management journals between 2000 and 2008 as there were in the 1990s. In this section we expand on the following topics: (a) the current consensus concerning the statistical issues surrounding data aggregation, (b) the burgeoning literature on focused climates, (c) some boundary conditions on climate– outcome relationships, and (d) some speculations on the relationships between generic climates for well-being and focused climates. Our goal is not to provide an exhaustive review of these topics but instead to provide key articles that best summarize or exemplify these trends in the literature. In addition, for the most part we do not present research on psychological climate. Statistical issues in data aggregation. The variety of issues and intricacies involved in the analysis of aggregated data would require much more space than we have available here. Therefore, our goal is to summarize the current state of the field and best practices for researchers analyzing climate data. Where appropriate, we note sources that address these issues in more depth. As we noted earlier, the basic concern for analyzing climate data is that the level of measurement, the individual, is different from the level of analysis, the unit (e.g., organization, department, group). To clarify how individual-level climate perceptions come to represent the climate of the unit, climate researchers should specify the composition model guiding their research. Drawing from Chan’s (1998) description of possible composition models in multilevel research, most climate research would fall

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

under either the direct consensus or the referent-shift consensus models. The main difference between the direct consensus and referent-shift consensus models is the wording of the items, or more specifically, whether respondents are asked about their own perspective or the perspective of their unit as a whole (Chan, 1998). As noted by Kozlowski and Klein (2000), for climate research the distinction between items presented as “I think my organization . . .” and items presented as “People in my organization think . . .” may not be as important as whether the items focus on describing the work environment as a whole (“The rewards received by employees . . .”) as opposed to the individual’s own experience (“The rewards I receive . . .”). We concur, and we recognize that most climate items do not actually specify whether employees are being asked for their own view or the view of their unit as a whole. In fact, there is little research examining such differences in the wording of items (see Klein, Conn, Smith, & Sorra, 2001, for an exception). There is consensus, though, that if one is studying the climate of the unit as a whole, one should frame items for respondents such that they describe the unit to which the items will be aggregated, and thus the level of analysis for analysis! Making sure that respondents are focused on the correct level of analysis by describing their units (instead of giving their own personal attitude or opinion) is the first step in enhancing the likelihood that respondents in a unit will reveal consensus in their perceptions. The idea of consensus is central to the conceptualization of organizational climate; in order for the mean of individual reports to meaningfully represent the unit as a whole, climate perceptions must be shared throughout the unit. Traditionally, assessments of this consensus or sharedness have focused on interrater agreement or interrater reliability (Kozlowski & Hattrup, 1992). In brief, interrater agreement addresses the extent to which raters provide similar absolute ratings of climate (i.e., the same numerical score on the measure), such that their ratings are interchangeable. The most common measure of agreement in climate research is rWG(J) (L. R. James, Demaree, & Wolf, 1984, 1993), although other alternatives, such as the average deviation

(AD) index (M. J. Burke, Finkelstein, & Dusig, 1999) and aWG (Brown & Hauenstein, 2005), have been proposed. The relative benefits of these different approaches are beyond the scope of this chapter (see LeBreton & Senter, 2008, for a review); the important issue here is that an adequate level of within-group agreement should be demonstrated prior to aggregation. Interrater reliability addresses the extent to which the rank ordering of the ratings is consistent across people within units (Bliese, 2000; LeBreton & Senter, 2008). Climate researchers typically provide at least one index of interrater reliability, although as LeBreton and Senter (2008) pointed out, typical measures of interrater reliability (including those discussed here) actually provide information on both interrater agreement and interrater reliability. For instance, climate researchers will often report the intraclass correlation coefficient, or ICC(1), a ratio of between-unit variance to total variance (Bliese, 2000). High ICC(1) values will be found when variability within units is low (i.e., there is within-group interrater agreement) and variability between groups is high; low ICC(1) values can result from too much within-group variability and/or too little betweengroup variability. ICC(1) values can be interpreted as a percentage of variance that can be explained by group membership (Bliese, 2000). L. R. James (1982) reported a median value of .12 among the studies in their review, and LeBreton and Senter (2008) suggested that a value of .10 might be considered a medium effect size for ICC(1). It is also common for researchers to report ICC(2), sometimes also referred to as ICC(K) (McGraw & Wong, 1996; LeBreton & Senter, 2008). ICC(2) is an index of the reliability of group means and is related to ICC(1) as a function of group size (Bliese, 2000) such that the more respondents per group, the more reliable the mean response (climate score) will be. ICC(2) values are commonly interpreted in line with other measures of reliability. Glick (1985) recommended a minimum cutoff of .60; others have said that .70 or higher is adequate (Klein et al., 2000; LeBreton & Senter, 2008). In general, the lower the ICC(1), the larger the average group size needs to be to reach adequate levels of ICC(2) and thus reliably distinguish group means (Bliese, 1998, 2000; LeBreton & Senter, 381

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

2008). As a result, it is often difficult to achieve high ICC(2) values when group size is small. Interpreting measures of interrater agreement and reliability is not without complexities. For instance, setting appropriate cutoff values has been controversial—see Lance, Butts, and Michel’s (2006) discussion of the arbitrary establishment of the ubiquitous .70 cutoff for rWG(J). We concur with LeBreton and Senter (2008) that such arbitrary cutoffs unnecessarily dichotomize decision-making processes that deserve more attention and thought; that agreement is better conceptualized as a continuum from no agreement to moderate agreement to very strong agreement; and that a number of issues, such as the context of the research and the measures being used, should be considered when determining what standard for level of agreement is appropriate for a particular research study. Of course, the same could be said for interpreting ICC(1) and ICC(2) values. Besides the interpretation of the values, there are a host of other decisions that must be made when aggregating individual-level climate data, and researchers should put thought and care into these decisions to ensure that they are appropriate and justifiable. Many of these issues are discussed by LeBreton and Senter (2008); we highlight only a few here. For example, how should researchers deal with individual units that demonstrate low withingroup agreement while others reveal high agreement? We concur with LeBreton and Senter (2008) in discouraging the practice of dropping lower agreement groups in favor of focusing on the overall pattern of agreement in the data set, or even modeling the level of agreement in the analyses (e.g., climate strength; see a later section of this chapter). Groups with low response rates present another challenge; Newman and Sin (2009) provided an in-depth analysis of that issue. They concluded, and we agree, that researchers should attempt to understand the nature of the missing data and its impact on their agreement statistics rather than deleting low response rate groups. An additional issue is the choice of the null distribution when calculating rWG(J), which has challenged researchers since L. R. James et al. (1984). LeBreton and Senter (2008) summarized current thinking on that topic and the issues researchers should take into account when making 382

those decisions and concluded that researchers should take care in identifying the appropriate null distribution (or multiple possible distributions) based on the construct being studied and possible response biases that could be at play. Finally, although most climate researchers report measures of both within-group agreement and interrater reliability, the relative importance of the two and the necessity of demonstrating adequate levels of both are not entirely clear (Chan, 1998; George & James, 1993). The demonstration of within-group agreement is most closely tied to the definition of the organizational climate construct as shared perceptions. In contrast, a lack of interrater reliability based on ICC(1) and ICC(2) may simply indicate a lack of between-group variability, thus limiting the ability to detect relationships at the unit level (Bliese, 2000). George and James (1993) argued that only within-group agreement, but not betweengroup variability, is needed to justify aggregation, as between-group variability is important for finding group-level relationships but not for justifying the appropriateness of aggregation. As pointed out by Chan (1998), however, a lack of between-group variability may indicate that the level of analysis is inappropriate (i.e., if all groups have similar means within a department, then the department level may be the appropriate level of analysis). We would advocate that climate researchers demonstrate withingroup agreement at a minimum. In situations in which adequate within-group agreement is coupled with lower interrater reliability, researchers should proceed with caution and ensure that they understand why the situation occurred and the effects it may have on their findings. Research on focused or strategic climates. Recall that Schneider (1975) proposed that one way to both reduce the number of dimensions to be studied and perhaps improve the validity of the climate construct was to focus climate research on one or more strategic outcomes of the organization. Thus, he argued that if climate sets the tone of an organization for the employees there, then the strategic outcomes organizations desire to achieve should be the focus of the items to which they respond. He and his colleagues (Schneider, Parkington, & Buxton, 1980) were the

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

first to demonstrate the validity of this approach to climate research showing that employee perceptions of service climate aggregated to the branch level of analysis were significantly correlated with the customer satisfaction data obtained from those same branches (and also aggregated to the branch level of analysis). Zohar’s (1980) work on safety climate soon followed in which he demonstrated that safety climate data collected from employees in diverse manufacturing organizations in Israel were significantly correlated with safety records and ratings by safety inspectors of those organizations. Subsequent research on these strategic outcomes has also revealed considerable validity (on customer satisfaction; for a literature review, see Schneider & White, 2004; for a meta-analysis on safety climate outcomes, see Clarke, 2006). There has also been research on the climate for innovation (Klein & Sorra, 1996) and, perhaps most interesting, research on the climates for various internal organizational processes. In this latter approach, an attribute of organizational process deemed to be important for organizational functioning is chosen for study and then it becomes the target of climate research. In this guise, the climate for something has focused on such diverse topics as the climate for fairness (Colquitt, Noe, & Jackson, 2002), the climate for burnout (Moliner, Martinez-Tur, Peiro, Ramos, & Cropanzano, 2005), the climate for ethics (K. D. Martin & Cullen, 2006), the climate for diversity within organizations (Hicks-Clarke & Iles, 2000), the climate for industrial relations (Dastmalchian, 2008), and the climate for sexual harassment by clients and customers (Gettman & Gelfand, 2007). In their review of the climate literature, Kuenzi and Schminke (2009) used Katz and Kahn’s (1966) four types of motivational patterns to organize all of these focused (or what they called facet-specific) climates in the following way: climates focused on behavioral guidance (includes ethics and justice), climates focused on involvement (includes participation, support, and empowerment), climates focused on development (includes innovation and creativity), and climates focused on core operations (includes service and safety). Their framework is the first attempt we know of to organize research on focused climates in such a way, and it may prove useful not only as an organizational tool but also as

a way to identify gaps that can be addressed in future climate research. Apparently, having a focus for the climate construct not only narrows the frame of reference for the range of variables to be studied but also produces validity for the outcome—be it safety, customer satisfaction, or process. This positive outcome of the approach is obviously a bandwidth phenomenon. The bandwidth phenomenon says that predictors that assess the relevant conceptual attributes of the outcome of interest are more likely to be valid predictors. Applied to climate research, this says that the variables or dimensions of climate studied must be chosen specifically because they are conceptually related to the outcome of interest—and of at least equal importance, the focus of the dimensions should be on the outcome of interest. A simple example would be the dimension of support that appears repeatedly in reviews of the climate literature (Campbell et al., 1970). If one measures support, the strategic focus asks, support for what? So the questions about support that are asked of employees should be on support for service or support for innovation or support for safety. The same applies to leadership if it is a dimension of climate—leadership toward what end or ends? The focus on a specific strategic outcome has led to some interesting conceptual work on what organizational climate is. As have we, most researchers have accepted a version of the definition of climate such as the one proposed by Ostroff et al. (2003, p. 566; after Schneider, White, & Paul, 1998): “Climate involves employees’ perceptions of what the organization is like in terms of practices, policies and procedures, routines and rewards.” Of course, if climate is defined in terms of the perceptions of these facets of organizations, then climate must also somehow include the meaning implicitly or explicitly derived from these perceived elements of organizational functioning. So, for example, in Schneider, White, and Paul (1998), service climate is defined by seven different facets of organizational functioning, all of which connote an emphasis on service quality (see the service climate items in Exhibit 12.1). Note in Exhibit 12.1 how each item refers to some organizational practice, routine, or reward and that each focuses on service quality either explicitly or 383

Schneider, Ehrhart, and Macey

Exhibit 12.1 Service Climate Survey Items ■

■

■

■

Copyright American Psychological Association. Not for further distribution.

■

■

■

How would you rate the job knowledge and skills of employees in your business to deliver superior quality service? How would you rate efforts to measure and track the quality of service in your business? How would you rate the recognition and rewards employees receive for the delivery of superior service? How would you rate the overall quality of service provided by your business? How would you rate the leadership shown by management in your business in supporting the service quality effort? How would you rate the effectiveness of our communications efforts to both employees and customers? How would you rate the tools, technology, and other resources provided to employees to support the delivery of superior quality service?

the behavior of supervisors vis-à-vis safety, Zohar believed he captured the essence of safety climate. Exhibit 12.2 lists the items used by Zohar to define safety climate and the way each item focuses on what the supervisor does vis-à-vis safety. In this case, employee reports of the behavior of supervisors with regard to safety provide information about the meaning employees derive from those behaviors and thus the safety climate of the organization. In the same way that service climate has shown validity for predicting service outcomes, safety climate has demonstrated validity for predicting safety outcomes. A meta-analysis by Clarke (2006) reported corrected validity coefficients of .43 for safety compliance, .50 for safety participation, and .35 for (lower) accident involvement (for prospective study designs with objective accident data collected after the measurement of safety climate).

Note. Adapted from “Linking Service Climate and Customer Perceptions of Service Quality: Test of a Causal Model,” by B. Schneider, S. S. White, and M. C. Paul, 1998, Journal of Applied Psychology, 83, p. 154. Copyright 1998 by the American Psychological Association.

Exhibit 12.2 Safety Climate Items ■

implicitly (by focusing on customers). The aggregate of those items focuses on service quality literally and operationally defines the meaning of service climate for employees. The validity of this scale for predicting customer satisfaction has been demonstrated repeatedly (Liao & Chuang, 2004; Schneider, Ehrhart, Mayer, Saltz, & Niles-Jolly, 2005; Schneider, Macey, Lee, & Young, 2009; Schneider, White, & Paul, 1998) with correlations at the unit level of analysis (e.g., supermarket departments) around .30 and at the firm level of analysis of about .45—even across a 2-year gap. Additional validity has been shown in predicting financial outcomes, although this relationship appears to be mediated by customer satisfaction (Schneider et al., 2009). Zohar (2000) provided another example in another domain. In his careful research spanning more than 25 years, he reached the conclusion that the supervisor at the work site is the major determinant of safety climate. This means that by capturing 384

■

■

■

■

■

■

■

■

■

My supervisor says a good word whenever he/she sees a job done according to the safety rules. My supervisor seriously considers any worker’s suggestions for improving safety. My supervisor approaches workers during work to discuss safety issues. My supervisor gets annoyed with any worker ignoring safety rules, even minor rules. My supervisor watches more often when a worker has violated some safety rule. As long as there is no accident, my supervisor doesn’t care how the work is done (R). Whenever pressure builds up, my supervisor wants us to work faster rather than by the rules (R). My supervisor pays less attention to safety problems than most other supervisors in this company (R). My supervisor only keeps track of major safety problems and overlooks routine problems (R). As long as work remains on schedule, my supervisor doesn’t care how this has been achieved (R).

Note. The letter R after an item indicates that it was reverse-scored such that a low score is positive. From “A Group Level Model of Safety Climate: Testing the Effect of Group Climate on Microaccidents in Manufacturing Jobs,” by D. Zohar, 2000, Journal of Applied Psychology, 85, p. 591. Copyright 2000 by the American Psychological Association.

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

The distinction noted above between the measurement of service climate and the measurement of safety climate raises the critical point that when measuring climate with reference to specific organizational outcomes, it is not essential to capture all possible facets, practices, or behaviors. From its origins, climate research has assumed a gestalt psychological position (Schneider, 1975). This presumes that when a reasonably complete set of indicators captures the meaning of a concept, then the missing indicators are filled in by the observer. At the same time, from a practitioner perspective a more thorough measure may be preferred to better isolate those aspects of the organization’s functioning that are not in alignment and thus that are negatively affecting the organization’s strategic climate. In summary, the focus on specific outcomes or processes has produced insights into the elements of climate, and in the research cited there is validity against external judgments (e.g., by customers) or external ratings (e.g., of safety practices) that confirm the validity of the data gathered from employees and aggregated to the appropriate unit of analysis. Boundary conditions on climate–outcome relationships. In the service climate literature, the relationship between service climate perceptions and customer satisfaction has come to be called linkage research (Wiley, 1996), and there is some research questioning the extent to which the linkage relationship will “always” be there. (See also Vol. 3, chap. 9, this handbook.) In other words, the question is whether there are boundary conditions (moderators) of the service climate–customer satisfaction relationship. Three kinds of research suggest that the answer is yes—there are boundary conditions. The first stream of research concerns the issue of climate strength (Colquitt, Noe, & Jackson, 2002; Gonzalez-Roma et al., 2002; Lindell & Brandt, 2000; Schneider, Salvaggio, & Subirats, 2002; Zohar & Luria, 2005). Climate strength returns us to the issue of aggregation and the climate researcher’s dread of finding that each individual had his or her own perceptions of the organization’s practices and procedures (psychological climate) negating the attempt to aggregate across respondents. Climate strength research asks the following question: In

organizations where there is higher consensus across respondents, does the aggregate mean derived from such scores differ in its predictive capability from scores aggregated across respondents where consensus was lower? In other words, when there are differences across units/organizations in consensus for psychological climate aggregates, are those differences meaningful? There is a paradox in climate strength research such that the better job the researcher does in achieving high levels of consensus in all the units studied, the less likely one is to show that climate strength moderates the relationship of interest. This is because if there is low variability across units in climate strength (i.e., there is consistent consensus), then climate strength cannot serve as a moderator since moderators require high variability. GonzalezRoma et al. (2002), Schneider et al. (2002), and Colquitt et al. (2002) found some support for climate strength serving as a moderator variable (relationships were stronger when strength was stronger), but there are also failures in the literature to find such relationships (Dawson, Gonzalez-Roma, Davis, & West, 2008; Sowinski, Fortmann, & Lezotte, 2008). Such failures invoke the aforementioned explanation of low variability for failure; support for the moderating effects of climate strength clearly requires considerable variability in climate strength across the units or organizations studied. In addition, the variability within organizations cannot be too extreme, because if it is, there will be no main effect for the climate mean! As a sidebar, it is interesting to note that research on climate strength has begun to expand in some new directions. For instance, Zohar and Luria (2005) introduced the idea of climate variability, which represents the between-group, within-organization variability. In other words, they raised the interesting idea that climate strength at the organizational level can be operationalized as the variability across individuals within the organization or the variability across groups, which raises a number of questions about the similarities and differences between the two, and which is most relevant in what situations. Researchers have also begun to investigate predictors of climate strength. For instance, larger teams and more demographically diverse teams tend to have 385

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

lower climate strength (Colquitt, Noe, & Jackson, 2002); organizations that are clearly mechanistic or clearly organic tend to have stronger climates (Dickson, Resick, & Hanges, 2006); teams that perform more sense-making activities tend to have higher climate strength (Roberson, 2006a), as do more interdependent teams with higher group identification (Roberson, 2006b); and more cohesive military units that are led by transformational leaders and that have denser communication networks tend to have stronger climates (Luria, 2008; Zohar & Tenne-Gazit, 2008). The second kind of boundary condition would be specific to the relationship of interest and not a generic issue like consensus within and across organizations. For example, one might propose that the relationship between diversity climate and career progress in organizations would be moderated by the number of hours workers actually work in a common physical space. In other words, in an age of “home work” diversity, climate might be more strongly related to career progress when presence at work in the workplace is a requirement. We do not know of any research on this specific topic, but that is the idea. There is research along these lines on boundary conditions surrounding the service climate–customer satisfaction relationship. For example, Dietz, Pugh, and Wiley (2004) found, in a study of 160 bank branches, that frequency of customer contact moderates the service climate–customer satisfaction relationship. That is, in branches where customers on average visited the branch more frequently, the relationship was significantly stronger. In a similar vein, Mayer, Ehrhart, and Schneider (2009) showed that departments within supermarkets (meat, deli, produce) differed in the extent to which employees had direct customer contact and that the more customer contact there was, the stronger was the relationship between department service climate and department customer satisfaction. At the same time, they also found that the tangibility of the product offered by departments—pharmacy is less tangible than produce—moderated the relationship such that low tangibility produced a stronger relationship between service climate and customer satisfaction. Finally, they also showed that in departments 386

requiring a team effort to serve customers (e.g., pharmacy), the relationship was stronger than in departments where the service delivery act was in the hands of fewer or even one person (e.g., cashier). Gittell (2002) provided similar findings with regard to interdependence and argued that it is not only interdependence among employees involved in the delivery of service to customers but also interdependence among employees serving those employees who serve customers. There are other examples from outside the domain of service climate in which moderators of the relationship between climate and outcomes are examined. For example, in the literature on procedural justice climate, Liao and Rupp (2005) found that justice orientation moderated the cross-level relationship between procedural justice climate and individual-level attitudes, and Yang, Mossholder, and Peng (2007) found that work-group power distance moderated the relationship between procedural justice climate and individual-level commitment and citizenship behavior. In the safety climate literature, Hofmann and Mark (2006) found that the relationship between safety climate and safety outcomes was moderated by the complexity of patient conditions, such that safety climate had stronger effects with more complex patient conditions. In another interesting study of safety climate, M. J. Burke, Chan-Serafin, Salvador, Smith, and Sarpy (2008) showed that safety climate moderated the relationship between safety training and injury and accidents such that safety training had a greater effect in a strong safety climate. As the reverse is also true, their research can also be interpreted to have shown that safety training moderated the relationship between safety climate and injuries and accidents, indicating that safety climate is most beneficial when employees have received adequate safety training. Such research on training is reminiscent of one of the first studies of climate by Fleishman (1953), who investigated leadership climate. This is a classic study of the transfer of training because what Fleishman showed was that supervisors who were trained to behave more in a human relations mode with subordinates did so only when they in turn worked for a leader who believed in (who created a climate for) the human relations approach to leadership.

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

Finally, we highlight one other study, on justice climate, by Spell and Arnold (2007). This study is unique because the authors examined the interaction between two different justice climates: procedural justice climate and distributive justice climate. They hypothesized and found that the negative relationship between distributive justice climate and individuallevel depression and anxiety would be moderated by procedural justice climate. More specifically, one type of climate, procedural justice climate, buffered the negative effects when levels of another type of climate, distributive justice climate, were low. Although they did not examine unit- or organizational-level outcomes, their research raises the question of how one type of climate may moderate the relationship between another climate and its outcomes. In summary with regard to boundary conditions, as a field of study matures, research on some boundary conditions of established relationships begins to emerge, and this has certainly been true of the linkage research relating climate to outcomes. Thus, we see that climate strength and the nature of the context (e.g., product tangibility, amount of customer contact, severity of patient condition) serve as moderators of the basic climate–outcome relationships. We predict that there will be an increase in studies similar to those just described. On mediator studies linking climate to outcomes. As opposed to moderator variables that affect the strength of relationships between two variables, mediators are variables that explain the relationships between two variables (i.e., why two variables are related to each other). Mediator studies have begun to enter climate research, especially research on the antecedents of climate using climate as the mediator between the antecedents and the outcomes. An increasingly common kind of research of this sort integrates what we have called generic climate or climate for well-being with focused climates in understanding outcomes. In this approach, the generic climate is seen to serve as a foundation or a frame of reference for the focused climate, and the focused climate is the mediator between generic climate and the outcome. Schneider, White, and Paul (1998) called the generic climate work facilitation and argued that when employees work in conditions that

generically facilitate their work, then a climate for service can be built on that foundation. They argued that work facilitation does not cause the service climate but provides a foundation on which such a climate can be built because the necessary management issues that have been attended to in turn permit the creation of a focused climate. In their work they showed that service climate mediated the relationship between work facilitation and customer satisfaction. Along similar lines, Salanova, Agut, and Peiró (2005) showed that service climate fully mediated the effects of organizational resources and work engagement on employee performance and customer loyalty. Wallace, Popp, and Mondore (2006) had a related conceptualization in their research on safety climate. In their work they conceptualized and assessed two foundation climates: managerial support and organizational rewards. They proposed and found on 253 work groups that these foundation climates are only indirectly related to occupational accidents and that safety climate mediates the relationship between them and occupational accidents. Bowen and Ostroff (2004) provided another perspective on climate as a mediator, this time focusing on the strength of the human resources management (HRM) practices in organizations as the key to producing a strategic climate. So, rather than having a generic climate or a climate for well-being as the antecedent variable, they proposed that HRM practices have their maximum impact when the various practices implemented act to create a strategic climate appropriate for the organization in its marketplace. They argued that when HRM practices in organizations have internal coherence, the consistency in messaging contributes to climate strength, which in turn yields the collective employee performance that yields competitive advantage. In their conceptualization, climate strength is a mediator and not a moderator because they argued strength is a direct predictor of the outcome(s) of interest. This work on generic, foundation, work facilitation, worker well-being, or HRM-focused climates is useful both conceptually and practically. Conceptually, it is useful because it provides one possible explanation for the high variability that has existed in the validity of climate–outcome relationships before the more focused approach existed. 387

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

That is, in some generic climate studies, perhaps more focus existed in surveys relevant to the outcome of interest—as in the University of Michigan survey items focusing on production processes and not just employee well-being. Recall that the Michigan studies of Likert (1967) and others consistently revealed validity when others did not (Campbell et al., 1970; Schneider, 1975). Practically, it is useful because it indicates that one cannot likely just implement a focused climate (e.g., for safety, for service) in a vacuum and that a certain foundation in terms of climate for well-being, work facilitation, and/or HRM practices must exist first on which such a focused climate can exist. So, for example, the findings suggest that organizations wishing to create a focused strategic climate may well have to pay attention to more fundamental issues in terms of the generic climate that already exists for people in the setting (W. W. Burke, 2008; Schneider, Brief, & Guzzo, 1996). A second antecedent receiving increasing attention concerns the issue of leadership. (See also chap. 7, this volume.) While leadership has been a focus of organizational culture thinking and research (Schein, 1985), it has paradoxically not been a focus of climate research. The paradox relates to the idea that it was the original Lewin et al. (1939) studies of leadership that yielded the term climate. Indeed, 2 decades ago Kozlowski and Doherty (1989) expressed some dismay that leadership as an antecedent of climate was not very much explored. In fact, if one looks at the dimensions of climate that emerged from early research (e.g., Campbell et al., 1970; Hellriegel & Slocum, 1974), leadership is seen as one of the dimensions and not the causal factor. Zohar and Tenne-Gazit (2008) put the paradox this way: “The notion of leadership as a climate antecedent has hardly changed over the last 50 years (Dragoni, 2005; Ostroff et al., 2003), although this has resulted in limited empirical work” (p. 745). In the extreme, as mentioned earlier, leadership is seen as both the cause of climate and the climate itself (see Table 2 from Zohar, 2000), but it is most typically viewed as antecedent to the creation of a strategic climate. For example, in Zohar and Tenne-Gazit (2008), transformational leadership is seen as a safety climate antecedent. As another example, Dragoni (2005) suggested and showed that 388

a leader’s goal orientation creates a specific form of goal-oriented climate in work groups that, in turn, determines the goal-oriented patterns of work group members. Schneider et al. (2005) proposed and showed that what they called strategic leadership when focused on service quality creates a service climate that, in turn, is related to customer satisfaction. Other related studies include Wallace et al.’s (2006) research focusing on management– employee relationships as a foundation for safety climate, Ehrhart’s (2004) study of servant–leadership as a predictor of procedural justice climate, and Klein, Conn, and Sorra’s (2001) finding of management support as a key antecedent of the climate for technology implementation. In a cross-level twist on leadership as an antecedent of climate, Mayer, Nishii, Schneider, and Goldstein (2007) proposed that particular leader personality attributes get reflected in employees’ fairness experiences. For example, they proposed and found that leaders who are higher on agreeableness and conscientiousness and lower on neuroticism would create more positive procedural justice climates. The dearth of research on the relationship between leadership and climate is surprising, but the almostcomplete absence of the inclusion of personality attributes of anyone in climate research is additionally surprising. This is so because of the work of Holland (1997), who for 50 years has shown that what he calls “career environments” are defined by the interests and personalities of the members of the career in an environment. As we see later, this issue has received more attention in the organizational culture literature, but the absence of personality variables in climate research is to say the least interesting when the field has been dominated by psychological researchers. Finally, there are some studies emerging on mediators of the link between climate and outcomes. It is one thing to hypothesize and show that, for example, service climate is related to customer satisfaction, but there is precious little research on how this link occurs. At the most obvious level, climate is thought to yield behavior consistent with the climate employees experience, and it is the behavior that yields the outcome. But where is the research on the behavioral piece? Schneider et al. (2005) proposed

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

and found that service climate yields service-oriented organizational citizenship behavior (OCB) from employees and it is the OCB that yields customer satisfaction. In safety research, Neal and Griffin (2006) used a longitudinal design to show that safety climate was positively related to safety motivation, safety motivation was positively related to safety behavior, and aggregated safety behavior was negatively related to accidents. Also from the safety literature, Wallace and Chen (2006) demonstrated in a cross-level study that employees’ promotion or prevention focus mediated the relationship between safety climate and employee performance. As a final example, Klein, Conn, and Sorra (2001) revealed that the relationship between implementation climate and innovation effectiveness was mediated by implementation effectiveness. More studies along these lines are needed.

Summary of the Climate Research Paradigm Early organizational climate research focused on individual employee perceptions, and, with few exceptions, climate was studied in relation to the individual behavior of respondents rather than organizational effectiveness. The early interpretation of climate and the scholarly debate that followed focused on aspects of the work environment largely representing what we would think of today as employee well-being, and thus contributed to much of the conceptual confusion entangling climate and affect, and in particular job satisfaction. Partly in response to this, climate research emerged in a form emphasizing strategic purpose and relevance, and the link between climate and organizational outcomes attracted attention. As such, climate research came to more clearly fulfill the expected maturation of a trend that Schein (1965) had early noted toward group- and systems-oriented organizational psychology. As this research paradigm evolved, attention on methodological issues intensified, reflecting the basic question of whether climate, as an organizationallevel construct, could be adequately represented in the aggregate responses of individual employees. In its present state, climate research is becoming increasingly more centered on the antecedents of specific strategic climates. In this way climate research has perhaps come full circle with the global social

and well-being climates characteristic of research 20 years ago being now conceptualized as providing foundations on which more specific strategic climates can be built. In what follows, we make a similar suggestion with regard to the eventual role the organizational culture perspective may play in a future integration of climate and culture research and practice. ORGANIZATIONAL CULTURE Organizational culture is the younger sibling. Not that research on culture had not been going for a very long time—it was just relatively new to the study of organizations (Van Muijen, 1998). When the younger sibling arrived, it was initially more popular than organizational climate was in that period (roughly 1980–1995). Our clear impression is that while organizational climate research was suffering through its struggles with the levels of analysis problem and producing relatively few publications, both academics and practitioners promoted organizational culture in that period of time. Pettigrew (1990, p. 416) made this observation in his conclusion to Schneider’s (1990) edited volume on climate and culture: “The chapters in this book create the impression that climate studies have been boxed in, even marginalized by the appearance in the nest of this rather overnourished, noisy, and enigmatic cuckoo called organizational culture.” Schneider (2000, p. xviii) documented several cases of attempts by culture researchers to marginalize the older sibling: ■

■

■

■

Trice and Beyer’s (1993) review of culture has a description of what culture is not and the first item (p. 19) is that culture is not climate. Ott (1989, p. 47) said culture is relatively permanent but climate is a transient state so not worthy of sustained study. J. Martin (1992) in the first version of her book did not even index the term climate, but in her revision (J. Martin, 2002, p. 112) she acknowledged that some culture researchers “would be more comfortable with viewing climate and culture as closely related.” Schein (1992, 2004) proposed that climate refers to the built environment and called climate artifacts 389

Schneider, Ehrhart, and Macey

Copyright American Psychological Association. Not for further distribution.

of organizations (though the word climate does not appear in the index of the 2004 edition). Artifacts are relics and refer to objects, works of art, and manufactured articles; we are unfamiliar with anyone studying organizational climate who ascribes these features to their work, though culture researchers might (Rafaeli & Worline, 2000). In a far more nuanced perspective on climate and culture, Schein (2000) wrote that climate concerns the meaning people derive from what they experience (including reward systems) but that the culture people experience is a function of the values and beliefs that lead to the creation of what employees experience. We have always thought that Schein’s (1985) listing of what he called culture-embedding mechanisms was the most succinct and important listing of how climates are created in organizations, and we are pleased he agreed. Thus, in the 2004 edition of his book, Schein wrote: “They [cultureembedding mechanisms] are visible artifacts of the emerging culture and they directly create what would typically be called the ‘climate’ of the organization (Schneider, 1990; Ashkanasy, Wilderom, and Peterson, 2000)” (p. 246). Schein’s listing of what leaders do to embed culture can be summarized as follows: ■

■ ■ ■

■

what leaders pay attention to, measure, and control; how leaders react to critical incidents and crises; the behaviors that leaders model for others; on what bases leaders allocate rewards and other scarce resources; and the standards used for recruitment, selection, promotion, and termination.

It is easy to see how the creation of a focused climate (for service, for safety, for anything) would profit from following this list of culture-embedding mechanisms. When Schein wrote this chapter, he was obviously thinking about how leaders reveal to others the values and beliefs they hold—and it is the focus of the revealed values and beliefs that constitute the creation of climate. As Schein (2000) put it, “To understand what goes on in organizations and why it happens in the way it does, one needs several concepts. 390

Climate and culture, if each is carefully defined, then become two crucial building blocks for organizational description and analysis” (pp. xxiv–xxv). We agree. In what follows, we look into the stream of writing and thinking that has illuminated insights into organizational culture. As we do so, we attempt to note ways we think the two concepts Reichers and Schneider (1990) saw as developing and evolving on parallel tracks might be integrated. It appears that perhaps there is a beginning of an overlap, and we would like to contribute to furthering that because as Schein noted, the two are crucial building blocks for organizational description and analysis. However, before we begin, we should note that even though organizational culture was initially the more popular sibling, research on climate has been much more prominent in the literature in the fields of industrial/organizational psychology, human resources management, and organizational behavior over the past 10 to 15 years. In fact, we could find only a handful of studies published this decade in top journals (e.g., Journal of Applied Psychology, Academy of Management Journal, Personnel Psychology, Administrative Science Quarterly) in which culture was a primary variable of study, with the literature on person–organization fit being the major exception. Thus, our overview tends to focus on the more traditional lines of thinking about organizational culture with some mention of recent research when we found it. We depend heavily here on two handbooks of organizational climate and culture, though both are far more dominated by culture than climate thinking and research so they play a larger role in this section than they did in the prior one on climate (Ashkanasy, Wilderom, & Peterson, 2000; Cooper, Cartwright, & Earley, 2000). Both volumes are international in their authorship, with the Cooper et al. volume perhaps being more so. To illustrate the emphasis on organizational culture, consider the following: The Cooper et al. volume has six major sections containing 27 chapters with perhaps only 2 or 3 chapters directly relevant to the issue of organizational climate; and there is a chapter explicitly called “Organizational Culture” but none explicitly called “Organizational Climate.”

Perspectives on Organizational Climate and Culture

Copyright American Psychological Association. Not for further distribution.

Conceptual Issues in Organizational Culture Research Pettigrew (1979) introduced the construct of culture to the study of organizational behavior so that organizational researchers would become familiar with the language and concepts of social anthropologists. His focus was on “the study of processes of leadership, commitment building, and change” (Pettigrew, 1990, p. 422) with a definite focus on the latter and the nexus of culture, strategy, and change. The idea was that if one could comprehend the dominating beliefs and ideologies of organizations, one would gain leverage on the source of the actions that yielded those beliefs and ideologies and thus access what needed to be changed to enhance the possibility of strategically relevant change. Pettigrew’s concern for the relationships among culture, strategy, and change are exemplified in a book he later edited (Pettigrew, 1987b) with the title The Management of Strategic Change. In that book, Pettigrew (1987a) clearly stated that his purpose was to understand the strategic competitiveness of organizations and especially the role of what he called the “inner context” of organizations. In this vein, Weick (1985) produced a most insightful chapter on the intimate relationships of culture and strategy, arguing in essence that one cannot discuss the one without the other. At the start of his chapter he presented a series of statements with the first word missing and asked the reader to fill in the blank with either the word culture or the word strategy. We reproduce the last sentence (Weick, 1985, p. 382, which he adapted from Greiner, 1983) and ask readers to do the same: “________ emerges out of the cumulative effects of many informed actions and decisions taken daily and over years by many employees—not a ‘one-shot’ statement developed exclusively by top management for distribution to the organization.” Greiner’s paper was about strategy, not culture. We go into detail on the early writings of Pettigrew and Weick because they clearly map onto the foundation writings relevant for the study of climate described earlier with their deep concern for organizational effectiveness. Thus, Pettigrew and Weick were interested in characterizing not only the culture of an organization but also how that culture related to

and was reflected in the strategy and competitiveness of the organization. Practitioners and management consultants loved the concept of organizational culture, and it caught on quickly as a key variable in trying to understand more effective from less effective organizations. Trice and Beyer (1993) attributed this interest to the uncertainties associated with the 1980s, especially with regard to Japanese inroads into the American auto industry and the different ways Japanese car companies were run. A number of popular management trade books (e.g., Deal & Kennedy, 1982; Peters & Waterman, 1982) appeared using concepts from the study of culture, like myths and taboos, and consultants went around the country trying to convince organizations that they needed to change the stories they told so that they could change their culture. What these consultants apparently misunderstood was that the signs of the cultures they were studying were reflections of those cultures and not causes. Perhaps they also misunderstood that these signs or indicators are a reflection of the metaphorical gravity that a culture has for all that emerges there and that the stories and underlying messages are not easily displaced much less modified. One of the problems eventually encountered in the organizational culture stream of research, like that in climate research, was the inability of culture researchers to demonstrate a relationship of their diagnoses to organizational effectiveness. Indeed, there were claims by culture researchers that connections of culture diagnoses to financial outcomes in organizations were not likely at all given the many variables intervening between culture and those outcomes (Siehl & Martin, 1990). As Miner (2002) put it in his massive review of organizational behavior research and theory: The study of organizational culture has been described as in a state of chaos at present (Martin & Frost, 1996), and with good reason. There is no science to sort out truth from fantasy, and stridency of protestation becomes the major criterion for fleeting acceptance. (p. 613) We think this state of affairs has emerged in organizational culture research because too many 391

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

culture researchers became enamored of the concept itself and lost their way in studying it for its own sake and not for its relationship to organizational effectiveness. Thus, a concern emerged for what culture is and not for how culture is related to organizational effectiveness. For example, Trice and Beyer (1993) did not have a chapter on the relationship between organizational culture and organizational effectiveness but did have numerous (very interesting and detailed) chapters on the features of organizational culture: beliefs, ideologies, and values, and the ways these are transmitted through symbols, language, narratives (myths, stories), and practices (rituals and taboos). The important role socialization tactics play in how newcomers acquire insight into the culture of the organization they join is emphasized, and the development of subcultures in organizations with all of their implications is also presented. We believe these to be important topics but not in and of themselves unless we see how they are related, directly or indirectly, to organizational effectiveness. Thus, just as climate researchers became obsessed with untangling statistical levels of analysis issues, there is evidence that culture researchers became obsessed with issues concerned with the varieties of ways culture might be conceptualized (Smircich, 1983). Note that these dire conclusions about organizational culture are with reference to “classic” organizational culture research defined by the qualitative case method. As we see later, this methodology has been substantially replaced in research by so-called culture surveys, which some (J. Martin, 1992; Schein, 1992, 2004; Trice & Beyer, 1993) decry as not being organizational culture research. With culture surveys, the evidence for the validity of such research is closer to the research on organizational climate (Wilderom, Glunk, & Maslowski, 2000); more on this transition in methodology later. In what follows, we present in turn some of the key conceptual issues surrounding organizational culture research. For each we attempt to note the contribution it might make if it were included in a climate–culture research paradigm focused on organizational effectiveness, one that includes the most promising features of each. 392

Beliefs, ideologies, language, and taboos. Language, including language used in myths and stories, is a key focus of cultural research regardless of whether it is in organizations or in the field with native peoples. Language in both form and content has not at all been a focus of organizational climate researchers, and it should be, because the language used in organizations reveals the patterns of thought processes that characterize a place. Language, for example, not only reveals the thoughts people have about others (e.g., women, African Americans, Jews) but also serves as an identity marker with regard to occupational subcultures in organizations. People in accounting just talk differently, the same as people in human resources do, and the two rarely share anything; as we will see when we discuss methodology, sharing is a big part of organizational culture research. We have more to say about subcultures, including occupational subcultures, later. The contents of myths, stories, and sagas reveal the history of an organization and thus influence how people understand what their organization values and believes. The history these myths and stories carry is almost never a focus of climate research, yet it is important information if one wishes to change climate. That is, unless one has a sense for the history of the current practices and procedures, one loses a sense of not only the culture of an organization but also how sense-making in the organization occurs precisely because people look for ways to interpret that history in order to move from what cannot easily be described and perhaps even recognized to what can be more easily translated and understood. So, it is in this way that culture is thought of by some as layered (e.g., Rousseau, 1990), and it is in thinking of it in this way that the correspondence of climate and culture becomes more easily apparent. In a very real and practical sense, change agents must be aware not only of what is but where it comes from and how it is transmitted so change can be facilitated by having a new language, a new set of stories, and new myths. As an example of the importance of language and stories to organizational functioning and change, one of us did a service diagnosis project for an auction house and noticed that the display of items for customers seemed inconsistent with the quality and

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

price for which items were being sold. That is, items were displayed in small pockets around the floor on boxes arrayed on a cement floor rather than on fancy tables set on plush carpet. When asked about this, employees related a story about how the auction house had begun as one involved in selling to the trade and not to individual customers, so the displays preserved the original “to the trade” look and feel of the business. Speaking of customers, what is the difference between a customer, a client, and a patient? All come for service to an organization but receive different kinds of treatment as a function of what they are called. Banks have customers, lawyers have clients, and doctors have patients, and the language used in these labels implies the type of service and the kind of the relationship expected. Language is also a methodological issue, but we discuss it here because the language of organizations is rarely considered in climate survey research. Culture researchers want to know what is unique about organizations so they focus on language, myth, and stories that are unique to settings—like the story we heard in the auction house. Climate researchers want to know how organizations compare on standardized questions so the uniqueness is in the quantitative differences revealed on a set of dimensions but not the content of the differences. Ways to combine the quantitative differences with the content differences would be important. In fact, since much of climate survey work begins with focus groups (Schneider, Wheeler, & Cox, 1992), it may not be difficult to combine the content with the quantitative data for a more complete understanding of organizational life. We note, however, that some perceive a “culture of organizational culture research” (e.g., Denison, 1996, p. 647) that seemingly works against such integrative approaches. Socialization. Organizational culture researchers have been concerned with what Louis (1990) called the native view of organizations. The native view concerns the newcomer as native anthropologist (in Louis’s terms, lay ethnographers) trying to learn the ropes of the new organization. The ropes are learned through the myths and stories newcomers hear, the behaviors they observe, and implicit attention to what Schein (2004) called culture-embedding mechanisms.

Research on organizational socialization has proceeded using the case study qualitative method— more on this when we discuss methodology—with a focus on in-depth understanding of the processes by which newcomers absorb the inner context of the organization. For an extended treatment of socialization, see Volume 3, chapter 2, in this handbook. In fact, socialization research may be the earliest form of organizational culture research, certainly predating Pettigrew’s (1979) calls for a more socialanthropological base for organizational behavior. Parsons (1951), for example, clearly outlined the purpose of socialization to be acquiring the necessary knowledge (social, political, and task-related) for functioning well in the new organizational role; his emphasis was on the total role in the organization, not just doing one’s tasks. Etzioni (1961) was also concerned with socialization, because he saw it as important for newcomers to understand the sources of power required to take action in organizations as critical to getting along there. Of potential greatest impact on contemporary organizational behavior were the Hawthorne studies, which after the initial lighting experiments moved on to attempt to understand the social and psychological life of workers. A key observational part of that research concerned socialization, especially as documented in the famous bank wiring room observations (Roethlisberger & Dickson, 1946). Here we see documented the “binging” (punches in the shoulders) suffered by newcomers who were thought to be rate-busters as they wired the boards for use by operators in companies. Van Maanen and Schein (1979) proposed a framework for understanding the likely effects of different socialization tactics on whether newcomers would adopt a firmly prescribed role for themselves or would be more innovative as they assumed their new roles. The tactics they illuminated were quite complex, but research on them seems to reduce to the following conclusions (Major, 2000, p. 364): (a) Tactics can be classified as being formal and institutional versus less formal and individual; (b) less formal and individual tactics produce improved feelings of adjustment and perhaps less rigid adherence to job specifications especially in high-performance work organizations; and (c) more formal and institutional tactics are effective in encouraging personal growth 393

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

and development. For organizational culture researchers, the observation of socialization tactics is a key route to understanding an organization’s culture because it is what the agents of socialization consider to be important that illuminates the different vantage points on the organization’s culture. The research on socialization focuses on newcomer adjustment and not on newcomer performance, so it is not clear who the most effective agents of socialization might be. For example, are supervisors more effective than coworkers or other newcomers? Or perhaps, as Louis (1990) suggested, it may be newcomers who are the best agent of socialization through the proactive behavior they display. In summary, there is broad agreement that socialization is an important area for “learning the ropes” and that informal and individual attention by supervisors and coworkers can facilitate feelings of adjustment. But it is not clear what the roles of different tactics or agents are in enhancing the performance of newcomers, nor is there research comparing the tactics and agents across organizations to more or less tear and compare them so far as organizational effectiveness is concerned. One possible exception worth noting in the domain of socialization research concerns the development of psychological contracts (Rousseau, 1990) between employers and employees. This work speaks to a mechanism—the violation of psychological contracts—by which the socializationrelated behavior of organizational agents does influence organizational effectiveness. However, the research on both individual in-role and extra-role performance is at the individual level of analysis; it would be useful to extend that work to unit and organizational effectiveness. Organizational culture and organizational subcultures. To this point we have written about organizational culture as if organizations have one. J. Martin (1992, 2002) sensitized us to the idea that organizations may have multiple cultures. While the idea of subcultures is not original with Martin, she has contributed greatly to our understanding of the issue by focusing heavily on the following question: How many ways can organizational culture be conceptualized and still have it be a meaningful and useful construct? 394

J. Martin (1992) elaborated three different perspectives on organizational culture: integration, differentiation, and fragmentation; we discuss the last in some detail after presenting the integration and differentiation perspectives. The integration perspective views organizational culture as one overall culture per organization. Organizations may differ in their culture, but each has one pervasive culture by which it may be categorized. The differentiation perspective proposes that an organization may have numerous subcultures. Sometimes these are a function of occupations (called occupational subcultures), but they can also be a function of hierarchical status in the organization. Psychologists are familiar with the idea of occupational subcultures based on Holland’s (1997) impressive work on career environments, in which he characterized what life would be like in those environments as a function of the kinds of people who enter different occupations (careers). So, within organizations, people would likely differ significantly in finance compared with laboratory researchers or salespeople on their beliefs and values, the language they use and the stories they tell, and the way they are socialized. Rentsch (1990) conducted some research relevant to differentiation in organizations based on hierarchy. She proposed that the same acts in organizations can have different meanings as a function of the status hierarchy they occupied. She did her research in an accounting firm and compared partners’ interpretations of themselves playing golf on Wednesday afternoons (business, work, necessary) with secretarial staff’s interpretation of the same act (fun, playtime, frivolous). As J. Martin (1992, 2002) noted, it is possible that people might differ in terms of specific beliefs, values, and language and may be socialized to their functional work units differently yet may also simultaneously share deeper beliefs and values. This would permit the differentiated and pervasive cultures to exist simultaneously. Similarly, in Rentsch’s (1990) work, the partners and the secretarial staff might agree on the strategic directions the firm is taking while disagreeing on the meaning of playing golf. So, one can view these alternative perspectives either as complementary and existing at different levels of abstraction or as competing. We agree

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

with J. Martin (2002) that they are complementary because this issue of the pervasive and the differentiated cultures of organizations is reminiscent of the generic climate (for well-being, facilitation, and so forth) and the focused-climate issue; both the generic and the focused climates can exist simultaneously. Not so incidentally, this also tells us that more than one focused climate can exist at a time in organizations, though little research exists on this issue and more needs to be accomplished. Now to the issue of fragmentation in organizational culture. Fragmentation exists because there is ambiguity about whether or not there is a culture and whether or not there are subcultures (J. Martin, 2002). Fragmentation cultures are metaphorically low on climate strength—either for a single climate or the multiple foci of multiple climates. In organizational culture research the most often used word is shared. Culture researchers do not just speak about values and beliefs, but they also speak about shared values and beliefs as if they are always shared. What Martin had so nicely done was explicate the situation in which there was ambiguity and she did so without saying whether it was positive or negative with regard to organizational effectiveness. Indeed, she wrote that occasionally ambiguity might be useful because it provides for ease in change because people are not mutually locked into a specific culture (J. Martin, 2002; see also Sørensen, 2002, who made the same point). It could be important for climate researchers to study the situations in which weak climate strength is a positive, for example, with regard to the readiness for change. In fact, ambiguity is precisely when leadership can produce change because people tend to not feel comfortable with ambiguity (Bass, 1990. p. 631). On the other hand, Kotter and Heskett (1992) described and tested the hypothesis that a strong organizational culture should be useful for organizations and, further, that strong cultures would be reflected in positive organizational performance. We have difficulty understanding the hypothesis that culture (or, for that matter, climate) strength should be linearly related to outcomes because one could obviously have a strong culture that is very negative as well as a strong culture that is very positive. Although Kotter and Heskett agreed with

our logic, they noted that proponents of this theory argue that most organizational cultures are basically positive or the business would have gone out of business, and it is therefore how strong the culture is that is important (see Cooke & Szumal’s, 2000, discussion of the defensive misattribution of success and the culture bypass for explanations for why negative cultures can still be successful). Kotter and Heskett (1992) designed a survey they sent to the top executives in 202 organizations asking them to describe the culture strength of their competitors. They defined a strong culture for respondents as one (1990) “usually associated with affirmative answers to the following three questions” (p. 159): 1. Have managers in competing firms commonly spoken of this company’s “style” or way of doing things? 2. Has this firm both made its values known through creed or credo and made a serious attempt to encourage managers to follow them? 3. Has the firm been managed according to longstanding policies and practices other than those of the incumbent CEO? Companies were then rated on a single item with a 5-point scale anchored by 1 = the presence over the last decade of a strong corporate culture and 5 = the presence over the last decade of a very weak or nonexistent corporate culture. Correlation coefficients ranging from .26 to .42 against various financial and market performance indices were reported. Note two important points about this study. First, the study was done with surveys and not case studies, thus permitting analyses across a sufficiently large sample of organizations. Second, the executives who made the culture strength ratings of their competitors were already familiar with the financial and market performance of those competitors and likely held the hypothesis that a strong culture is what made them effective and this influenced their ratings (Wilderom et al., 2000, p. 196, called this a “management illusion”). In any case, the results can be interpreted as suggesting that an integrative culture is superior to a fragmented culture. In this regard, perhaps what is true is that culture strength is reflected in organizational performance under only some specific boundary conditions. Thus, 395

Schneider, Ehrhart, and Macey

Copyright American Psychological Association. Not for further distribution.

earlier we showed that climate strength moderates the relationship between service climate and customer satisfaction, but, as with any moderator variable, the effect goes both ways. That is, it is also true that service climate moderates the relationship between climate strength and customer satisfaction, meaning that under more positive service climate conditions one would expect climate strength to relate to customer satisfaction. Sørensen (2002) made this possibility for culture strength as a correlate of organizational performance quite clear in his discussion of findings with regard to culture strength: The results indicate that the strength of corporate culture affects the variability of firm performance and that this relationship is contingent on the level of industry volatility. In stable environments, firms perceived to have strong corporate cultures exhibit superior and more reliable performance. This suggests that in these environments, the consensus surrounding organizational goals and values characteristic of strong-culture firms enhances their ability to exploit established competencies. The benefits of a strong culture carry a cost with respect to adaptation in volatile environments, however, as the reliability benefits of strong cultures attenuate as industry volatility increases. (pp. 85–86) J. Martin’s work, both in her 1992 and 2002 versions, portrays culture research in numerous guises, as we hope is clear from the above. The fact of the matter is that her work has received scant attention in the I/O–OB–HR (industrial/organizational, organizational behavior, human resources) literatures with no explicit tests of her three different frameworks against organizational performance (or any other outcomes like customer satisfaction, turnover, or drug abuse so far as we can find). In our review of the literature through 2008, for example, we found only about 40 citations of the 2002 book in I/O–OB–HR-related journals, and no comparative tests of the tripartite conceptualization. We think this is unfortunate because as J. Martin (2002) suggested, the three issues likely exist simultaneously at differ396

ent levels of generality and levels of analysis in most organizations depending on where one looks and at what one looks. In summary, organizations appear to be able to simultaneously have macrocultures, subcultures, and ambiguous cultures depending on the lens through which organizations are viewed. The lens is a useful metaphor because one sees different levels of specificity as one increases lens strength. Thus, in addition to occupation and hierarchical level subcultures, there has also been research on gender, racial, disability, and personal attractiveness subcultures in organizations (see Dipboye & Colella, 2005), but no work on what happens when all of those simultaneously exist. Leadership and organizational culture. As noted earlier, for Schein (1985, 1992, 2004) leadership is the source of organizational culture in organizations: What the leader pays attention to and devotes resources to establishes what is valued in the organization. Indeed Schein put the burden on leadership by noting that it is what founders of organizations pay attention to that drives culture both early and later because early decisions by the founder have a lasting impact. Leaders theoretically have this lasting impact because, as Schneider (1987; Schneider, Smith, Taylor & Fleenor, 1998) proposed, they determine the kind of people who are attracted to, selected by, and stay with an organization. A conceptualization of the effects of leadership on culture that has a similar foundation to Schein’s (2004) is one proposed by Quinn and his colleagues in the late 1970s and early 1980s called the competing values framework (Quinn, 1988; Quinn & McGrath, 1985; Quinn & Rohrbaugh, 1983). It is similar in that the focus is on the values of management in the ways organizations get designed, but it is different in that the values of management that are determinant are the ideologies they hold with regard to the outcomes the managers are trying to achieve. These ideologies or values are competing in the sense that achieving balance among them is very difficult. Two major dimensions define the model, yielding four quadrants of competing means-end values. The two dimensions are the Flexibility–Control dimension and the Internal Focus–External Focus dimension.

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

A flavor for the competing values can be had by descriptions of two of the quadrants. The Human Relations model (high Flexibility and high Internal Focus) has as the primary ends to be achieved cohesion and morale, and the means by which those ends are achieved are through, for example, training and the development of human resources. The Rational Goals model (high Control and high External Focus), on the other hand, has as the primary ends to be achieved productivity and efficiency, and the means by which these are achieved are through planning and goals setting (see Quinn & Rohrbaugh, 1983; Zammuto, Gifford, & Goodman, 2000, especially pp. 266–267). The model predicts that when the organization does anything, the people in it will respond within the primary ideology that defines it—one might say the culture that defines it. To the degree that the anything is not in keeping with the primary ideology, it will be rejected at worst and reinterpreted to fit the ideology at best. The model has huge implications for organizational change, a topic we discuss in some detail later. We agree with the emphasis Schein and Quinn gave to leadership as a determinant of culture—as we also agree with leadership as a determinant of climate—but it is possible that this emphasis on leadership is an extreme vantage point. On the one hand, Trice and Beyer (1993) devoted an entire chapter to the issue of leadership, ascribing to it important early influences and later challenges when carried out effectively leading to important adaptations to environmental changes. On the other hand, in the Ashkanasy, Wilderom, and Peterson (2000) handbook, for example, there is no chapter on leadership though there is one on the competing values framework (Zammuto et al., 2000). J. Martin (2002) indexed just a single page in her book to the topic of leadership, proposing that it is only for the integrationist perspective on culture that leadership plays a significant role. As Meindl, Ehrlich, and Dukerich (1985) noted, perhaps we have created a “romance of leadership” whereby we attribute all of these wonderful traits and behaviors to leaders, but in reality they have less impact than we suppose. Even Schein, who at times has clearly emphasized the role of leadership in the development and change of organizational culture (Schein, 1992, 2004), at

other times (e.g., Schein, 2000) pays less attention to the role of the leader, writing instead as if the culture persists over time as a thing in and of itself. We think both conceptualizations of the role of leaders are correct. Early on in the life of the organization, leaders/founders determine the culture of organizations because of the decisions they must make (culture-embedding mechanisms), and the more effective they are, the more determinant those early decisions are of the future culture of the organization. The founding of an organization is a critical period in the life of the organization, and we agree with Schein (1992, 2004) that it is in these critical periods that leadership likely has its greatest impact. The problem, as Miner (2002) noted, is that we have no empirical evidence of when and how the role of the leader gets played out in terms of organizational culture as behavior at various stages in the life of organizations (on this issue, also see Aldrich & Ruef, 2006). Summary of conceptual issues related to organizational culture. Organizational culture, as a construct of interest to organizational psychologists, emerged from thinking more closely aligned with anthropology and sociology. The original link between organizational success and culture was embedded in the thinking of those who recognized that the sustainability of organizational success is inextricably linked to the values possessed by those organizations that allow for the vision and expectations of the founder(s) to persevere and guide individual behavior and decision making. Despite the strongly implicit link between culture and organizational effectiveness, the balance of thought and research has been weighted more toward the manifestations of culture than toward understanding the links to organizational effectiveness. Other questions of conceptual interest have included how socialization processes contribute to the sustainability of what is regarded as culture and whether it is conceivable that an organization can be described as multiple cultures, even if doing so seems inconsistent with the shared notion commonly accepted as an essential characteristic of what culture is. Despite these seemingly controvertible threads, this form of thinking has helped to shape a 397

Schneider, Ehrhart, and Macey

perspective on how cultures can have ambiguous interpretations and meanings, and in turn leading to a more fully developed understanding of how culture is sustained through various embedding mechanisms including leadership practices and decisions.

Copyright American Psychological Association. Not for further distribution.

Methodological Issues and Organizational Culture The major methodological issues in research on and about organizational culture concern the approaches to culture assessment (case methods compared with survey methods) and levels issues (subcultures vs. single cultures in organizations). On the former issue, we review several organizational culture surveys, noting the degree to which they address central organizational culture issues (socialization, beliefs, values) and their validity against organizational outcomes. On the latter issue, we focus on levels, this time concerning the dilemma over whether organizations have a single macroculture or numerous subcultures characterized, for example, by careers or levels in the organization (e.g., executive compared with others). On surveys versus case studies. We have mentioned several times that some organizational culture researchers propose that research can only be “real” culture research if it is done through qualitative case analysis (Schein, 1985; Trice & Beyer, 1993). What we did not say earlier is that J. Martin (2002) now has an even more nuanced view than previously (J. Martin, 1992) of the pluses and minuses of alternative methodological approaches. She said that the key issues are the fit of the method to the problem of interest, whether the method is carried out well, and that each choice in method has its benefits and liabilities. Further, she said it is silly to dichotomize methods into qualitative and quantitative because there is so much variety in each and even hybrid methods, wherein the features of both are used simultaneously. She put it this way: “My multimethod views are not unusual anymore; the field is changing and openness to various theories, interests, and methods is growing” (J. Martin, 2002, p. 237). We agree. Indeed, Beyer also revealed a more nuanced perspective in her chapter (Beyer, Hannah, & Milton, 2000) on culture and attachment to organi398

zations where she freely cited both qualitative and survey research to make her points. Many researchers closely associated with organizational culture research have adopted the survey methodology completely; for a review of 108 (!) different culture inventories, see Taras (2006). For example, Denison (e.g., Denison, 1990; Denison & Neale, 2000) used survey methods in an early demonstration of the relationship between organizational culture and organizational performance. Not well known is the fact that much of the survey method Denison used is the latter-day version of the original work by the University of Michigan researchers, like Likert (1961, 1967), mentioned earlier in discussing the origins of climate research. Thus, Denison (1996) himself has been an important commentator on the similarities and differences of climate and culture survey research, which is not surprising given his feet in both camps. Denison’s measure assesses four major facets of organizational culture (Involvement, Consistency, Adaptability, and Mission) with each of the four macrodimensions subsuming three more specific indicators (e.g., under Involvement, the subscales are Empowerment, Team Orientation, and Capability Development). The 60-item measure assesses values, beliefs, and norms governing important organizational attempts to be effective but not the role of myths and stories, nor the presence of subcultures. Some sample items for assessment reveal the outward-looking feature characteristic of both climate and culture surveys, for example: (a) in the assessment of Empowerment, the item “Decisions are usually made where the best information is available,” (b) for Team Orientation, “Cooperation across different parts of the organization is actively encouraged,” and (c) for Capability Development, “There is continuous investment in the skills of employees.” Considerable psychometric research lies behind the current version of the measure (see http://www. Denisonconsulting.com for citations to papers). For example, Denison, Janovics, Young, and Cho (2006) summarized the research they had done in 160 companies on the reliability and validity of the measure. The scales all had internal consistency reliability estimates of .75 and higher, the subscales were modestly intercorrelated (about .50 on average across com-

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

panies), and the subscales also revealed statistically significant validity coefficients against employees’ ratings of outcomes like sales growth (.30), profits (.25), and quality (.35). In a more recent paper, Smerek and Denison (2007) presented validity evidence against hard financial outcome data including return on assets (ROA), sales growth, and market-to-book value using the same measure as in Denison et al. (2006). For this study they were able to gather follow-up outcome data for samples of the 102 companies for 3 years (e.g., N = 81 for the 1st-year follow-up, called Year 1; N = 51 for the 3rd year follow-up, called Year 3). The results can be summarized as follows: ■

■

■

All four dimensions significantly predict sales growth Year 0–1 and Year 1–2 (r averages approximately .32). Adaptability and mission significantly and consistently predict ROA across all years (r about .25). All but consistency significantly predict marketto-book value in the 2nd- and 3rd-year follow-ups based on Year 0 survey data (r approximately .35).

And in an even more recent paper (Gillespie, Denison, Haaland, Smerek, & Neale, 2008), there is similarly significant validity against customer satisfaction ratings. The work by Denison and his colleagues is done from primarily a consulting base, and to their credit, they have been rigorous in documenting the psychometrics and validity of their approach and measure not only in the United States but internationally as well. A second important survey methodological approach to culture assessment is the one developed by Chatman and her colleagues (Chatman, 1991; O’Reilly, Chatman, & Caldwell, 1991), called the Organizational Culture Profile (OCP). The original method used was a Q-sort (nine categories) of 57 adjectives by which newcomers described their own values and incumbents described the environment of the firms in which they worked (accounting firms in the original research). There are eight subscales of the OCP; sample items (with their respective subscale) include the following: analytical (Attention to Detail), demanding (Outcome Orientation), socially responsible (Aggressiveness), and high pay for performance (Emphasis on Rewards). The fit of

the new employee to the existing culture (aggregated across incumbents) was a statistically significant predictor of newcomer adjustment, commitment, and turnover. As far as we know, the main effects of culture on these outcomes were not explored. The OCP assesses what organizations are perceived to value but not issues related to how those values get played out, the myths and stories that communicate the values, socialization practices, nor the presence of subcultures; we classify the measure as a partial measure of organizational culture strictly because of its focus on values. This measure is fairly commonly used in studies of person–organization fit; for other recent applications of the OCP, see Chatman and Spataro’s (2005) research on relational demography and Erdogan, Liden, and Kraimer’s (2006) study of culture as a moderator of the relationship between justice and leader–member exchange. A third frequently used and cited measure of organizational culture is the Organizational Culture Inventory (OCI; Cooke & Rousseau, 1988; Cooke & Szumal, 2000). This measure focuses on 10 behavioral norms of firms and conceptualizes three major foci of such norms: Constructive Cultures (achieving, self-actualizing, humanistic, and affiliative norms), Passive/Defensive Cultures (approval, conventional, dependent, and avoidance norms), and Aggressive/ Defensive Cultures (oppositional, power, competitive, and perfectionist norms). Using 78 items, respondents report on the strength of the local expectations for adherence to norms as a way of reporting the thinking and behavior that characterize their workplace. For example, to assess the affiliative norm, respondents report on the degree to which members are expected to be friendly, cooperative, and sensitive to the satisfaction of their work group members. Cooke and Szumal (2000) summarized the research that has been accomplished with the OCI against group and organizational outcomes suggesting that one or more of the three major dimensions noted earlier are significantly correlated with outcomes as follows: Constructive norms relate consistently positively to teamwork, quality of work relations, product/service quality, and customer satisfaction; Passive/Defensive norms relate negatively but weakly to the same outcomes; and Aggressive/ Defensive norms relate negatively but weakly to 399

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

teamwork and quality of work relations but not to the other dimensions. Note that it is difficult to find peer-reviewed publications or Internet access to research papers that report the psychometrics and/or validity evidence cited in the Cooke and Szumal (2000) chapter, likely because the OCI is a proprietary instrument used in consulting practice and, with special permission, in dissertation research. One exception is a paper by Balthazard, Cooke, and Potter (2006), which reports on OCI data from a sample of over 60,000 respondents collected over 3.5 years at Human Synergistics. They reported alpha reliabilities of .73 or higher across all the OCI’s subscales and correlations between the OCI and what they called performance drivers (including fit, satisfaction, quality, and turnover intentions, among others). Although most correlations were statistically significant, all of the drivers were from the same source as the culture measures, and all analyses were conducted at the individual level. The OCI assesses norms but not values, beliefs, socialization practices, or language in any form (myths, stories), though it does present an interesting perspective on how organizations function. It would probably best be classified as a measure of the outer layers (Rousseau, 1990) or the artifacts (Schein, 1992, 2004) of culture. Ashkanasy, Broadfoot, and Falkus (2000) provided a useful review of 18 measures they used as a basis for the development of a hoped-for new measure called the Organizational Culture Profile (OCP). The purpose of the OCP was to provide the research community with a single measure to tap the 10 major dimensions of organizational culture as revealed in their review (see Table 8.1 and Table 8.2 in Ashkanasy, Broadfoot, & Falkus, 2000). The problem was that they discovered a 2-factor solution fit the data better than any other approximation of the factor structure, although there was some criterionrelated validity for the measure in both the 2-factor and the 10-factor model (validity indicators were not provided). We have already noted that there is some pessimism among scholars about the relationship between organizational culture and organizational performance—at least when organizational culture has been studied through the qualitative case method. However, when surveys are used, the story as we 400

have already shown here is somewhat altered. In Wilderom et al.’s (2000) review of 10 studies that investigated the relationship between culture and performance and that they deemed to be sufficiently adequate methodologically, all of them used culture surveys and not case methods and all of them revealed significant relationships with one or more hard data organizational performance outcomes. This was interesting because Wilderom et al. noted that all of the studies used different operationalizations of culture, and all of them used different organizational performance indicators (see Wilderom et al., 2000, Table 12-2, pp. 198–199). We wish to note here that culture studies have tended to be generic in that the culture surveys themselves appear to not have a particular strategic focus. As we argued when presenting the climate research, we think that more research focused on specific outcomes could prove useful. At a minimum, it is imperative that the surveys used contain items that have a conceptually clear relevance to organizational performance or the likelihood is reduced that significant direct relationships with specific outcomes will be observed. For recent examples along these lines, see Van Dyck, Frese, Baer, and Sonnentag’s (2005) empirical study of organizational error management culture and Jones, Felps, and Bigley’s (2007) theoretical discussion of stakeholder culture. In summary, it is of great interest that such different assessments of culture—focusing on such different possible facets and indicators of culture— appear to yield significant relationships with outcomes. This suggests to us that if the measures were more focused on specific outcomes as well as the generic dimensions now assessed, then the relationship with outcomes could be enhanced. Further, this also suggests to us that if more attention was paid to the importance of myths and stories, for example, as ways of socializing individuals and socialization tactics themselves were also assessed, then a fuller picture of an organization’s culture could be illuminated using quantitative survey techniques. Finally, it would appear useful if such measures also explored the presence of subcultures and the rapport/conflict between them; the measures might go well toward illuminating how these cultural manifestations may yield organizational effectiveness.

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

On levels and lenses for organizational culture research. In the earlier section on Organizational Culture and Organizational Subcultures, it became clear that culture is a many-levels phenomenon in organizations. This was especially true when we illuminated the insights of J. Martin (2002) with regard to the fragmentation model of organizational culture. We used the lens metaphor to make the point that depending on the lens through which one views organizational culture, different features of culture and different cultures become clear—career subcultures, demographic subcultures, organizational levels subcultures, and so forth. In organizational culture research, in contrast to organizational climate research, these levels and lens issues are seemingly accepted at face value without formal quantitative tests of them. For example, Lundberg (2000) wrote: While most organizations do have a more or less well-developed “umbrella” culture, it is also common that most organizational subsystems have cultures also, e.g., division or departments, major layers of the organization, as do the major occupations of the organization. And there is always the question of congruence among subsystem cultures and between them and the umbrella culture . . . and subsystems cultures may vary considerably along a continuum of intra-culture clarity and consensus. (p. 329) But there is precious little empirical research on these levels, subsystems, congruence, or consensus issues in the culture research literature whether qualitative or quantitative. There are, of course, a few notable exceptions. For example, Jermier, Slocum, Fry, and Gaines (1991) used a mix of quantitative and qualitative methods to identify five distinct subcultures within a police organization, emphasizing the similarities and differences of each with the organization’s “official” culture. In addition, to investigate the effects of culture strength and performance, as we noted earlier in the discussion of culture strength, Sørensen (2002) showed that high culture strength was associated with more reliable

organizational performance over time and that this benefit decreased as industry volatility increased. In contrast to these studies, the organizational culture surveys we reviewed earlier all tested their validity against outcomes at the firm level of analysis and did not test for levels, subcultures, consensus, or congruence of any kind. Especially given the use of surveys, it seems obvious that culture researchers might begin to explore such differences that might exist from the different levels perspectives to see where, when, and why they exist. Sparrow (2000) in the Cooper et al. (2000) volume made this point clearly in his excellent treatment of levels issues, but the other chapters in Cooper et al. (the one by Payne, 2000, is an exception) appear relatively immune to his suggestions. One levels issue almost no one is discussing in either the climate or culture literatures concerns cross-cultural effects on organizational culture (or climate for that matter). The most complete statement of the issue we found is in a series of chapters written for the GLOBE (House, Hanges, Javidan, Dorfman, & Gupta, 2004) report of research on culture and leadership in 62 societies around the world. GLOBE was a massive effort, developing measurement of both organizational culture and society culture, providing an opportunity to explore their interrelationships. The basic model under study was one proposed by Kanungo and Jaeger (1990) stating in brief that societal-level culture is likely to influence organizational culture by the shared assumptions managers in a society have about the nature of employees and how organizations must be structured given those employee attributes. Readers can easily see the challenges connected with such a proposal, and the lens metaphor makes it easier. That is, a lens focused at the level of the basic cultures around the world can see the differences among societies; there is a main effect for societies. At the same strength, though, differences across organizations within a society will not be obvious because the lens is not strong enough to see such differences; there is no main effect for organizations. So, one must have a stronger lens to see the differences within a society. By using both lenses at the same time, one may be able to generate data that permit examination of the relationship between what one 401

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

sees with the weaker lens and what one sees with the stronger lens. GLOBE was able to do this in the development of the measures used, a wonderful accomplishment to say the least. In a series of chapters in the GLOBE book, various research teams were asked to explore the resultant collected data along nine dimensions of societal and organizational culture (e.g., power–distance, individualism–collectivism, gender egalitarianism, uncertainty avoidance). In all cases the data were analyzed using hierarchical linear modeling so that the effects could be examined at multiple levels of analysis. In brief, the results may be summarized as follows: First, societal cultural values and societal practices are significantly related to organizational values and practices, and this is especially true for societal values and their influence on organizational culture values. Second, across societies there were significant Industry × Organizational Culture values effects for four of the nine cultural values studied. However, when controlling for the societal culture effects on organizational culture, all but one of those interactions became nonsignificant. In short, these results indicate that, as hypothesized in the GLOBE effort, societal culture has a pervasive effect on the organizational cultures of firms in that society. Multilevel approaches to the study of organizational climate and culture are a future need in both fields, and the development of multilevel analysis programs, like hierarchical linear modeling, makes possible the study of such cross-levels effects on individual, group, unit, organizational, and societal outcomes. In this vein it is interesting to recall that McClelland (1961) did a variation on this 50 years ago, described in the wonderful book The Achieving Society. McClelland showed that societal practices (e.g., the readers used to teach students to read) had influences on aggregated individual levels of achievement motivation in a society and were useful predictors of societal accomplishments (at least in terms of various economic indicators of success). In summary, it is clear that culture is a multilayered and multilevel construct and, depending on the lens or lenses with it is viewed, it will emerge in perhaps different forms and with perhaps different main and interactional effects. More attention focused on the specific level at which culture research is 402

being conducted, as in climate research, might produce more consensus on the status and correlates of culture, including performance correlates.

Culture and Organizational Change Organizational change can be viewed from both conceptual and methodological perspectives and there are various models of change (W. W. Burke, 2008), but they all begin with an assessment of the current state of the organization. This baseline is used both for hints about what requires change and as a standard against which the effects of changes that are implemented can be judged. Because most organizations are changed one at a time, much of the baseline work is done using a hybrid model so that (a) results of the survey diagnosis can be judged against norms for the measure, and (b) the changes to be implemented are specific to the organization’s unique language, myths, stories, subcultures, and socialization practices. Our perspective on organizational change is one we share with Michela and Burke (2000), namely, that the more facets of culture and climate that one can assess, the more likely it is that key facets inhibiting organizational effectiveness can be identified. Perhaps of equal importance is that the key factors that have promoted effectiveness where it exists can also be identified and appreciated so that change does not alter them except through possible enhancements (Srivistava & Cooperrider, 1990). Companies appear to be increasingly concerned with the necessity to focus on change. This increase has several causes, beginning with the desire to become more efficient (to produce more with fewer people and their associated expenses), to respond to the ever-increasing rapidity of changes in the marketplace (at a minimum keeping up with Joneses), to deal with the expected retirements of the baby boomers (and the associated loss of expertise and knowledge), and to cope with the integrations required as a result of mergers and acquisitions, joint ventures, strategic alliances, and the like (W. W. Burke, 2008). Culture diagnoses help organizations at least know where they are before changing for any of these reasons, and all of these changes have cultural implications. For instance, a focus on efficiency may result in changes in core assumptions about people and the implicit obliga-

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

tions between the organization and its employees. Literature on differences between the generations (e.g., Howe & Strauss, 2000) suggests that the younger employees who will come to dominate the workplace because of the retirement of baby boomers will bring a different set of values to their work; over time, those values will have more and more of an impact on the culture of the organization. The literature on mergers and acquisitions is particularly interesting with regard to the implications of organizational culture. For instance, research on these issues has identified key reasons for failure of change, with a particular focus on differences that exist between the parties trying to integrate (W. W. Burke & Biggart, 1997, as summarized in W. W. Burke, 2008, pp. 106–107): (a) imbalance of power and control between organizations; (b) imbalance of expertise, status, or prestige between the two parties; and (c) lack of perceived equity, for example, in distributions of key roles and jobs. A significant problem with the culture research on organizational integrations is that it is meager at best; there just is not much research. In his extensive review, Weber (2000, p. 309) put the state of affairs this way: “Most of it [the literature] is based on observations by practitioners and consultants, with little theoretical or empirical support.” When we read the literature, we agree; even the research (like the W. W. Burke and Biggart, 1997, study cited earlier) fails for the most part to actively invoke the issues that make the study of organizational culture such a rich possible vehicle for understanding why and how these integrations might succeed. That is, if the diagnoses focused on the language and stories, the socialization practices, and the subcultures of the parties to the integration, they would have an increased appreciation for how they are different and perhaps ways they are similar. It is not just imbalances in jobs and roles or imbalances in expertise or status that a cultural lens offers here but how organizations differ in the deepest parts of their identity. As noted by Schein (2004), “a cultural mismatch in an acquisition or merger is as great a risk as a financial, product, or market mismatch” (p. 411). The ultimate conclusion of a number of culture researchers is that the best way for organizations to handle organizational change is to create a culture that

has at its core the values of constant improvement and adaptation. Schein (2004) dedicated the final chapter of his book to learning cultures and learning leaders, drawing from such sources as Senge (1990) to emphasize how the learning culture represents a way to take the tendency for culture to be a force for stability and predictability and turn it on its head. Through a set of assumptions that emphasize the importance of learning and innovation above all else, the organization’s culture will not by necessity become a source of rigidity and constraint, but will allow the organization to effectively adapt to an ever-changing environment. Kotter and Heskett (1992) also described how strong cultures and strategically appropriate cultures (those designed to best fit environmental conditions) will ultimately not be as effective as adaptive cultures that promote flexibility, initiative, and risk taking. Moreover, they specifically noted that the way to create such a culture is to follow the philosophy that the organization must take care of its core stakeholders (customers, employees, and stockholders). By paying attention to all three of these groups, the organization does not become beholden to one certain way of doing things but is willing to change to best serve all of these groups. Note that this approach fits well with what we have been advocating; by focusing on the outcomes pertinent to each of these groups, cultures with a particular strategic focus can be developed that will enhance organizational effectiveness. Aside from the quote from Weick (1985) we presented earlier equating culture with strategy, it is very hard to find a mention of strategy in the culture literature. So, in addition to a relative absence of research on levels and subcultures, there is a lack of research on strategic cultures as well. Nadler, Thies, and Nadler (2000) argued that culture researchers must begin to conceptualize organizations as strategic enterprises and asked the degree to which organizational cultures achieve focus and leverage through the ways they organize their subsystems. They meant by “organize” a very specific way of doing things: focusing on activities that simultaneously achieve new strategic foci and capitalize on core capabilities of the organization. As with climate research, such strategic foci might serve culture research (and practice) well precisely because such foci will target the issues to 403

Schneider, Ehrhart, and Macey

be studied, the degree to which subsystems of the organization are appropriately and congruently focused, the language and myths that surround such strategic foci, and the socialization of newcomers to the values of the organization.

Copyright American Psychological Association. Not for further distribution.

Summary of the Organizational Culture Paradigm Organizational culture, as a construct of interest to organizational psychologists, emerged from thinking more closely aligned with anthropology and sociology. The original link between organizational success and culture was embedded in the thinking of those who recognized that the sustainability of organizational success is inextricably linked to the values possessed by those organizations that allow for the vision and expectations of the founder(s) to persevere and guide aggregate behavior and effective decision making. Despite the strong implicit link between culture and organizational effectiveness, the early thought and research was weighted more toward how organizational culture is created and its various manifestations in organizations than toward understanding the links of culture to organizational effectiveness. This has somewhat changed with the introduction and acceptance of survey research that gets at one or more aspects of culture (e.g., values, socialization practices). This content differs from the content of climate surveys that focus more on policies, practices, and procedures, but like climate surveys, can be linked directly to various indicators of organizational performance. So, the questions of conceptual interest from the culture paradigm have included how socialization processes contribute to the sustainability of what is regarded as “culture,” the role of leadership in “embedding” culture in organizations, organizational change, and whether an organization can be described as being simultaneously multiple cultures and being a “shared” culture, the more usual or commonly accepted notion of what culture is. Despite these conceptual advances in understanding the psychology of organizations, the direct empirical link of organizational culture to organizational performance and effectiveness had been somewhat elusive until the more recent survey approaches to 404

assessing culture demonstrated such relationships. Our conclusion is that the culture perspective, with its focus on the origins of culture (leadership, myths, and stories) and the sustaining of culture (socialization), offers important insights into not only understanding organizational behavior but, if focused on strategic outcomes, understanding organizational effectiveness and change as well. SUMMARY AND INTEGRATION: TOWARD A NEW PARADIGM FOR RESEARCH AND PRACTICE Climate and culture research and thinking have existed on seemingly parallel tracks with little explicit conceptual overlap (Ashkanasy, Wilderom, & Peterson, 2000; Reichers & Schneider, 1990); the siblings have not been communicating with each other much, at least not in public. The present review indicates that there has been much happening out of the public eye on which we might build an integrated model of the psychology of organizations. One approach to creating an integrated model emerges from culture researchers’ adoption of the survey method for research and practice. A second approach emerges from consideration of the simultaneous study of both strategic or focused climates and the foundation issues that support the focus (Schneider, White, & Paul, 1998). It is not a huge conceptual leap to think of the foundation climate issues as the values and beliefs about the human element in organizations about which culture researchers speak. Especially when supported by effective leadership to ensure those foundation values and beliefs are embedded (Schein, 2004), the combination of climate and culture can make the importance of human resources for organizational competitiveness as clear today as they were when the field of organizational psychology began (Likert, 1961, 1967; McGregor, 1960; Schein, 1965). The simple model we propose is shown in Figure 12.1; let’s call it the climcult model. Positive values and beliefs about humans are embedded in organizations as the basis for a positive organizational culture of well-being. The implication of this is that executives of the company attend to, make decisions supporting, and monitor and measure the well-being

Perspectives on Organizational Climate and Culture

Positive values about people

Culture of well-being

Leadership simultaneously values people and promotes strategy

Success in competing for and retaining talent Organizational effectiveness

Policies, practices and procedures

Strategically relevant climate(s)

Success in the competitive marketplace

Copyright American Psychological Association. Not for further distribution.

FIGURE 12.1. The “climcult model”: An integrated research paradigm for organizational climate and culture.

of the firm’s human assets. This base provides a foundation on which strategically relevant climates can be built. It is important to note that the foundation culture of well-being is not the cause of the strategic climate; it provides the foundation on which such a climate can be built. The implication is that actions focused on employee well-being will not necessarily produce strategic accomplishments; more is required if, for example, safety or service or innovation is to be promoted as a strategic path to success. But, as shown in the figure, a direct positive consequence of a focus on employee well-being will be the attraction and retention of the talent organizations require for competitive advantage (Lawler, 2008). Leadership plays a central role in this model, first by ensuring the embedding of the HR practices that connote valuing people (Bowen & Ostroff, 2004), and second by implementing the practices, policies, and procedures necessary to attain strategic objectives by rewarding, supporting, and expecting strategically appropriate behavior (Schneider et al., 2005). We learn from the climate literature the importance of focus, and we learn from the culture literature not only the possible positive pervasive effects of positive values and beliefs vis-à-vis humans at work but also the importance of ensuring that both focus and values are communicated through the socialization experiences of newcomers to organizations. Thus, we propose that future “climcult” research must address not only the climate for well-being and one or more strategic climates but also the socialization experiences newcomers have as they move from newcomer to old-timer in the organization. We also learn from the socialization literature that newcomers

must learn the ropes not only of the way the organization functions vis-à-vis humans at work but also the strategic goals toward which they should be focusing their energies and competencies (Louis, 1990). Culture researchers, according to our review and model, have been deficient in their strategic focus on organizational effectiveness, and we propose that future culture research also include the degree to which the culture has the strategic focus or foci most likely to enhance the probability of success in a competitive marketplace environment. Recall that Weick (1985) noted that there is considerable overlap between culture and strategy, but we believe this is because both emerge implicitly over time with no focus. In a world requiring organizations to choose their strategic focus or foci, they also must choose how they will reveal their valuing of their employees— and culture researchers must document both. Our review also suggests the likelihood that organizations may simultaneously have unified and differentiated cultures (J. Martin, 2002); that is, they differ in the strength with which the culture is shared. Thus rather than it being a dichotomy— unified cultures or differentiated subcultures— both can exist simultaneously. That is, leaders can imbue throughout their organizations positive values about people (a unified culture) and approach their common strategic vision through subcultures that are appropriate for the various functions in organizations, levels in the hierarchy, and so forth (a differentiated culture). Two conditions make this possible: (a) consensus on how positively valued people are, and (b) consensus on the strategic imperative(s) of the organization. When a strong 405

Copyright American Psychological Association. Not for further distribution.

Schneider, Ehrhart, and Macey

culture of human well-being and a strong strategic climate both exist, the potential for organizational effectiveness is considerably enhanced. This is true, we propose, even though simultaneously in different functions and at different levels of the organization, how the strategy is implemented may differ. For example, in a restaurant a strong culture of employee well-being can exist simultaneously with a strong service culture for wait-staff and a strong efficiency/ reliability climate in the kitchen. Finally, the myths, stories, and history by which culture emerges and is transmitted in organizations are a key construct that is worth borrowing for the integrated climcult model as well. Climate researchers have been almost completely nonhistorical in their thinking and research. Failure to consider whence cometh the experienced climate— how did it come to be the way it is?—has been a strength of the culture research paradigm, although as we showed in our review the survey approaches to culture measurement have also tended to ignore these facets of culture history and transmission. But climate and culture have not been of interest just to academic researchers; the topics have been of considerable import to practitioners as well. The important thread practitioners have taken from these ideas is the notion of focusing on alignment of organization members with each other and with a common purpose or goal related to organizational effectiveness. We think the emphasis on alignment in the organizational culture camp (e.g., “strong cultures” through strong socialization practices and an emphasis on shared myths and stories) was one of the keys to its emergence in the practice literature. Our sense is that neither culture nor climate researchers have focused recently on the processes by which sharedness occurs—the processes by which strong climates or cultures are created—and this has been a deficit in both literatures (for an exception, see Bowen & Ostroff, 2004). But we doubt that organizational leaders care about some of the issues that worry the research community, particularly with regard to the refinements of what is culture or climate or alignment or fragmentation or, especially, ICCs. Executives understand that it is important for people to be “on the same page” about their human assets (have strong cultures of 406

well-being) and “do what is right” (have strong strategic climates). One area in which executives need help is in understanding how strong cultures and climates happen. This is something more than simply revealing the myths and stories that are the relevant signposts to the culture researcher. It requires understanding how those myths and stories emerge over time. At the extreme, it is said that history is written by the survivors, and it may be that the stories people tell newcomers represent what old-timers in retrospect see as consistent with today’s success (Schein, 2004). So, we would not suggest to an interested executive that by telling the right kind of stories, the executive is likely to promote the kind of climcult he or she wants. Rather, we would help the executive realize that his or her decisions, when consistent and aligned with the stories he or she wants to be told, and when consistent with what others are doing and telling, help to create the climcult the executive wants. This suggests that the culture perspective, which focuses on the past, is insufficient in a dynamic and changing world. By this we mean that what is required for a climcult paradigm of the future is a change paradigm, one that focuses on the creation of climcults that can be successful through a focus on people and strategy leading to the new myths and stories told by incumbents to newcomers. This would be especially true as organizations become multinational in focus and as the “same” work for the same company (e.g., call center work) is being done 24 hours per day across the globe in New York, Ireland, India, Singapore, and San Francisco. If societal cultures dominate organizational cultures (House et al., 2004), then far-flung companies have an especially difficult problem if they attempt to impose their ways of doing things on units in other countries. Further, if the goal of a company is to create a strong culture, this is just the opposite of what is required if the company needs to be adaptable to changing times and operating simultaneously in different cultures (Kotter & Heskett, 1992; Sørensen, 2002). What myths and stories one shares, and how socialization should play out around the world, are unanswered questions right now. Perhaps the most appropriate advice is to ensure as best as possible a strong strategically focused climate

Perspectives on Organizational Climate and Culture

in each location and create the culture of well-being that is most appropriate for each location.

Copyright American Psychological Association. Not for further distribution.

CONCLUSION As authors who have scanned both the organizational climate and the organizational culture literatures, we are impressed by the depth and breadth of knowledge there is about these important organizational constructs. The review in this chapter reveals how important these are for effectiveness in organizations both for the people who work in them and for the accomplishments those organizations may have in the marketplace. Companies do not much care about whether it is climate or culture, but our review shows it does matter. Focusing just on values is not good enough; one must also focus on strategic behavior. Focusing on people is not good enough; one must also focus on the important strategic outcomes. Focusing on polices, practices, and procedures is not good enough; one must also focus on socialization practices, myths, and stories. In the climcult model we can see the siblings organizational climate and organizational culture coming to grips with who they are and what they are—their identities have become sufficiently solidified for them to learn from and work together. This certainly does not mean that their work is done, because now they must integrate across their own built-up myths and stories to create a new, more vibrant field in which they simultaneously study the values and beliefs of organizations vis-à-vis people and the creation and maintenance of the climates required for success; if they succeed, a vibrant climcult paradigm for both science and practice is possible.

References Aldrich, H. E., & Ruef, M. (2006). Organizations evolving (2nd ed.). Thousand Oaks, CA: Sage. Argyris, C. (1957). Personality and organization. New York: Harper. Ashforth, B. (1985). Climate formation: Issues and extensions. Academy of Management Review, 10, 837–847. doi:10.2307/258051 Ashkanasy, N. M., Broadfoot, L. E., & Falkus, S. (2000). Questionnaire measures of organizational culture. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F.

Peterson (Eds.), Handbook of organizational culture and climate (pp. 131–146). Thousand Oaks, CA: Sage. Ashkanasy, N. M., Wilderom, C. P. M., & Peterson, M. F. (Eds.). (2000). Handbook of organizational culture and climate. Thousand Oaks, CA: Sage. Balthazard, P. A., Cooke, R. E., & Potter, R. A. (2006). Dysfunctional culture, dysfunctional organization: Capturing the behavioral norms that form organizational culture and drive performance. Journal of Managerial Psychology, 21, 709–732. Bass, B. M. (1990). Bass and Stogdill’s handbook of leadership: Theory, research and managerial applications (3rd ed.). New York: Free Press. Beyer, J. M., Hannah, D. R., & Milton, L. P. (2000). Ties that bind: Culture and attachment to organizations. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 323–338). Thousand Oaks, CA: Sage. Bliese, P. D. (1998). Group size, ICC values, and group level correlations: A simulation. Organizational Research Methods, 1, 355–373. doi:10.1177/ 109442819814001 Bliese, P. D. (2000). Within-group agreement, nonindependence, and reliability: Implications for data aggregation and analyses. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research and methods in organizations: Foundations, extensions, and new directions (pp. 349–381). San Francisco: Jossey-Bass. Bowen, D. E., & Ostroff, C. (2004). Understanding HRM–firm performance linkages: The role of the “strength” of the HRM system. Academy of Management Review, 29, 203–221. Brown, R. D., & Hauenstein, N. M. A. (2005). Interrater agreement reconsidered: An alternative to the rWG indices. Organizational Research Methods, 8, 165–184. doi:10.1177/1094428105275376 Burke, M. J., Borucki, C., & Hurley, A. (1992). Reconceptualizing psychological climate in a retail service environment: A multiple stakeholder perspective. Journal of Applied Psychology, 77, 717–729. doi:10.1037/0021-9010.77.5.717 Burke, M. J., Chan-Serafin, S., Salvador, R., Smith, A., & Sarpy, S. A. (2008). The role of national culture and organizational climate in safety training effectiveness. European Journal of Work and Organizational Psychology, 17, 133–152. doi:10.1080/ 13594320701307503 Burke, M. J., Finkelstein, L. M., & Dusig, M. S. (1999). On average deviation indices for estimating interrater agreement. Organizational Research Methods, 2, 49–68. doi:10.1177/109442819921004 Burke, W. W. (2008). Organization change: Theory and practice (2nd ed.). Thousand Oaks, CA: Sage. 407

Schneider, Ehrhart, and Macey

Burke, W. W., & Biggart, N. W. (1997). Interorganizational relations. In D. Druckman, J. E. Singer, & H. Van Cott (Eds.), Enhancing organizational performance (pp. 120–149). Washington, DC: National Academy Press. Campbell, J. P., Dunnette, M. D., Lawler, E. E., III, & Weick, K. E. (1970). Managerial behavior, performance, and effectiveness. New York: McGraw-Hill.

Copyright American Psychological Association. Not for further distribution.

Carr, J. Z., Schmidt, A. M., Ford, J. K., & DeShon, R. P. (2003). Climate perceptions matter: A meta-analytic path analysis relating molar climate, cognitive and affective states, and individual level work outcomes. Journal of Applied Psychology, 88, 605–619. doi:10.1037/0021-9010.88.4.605 Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. doi:10.1037/ 0021-9010.83.2.234 Chatman, J. (1991). Matching people and organizations: Selection and socialization in public accounting firms. Administrative Science Quarterly, 36, 459–484. doi:10.2307/2393204 Chatman, J., & Spataro, S. (2005). Getting people to cooperate: Understanding relational demographybased variations in people’s responsiveness to organizational inducements. Academy of Management Journal, 48, 321–331. Clarke, S. (2006). The relationship between safety climate and safety performance: A meta-analytic review. Journal of Occupational Health Psychology, 11, 315–327. doi:10.1037/1076-8998.11.4.315 Colquitt, J. A., Noe, R. A., & Jackson, C. L. (2002). Justice in teams: Antecedents and consequences of procedural justice climate. Personnel Psychology, 55, 83–109. doi:10.1111/j.1744-6570.2002.tb00104.x

Dastmalchian, A. (2008). Industrial relations climate. In P. Blyton, N. Bacon, J. Fiorito, & E. Heery (Eds.), Sage handbook of industrial relations (pp. 548–571). Thousand Oaks, CA: Sage. Dawson, J. F., Gonzalez-Roma, V., Davis, A., & West, M. A. (2008). Organizational climate and climate strength in UK hospitals. European Journal of Work and Organizational Psychology, 17, 89–111. doi:10.1080/ 13594320601046664 Deal, T. E., & Kennedy, A. A. (1982). Corporate cultures. Reading, MA: Addison-Wesley. Denison, D. R. (1990). Corporate culture and organizational effectiveness. New York: Wiley. Denison, D. R. (1996). What is the difference between organizational culture and organizational climate? A native’s point of view on a decade of paradigm wars. Academy of Management Review, 21, 619–654. doi:10.2307/258997 Denison, D. R., Janovics, J., Young, J., & Cho, H. J. (2006). Diagnosing organizational cultures. Lausanne, Switzerland: International Institute for Management Development. Denison, D. R., & Neale, W. (2000). Denison organizational culture survey. Ann Arbor, MI: Denison Consulting. Dickson, M. W., Resick, C. J., & Hanges, P. J. (2006). When organizational climate is unambiguous it is also strong. Journal of Applied Psychology, 91, 351–364. doi:10.1037/0021-9010.91.2.351 Dietz, J., Pugh, S. D., & Wiley, J. W. (2004). Service climate effects on customer attitudes: An examination of boundary conditions. Academy of Management Journal, 47, 81–92. Dipboye, R. L., & Colella, A. (Eds.). (2005). Discrimination at work: The psychological and organizational bases. Mahwah, NJ: Erlbaum.

Cooke, R. A., & Rousseau, D. M. (1988). Behavioral norms and expectations: A quantitative approach to the assessment of organizational culture. Group and Organization Studies, 13, 245–273. doi:10.1177/ 105960118801300302

Dragoni, L. (2005). Understanding the emergence of state goal orientation in organizational work groups: The role of leadership and multi-level climate perceptions. Journal of Applied Psychology, 90, 1084–1095. doi:10.1037/0021-9010.90.6.1084

Cooke, R. A., & Szumal, J. L. (2000). Using the organizational culture inventory to understand the operating couture of organizations. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 147–162). Thousand Oaks, CA: Sage.

Ehrhart, M. G. (2004). Leadership and procedural justice climate as antecedents of unit-level organizational citizenship behavior. Personnel Psychology, 57, 61–94. doi:10.1111/j.1744-6570.2004.tb02484.x

Cooper, C. L., Cartwright, S., & Earley, P. C. (Eds.). (2000). The international handbook of organizational culture and climate. Chichester, England: Wiley. D’Amato, A., & Zijlstra, F. R. H. (2008). Psychological climate and individual factors as antecedents of work outcomes. European Journal of Work and Organizational Psychology, 17, 33–54. doi:10.1080/ 13594320701307420 408

Erdogan, B., Liden, R. C., & Kraimer, M. L. (2006). Justice and leader–member exchange: The moderating role of organizational culture. Academy of Management Journal, 49, 395–406. Etzioni, A. (1961). A comparative analysis of complex organizations. New York: Free Press. Fleishman, E. A. (1953). Leadership climate, human relations training and supervisory behavior. Personnel Psychology, 6, 205–222. doi:10.1111/j.1744-6570. 1953.tb01040.x

Copyright American Psychological Association. Not for further distribution.

Perspectives on Organizational Climate and Culture

Forehand, G. A., & Gilmer, B. H. (1964). Environmental variation in studies of organizational behavior. Psychological Bulletin, 62, 361–382. doi:10.1037/ h0045960

House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V. (2004). Culture, leadership and organizations: The GLOBE study of 62 societies. Thousand Oaks, CA: Sage.

George, J. M., & James, L. R. (1993). Personality, affect, and behavior in groups revisited: Comment on aggregation, levels of analysis, and a recent application of within and between analysis. Journal of Applied Psychology, 78, 798–804. doi:10.1037/0021-9010. 78.5.798

Howe, N., & Strauss, W. (2000). Millennials rising: The next great generation. New York: Vintage Books.

Gettman, H. J., & Gelfand, M. J. (2007). When the customer shouldn’t be king: Antecedents and consequences of sexual harassment by clients and customers. Journal of Applied Psychology, 92, 757–770. doi:10.1037/0021-9010.92.3.757 Gillespie, M. A., Denison, D. R., Haaland, S., Smerek, R., & Neale, W. S. (2008). Linking organizational culture and customer satisfaction: Results from two companies in different industries. European Journal of Work and Organizational Psychology, 17, 112–132. doi:10.1080/ 13594320701560820 Gittell, J. H. (2002). Relationships between service providers and their impact on customers. Journal of Service Research, 4, 299–311. doi:10.1177/ 1094670502004004007 Glick, W. H. (1985). Conceptualizing and measuring organizational and psychological climate: Pitfalls of multilevel research. Academy of Management Review, 10, 601–616. doi:10.2307/258140 Gonzalez-Roma, V., Peiro, J. M., & Tordera, N. (2002). An examination of the antecedents and moderator influences of climate strength. Journal of Applied Psychology, 87, 465–473. doi:10.1037/0021-9010. 87.3.465 Greiner, L. E. (1983). Senior executives as strategic actors. Nursing Management, 1, 11–15.

James, L. A., & James, L. R. (1989). Integrating work environment perceptions: Explorations into the measurement of meaning. Journal of Applied Psychology, 74, 739–751. doi:10.1037/0021-9010.74.5.739 James, L. R. (1982). Aggregation bias in estimates of perceptual agreement. Journal of Applied Psychology, 67, 219–229. doi:10.1037/0021-9010.67.2.219 James, L. R., Choi, C. C., Ko, C. H. E., McNeil, P. K., Minton, M. K., Wright, M. A., & Kim, K. (2008). Organizational and psychological climate: A review of theory and research. European Journal of Work and Organizational Psychology, 17, 5–32. doi:10.1080/ 13594320701662550 James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85–98. doi:10.1037/0021-9010.69.1.85 James, L. R., Demaree, R. G., & Wolf, G. (1993). rWG: An assessment of within-group interrater agreement. Journal of Applied Psychology, 78, 306–309. doi:10.1037/0021-9010.78.2.306 James, L. R., & Jones, A. P. (1974). Organizational climate: A review of theory and research. Psychological Bulletin, 81, 1096–1112. doi:10.1037/h0037511 James, L. R., & Tetrick, L. E. (1986). Confirmatory analytic tests of three causal models relating job perceptions to job satisfaction. Journal of Applied Psychology, 71, 77–82. doi:10.1037/0021-9010.71.1.77

Guion, R. M. (1973). A note on organizational climate. Organizational Behavior and Human Performance, 9, 120–125. doi:10.1016/0030-5073(73)90041-X

Jermier, J. M., Slocum, J. W., Fry, L. W., & Gaines, J. (1991). Organizational subcultures in a soft bureaucracy: Resistance behind the myth and façade of an official culture. Organization Science, 2, 170–194. doi:10.1287/orsc.2.2.170

Hellriegel, D., & Slocum, J. W., Jr. (1974). Organizational climate: Measures, research, and contingencies. Academy of Management Journal, 17, 255–280. doi:10.2307/254979

Jones, T. M., Felps, W., & Bigley, G. (2007). Ethical theory and stakeholder-related decisions: The role of stakeholder culture. Academy of Management Review, 32, 137–155.

Hicks-Clarke, D., & Iles, P. (2000). Climate for diversity and effects on career and organizational attitudes and perceptions. Personnel Review, 29, 324–345. doi:10.1108/00483480010324689

Kanungo, R. N., & Jaeger, A. M. (1990). Introduction: The need for indigenous management in developing countries. In A. M. Jaeger & R. N. Kanungo (Eds.), Management in developing countries (pp. 1–23). London: Routledge.

Hofmann, D. A., & Mark, B. (2006). An investigation of the relationship between safety climate and medication errors as well as other nurse and patient outcomes. Personnel Psychology, 59, 847–869. doi:10.1111/ j.1744-6570.2006.00056.x

Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.). New York: Wiley.

Holland, J. L. (1997). Making vocational choices (3rd ed.). Odessa, FL: Psychological Assessment Resources.

Klein, K. J., Bliese, P. D., Kozlowski, S. W. J., Dansereau, F., Gavin, M. B., Griffin, M. A., et al. (2000).

Katz, D., & Kahn, R. L. (1966). The social psychology of organizations. New York: Wiley.

409

Schneider, Ehrhart, and Macey

Multilevel analytic techniques: Commonalities, differences, and continuing questions. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research and methods in organizations: Foundations, extensions, and new directions (pp. 512–553). San Francisco: Jossey-Bass.

Copyright American Psychological Association. Not for further distribution.

Klein, K. J., Conn, A. B., Smith, D. B., & Sorra, J. S. (2001). Is everyone on agreement? Exploring the determinants of within-group agreement in survey responses. Journal of Applied Psychology, 86, 3–16. doi:10.1037/ 0021-9010.86.1.3 Klein, K. J., Conn, A. B., & Sorra, J. S. (2001). Implementing computerized technology: An organizational analysis. Journal of Applied Psychology, 86, 811–824. doi:10.1037/0021-9010.86.5.811 Klein, K. J., & Sorra, J. S. (1996). The challenge of innovation implementation. Academy of Management Review, 21, 1055–1080. doi:10.2307/259164 Kotter, J. P., & Heskett, J. L. (1992). Corporate culture and performance. New York: Free Press. Kozlowski, S. W. J., & Doherty, M. L. (1989). An integration of climate and leadership: Examination of a neglected issue. Journal of Applied Psychology, 74, 546–553. doi:10.1037/0021-9010.74.4.546 Kozlowski, S. W. J., & Hattrup, K. (1992). A disagreement about within-group agreement: Disentangling issues of consistency versus consensus. Journal of Applied Psychology, 77, 161–167. doi:10.1037/0021-9010. 77.2.161 Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research and methods in organizations: Foundations, extensions, and new directions (pp. 3–90). San Francisco: Jossey-Bass.

Lewin, K., Lippitt, R., & White, R. K. (1939). Patterns of aggressive behavior in experimentally created “social climates.” Journal of Social Psychology, 10, 271–299. Liao, H., & Chuang, A. (2004). A multilevel investigation of factors influencing employee service performance and customer outcomes. Academy of Management Journal, 47, 41–58. Liao, H., & Rupp, D. E. (2005). The impact of justice climate and justice orientation on work outcomes: A cross-level multifoci framework. Journal of Applied Psychology, 90, 242–256. doi:10.1037/0021-9010. 90.2.242 Likert, R. (1961). New patterns of management. New York: McGraw-Hill. Likert, R. (1967). The human organization. New York: McGraw-Hill. Lindell, M. K., & Brandt, C. J. (2000). Climate quality and climate consensus as mediators of the relationship between organizational antecedents and outcomes. Journal of Applied Psychology, 85, 331–348. doi:10.1037/0021-9010.85.3.331 Litwin, G. H., & Stringer, R. A. (1968). Motivation and organizational climate. Cambridge, MA: Harvard Business School, Division of Research. Louis, M. (1990). Acculturation in the work place: Newcomers as lay ethnographers. In B. Schneider (Ed.), Organizational climate and culture (pp. 85–129). San Francisco: Jossey-Bass. Lundberg, C. C. (2000). Working with cultures in organizations. In C. L. Cooper, S. Cartwright, & P. C. Earley (Eds.), The international handbook of organizational culture and climate (pp. 325–346). Chichester, England: Wiley. Luria, G. (2008). Climate strength: How leaders form consensus. The Leadership Quarterly, 19, 42–53. doi:10.1016/j.leaqua.2007.12.004

Kuenzi, M., & Schminke, M. (2009). Assembling fragments into a lens: A review, critique, and proposed research agenda for the organizational work climate literature. Journal of Management. doi:10.1177/ 0149206308330559.

Major, D. A. (2000). Effective newcomer socialization into high performance organizational cultures. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 355–368). Thousand Oaks, CA: Sage.

LaFollette, W. R., & Sims, H. P., Jr. (1975). Is satisfaction redundant with climate? Organizational Behavior and Human Performance, 10, 118–144.

Martin, J. (1992). Cultures in organizations: Three perspectives. New York: Oxford University Press.

Lance, C. E., Butts, M. M., & Michels, L. C. (2006). The sources of four commonly reported cutoff criteria: What did they really say? Organizational Research Methods, 9, 202–220. doi:10.1177/1094428105284919

Martin, J. (2002). Organizational culture: Mapping the terrain. Thousand Oaks, CA: Sage.

Lawler, E. E., III. (2008). Talent: Making people your competitive advantage. San Francisco: Jossey-Bass.

Martin, J., & Frost, P. (1996). The organizational culture war games: A struggle for intellectual dominance. In S. R. Clegg, C. Hardy, & W. R. Nord (Eds.), Handbook of organizational studies (pp. 599–621). Thousand Oaks, CA: Sage.

LeBreton, J. M., & Senter, J. L. (2008). Answers to twenty questions about interrater reliability and interrater agreement. Organizational Research Methods, 11, 815–852. doi:10.1177/1094428106296642

Martin, K. D., & Cullen, J. B. (2006). Continuities and extensions of ethical climate theory: A meta-analytic review. Journal of Business Ethics, 69, 175–194. doi:10.1007/s10551-006-9084-7

410

Perspectives on Organizational Climate and Culture

Mayer, D. M., Ehrhart, M. G., & Schneider, B. (2009). Service attribute boundary conditions of the service climate–customer satisfaction link. Academy of Management Journal, 52, 1034–1050. Mayer, D. M., Nishii, L. H., Schneider, B., & Goldstein, H. W. (2007). The precursors and products of justice climates: Group leader antecedents and employee attitudinal consequences. Personnel Psychology, 60, 929–963. doi:10.1111/j.1744-6570.2007.00096.x

Copyright American Psychological Association. Not for further distribution.

McClelland, D. C. (1961). The achieving society. New York: Van Nostrand. McGraw, K. O., & Wong, S. P. (1996). Forming inferences about some intraclass correlation coefficients. Psychological Methods, 1, 30–46. doi:10.1037/ 1082-989X.1.1.30

Ostroff, C. (1993). The effects of climate and personal influences on individual behavior and attitudes in organizations. Organizational Behavior and Human Decision Processes, 56, 56–90. doi:10.1006/obhd.1993.1045 Ostroff, C., Kinicki, A. J., & Tamkins, M. M. (2003). Organizational culture and climate. In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.), Handbook of psychology: Vol. 12. Industrial and organizational psychology (pp. 565–593). New York: Wiley. Ott, J. S. (1989). The organizational culture perspective. Pacific Grove, CA: Brooks-Cole. Parsons, T. (1951). The social system. Glencoe, IL: Free Press.

Meindl, J. R., Ehrlich, S. B., & Dukerich, J. M. (1985). The romance of leadership. Administrative Science Quarterly, 30, 78–102. doi:10.2307/2392813

Patterson, M. G., West, M. A., Shackleton, V. J., Dawson, J. F., Lawthom, R., Maitlis, S., et al. (2005). Validating the organizational climate measure: Links to managerial practices, productivity and innovation. Journal of Organizational Behavior, 26, 379–408. doi:10.1002/job.312

Michela, J. L., & Burke, W. W. (2000). Organizational culture and climate in transformations for quality and innovation. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 225–244). Thousand Oaks, CA: Sage.

Payne, R. L. (2000). A three dimensional framework for analyzing and assessing culture/climate and its relevance for cultural change. In C. L. Cooper, S. Cartwright, & P. C. Earley (Eds.), The international handbook of organizational culture and climate (pp. 107–122). Chichester, England: Wiley.

Miner, J. B. (2002). Organizational behavior: Foundations, theories and analyses. New York: Oxford University Press.

Payne, R. L., & Pugh, D. S. (1976). Organizational structure and climate. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1125–1173). Chicago: Rand McNally.

McGregor, D. M. (1960). The human side of enterprise. New York: McGraw-Hill.

Moliner, C., Martinez-Tur, V., Peiro, J. M., Ramos, J., & Cropanzano, R. (2005). Relationship between organizational justice and burnout at the work-unit level. International Journal of Stress Management, 12, 99–116. doi:10.1037/1072-5245.12.2.99 Nadler, D. A., Thies, R. K., & Nadler, M. A. (2000). Culture change in the strategic enterprise: Lessons from the field. In C. L. Cooper, S. Cartwright, & P. C. Earley (Eds.), The international handbook of organizational culture and climate (pp. 347–376). Chichester, England: Wiley. Neal, A., & Griffin, M. A. (2006). A study of the lagged relationships among safety climate, safety motivation, safety behavior, and accidents at the individual and group levels. Journal of Applied Psychology, 91, 946–953. doi:10.1037/0021-9010.91.4.946 Newman, D. A., & Sin, H. P. (2009). How do missing data bias estimates of within-group agreement?: Sensitivity of SDWG, CVWG, rWG(J), rWG(J)*, and ICC to systematic nonresponse. Organizational Research Methods, 12, 113–147. doi:10.1177/1094428106298969 O’Reilly, C. A., Chatman, J., & Caldwell, D. F. (1991). People and organizational culture: A profile comparison approach to assessing person-organization fit. Academy of Management Journal, 34, 487–516. doi:10.2307/256404

Peters, T., & Waterman, R. H. (1982). In search of excellence. New York: Harper & Row. Pettigrew, A. M. (1979). On studying organizational cultures. Administrative Science Quarterly, 24, 570–581. doi:10.2307/2392363 Pettigrew, A. M. (1987a). Introduction: Researching strategic change. In A. M. Pettigrew (Ed.), The management of strategic change (pp. 1–13). Oxford, England: Blackwell. Pettigrew, A. M. (Ed.). (1987b). The management of strategic change. Oxford, England: Blackwell. Pettigrew, A. M. (1990). Organizational climate and culture: Two constructs in search of a role. In B. Schneider (Ed.), Organizational climate and culture (pp. 413–434). San Francisco: Jossey-Bass. Pritchard, R. D., & Karasick, B. (1973). The effects of organizational climate on managerial job performance and job satisfaction. Organizational Behavior and Human Performance, 9, 126–146. doi:10.1016/ 0030-5073(73)90042-1 Pugh, D. S. (1966). Modern organization theory: A psychological and sociological study. Psychological Bulletin, 66, 235–251. doi:10.1037/h0023853 411

Schneider, Ehrhart, and Macey

Quinn, R. E. (1988). Beyond rational management: Mastering the paradoxes and competing demands of high performance. San Francisco: Jossey-Bass. Quinn, R. E., & McGrath, M. R. (1985). The transformation of organizational cultures: A competing values perspective. In P. J. Frost, L. F. Moore, M. R. Louis, C. C. Lundberg, & J. Martin (Eds.), Organizational culture (pp. 315–334). Beverly Hills, CA: Sage.

Copyright American Psychological Association. Not for further distribution.

Quinn, R. E., & Rohrbaugh, J. (1983). A special model of effectiveness criteria: Toward a competing values approach to organizational analysis. Management Science, 29, 363–377. doi:10.1287/mnsc.29.3.363 Rafaeli, A., & Worline, M. (2000). Symbols in organizational culture. In N. M. Ashkanasy, C. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 71–84). Thousand Oaks, C.A.: Sage. Reichers, A. E., & Schneider, B. (1990). Climate and culture: An evolution of constructs. In B. Schneider (Ed.), Organizational climate and culture (pp. 5–39). San Francisco: Jossey-Bass. Rentsch, J. R. (1990). Climate and culture: Interaction and qualitative differences in organizational meanings. Journal of Applied Psychology, 75, 668–681. doi:10.1037/0021-9010.75.6.668 Roberson, Q. M. (2006a). Justice in teams: The activation and role of sensemaking in the emergence of justice climates. Organizational Behavior and Human Decision Processes, 100, 177–192. doi:10.1016/ j.obhdp.2006.02.006 Roberson, Q. M. (2006b). Justice in teams: The effects of interdependence and identification on referent choice and justice climate strength. Social Justice Research, 19, 323–344. doi:10.1007/s11211-006-0010-z Roethlisberger, F. J., & Dickson, W. J. (1946). Management and the worker. Cambridge, MA: Harvard University Press. Rousseau, D. M. (1990). Normative beliefs in fund-raising organizations: Linking culture to organizational performance and individual responses. Group and Organizational Studies, 15, 448–460. doi:10.1177/ 105960119001500408 Salanova, M., Agut, S., & Peiró, J. M. (2005). Linking organizational resources and work engagement to employee performance and customer loyalty: The mediation of service climate. Journal of Applied Psychology, 90, 1217–1227. doi:10.1037/0021-9010. 90.6.1217 Schein, E. H. (1965). Organizational psychology. Englewood Cliffs, NJ: Prentice-Hall. Schein, E. H. (1985). Organizational culture and leadership. San Francisco: Jossey-Bass. Schein, E. H. (1992). Organizational culture and leadership (2nd ed.). San Francisco: Jossey-Bass. 412

Schein, E. H. (2000). Sense and nonsense about culture and climate. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. xxiii–xxx). Thousand Oaks, CA: Sage. Schein, E. H. (2004). Organizational culture and leadership (3rd ed.). San Francisco: Jossey-Bass. Schneider, B. (1975). Organizational climates: An essay. Personnel Psychology, 28, 447–479. doi:10.1111/ j.1744-6570.1975.tb01386.x Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437–453. doi:10.1111/j.1744-6570. 1987.tb00609.x Schneider, B. (Ed.). (1990). Organizational climate and culture. San Francisco: Jossey-Bass. Schneider, B. (2000). The psychological life of organizations. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. xvii–xxii). Thousand Oaks, CA: Sage. Schneider, B., & Bartlett, C. J. (1968). Individual differences and organizational climate: I. The research plan and questionnaire development. Personnel Psychology, 21, 323–333. doi:10.1111/j.1744-6570. 1968.tb02033.x Schneider, B., & Bartlett, C. J. (1970). Individual differences and organizational climate: II. Measurement of organizational climate by the multitrait–multirater matrix. Personnel Psychology, 23, 493–512. doi:10.1111/j.1744-6570.1970.tb01368.x Schneider, B., Bowen, D. E., Ehrhart, M. G., & Holcombe, K. M. (2000). The climate for service: Evolution of a construct. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 21–36). Thousand Oaks, CA: Sage. Schneider, B., Brief, A. P., & Guzzo, R. A. (1996). Creating a climate and culture for sustainable organization change. Organizational Dynamics, 24, 7–19. doi:10.1016/S0090-2616(96)90010-8 Schneider, B., Ehrhart, M. G., Mayer, D. M., Saltz, J. L., & Niles-Jolly, K. (2005). Understanding organization–customer links in service settings. Academy of Management Journal, 48, 1017–1032. Schneider, B., Macey, W. H., Lee, W., & Young, S. A. (2009). Organizational service climate drivers of the American Customer Satisfaction Index (ACSI) and financial and market performance. Journal of Service Research, 12, 3–14. Schneider, B., Parkington, J. P., & Buxton, V. M. (1980). Employee and customer perceptions of service in banks. Administrative Science Quarterly, 25, 252–267. doi:10.2307/2392454

Perspectives on Organizational Climate and Culture

Schneider, B., & Reichers, A. E. (1983). On the etiology of climates. Personnel Psychology, 36, 19–39. doi:10.1111/j.1744-6570.1983.tb00500.x Schneider, B., Salvaggio, A. N., & Subirats, M. (2002). Climate strength: A new direction for climate research. Journal of Applied Psychology, 87, 220–229. doi:10.1037/0021-9010.87.2.220

Copyright American Psychological Association. Not for further distribution.

Schneider, B., Smith, D. B., Taylor, S., & Fleenor, J. (1998). Personality and organizations: A test of the homogeneity of personality hypothesis. Journal of Applied Psychology, 83, 462–470. doi:10.1037/0021-9010. 83.3.462 Schneider, B., & Snyder, R. A. (1975). Some relationships between job satisfaction and organizational climate. Journal of Applied Psychology, 60, 318–328. doi:10.1037/h0076756 Schneider, B., Wheeler, J. K., & Cox, J. F. (1992). A passion for service: Using content analysis to explicate service climate themes. Journal of Applied Psychology, 77, 705–716. doi:10.1037/0021-9010.77.5.705 Schneider, B., & White, S. S. (2004). Service quality: Research perspectives. Thousand Oaks, CA: Sage. Schneider, B., White, S. S., & Paul, M. C. (1998). Linking service climate and customer perceptions of service quality: Test of a causal model. Journal of Applied Psychology, 83, 150–163. doi:10.1037/0021-9010. 83.2.150 Seashore, S. E. (1964). Field experiments with formal organizations. Human Organization, 23, 164–170. Senge, P. (1990). The fifth discipline: The art and practice of learning organization. New York: Doubleday. Siehl, C., & Martin, J. (1990). Organizational culture: A key to financial performance? In B. Schneider (Ed.), Organizational climate and culture (pp. 241–281). San Francisco: Jossey-Bass. Smerek, R. E., & Denison, D. R. (2007). Social capital in organizations: Understanding the link to firm performance. In G. T. Solomon (Ed.), Proceedings of the sixty-sixth annual meeting of the academy of management [CD]. ISSN 1543-8643. Smircich, L. (1983). Concepts of culture and organizational analysis. Administrative Science Quarterly, 28, 339–358. doi:10.2307/2392246 Smith, P. C., Kendall, L. M., & Hulin, C. L. (1969). The measurement of satisfaction in work and retirement. Chicago: Rand-McNally. Sørensen, J. B. (2002). The strength of corporate culture and the reliability of firm performance. Administrative Science Quarterly, 47, 70–91. doi:10.2307/3094891 Sowinski, D. R., Fortmann, K. A., & Lezotte, D. V. (2008). Climate for service and the moderating effects of climate strength on customer satisfaction, voluntary turnover, and profitability. European Journal of

Work and Organizational Psychology, 17, 73–88. doi:10.1080/13594320701473065 Sparrow, P. R. (2000). Developing diagnostics for high performance organizational cultures. In C. L. Cooper, S. Cartwright, & P. C. Earley (Eds.), The international handbook of organizational culture and climate (pp. 85–106). Chichester, England: Wiley. Spell, C. S., & Arnold, T. J. (2007). A multi-level analysis of organizational justice climate, structure, and employee mental health. Journal of Management, 33, 724–751. doi:10.1177/0149206307305560 Srivistava, S., & Cooperrider, D. L. (Eds.). (1990). Appreciative management and leadership: The power of positive thought and action in organizations. San Francisco: Jossey-Bass. Tannenbaum, A. S. (1962). Control in organizations. Administrative Science Quarterly, 7, 236–257. doi:10.2307/2390857 Taras, V. (2006). Culture survey catalogue. Retrieved from http://ucalgary.ca/∼taras/_private/Culture_ Survey_Catalogue.pdf Tiffin, J. (1946). Industrial psychology. Englewood Cliffs, NJ: Prentice-Hall. Trice, H. M., & Beyer, J. M. (1993). The cultures of work organizations. Englewood Cliffs, NJ: Prentice-Hall. van Dyck, C., Frese, M., Baer, M., & Sonnentag, S. (2005). Organizational error management culture and its impact on performance: A two-study replication. Journal of Applied Psychology, 90, 1228–1240. Van Maanen, J., & Schein, E. H. (1979). Toward a theory of organizational socialization. In B. M. Staw (Ed.), Research in organizational behavior (Vol. 1, pp. 209–264). Greenwich, CT: JAI. Van Muijen, J. J. (1998). Organizational culture. In P. J. D. Drenth, H. Thiery, & C. J. de Wolff (Eds.), Handbook of work and organizational psychology: Vol. 4. Organizational psychology (2nd ed., pp. 113–128). London: Psychology Press. Viteles, M. S. (1953). Motivation and morale in industry. New York: Norton. Wallace, J. C., & Chen, G. (2006). A multilevel integration of personality, climate, self-regulation, and performance. Personnel Psychology, 59, 529–557. doi:10.1111/j.1744-6570.2006.00046.x Wallace, J. C., Popp, E., & Mondore, S. (2006). Safety climate as a mediator between foundation climates and occupational accidents: A group-level investigation. Journal of Applied Psychology, 91, 681–688. doi:10.1037/0021-9010.91.3.681 Weber, Y. (2000). Measuring cultural fit in mergers and acquisitions. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 309–320). Thousand Oaks, CA: Sage. 413

Schneider, Ehrhart, and Macey

Weick, K. E. (1985). The significance of corporate culture. In P. J. Frost, L. F. Moore, M. R. Louis, C. C. Lundberg, & J. Martin (Eds.), Organizational culture (pp. 381–390). Beverly Hills, CA: Sage.

Copyright American Psychological Association. Not for further distribution.

Wilderom, C. P. M., Glunk, U., & Maslowski, R. (2000). Organizational culture as a predictor of organizational performance. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 193–209). Thousand Oaks, CA: Sage. Wiley, J. W. (1996). Linking survey results to customer satisfaction and business performance. In A. I. Kraut (Ed.), Organizational surveys: Tools for assessment and change (pp. 350–369). San Francisco: JosseyBass. Yang, J., Mossholder, K. W., & Peng, T. K. (2007). Procedural justice climate and group power distance: An examination of cross-level interaction effects. Journal of Applied Psychology, 92, 681–692. doi:10.1037/0021-9010.92.3.681 Zammuto, R. F., Gifford, B., & Goodman, E. A. (2000). Managerial ideologies, organization culture, and the

414

outcomes of innovation. In N. M. Ashkanasy, C. P. M. Wilderom, & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 261–278). Thousand Oaks, CA: Sage. Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and applied implications. Journal of Applied Psychology, 65, 96–102. doi:10.1037/00219010.65.1.96 Zohar, D. (2000). A group level model of safety climate: Testing the effect of group climate on microaccidents in manufacturing jobs. Journal of Applied Psychology, 85, 587–596. doi:10.1037/0021-9010.85.4.587 Zohar, D., & Luria, G. (2005). A multi-level model of safety climate: Cross-level relationships between organization and group-level climates. Journal of Applied Psychology, 90, 616–628. doi:10.1037/00219010.90.4.616 Zohar, D., & Tenne-Gazit, O. (2008). Transformational leadership and group interaction as climate antecedents: A social network analysis. Journal of Applied Psychology, 93, 744–757. doi:10.1037/00219010.93.4.744

CHAPTER 13

WORK MATTERS: JOB DESIGN IN CLASSIC AND CONTEMPORARY PERSPECTIVES

Copyright American Psychological Association. Not for further distribution.

Adam M. Grant, Yitzhak Fried, and Tina Juillerat

“We are nothing more than glorified clerks.” The bank tellers at a multibillion dollar corporation in the Southwest United States were dissatisfied with their jobs. They described their tasks as boring, and felt micromanaged by supervisors when making minor decisions (Griffin, 1991). In an effort to improve the situation, managers decided to intervene by redesigning the bank tellers’ jobs. To reduce boredom, managers added new tasks to the jobs, providing tellers with greater variety and opportunities to use a broader range of skills. Along with their original tasks of cashing checks and accepting deposits and loan payments, tellers were now trained to handle commercial and travelers’ checks and post transactions in an online computer terminal. To reduce micromanagement, managers provided tellers with more autonomy. Managers also delegated decision-making responsibilities: Instead of requiring tellers to obtain supervisors’ signatures to credit deposits and approve withdrawals over $100, they gave tellers the authority to post checks immediately and approve their own withdrawals when the customer’s account had sufficient funds. Managers also provided feedback on transactions and errors, giving tellers increased ability to monitor their own work processes. Finally, managers modified transaction receipts to include the name and contact information for the teller who handled the transaction. This allowed customers to contact tellers directly to ask questions or report errors, enabling tellers to take responsibility for their own customers.

These efforts to redesign and enrich the tellers’ jobs produced lasting effects on their attitudes and behaviors (Griffin, 1991). Six months later, the tellers were more satisfied with their jobs and more committed to the company, whereas tellers at a comparison bank whose jobs were not enriched achieved no increases in satisfaction or commitment. The effects on performance were more remarkable. Griffin asked supervisors to evaluate tellers’ performance in terms of both quality and quantity. After a period of adaptation, tellers whose jobs were enriched were rated by supervisors as displaying significantly better performance, with the effects lasting at least 4 years. This study demonstrated that enriching jobs to provide variety, feedback, and autonomy can improve attitudes and performance. (See Vol. 3, chap. 3, this handbook.) “This may not be a great place to study motivation, well, because there isn’t any. Then again, we could use some help.” The managers of a call center in the Midwest United States were facing annual staff turnover exceeding 400%: Over the course of each 3-month cycle, the entire staff quit. The hiring and training costs resulted in performance challenges: The call center employed fundraisers to solicit alumni donations to a large public university, and the total funds raised were falling below expectations. A team of organizational psychologists entered the call center hoping to use principles of job redesign to increase caller motivation and performance (Grant, Campbell, Chen, Cottone, Lapedis, & Lee, 2007). Their initial

We thank Sheldon Zedeck and Sharon Parker for generative feedback on our outline and draft.

http://dx.doi.org/10.1037/12169-013 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

417

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

diagnosis was that the callers might benefit from a job enrichment process similar to what the bank tellers encountered. Noticing that the callers were required to make repetitive calls using standardized scripts, the researchers proposed to redesign callers’ jobs to provide greater variety. With a forlorn grin, the manager replied, “Variety is not an option. This job only involves one task: calling as many alumni as possible to convince them to give their hard-earned money to their alma mater.” The researchers returned to the drawing board and continued interviewing, surveying, and observing the callers. They soon discovered that callers reported being in the dark about how the alumni donations were used. The majority of the funds raised were funneled directly into scholarships for students to attend the university. The researchers proposed to enrich the callers’ jobs by placing them in contact with scholarship recipients who had benefited from the funds raised, which was expected to increase task significance by providing a vivid illustration of the impact of callers’ jobs on others (Grant, 2007). The researchers recruited scholarship recipients to help the callers understand how their efforts made a difference in scholarship students’ lives. The researchers then designed a series of field experiments and quasiexperiments in which they connected the callers to scholarship recipients through face-to-face meetings or written letters. They measured the callers’ weekly persistence (calls made and minutes on the phone) and performance (pledges obtained and donation money raised) before and after the interventions. In the first experiment, the researchers were surprised to discover that a full month after the interventions, callers who had contact with scholarship recipients had increased dramatically in their persistence and performance. Relative to baseline levels prior to the intervention, the average caller was spending more than twice as many minutes on the phone—and raising more than twice as much money—per week (Grant et al., 2007). Subsequent experiments replicated these effects with different samples of callers, different scholarship recipients, different measures of persistence and performance, and both managersupervised and researcher-supervised interventions (Grant, 2008a). In one version of the intervention,

418

hearing a story from one scholarship recipient led to fivefold increases in the amount of donation money that callers raised per week (Grant, 2008b). Similar effects emerged with newcomers to the job: When callers were connected to their impact on scholarship recipients during training, they secured nearly twice as many pledges as a control group in their very first week on the job (Grant, 2008a). Across all of these experiments, callers in pure control and alternative treatment groups did not change significantly on any of the performance measures. These findings highlight the motivating power of enriching jobs to connect employees to the people who benefit from their work. JOB DESIGN Researchers originally defined job design as the set of opportunities and constraints structured into assigned tasks and responsibilities that affect how an employee accomplishes and experiences work (Hackman & Oldham, 1980). Today, job design is defined more broadly as encapsulating the processes and outcomes of how work is structured, organized, experienced, and enacted (Morgeson & Humphrey, 2008; Parker & Wall, 1998). (See also Vol. 2, chap. 1, this handbook.) This broader definition opens the door to include dynamic, emergent roles and changes in work from project to project, as opposed to merely emphasizing static job descriptions composed of fixed tasks assigned from above (Ilgen & Hollenbeck, 1991; Parker, Wall, & Cordery, 2001; Wrzesniewski & Dutton, 2001). We will return to these definitional issues throughout the chapter. Job design has played a central role in the history of research in applied psychology and organizational behavior, and it continues to be a key topic for several reasons. First, in past and recent decades, job design has been one of only a handful of organizational theories rated as simultaneously high in validity, importance, and usefulness (Miner, 1984, 2003). As illustrated by our opening stories, job design theory and research has enabled applied psychologists, organizational scholars, and practitioners to describe, diagnose, and resolve important practical problems in organizations. Second, because it is a fundamental component of the execution and experience of work,

Copyright American Psychological Association. Not for further distribution.

Work Matters

job design is as old as work itself. Job design exerts a foundational influence on the actions and experiences of employees in every type of work, occupation, and organization. Third, job design is an actionable feature of organizational contexts. Managers typically have more influence and control over job design than they do over culture, structure, relationships, technology, and people themselves (Hackman & Oldham, 1980). As such, most managers are responsible for making decisions about how to design jobs for employees (Mintzberg, 1973). Job design therefore commands what some have described as an organization’s most valuable and scarce resource: the time and attention of managers (Dutton & Ashford, 1993). Depending on how managers make decisions about job design, it can be a liability or a potential source of competitive advantage for organizations (Pfeffer, 1994). Unfortunately, however, many managers often use simplified work as the default approach to designing jobs (Campion & Stevens, 1991). Fourth, job design is receiving a resurgence of attention as dramatic changes in domestic and international landscapes of work have created new types of jobs, particularly in service and knowledge/creative sectors (Elsbach & Hargadon, 2006; Grant & Parker, 2009; Parker et al., 2001; Rousseau & Fried, 2001). These changes have spawned rapid increases in autonomy, professionalization, and service customization, providing employees with growing amounts of latitude and discretion to alter their own job designs. As organizations flatten, employees have opportunities to craft their jobs (Wrzesniewski & Dutton, 2001), expand their roles (Parker, Wall, & Jackson, 1997), revise their tasks (Staw & Boettger, 1990), and negotiate new roles and idiosyncratic deals (Ilgen & Hollenbeck, 1991; Rousseau, Ho, & Greenberg, 2006). Moreover, technological advances have increasingly made information available that is conducive to autonomy and empowerment (Sinha & Van de Ven, 2005). Integrating these final two points suggests that job design is especially important in theory and practice because—unlike more intractable factors such as culture and structure— both managers and employees have the opportunity to change job designs on a regular basis.

WHERE HAVE WE BEEN? A SELECTIVE HISTORY OF JOB DESIGN RESEARCH FROM PAST TO PRESENT Having highlighted the importance of job design in scholarship and practice, we now provide a selective overview of the major theoretical perspectives and empirical findings in the job design literature. Our review includes economic perspectives on the division of labor, the human relations movement and the emergence of job enrichment, the job characteristics model, the social information processing perspective, sociotechnical systems theory, interdisciplinary frameworks, and models of job demands. For further details, we refer the reader to several excellent reviews of the job design literature (e.g., Fried, Levi, & Laurence, 2008; Griffin, 1987; Morgeson & Campion, 2003; Morgeson & Humphrey, 2008; Oldham, 1996; Parker & Ohly, 2008; Parker & Wall, 1998; Wall & Martin, 1987).

Economic Theories of Division of Labor Job design theory and research has its roots in economic perspectives on the division of labor (Babbage, 1835; Smith, 1776). Economists such as Smith and Babbage proposed that productivity could be increased if jobs were broken down into simple tasks. The premise behind this thinking was that division of labor and simplification would allow employees to develop specialized skills and efficient techniques for completing tasks, as well as to eliminate distractions and reduce time wasted while switching tasks (Morgeson & Campion, 2003). In the beginning of the 20th century, proponents of “scientific management” sought to test and apply this logic. For example, Taylor (1911) conducted time and motion studies in an effort to systematize efficient division of labor by managers.

Human Relations Movement Although researchers continue to debate about whether Taylor’s motives were benevolent, malevolent, or indifferent toward employees (WagnerTsukamoto, 2007), scientific management sparked a reactionary movement. Researchers began to observe that attempts to achieve efficiency were pursued at

419

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

the expense of employee satisfaction and motivation. To address these problems, the human relations movement was born. Advocates of this movement were deeply concerned about the well-being, satisfaction, and motivation of employees. They launched the classic Hawthorne studies to improve environmental conditions, such as lighting, in ways that they expected to be conducive to both comfort and productivity (Mayo, 1933, 1945; Roethlisberger & Dickson, 1939). They learned that taking an interest in employees’ opinions, rather than the lighting conditions themselves, appeared to drive productivity increases. They then investigated the effects of other changes to employees’ job designs and schedules, such as varying break intervals, working hours, and vacations. As productivity continued to increase, the researchers came to recognize the importance of employees’ attitudes in shaping their behavior (Hsueh, 2002). They began to interview employees to understand their feelings about their jobs, supervision, and working conditions. This focus on jobs—and the supervision and working conditions that affect how employees carry out their jobs—paved the way for a full-blown research agenda on the design of jobs to satisfy and fulfill employees’ basic motives and psychological needs. While Likert (1961, 1967) emphasized the importance of participative management, McGregor (1960) distinguished between two theories that leaders and managers can hold. “Theory X” leaders believe that employees are inherently lazy: They dislike work and responsibility and will avoid it if possible, preferring to follow rather than lead. When designing jobs, Theory X leaders micromanage employees, restricting their autonomy and freedom. “Theory Y” leaders, on the other hand, believe that work can be as naturally enjoyable as play or rest, and that doing a good job can be a source of satisfaction in and of itself. Theory Y leaders therefore believe that if employees are given freedom, they will be self-motivated and ambitious, seek responsibility, exercise self-control and selfdirection, and pursue goals that benefit themselves and the organization. When designing jobs, Theory Y leaders advocate empowerment and participative 1

management, giving employees considerable autonomy and freedom in their work. Both Likert and McGregor emphasized the potential value of reducing managerial control in designing jobs to provide employees with freedom to fulfill their psychological needs. Their perspectives dovetailed with the work of Herzberg and colleagues, who introduced the notion of job enrichment to applied psychology and organizational behavior. These authors proposed motivator–hygiene theory, which argues that job satisfaction and dissatisfaction are distinct states caused by different forces (Herzberg, 1966; Herzberg, Mausner, & Snyderman, 1967). According to this theory, satisfaction is caused by “motivators” intrinsic to the nature and content of a job: opportunities to achieve, receive recognition, perform interesting work, be responsible, grow, and advance. Dissatisfaction, on the other hand, arises not from the job itself but rather from “hygiene” factors related to the context of the job: policy and administration, supervision, interpersonal relations, working conditions, salary, status, and security.

Job Design and Enrichment Subsequent research challenged the validity of distinguishing between motivators and hygiene factors and between satisfaction and dissatisfaction, revealing that the two-factor theory is method bound and has little empirical support for predicting satisfaction (Ambrose & Kulik, 1999; Locke & Henne, 1986). However, the thrust of Herzberg’s contribution is conveyed by a reflection from Terkel (1972): “Most of us have jobs that are too small for our spirit. Jobs are not big enough for people” (p. 29). Herzberg’s work proved influential in drawing researchers’ attention to the potential for jobs to be redesigned, enlarged, and enriched to increase motivation and satisfaction.1 Building on this notion, Turner and Lawrence (1965) sought to develop a more systematic classification of the task attributes that influence employees’ attitudes and behaviors. Informed by the works of Herzberg, as well as others focusing on job enlargement, task attributes, and the interaction of technology, people, and work (e.g., Trist & Bamforth,

Job enlargement refers to adding requirements at the same level to expand the scope of the job, while job enrichment refers to adding higher-level responsibilities to increase intrinsic motivation (e.g., Campion & McClelland, 1993; Hackman & Oldham, 1980; Herzberg, 1966).

420

Copyright American Psychological Association. Not for further distribution.

Work Matters

1951; Walker & Guest, 1952), Turner and Lawrence (1965) argued that “workers’ response to task attributes could and should become a more important factor in job design” (p. 2). They raised the possibility that jobs could be described from behavioral (what behaviors need to be enacted for the work to be completed), technical (physical and mechanical operations to be performed), organizational (function of the job in combining with other jobs to achieve the organization’s goals), social (the social desirability and status of the work), and personal (expected career progression) perspectives. Focusing primarily on the behavioral perspective, Turner and Lawrence introduced six multidimensional task attributes that could be required to varying degrees by the intrinsic nature of the work itself: variety, autonomy, required interaction, optional interaction on and off the job, required knowledge and skill, and responsibility. They also examined several additional “associated task attributes” that are part of the job but not essential to its performance: task identity, pay, working conditions, cycle time, level of mechanization, and capital investment.2 With a sample of 470 employees in 47 different jobs, Turner and Lawrence measured these task attributes and provided an initial examination of their relationships with job satisfaction and attendance. They found that the requisite task attributes predicted higher satisfaction and attendance only among employees from factories in small towns, but not in urban settings, suggesting that cultural backgrounds may shape employees’ task preferences.

Job Characteristics Model Setting the stage for contemporary perspectives on job design, Hackman and Lawler (1971) sought to investigate the influence of job characteristics on attitudes and behaviors. They developed a conceptual framework with roots in Turner and Lawrence’s (1965) work, as well as in classic formulations of expectancy theory (Porter & Lawler, 1968; Vroom, 1964). The framework specified that four core job dimensions of variety, autonomy, task identity, and feedback would be associated with higher motivation, 2

job satisfaction, and performance, as well as lower absenteeism for employees with strong “higher order needs” for accomplishment and personal growth. Using data from telephone company employees, Hackman and Lawler found general support for these hypotheses. This paper laid the groundwork for the development of a framework that has fueled three decades of research and remains the dominant model of job design today: the job characteristics model (JCM; Hackman & Oldham, 1975, 1976, 1980; for a reflection on how the model developed, see Oldham & Hackman, 2005). The JCM focuses on five core job characteristics: task significance, task identity, skill variety, autonomy, and job feedback. Task significance is the extent to which the job provides opportunities to have a positive impact on the well-being of other people; task identity is the extent to which the job allows individuals to complete a whole, identifiable, visible piece of work from start to finish; skill variety is the extent to which the job involves a wide range of capabilities and talents; autonomy is the extent to which the job provides freedom and discretion in how and when to do the work; and feedback is the extent to which the job itself provides clear, direct information about performance effectiveness. Hackman and Oldham (1975, 1976) argued that these five core job characteristics are objective properties of the structure of employees’ assigned tasks that influence their job perceptions. They proposed that the five core job characteristics lead to three critical psychological states: experienced meaningfulness, responsibility, and knowledge of results. More specifically, they predicted that task significance, task identity, and skill variety would contribute in an additive or compensatory fashion to experienced meaningfulness: When these characteristics were present, employees would perceive their work as more worthwhile and valuable. They further predicted that autonomy would lead employees to experience greater personal responsibility or ownership over their work, and that job feedback would lead employees to experience greater knowledge of results, or awareness of effectiveness. Hackman and Oldham (1976) proposed that the

Rarely mentioned is that Turner and Lawrence (1965) developed four additional task attributes that they eliminated from their classification due to measurement difficulties: requisite interdependence, strategic position (the extent to which a job was strategic to the overall work process), direction of interaction (initiated vs. received), and variety of jobs in the working area.

421

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

core job characteristics could be combined, when grouped according to the critical psychological states, to create a score for the motivating potential of a job. The motivating potential of a job was defined as the product of (a) autonomy, (b) job feedback, and (c) the average of task significance, task identity, and skill variety (the meaningfulness-related dimensions), such that the motivating potential of a job = autonomy × job feedback × 1/3(task significance + task identity + skill variety). Hackman and Oldham (1976) argued that the critical psychological states would mediate the positive association between the core job characteristics and the outcomes of internal work motivation, performance quality, job and growth satisfaction, and low absenteeism and turnover. They further proposed, in line with Hackman and Lawler’s (1971) arguments, that these relationships would be moderated by employees’ growth need strength at two stages in the model.3 First, the stronger the employees’ growth needs, the more likely the core job characteristics would be to cultivate the critical psychological states. Second, the stronger the employees’ growth needs, the more likely the critical psychological states would be to shape the motivation, attitude, and behavior and performance outcomes. These moderating hypotheses were again based on the logic of expectancy theory (Vroom, 1964; Porter & Lawler, 1968). Employees with strong growth needs would be more dependent on enriched job characteristics to experience meaningfulness, responsibility, and knowledge of results, as well as more dependent on the critical psychological states to experience enhanced motivation and more positive attitudes and display higher performance quality and fewer withdrawal behaviors. Researchers have conducted several hundred studies to test the JCM and its central propositions. The majority of studies have relied on cross-sectional designs, using self-reports, observer-reports, or occupational title classifications to evaluate job characteristics and self-reports, observer reports, or objective behavioral measures to assess motivation, satisfac3

tion, performance, and withdrawal behaviors. Metaanalyses provide general support for the hypotheses that the core job characteristics are associated with favorable attitudinal and behavioral reactions, as mediated by the critical psychological states (Fried, 1991; Fried & Ferris, 1987; Humphrey, Nahrgang, & Morgeson, 2007; Johns et al., 1992). Generally speaking, these meta-analyses have revealed stronger relationships of job characteristics with psychological–attitudinal outcomes than with behavior and performance outcomes. For example, Humphrey et al. (2007) reported mean correlations (p, corrected for unreliability) for the five core job characteristics (autonomy, skill variety, task identity, task significance, and job feedback) of .41, .55, and .39 with job satisfaction, growth satisfaction, and internal work motivation, respectively. They found a weaker relationship between the job characteristics and absenteeism, with corrected correlations of −.15 for autonomy, −.09 for task identity, and −.10 for job feedback. The only one of the five motivational characteristics that was significantly correlated with objective performance was autonomy (p = .17). On the other hand, research testing the moderating role of growth need strength has returned mixed results. While some studies have found support, others have not (Johns et al., 1992; Tiegs, Tetrick, & Fried, 1992). It is not yet clear whether these conflicting findings are an artifact of range restriction and other measurement limitations or whether they are due to the theoretical possibility that growth need strength may be more relevant to some outcomes than others (Fried & Ferris, 1987; Johns et al., 1992; Loher, Noe, Moeller, & Fitzgerald, 1985; Spector, 1985). Researchers have also extended the JCM by examining the distinction between enriched tasks and enriched jobs. Wong and Campion (1991) argued that although researchers have defined a job as a group of tasks designed for one employee to complete (Griffin, 1987), the JCM is ambiguous about whether the five core job characteristics are motivating at the level of individual tasks or at the aggregate level of the job itself. On one hand, several of the job charac-

Researchers expanded the model to include two additional classes of moderators: individual knowledge and skill and context satisfaction (Hackman & Oldham, 1980; Oldham, Hackman, & Pearce, 1976). They proposed that the core job characteristics would be more likely to cultivate critical psychological states and favorable psychological and behavioral reactions when individuals were capable of performing their jobs and when they were satisfied with their supervisors, coworkers, compensation, and job security. These two categories of moderators have received little theoretical and empirical attention (Johns, Xie, & Fang, 1992).

422

Copyright American Psychological Association. Not for further distribution.

Work Matters

teristics are labeled as features of tasks (task identity, task significance). On the other hand, the characteristics are defined and measured as features of jobs. To resolve these issues, Wong and Campion (1991) developed a mediational model proposing that tasklevel characteristics influence job-level characteristics, which in turn influence attitudinal reactions. Their data provided only partial support for the role of job characteristics in mediating the association between task characteristics and attitudinal outcomes. Subsequent research by Taber and Alliger (1995) shed light on these mixed results by revealing that employees use different psychological processes to evaluate their tasks versus their more global jobs, and that because tasks and jobs are defined externally, they may not fully capture employees’ own task and job perceptions (see also Dierdorff & Morgeson, 2007; Ilgen & Hollenbeck, 1991; Morrison, 1994). These findings suggest that although focusing on the job level may be the most parsimonious way to understand employees’ work experiences and behaviors, we can deepen our knowledge by incorporating more molecular, personalized units of work such as tasks, roles, duties, activities, and projects. In spite of—or perhaps more accurately in response to—its popularity, the JCM has attracted criticism from a number of theoretical and empirical perspectives (e.g., Roberts & Glick, 1981). Researchers have debated about whether jobs have objective characteristics (Griffin, 1987; Morgeson & Campion, 2003; Oldham, 1996), as well as whether the five core job characteristics are distinct properties of jobs, can be subsumed by a smaller set of characteristics, or can even be reduced to a single characteristic of job complexity (e.g., Taber & Taylor, 1990), although more recent work has revealed that the characteristics are distinct (Edwards, Scully, & Brtek, 2000). Researchers have found that eliminating negatively worded items in scales can improve the factor structure, but not necessarily the predictive validity, of the measures of job characteristics in Hackman and Oldham’s (1975) Job Diagnostic Survey (Cordery & Sevastos, 1993; Kulik, Oldham, & Langner, 1988). However, these methodological critiques have tended to focus more heavily on the instruments used to test the JCM than on the core premises of the JCM itself. More recently, researchers have begun to devote

greater attention to extending the JCM conceptually to include a broader range of job characteristics, outcomes, mediators, moderators, and antecedents (e.g., Grant & Parker, 2009; Morgeson & Humphrey, 2006; Parker et al., 2001).

Social Information Processing Perspective The foundational assumptions of the JCM were challenged by Salancik and Pfeffer (1978), who offered the social information processing perspective as an alternative. Salancik and Pfeffer argued that employees’ job perceptions and attitudes derive not from objective structural properties of the work itself, but rather from how the work is socially constructed by cues from coworkers, supervisors, customers, family members, and other sources, as well as by their own past behaviors and experiences (for reviews, see Blau & Katerberg, 1982; Griffin, 1987; Wall & Martin, 1987; Zalesny & Ford, 1990). Salancik and Pfeffer proposed that social cues can affect employees’ job perceptions and attitudes through four different pathways. First, through a direct pathway, social cues can serve as a form of social influence, such that overt statements from other people about a job affect employees’ perceptions and attitudes. Second, through an attentional pathway, social cues can make particular aspects of a job salient, shaping the dimensions on which employees assess their perceptions and attitudes. Third, through an interpretation pathway, social cues can provide frames for assessing ambiguous job properties, shaping the interpretations that employees make of their jobs. Fourth, through a learning pathway, social cues can provide information about what needs or values are important, shaping employees’ judgments about what they want in a job. Research has provided mixed support for the social information processing perspective (Zalesny & Ford, 1990). Some field studies have shown that social comparisons are related to employees’ attitudinal and behavioral reactions to job design, with employees who perceive inequity displaying less favorable responses (Oldham, Kulik, Ambrose, Stepina, & Brand, 1986; see also Oldham, Kulik, Stepina, & Ambrose, 1986). The majority of investigations of the effects of social cues on perceptions and performance have taken the form of short-term laboratory experiments, which have generally shown 423

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

that positive social cues about a job result in more favorable task perceptions (Griffin, Bateman, Wayne, & Head, 1987; Thomas & Griffin, 1983; Zalesny & Ford, 1990). However, laboratory experiments have displayed inconclusive results for the effects of social cues on performance. Some experiments have shown that positive social cues can increase performance and productivity (e.g., White & Mitchell, 1979; White, Mitchell, & Bell, 1977), whereas others have returned null effects on behavior (e.g., Kilduff & Regan, 1988; Shaw & Weekley, 1981). Moreover, field experiments have called into question whether social cues can have lasting effects on the job perceptions and performance of employees in work organizations. For example, Jex and Spector (1989) conducted two field experiments directly applying social cues manipulations used in the laboratory, and found no changes in job perceptions and attitudes. Griffin (1983) conducted a field experiment in which he trained supervisors to provide positive social cues to manufacturing employees about specific task characteristics; the results indicated that social information affected task perceptions, but not productivity. Griffin’s (1983, 1987) theoretical and empirical integrations of job design and social information processing perspectives suggest that social cues can have effects on attitudes and behaviors, but these effects are generally weaker than those of job design itself. Thus, whereas social information processing theorists argued that scholars and practitioners should pay less attention to objective job characteristics than to social cues, research points to the opposite conclusion, accentuating the value of considering how jobs are objectively designed and structured. However, researchers continue to debate whether we should study objective job characteristics or individual perceptions of job characteristics (for reviews, see Morgeson & Campion, 2003; Oldham, 1996; Parker & Wall, 1998).

Sociotechnical Systems Theory Sociotechnical systems theory, developed primarily at the Tavistock Institute in the United Kingdom, is closely linked to job design theory and research (Rousseau, 1977). A core premise of sociotechnical systems theory is that individual and organizational effectiveness depend on the joint optimization of 424

human and mechanical–technological components of organizations (Trist, 1981; Trist & Bamforth, 1951). Sociotechnical systems theory proposes that creating autonomous workgroups can help to accomplish such optimization. Providing groups with the autonomy to manage their own work processes is believed to facilitate communication and problem solving, thereby enhancing productivity and well-being. Researchers have conducted numerous experiments to apply and test principles of sociotechnical systems theory (for reviews, see Pasmore, Francis, Haldeman, & Shani, 1982; Parker & Wall, 1998; see also Cummings, 1986). For example, in a longitudinal quasi-experiment, Wall, Kemp, Jackson, and Clegg (1986) found that the introduction of autonomous workgroups in a manufacturing company produced mixed effects. At the individual level, autonomous workgroups achieved lasting increases in intrinsic job satisfaction and fleeting increases in extrinsic job satisfaction but did not influence individual work motivation or performance. At the organizational level, autonomous workgroups enhanced productivity by eliminating unnecessary managerial positions but also increased voluntary labor turnover. In the past 2 decades, sociotechnical systems theory has seen few empirical tests and conceptual developments, in large part because the core propositions lack specificity (Parker & Wall, 1998; Parker et al., 2001). However, the theory continues to provide a meta-theoretical perspective that informs ongoing job design research, especially that which is related to autonomous workgroups.

Interdisciplinary Models of Job Design As of the 1980s, research on job design in industrial and organizational (I/O) psychology and organizational behavior was dominated by Hackman and Oldham’s motivational perspective. To broaden the job design literature and integrate it with principles from other disciplines, Campion and colleagues introduced an interdisciplinary perspective that theoretically integrates four different approaches to job design (Campion, 1988; Campion & Thayer, 1985; for a review, see Campion, Mumford, Morgeson, & Nahrgang, 2005). The motivational perspective emphasizes JCM principles such as vari-

Copyright American Psychological Association. Not for further distribution.

Work Matters

ety, autonomy, task identity and feedback. The motivational approach to job design offers benefits of motivation, satisfaction, and retention, but often involves enhanced training costs and stress. The mechanistic perspective, rooted in industrial engineering, emphasizes economic and scientific management principles such as specialization, simplification, and repetition. The mechanistic approach to job design offers benefits for efficiency and staffing and training, but tends to produce lower levels of motivation and satisfaction. The perceptual-motor perspective, rooted in human factors and cognitive psychology, emphasizes the principle of reducing information processing requirements (e.g., by operating and monitoring technology rather than performing tasks manually). The perceptual-motor approach to job design offers the benefits of reducing errors, accidents, and mental overload but tends to result in boredom and decreased motivation and satisfaction. Finally, the biological perspective, rooted in biology and medicine, emphasizes principles of physical comfort. The biological approach to job design offers benefits for health, stress, and fatigue, but it tends to involve considerable financial resources and low levels of physical activity. To examine the benefits and costs of these four general approaches to job design, Campion and McClelland (1991, 1993) conducted longitudinal quasi-experiments with clerical employees and managers in a financial services organization. Their initial results suggested that when jobs were enlarged by adding tasks and combining jobs, motivational principles improved, whereas mechanistic principles declined. The enlarged jobs were generally associated with higher satisfaction, lower boredom, greater probability of detecting errors, and improved customer service but required more training, higher skills, and higher compensation (Campion & McClelland, 1991). A follow-up study 2 years later suggested that the benefits and costs of job redesign changed over time as a function of how the redesign was conducted (Campion & McClelland, 1993). Enlarging jobs by adding tasks and combining jobs was increasingly costly over time: Employees were less satisfied and 4

efficient, experienced greater overload, made more errors, and provided poorer customer service. However, the organization had used a second approach to job redesign. For some jobs, instead of enlarging them by adding more low-level tasks, they enriched them by adding higher-level responsibilities for understanding procedures and rules for the organization’s products. When jobs were enriched in this fashion, the majority of the effects were beneficial over time: Employees were more satisfied, experienced less overload, made fewer errors, and provided better customer service. These findings supported the original arguments by Herzberg (1966) and Hackman and Oldham (1980) that organizations and their employees may achieve greater benefits from job enrichment than job enlargement.4 In subsequent research, Morgeson and Campion (2002) sought to address the trade-off between satisfaction and efficiency that frequently emerged between motivational and mechanistic approaches to job design. They proposed that jobs could be designed to be both satisfying and efficient by focusing on task clusters, “the smallest collection of logically related tasks that are normally performed by a single person such that they form a whole or natural work process” (Morgeson & Campion, 2002, p. 593). By increasing specialization, employees can work on clusters of tasks that allow for both skill utilization and efficiency. This idea was informed by the research of Edwards, Scully, and Brtek (1999, 2000), who showed that each of the four interdisciplinary approaches to job design is multidimensional. Their analyses revealed that the negative relationship between motivational and mechanistic job design was primarily due to the common trade-off between skill usage and simplicity: As one increases, the other tends to decrease. However, some forms of specialization enhance skill requirements without reducing complexity, making it possible to increase specialization in ways that are both mechanistically and motivationally sound. In a longitudinal quasi-experiment in a pharmaceutical company, Morgeson and Campion found support for the hypothesis that trade-offs between motivational

Campion and McClelland (1993) also found that job enlargement tends to lead to poorer biological designs, reducing physical comfort. Although the biological perspective has received less attention in job design research, it deserves further attention in light of its potential to improve physical health and protect against stress.

425

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

and mechanistic approaches can be avoided by enhancing specialization in task clusters. Employees whose jobs were redesigned in this fashion displayed increased satisfaction without increasing training costs or perceptions of simplicity. The interdisciplinary perspective has been generative in introducing new job characteristics and outcomes that had not previously been documented in psychological and organizational research on job design, especially from the standpoints of ergonomics, human factors, and industrial engineering. Researchers now recognize the importance of considering mechanistic, perceptual-motor, and biological perspectives, as well as traditional motivational perspectives, on job design. In addition, the interdisciplinary perspective has provided scholars and practitioners with new tools for diagnosing, planning, implementing, and evaluating job redesign interventions.

Job Demands-Control-Support and Job Demands–Resources Models Although it is not always included in reviews of the job design literature, another perspective on job design was developed by Karasek and colleagues (Karasek, 1979; Karasek & Theorell, 1990; for a review, see Vol. 3, chap. 13, this handbook). These authors were interested in understanding and reducing the deleterious effects of job demands on stress, strain, burnout, and physical illnesses such as heart disease. They proposed that providing greater job control to employees could buffer against these detrimental effects of job demands. Enhanced job control, or decision latitude, was hypothesized to allow employees to develop a sense of mastery and learn to cope with their job demands (e.g., Sonnentag & Zijlstra, 2006; Theorell & Karasek, 1996). Discovering that social support also helped to buffer against job demands, researchers expanded the model into the job demands-control-support model (Karasek & Theorell, 1990) and explored the possibility that control and support are interchangeable (e.g., Van Yperen & Hagedoorn, 2003). Researchers have discovered mixed evidence for the predicted two-way (demand-control and demandsupport) and three-way (demand-control-support) 426

interactions (van der Doef & Maes, 1999). In light of this mixed evidence, European researchers have recently proposed a job demands–resources model that focuses on independent effects of job demands and resources on different aspects of burnout (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). Job demands—characteristics that require effort— are proposed to contribute to the emotional exhaustion dimension of burnout. Job resources— characteristics that facilitate goal achievement, demand reduction, or personal growth—reduce disengagement or depersonalization (Bakker & Demerouti, 2007; Halbesleben & Buckley, 2004). Together, the job demands-control-support and job demands–resources models encourage job design researchers to study additional job characteristics and consider their implications for occupational health outcomes such as stress, strain, burnout, and illness. WHERE ARE WE NOW? CONTEMPORARY PERSPECTIVES ON JOB DESIGN Now that we have traced the history of job design research, we turn to contemporary perspectives that have emerged in recent years and are continuing to receive attention. These contemporary perspectives can be divided into two general categories: (a) new job characteristics and (b) new moderators, mediators, and outcomes of job design. These developments are directed toward overcoming the narrow focus of the JCM on only five job characteristics, three psychological mechanisms, and four outcomes of motivation, satisfaction, performance, and withdrawal behaviors (Parker et al., 2001). Figure 13.1 provides a summary model to integrate both the classic and contemporary perspectives that we discuss in the chapter.

New Job Characteristics: Including the Physical, Knowledge, and Social Morgeson and colleagues have developed an integrative typology and Work Design Questionnaire that divides job characteristics into four broad categories: task, physical, knowledge, and social (Humphrey et al., 2007; Morgeson & Humphrey, 2006). The task characteristics focus on the five JCM characteristics of

Copyright American Psychological Association. Not for further distribution.

Work Matters

FIGURE 13.1. An integrative model of job design.

autonomy, variety, task significance, task identity, and job feedback.5 The physical, knowledge, and social characteristics build on the efforts of several teams of researchers to broaden the scope of job design research beyond a relatively narrow focus on the characteristics of the tasks that employees perform. Physical Characteristics of Jobs. One of the earliest extensions expanded job design research beyond a restricted emphasis on the tasks that employees perform toward a consideration of the physical contexts in which they work. This choice was supported by a multidimensional scaling study by Stone and Gueutal (1985), who found that physical demands are one of the three core dimensions along which individuals perceive jobs. Two years earlier, Oldham and Rotchford (1983) introduced a typology of office characteristics specifying that 5

physical environments in which employees perform their jobs vary in terms of openness, office density, workspace density, accessibility, and darkness. In a study of university employees, they found that these office characteristics were related to satisfaction and discretionary behavior through employees’ interpersonal, job, and environmental experiences. Subsequent research has underscored the performance and well-being costs of crowding via high spatial density, low interpersonal distance, or a lack of physical enclosures such as partitions—especially if employees work in simple jobs, have poor stimulus-screening skills (Fried, 1990; Oldham, Kulik, & Stepina, 1991), or have high privacy needs (Oldham, 1988). Thus, physical characteristics of jobs can refer to both the physical features of tasks as well as the broader

In the Morgeson and Humphrey (2006) typology and questionnaire, autonomy is divided into three dimensions (decision making, scheduling, and work methods), and variety is divided into two types (task variety, a task characteristic, and skill variety, a knowledge characteristic).

427

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

physical environments in which employees perform their tasks. Other studies have demonstrated that employees adapt more favorably to high-density physical environments when they spent their childhoods in dense residential environments (Zhou, Oldham, & Cummings, 1998), that employees with low selfefficacy or external loci of health control are more responsive to physical conditions (May, Schwoerer, Reed, & Potter, 1997), and that the physical context of job design influences not only employees’ reactions but customers’ reactions as well (Conlon, Van Dyne, Milner, & Ng, 2004). Researchers have also begun to study physical danger (Jermier, Gaines, & McIntosh, 1989), physical taint (Ashforth & Kreiner, 1999), and noise as other physical characteristics along which jobs vary. With respect to noise, some researchers have identified music as a source of relaxation in simple jobs (Oldham, Cummings, Mischel, Schmidtke, & Zhou, 1995), whereas others have discovered that ambient noise is a source of job dissatisfaction, stress, high blood pressure, and absenteeism in complex jobs (Fried, Melamed, & Ben-David, 2002; Melamed, Fried, & Froom, 2001). Although this emerging literature on the physical context of job design is reminiscent of the early Hawthorne studies on environmental conditions, these studies have provided new insights into the important impact that physical characteristics of jobs can have on psychological, behavioral, and health outcomes. Researchers have developed questionnaires to capture the physical context of job design, which includes dimensions such as ergonomics, physical demands, work conditions, and equipment use (Morgeson & Humphrey, 2006), and environmental design, facilities, workload and activity levels, equipment and tools, and health and safety (Carlopio, 1996). Field experiments and quasi-experiments conducted by Oldham and colleagues (Oldham, 1988; Oldham et al., 1995) and May and colleagues (May, Reed, Schwoerer, & Potter, 2004; May & Schwoerer, 1994) have helped to strengthen causal inferences and illuminate factors that moderate individuals’ reactions to the physical context of jobs. Moreover, in a recent meta-analysis, Humphrey et al. (2007) found that after controlling for task, knowledge, and social 428

characteristics, physical job characteristics explained 16% incremental variance in stress and 4% incremental variance in job satisfaction. As such, scholars now agree that to gain a complete understanding of job design, we need to study the physical environment as well as the task environment (Fried, Slowik, Ben-David, & Tiegs, 2001; May, Oldham, & Rathert, 2005; Morgeson & Humphrey, 2006). Knowledge characteristics of jobs. Researchers have also called attention to the fact that jobs vary in terms of the knowledge that they require employees to acquire, retain, and utilize. This focus on knowledge characteristics was spearheaded primarily by the efforts of Campion and colleagues (e.g., Campion, 1988) and Wall and colleagues (e.g., Wall, Jackson, & Mullarkey, 1995). Morgeson and Humphrey (2006) synthesized research on five knowledge characteristics of jobs: complexity, information processing, problem solving, skill variety, and specialization. They found that knowledge characteristics predict job satisfaction, and unlike task characteristics, knowledge characteristics are related to training and compensation requirements. Complexity, which describes the difficulty versus simplicity of a job, was one of the first knowledge characteristics to receive attention in the literature (Campion, 1988; Edwards et al., 2000). Information processing is a related knowledge characteristic that captures the extent to which a job requires employees to pay attention to events, monitor data, and actively use cognitive abilities for sense-making and decisionmaking purposes (Jackson, Wall, Martin, & Davids, 1993; Martin & Wall, 1989). Problem solving, a third knowledge characteristic, focuses on the degree to which the job involves generating ideas, implementing solutions, and diagnosing and resolving errors (Wall, Corbert, Martin, Clegg, & Jackson, 1990), activities which are especially common in jobs with high creativity requirements (Morgeson & Humphrey, 2006; Unsworth, Wall, & Carter, 2005). Skill variety, a fourth knowledge characteristic, is drawn directly from Hackman and Oldham’s (1976) conceptualization of the breadth of capabilities needed to carry out the work. Specialization, a fifth knowledge characteristic, differs from skill variety in

Copyright American Psychological Association. Not for further distribution.

Work Matters

that it captures the depth, rather than breadth, of skills required to perform the work (Campion, 1988; Edwards et al., 2000; Morgeson & Humphrey, 2006). Research on knowledge characteristics has challenged scholars to recognize that jobs vary in their learning and skill requirements as well as in their motivational opportunities and physical conditions. As hinted previously, knowledge characteristics often present trade-offs between simplicity and skill usage, and thus between efficiency and satisfaction (Morgeson & Campion, 2002; Morgeson & Humphrey, 2006). For example, research suggests that jobs high in complexity and information processing involve considerable mental demands and challenges, and can thus serve as sources of both stress and satisfaction (e.g., Xie & Johns, 1995; see also Little, 1989). Researchers have only begun to study the conditions under which these trade-offs can be minimized or even eliminated (Morgeson & Campion, 2002; see also Drach-Zahavy, 2004). Social characteristics of jobs. Researchers have also begun to call attention to the social characteristics along which jobs vary—the interpersonal connections, interactions, and relationships that are embedded in assigned responsibilities (Grant, 2007; Morgeson & Humphrey, 2006). Although early research on job design included interpersonal components of jobs such as social structure (Trist & Bamforth, 1951), requisite interdependence, required and optional interaction, and received versus initiated directions of interaction (Turner & Lawrence, 1965), and dealing with others, feedback from others, and friendship opportunities (Hackman & Lawler, 1971; Hackman & Oldham, 1976), as noted earlier, these social characteristics disappeared from subsequent research (Grant, 2007; Grant et al., 2007; Latham & Pinder, 2005; Morgeson & Campion, 2003).6 This is surprising given that Stone and Gueutal (1985) identified service to the public, a social characteristic, as one of the three core dimensions on which individuals perceive jobs. Recently, we have witnessed a resurgence of attention to the social context of job design. Morgeson and Humphrey (2006) consid6

ered jobs as varying in terms of four social characteristics: social support, interdependence, interaction outside the organization, and feedback from others. Morgeson and Humphrey’s view of social support is based on the aforementioned research by Karasek and colleagues, which highlights that jobs differ in the degree to which they allow employees to receive assistance from supervisors and coworkers (Karasek, 1979; Karasek & Theorell, 1990), as well as early conceptualizations of friendship opportunities (Hackman & Lawler, 1971; Sims, Szilagyi, & Keller, 1976). Interdependence emphasizes the extent to which employees rely on each other to complete work, and can be divided into two types: initiated interdependence, where employees pass their work along to others, and received interdependence, where others’ work is passed along to employees (Kiggundu, 1981, 1983; Morgeson & Humphrey, 2006). Interaction outside the organization describes the extent to which the job enables employees to communicate and interrelate with people external to the organization’s boundaries, such as clients, customers, or suppliers (Morgeson & Humphrey, 2006). Finally, feedback from others captures the extent to which employees receive information from other people about their performance (Hackman & Lawler, 1971; Hackman & Oldham, 1980; Morgeson & Humphrey, 2006). These social characteristics of jobs appear to play an important role in employees’ attitudes and experiences. In a meta-analysis, Humphrey et al. (2007) found that all four social characteristics were associated with job satisfaction (mean p = .36). Moreover, they found that even after controlling for task and knowledge characteristics, these four social characteristics explained incremental variance of 17% in job satisfaction, 18% in role ambiguity and conflict, 40% in organizational commitment, 24% in turnover intentions, and 9% in subjective performance. Together, they found that task, knowledge, physical, and social characteristics explained 55% of the variance in job satisfaction, 54% in role ambiguity, 38% in stress, and 23% in burnout.

Social characteristics of jobs are distinct from the social cues discussed in Salancik and Pfeffer’s (1978) social information processing perspective. Whereas Salancik and Pfeffer focused on social cues that are independent of the objective structure of the job itself, social characteristics capture the connections, interactions, and relationships that are structured into the job. For further explanation, see Grant (2008a).

429

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

A different view of the social characteristics of jobs is offered by Grant and colleagues. These researchers have focused on the design of jobs to fuel prosocial, rather than intrinsic, motivation: to motivate employees to care about protecting and promoting the wellbeing of beneficiaries (Grant, 2007, 2008c). These researchers have proposed that when jobs are high in both task significance and contact with beneficiaries, employees will experience higher perceptions of impact on beneficiaries and affective commitments to beneficiaries. These experiences will trigger prosocial motivation, which will drive employees to display additional effort, persistence, and helping behavior. These predictions have been tested in a series of recent studies. For instance, Grant et al. (2007) found significant effects of contact with beneficiaries on persistence that were (a) mediated by higher levels of perceived impact on and affective commitment to beneficiaries and (b) moderated by task significance, which strengthened the effect of contact with beneficiaries on persistence. Grant (2008a) expanded on this research by examining new mechanisms and boundary conditions of the performance effects of task significance. Noting that previous research had yet to establish a causal impact of task significance on job performance, Grant (2008a) sought to shed new light on this relationship, as well as its relational mediators and individual moderators. Whereas past research had treated task significance as a characteristic of the work itself that enables employees to experience their tasks as more meaningful (Hackman & Oldham, 1976; Morgeson & Humphrey, 2006), Grant (2007) proposed that task significance is also a relational job characteristic because it connects employees to the impact of their actions on other people. Grant (2008a) drew on this notion to propose that task significance increases job performance by strengthening employees’ perceptions of impact on beneficiaries, as well as by enabling employees to feel valued and appreciated by beneficiaries. To test these mechanisms and investigate their boundary conditions, Grant (2008a) conducted three field experiments. In the first experiment, he found that task significance cues increased the performance of fundraisers, relative to two control groups and their own baselines. In the second experiment, he 430

found that task significance cues increased the job dedication and helping behavior of lifeguards, relative to a control group and their own baselines. These effects were mediated by lifeguards’ heightened perceptions of impact on and appreciation from the guests in their pool. In the third experiment, he found that task significance cues led new fundraising callers to raise more pledges in their first week on the job than callers in a control group. He further found that these effects were independently moderated by individual differences in conscientiousness and prosocial values. Task significance had stronger performance effects for employees with low levels of conscientiousness, whose effort is more dependent on external signals, and employees with prosocial values, who are more concerned about doing work that protects and promotes the welfare of others. These experiments highlight the causal impact that task significance can have on job performance and introduce new relational mediators and individual moderators of these effects. A third perspective on social characteristics of jobs has been presented by researchers studying “necessary evils”—that is, tasks that require employees to harm others in the interest of a perceived greater good (Molinsky & Margolis, 2005). These researchers have offered an innovative theoretical perspective on how task structures affect the emotional drama of performing work that simultaneously does good and harm, as well as employees’ efforts to express compassion and sensitivity to the victims harmed by their efforts. Such tasks are especially common in the daily lives of health care professionals performing painful medical procedures, attorneys and judges determining the fates of accused criminals, and managers performing downsizings. Molinsky and Margolis proposed that necessarily evils vary in terms of task dimensions (complexity and frequency), agency dimensions (causality, task identity, legitimacy), and impact dimensions (magnitude and salience of harm, ratio of harm to benefit). One of the more fascinating issues raised by a focus on necessary evils is that some task designs may make the harm easier to deliver but undermine the employee’s motivation to express compassion and cause moral disengagement (Bandura, 1999) by shielding the employee from the harm being done. For example, Molinsky

Copyright American Psychological Association. Not for further distribution.

Work Matters

and Margolis (2005) suggest that complex or fragmented tasks involve less emotional drama but also invite less compassion and moral awareness. On the other hand, exposing employees directly to the victims and giving them responsibility for the entire process of harmdoing may facilitate expressions of compassion and protect moral sensibilities, but it tends to place severe emotional burdens on employees. Recently, researchers have begun to empirically investigate the conditions under which employees engage psychologically to express compassion while performing necessary evils (Margolis & Molinsky, 2008), as well as how the experience of doing good offsets the job dissatisfaction and burnout costs of the experience of doing harm (Grant & Campbell, 2007). Together, the studies highlighted above have challenged Hackman and Oldham’s (1976) findings about the weak predictive validity of dealing with others and friendship opportunities,7 corroborating earlier intuitions about the importance of social characteristics of jobs (Hackman & Lawler, 1971; Trist & Bamforth, 1951; Turner & Lawrence, 1965). The research programs advanced by Morgeson and colleagues, Grant and colleagues, and Molinsky and Margolis have accentuated the significant impact that social job characteristics can have on employees’ experiences, attitudes, behaviors, and performance. However, researchers have yet to explore how each social characteristic interacts with task, physical, and knowledge characteristics, and we see this as a promising opportunity for future research. In addition, there are other job characteristics that do not fit neatly into these four categories, and we cover them in a subsequent section on directions for future research.

New Moderators, Mediators, and Outcomes: Uncertainty, Proactivity, Dynamism, and Creativity As researchers have broadened job design theories to include task, physical, knowledge, and social characteristics, they have also presented new perspectives on the boundaries, processes and outcomes of job design. These developments move beyond the tradi7

tional focus on motivational processes and satisfaction, performance, and withdrawal outcomes, and they can be classified into four major categories: uncertainty, proactivity, dynamism, and creativity. Uncertainty. Scholars have pointed out that the majority of job design research has failed to attend to uncertainty, a contextual variable that plays a central role in psychological and organizational research (Johns, 2006). Wall and Jackson (1995) noted that conflicting evidence for the effects of job control might be resolved by incorporating uncertainty as a moderator, proposing a contingency perspective suggesting that job control is most likely to achieve beneficial outcomes when uncertainty is high. As Wright and Cordery (1999) summarized: Although both sociotechnical systems and job characteristics theorists stress job control as a primary causal factor influencing performance and job attitudes . . . neither explicitly predicts that the strength of these relationships will vary with the degree of contextual uncertainty . . . According to the contingency view, job redesign may fail to lead to improvements in performance simply because there are no system control benefits to be had from transferring decision-making control from supervisors to employees in simple, stable, and predictable operating environments. Conversely, job redesign programs may well succeed because they increase job control to suit the level of uncertainty at the job level or, alternatively, because they increase both uncertainty and job control simultaneously, such as through changes to workflow and technology. (p. 456) In an empirical study of production operators in a wastewater treatment company, Wright and Cordery (1999) found evidence that the association between job control and attitudinal outcomes was moderated by production uncertainty. More specifically, when

Future research is needed to explain why Hackman and Oldham (1976) returned weak results. Their findings may have been due to methodological artifacts such as range restriction and unreliable measures, attention to a limited range of social characteristics, or increases in the importance of social characteristics over time.

431

Grant, Fried, and Juillerat

Copyright American Psychological Association. Not for further distribution.

production uncertainty was low, job control was negatively associated with satisfaction and intrinsic motivation, but when production uncertainty was high, job control was positively associated with satisfaction and intrinsic motivation. These findings suggest that job control is most likely to offer psychological benefits to employees when they work in environments characterized by high levels of uncertainty, helping to position uncertainty as a key variable in job design theory and research. Proactivity. A number of researchers have challenged the assumption that jobs are static objects designed by managers. Ilgen and Hollenbeck (1991) recommended that we move away from our focus on jobs and toward an emphasis on roles, which capture both the formal and more informal, emergent attributes of work that are not always included in job descriptions. On the basis of an excellent synthesis of the largely separate literatures on job design and roles, Ilgen and Hollenbeck argued that jobs are created by managers, who identify a set of required task elements for employees to perform. However, as employees enact their jobs, they become aware of additional elements that need to be incorporated in order to perform them effectively in context. Ilgen and Hollenbeck defined the role as the combination of the formal, assigned and informal, emergent task elements. They pointed out that employees often take initiative to incorporate new task elements into their roles and negotiate altered roles with supervisors (see also Graen, 1976). Other researchers have elaborated on Ilgen and Hollenbeck’s ideas to capture the ways in which employees’ responsibilities change over time. (See also Vol. 2, chap. 19, this handbook.) Researchers have increasingly recognized that rather than passively reacting to the jobs that managers assign to them, employees proactively take initiative to alter their own roles and jobs (Frese & Fay, 2001). This general viewpoint has been expressed by a number of different scholars (for reviews, see Grant & Ashford, 2008; Grant & Parker, 2009). For example, Staw and Boettger (1990) introduced the concept of task revision to capture how employees proactively improve flawed task structures, and Black and Ashford (1995) 432

studied how new employees change their own roles to “make jobs fit” during the adjustment process. Similarly, Parker, Wall, and Jackson (1997) asserted that as organizational structures flatten, employees are given increased autonomy and latitude to change their own jobs. They collected data suggesting that modern manufacturing and production practices result in enhanced autonomy, which gives employees the freedom to expand their own roles. As the authors summarize, “Autonomy allows hands-on learning in which people have the opportunity to interact with the environment and become more involved in, and more knowledgeable about, the wider production process. This experience might then lead to broader ownership of problems and a more proactive view of performance” (Parker et al., 1997, p. 923). Thus, Parker and colleagues identified learning as a new mechanism through which autonomy enhances job performance (see also Frese, Kring, Soose, & Zempell, 1996; Langfred & Moye, 2004; Liden, Wayne, & Sparrowe, 2000; Wall, Jackson, & Davids, 1992). In subsequent research, Parker and colleagues have sought to investigate the psychological processes through which autonomy facilitates role expansion and thereby more proactive behaviors. They have argued that proactive behaviors emerge when autonomy cultivates a psychological state of role-breadth self-efficacy (RBSE), or feeling capable of taking on a broader, more proactive set of responsibilities (e.g., Parker, 2000, 2007). For example, Parker, Williams, and Turner (2006) found that individuals with higher levels of RBSE were more likely to be proactive in implementing ideas and solving problems, and Griffin, Neal, and Parker (2007) found that RBSE predicted proactive behaviors directed toward one’s task, one’s team, and one’s broader organization. In a series of studies, Parker and colleagues have found that autonomy and control are important facilitators of RBSE. Across two field studies, Parker (1998) found that autonomy can contribute to the development of RBSE by signaling to employees that they are capable of handling larger responsibilities, a finding replicated by Morgeson, Delaney-Klinger, and Hemingway (2005). In another field study, Parker and Sprigg (1999) discovered that job control and job demands interact to predict higher levels of RBSE

Copyright American Psychological Association. Not for further distribution.

Work Matters

only for employees with proactive personalities, who are motivated and able to take advantage of job control to cope with and learn from their job demands. Reinforcing the importance of autonomy for promoting RBSE, Axtell and Parker (2003) conducted a longitudinal study revealing that enlarging jobs without increasing autonomy was associated with decreases in RBSE, and Parker (2003) found in a longitudinal quasi-experiment that the introduction of lean production practices reduced RBSE by undermining employees’ perceptions of autonomy, skill utilization, and participation in decision making. Together, these studies underscore the value of considering knowledge, skill development, and learning mechanisms— not only motivational mechanisms—as mediators of the effects of job characteristics on employees’ attitudes, behaviors, and well-being (see also Holman & Wall, 2002). Building on this emphasis on proactivity, researchers have begun to examine the role of job design in shaping whether roles can be formalized or must emerge more proactively. Griffin et al. (2007) proposed that as interdependence rises, role performance depends on contributions to the broader team and organization rather than to individual tasks, and as uncertainty rises, role performance depends on adaptive and proactive behaviors rather than merely completing tasks proficiently. Their theoretical model highlights the importance of interdependence and uncertainty in encouraging employees to take on more proactive, emergent roles as opposed to merely carrying out formalized jobs (see also Dierdorff & Morgeson, 2007). This focus on proactivity also appears in Wrzesniewski and Dutton’s (2001) theoretical model of job crafting. Wrzesniewski and Dutton developed the concept of job crafting to describe the process through which employees proactively alter the boundaries of their own tasks and relationships. They proposed that employees can change physical task boundaries by altering the number or type of tasks that they complete, cognitive task boundaries by reframing their views of their tasks, and relational boundaries by altering with whom and how they 8

interact and communicate at work. They described how employees are motivated to engage in job crafting by desires for control, work meaning, positive identities, and interpersonal connections, and how the effect of these motives on job crafting depends on perceived opportunities for crafting, job features, and individual work and motivational orientations. They further suggested that by crafting their jobs, employees are able to change the meaning of their work and their identities at work. For example, they described how a group of hospital cleaners crafted their jobs by actively caring for patients and their families, even though this was not part of their job descriptions. A focus on job crafting suggests that employees are active architects, not merely passive recipients, of jobs. The job crafting concept has been generative in integrating different views of how employees proactively take initiative to alter their own jobs, roles, and tasks, and in inviting a broader consideration of the ways in which they do so and the work meaning and identity functions that it serves. In a more recent conceptual paper, Rousseau et al. (2006) suggested that job crafting may even occur prior to accepting a job. They proposed that employees often negotiate idiosyncratic deals, or “i-deals,” in which supervisors agree to unique job expectations or employment arrangements that differ from those given to other employees performing the same job.8 Combining these different perspectives, it is now clear that employees play a proactive role in shaping their own job designs. Dynamism. A recent advancement in job design theory was offered by Clegg and Spencer (2007). These authors criticized prior research for its static focus on fixed job designs, building on the proactivity research cited previously to propose a more flexible view that culminates in a “circular and dynamic” model of the job design process. They proposed that when employees perform effectively, supervisors interpret this performance as a sign of competence and develop higher levels of trust in employees. In addition, employees themselves interpret this performance as a sign of

Moreover, researchers have suggested that job crafting can involve negotiation with peers as well as supervisors (e.g., Fried, Levi, & Laurence, 2007). For example, Langfred (2007) suggested that when trust among team members is reduced due to conflict, team members are less willing to grant work autonomy to other team members. In contrast, when trust is high, team members are willing to allow and facilitate job crafting.

433

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

competence and develop higher levels of trust in themselves. These enhanced levels of interpersonal and intrapersonal trust lead to role expansion, which can be initiated by supervisors or by employees themselves through job crafting. Role expansion enhances employees’ motivation and opportunity to learn and develop new knowledge, thereby fueling higher performance, and the cycle begins again. The logic of the model also applies in reverse to poor performance. Supervisors and employees themselves interpret poor performance as a signal of incompetence, which reduces interpersonal and intrapersonal trust and leads to role constriction, through smaller assignments and less autonomy from supervisors or through employees’ own efforts to craft simpler jobs. This constricted role decreases employees’ motivations and opportunities to learn, decreasing performance, and the cycle repeats itself. Clegg and Spencer’s (2007) model presents several promising contributions to our understanding of job design. First, rather than treating job design solely as a predictor variable and performance as an outcome variable, they conceptualized both variables as predictors and outcomes that are dynamically interrelated. Second, by incorporating job crafting and other forms of proactivity, they moved beyond static perspectives by highlighting the flexibility and malleability of job design. Third, they integrated knowledge and motivational mechanisms through which role expansion and autonomy may facilitate performance. Despite these strengths, there are theoretical and methodological challenges that merit attention in further conceptual and empirical work. For example, Clegg and Spencer (2007) wisely noted that the model assumes that performance triggers self-fueling spirals or “deviation-amplifying loops” (Weick, 1979; see also Lindsley, Brass, & Thomas, 1995), but virtuous or vicious cycles are unlikely to continue into perpetuity. For example, at very high or low levels of performance, employees may reach “performance ceilings” or “performance floors” in which it is no longer possible for performance to continue escalating in positive or negative directions. Moreover, poor performance in and of itself may motivate supervisors to provide employees with further training and motivate employees themselves to proactively seek out 434

feedback and learning opportunities (e.g., Ashford, Blatt, & VandeWalle, 2003; Kluger & DeNisi, 1996). We hope to see researchers incorporate new mediators and moderators that explain how Clegg and Spencer’s virtuous and vicious cycles are counteracted. Nevertheless, we applaud the development of a dynamic, cyclical, reciprocal model that prompts researchers to examine the multiple causal pathways through which job designs, roles, and performance interrelate. To test their model, multiwave longitudinal studies will be critical (e.g., Frese, Garst, & Fay, 2007), and we are especially enthusiastic about the prospects for cross-lagged designs that can adjudicate questions about temporal order by facilitating comparisons of reciprocal relationships. We also hope to see researchers conduct growth modeling and nonlinear analyses to begin to explore the spirals proposed by Clegg and Spencer. Creativity and workday cycles. Job design researchers have also begun to consider creativity as an outcome. (See also chap. 9, this volume.) Oldham and Cummings (1996), for example, found that employees working in enriched jobs (i.e., high scores on the JCM attributes) were rated as more creative, produced more patents, and offered more suggestions. Enriched jobs were stronger predictors of several of these creativityrelevant outcomes when employees had creative personalities or supportive or noncontrolling supervision. Elsbach and Hargadon (2006) extended our understanding of job design and creativity by introducing a framework of “workday design” for knowledge workers. They asserted that many knowledge workers are chronically overloaded, facing daily demands and obstacles that undermine their creativity. They proposed that the creativity of knowledge workers can be enhanced by identifying and regularly scheduling simple, easily mastered tasks that involve low cognitive difficulty and low performance pressure (see also Ohly, Sonnentag, & Pluntke, 2006). They suggested that daily doses of “legitimate and scheduled mindless work” may enhance employees’ cognitive capacity, feelings of psychological safety, and positive affect, and that these psychological states will in turn fuel creativity. Elsbach and Hargadon’s (2006) framework offers at least three noteworthy contributions to job design

Copyright American Psychological Association. Not for further distribution.

Work Matters

theory and research. First, they shifted our unit of analysis by suggesting that researchers should focus on designing workdays rather than jobs or tasks; this draws our attention to the importance of considering how tasks are sequenced throughout the course of a day, an issue long neglected in job design research. Second, consistent with the predictions and findings presented by Xie and Johns (1995), Elsbach and Hargadon challenged the long-held assumption that reduced variety and complexity undermine motivation: When employees work in very complex, high-pressure knowledge jobs, tasks that would traditionally be described as dull and monotonous may provide a welcome break. Third, they offered new ideas for managing commonly observed tradeoffs in job design research (see Morgeson & Campion, 2002): By alternating complex, challenging tasks with routine, mindless tasks, employees may achieve a balance of pressure and relaxation that is conducive to high creativity and relatively low stress. Summary. These perspectives on uncertainty, proactivity, dynamism, and creativity break new ground in job design theory and research. Research on uncertainty has helped us understand how the effects of job control are contingent on organizational and industrial contexts. Research on proactivity has helped us understand how employees take initiative to shape their own job designs. Research on dynamism has illuminated how such initiative results in spirals of changes in job characteristics, relationships, and performance over time. Research on creativity has helped us understand how tasks can be sequenced within workdays to stimulate original, flexible thinking. Together, these viewpoints have expanded the scope of moderators, mediators, and outcomes beyond those traditionally considered in job design research. WHERE ARE WE GOING? FUTURE DIRECTIONS Now that we have covered the past and the present of job design theory and research, we turn our focus to the future. Our emphasis in this section is on unanswered questions and further directions that merit attention in ongoing conceptual and empirical

inquiry. We focus on two key themes: taking context seriously and unanswered questions.

Taking Context Seriously To paraphrase Bob Dylan, “Jobs, they are a-changin’.” Recent changes in the nature of work present both opportunities and challenges for job design research. A number of scholars have pointed out that the job design literature has largely neglected the dramatic changes in work contexts and job environments that have occurred over the past few decades (e.g., Johns, 2006; Holman, Clegg, & Waterson, 2002; Parker et al., 2001; Rousseau & Fried, 2001). We see several valuable steps that researchers can take to incorporate these contextual changes: continue studying new social and knowledge characteristics of jobs, consider temporal characteristics of jobs, and explore more macroscopic environmental variables as antecedents of job design and moderators of its effects. New social characteristics of jobs. Social characteristics of jobs are changing at a rapid pace. As we shift from a manufacturing economy to a service economy, and we continue to see increases in task interdependence and the use of teams, employees’ jobs may be more embedded in and interconnected to interpersonal relationships than ever before (e.g., Grant, 2007; Parker et al., 2001). The time is ripe for researchers to examine new social characteristics of jobs, revisit forgotten characteristics, or consider dimensions that have received little attention in prior research. For example, Turner and Lawrence (1965) suggested that jobs vary in their social desirability and status. Although social status and stigma have been central themes in research on dirty work (e.g., Ashforth & Kreiner, 1999), job design researchers have scarcely taken notice of these important variables. As a second example, Turner and Lawrence (1965) originally defined task identity as the extent to which a job involved work that was clearly differentiated as a unique and visible assignment. Similarly, Ariely, Kamenica, and Prelec (2008) found that having one’s products destroyed by others—seeing one’s written work put through a paper shredder or watching the experimenter disassemble a machine that one has built—may threaten meaning by 435

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

challenging individuals’ beliefs that the work will last as a whole, identifiable product that is visible to others. These ideas and findings suggest that task identity may be a social characteristic of jobs, in that task identity is higher in jobs that are more distinct from those of others and permanently observable to others. Turner and Lawrence also identified responsibility as a potential job characteristic that encompasses the probability of serious error, the ambiguity of remedial action (the clarity of the solution), and the time span of discretion (the delay needed to detect mistakes). All of these dimensions of responsibility may be social characteristics—not only task characteristics—in that they have implications for the harm that employees may do to others as a result of making errors. As a third example, friendship opportunities may become less prevalent as a social job characteristic. The advent of virtual work, global operations, temporary project work, and independent contracting may reduce opportunities for social interactions and interpersonal relations (Shamir & Salamon, 1985), as well as for building trust and strong ties. Therefore, both employers and employees are facing challenges in developing meaningful interpersonal relationships on the job. In response to these challenges, the phenomenon of “coworking” has emerged, whereby independent workers in different jobs work in a common space for a sense of community (Fost, 2008). This new form of working is ripe for theoretical and empirical attention. As a final example, researchers have begun to consider the social features of virtual work, with evidence suggesting that empowerment may be particularly important in virtual teams with little face-to-face interaction (Kirkman, Rosen, Tesluk, & Gibson, 2004). In addition, in the service industry, for example, as technology improves, we expect increases in the opportunity for virtual interaction between the service employees and their customers, regardless of geographical location. These increased opportunities for visual contact with beneficiaries are expected to enhance employees’ experience of task significance (Grant, 2007). Along these lines, we hope to see further research on new and forgotten social job characteristics. 436

New knowledge characteristics of jobs. Knowledge characteristics of jobs may be expanding and changing at similar rates. Recent years have brought continued increases in the scope and importance of knowledge work, significant growth in globalization and global operations, greater employee involvement in job design and greater autonomy for job crafting, and the enhanced use of continued information technology and flexible work methods, ranging from virtual teams to teleworking (e.g., Elsbach & Hargadon, 2006; Parker et al., 2001; Rousseau & Fried, 2001; Sinha & Van de Ven, 2005). Many of these changes are associated with increased unpredictability and uncertainty. As such, researchers have recommended that we devote greater attention to the design of knowledge and creative jobs and their creative requirements (Elsbach & Hargadon, 2006; Unsworth et al., 2005), as well as the design of the knowledge-intensive jobs held by executives (Hambrick, Finkelstein, & Mooney, 2005) and white-collar employees and managers (Xie & Johns, 1995). Shamir (1992) has even called for a “nonorganizational work psychology” that focuses on the dynamics of working from home, which are especially salient for employees performing virtual work. Along these lines, we expect to see researchers continue to uncover new knowledge characteristics of jobs and explore how their effects are contingent on moderators at the job, individual, and organizational levels. Changing knowledge characteristics of jobs may affect task characteristics as well. Autonomy is particularly important in knowledge work (e.g., Janz & Prasarnphanich, 2003), and knowledge workers are increasingly being given freedom not only in terms of “when to do” and “how to do” (Hackman & Oldham, 1980), but also in terms of “what to do,” “with whom to do,” and “from where to do” (e.g., Breaugh & Becker, 1987; Morgeson & Humphrey, 2006). In knowledge-based organizations, such as high-tech startups, the premium placed on innovation often leaves employees with discretion about what specific goals and tasks to pursue (Fried et al., 2008). Further, flexibility in work locations may have both benefits and costs for knowledge workers. On the one hand, increased location autonomy increases control over

Copyright American Psychological Association. Not for further distribution.

Work Matters

job performance; on the other hand, being able to work from home or away from work in nonstandardized hours may increase role overload and burnout (Fried et al., 2008). We clearly need more research on the effects of knowledge characteristics on task characteristics and outcomes in changing work environments. As another example of knowledge characteristics influencing task characteristics, the growing use of technology to provide electronic performance feedback and monitoring may lead to cognitive overload, burnout, reduced control and lower performance (Kluger & DeNisi, 1996; Parker et al., 2001). As technology progresses, we expect ongoing advances in opportunities for immediate and timely feedback, which may exacerbate the problem of excessive feedback. How can organizations design knowledge characteristics of jobs to create an optimal level of timely and detailed feedback? Finally, in addition to knowledge characteristics, there is a need to develop a theoretical conceptualization of skill and ability characteristics, which will capture what employees are trained and able to do, as opposed to simply what they know. Temporal job characteristics. We also hope to see researchers investigate whether Morgeson and Humphrey’s (2006) four categories of task, physical, knowledge, and social characteristics comprehensively capture the full set of categories that should be used to describe jobs. Temporal job characteristics—job features that influence the time horizons on which employees complete work—may be one category worth adding, especially as technological advances continue to fuel faster performance and cycle times. Such variables as time pressure (Elsbach & Hargadon, 2006) and work cycles, time-toaccomplishment, and required delay of gratification (Fried, Grant, Levi, Hadani, & Slowik, 2007) may qualify as temporal job characteristics. Existing temporal perspectives have focused on dynamic relationships among task and knowledge characteristics (Clegg & Spencer, 2007; Mathieu, Hofmann, & Farr, 1993) but have not yet fully captured temporal characteristics themselves.

New macroscopic environmental variables and cultural differences. The nature of the workforce itself is changing considerably, with more women, greater ethnic diversity, more educated employees, altered psychological contracts between employers and employees (Fried et al., 2008), and an aging population (e.g., Kanfer & Ackerman, 2004). These contextual changes give rise to new questions about the design, experience, and effects of jobs. Although the majority of job design models have been rooted in psychological frameworks focusing on individual motivation, satisfaction, and performance, researchers have offered hints that job designs are also embedded in national cultures, institutional fields, organizational structures, and emerging technologies (e.g., Brass, 1981; Dean & Snell, 1991; Oldham & Hackman, 1981; Parker et al., 2001; Robert, Probst, Martocchio, Drasgow, & Lawler, 2000; Spreitzer, 1996). For example, Robert et al. (2000) reported a negative relationship between empowerment and job satisfaction in India, which appears to be attributable to the lack of fit between empowering employees to make their own decisions and the Indian cultural values of power distance, which emphasize hierarchy and status. Similarly, Roe, Zinovieva, Dienes, and Ten Horn (2000) found a weaker relationship between autonomy and the JCM critical psychological states in Bulgaria and Hungary than in the Netherlands, which is characterized by a more individualistic culture (see also Gelfand, Erez, & Aycan, 2007). Furthermore, in a sample of more than 100,000 employees from 49 countries, Huang and Van De Vliert (2003) found that enriched job characteristics are related more strongly to job satisfaction in countries characterized by high wealth, high individualism, strong governmental social welfare programs, and low power distance. Finally, researchers have proposed that job design may have stronger effects in cultures characterized by high power distance, where employees are more likely to conform to supervisors’ expectations (Leung, 2001), and found that helping coworkers is more likely to be viewed as part of one’s job in collectivistic than individualistic cultures (Perlow & Weeks, 2002). These studies support the notion that the effect of job characteristics on 437

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

individual reactions will be affected by the national culture in which the organization is embedded. Although the findings on job characteristics in the context of culture are promising, we need more theoretical development and systematic research on the effect of particular job characteristics on specific outcome variables in different cultures and macroscopic contexts. Of particular value will be investigations of how autonomy and control unfold in different cultures. Some researchers have argued that autonomy is a universal psychological need across cultures that can be differentiated from individualism and independence: Autonomy involves choice, whereas individualism and independence involve separation from other people (Chirkov, Ryan, Kim, & Kaplan, 2003). Other researchers, however, have argued that autonomy is still more important in individualistic than collectivistic cultures (Chua & Iyengar, 2006). Further studies are needed to resolve this debate.

Unanswered Questions Job design researchers have only begun to scratch the surface of several important areas of inquiry. Next, we call attention to unanswered questions about the role of individual differences and job design, job design as a decision-making process, interactions among job characteristics, curvilinear effects, units of analysis, and multidimensionality of characteristics. Individual differences and job design. We believe it is time for researchers to move beyond growth need strength as the primary individual difference moderator of reactions to job characteristics. Although the five-factor model has been the dominant taxonomy of personality for nearly 2 decades (e.g., Barrick & Mount, 1991), surprisingly little research has investigated whether the Big Five personality traits of extraversion, neuroticism, conscientiousness, agreeableness, and openness moderate individuals’ attitudinal and behavioral reactions to job characteristics. There is evidence, however, that individual differences in conscientiousness and prosocial values moderate the effects of task significance on performance, with employees low in conscientiousness and high in prosocial values responding most favorably (Grant, 2008a). There is also evidence that 438

positive affectivity moderates the effect of objective task enrichment on task perceptions, with employees high rather than low in positive affectivity responding more favorably to moderately enriched tasks (Fortunato & Stone-Romero, 2001), and that psychologically flexible employees respond more favorably to enhanced job control (Bond, Flaxman, & Bunce, 2008). We hope to see attention to a broader range of individual differences as moderators. In addition to the Big Five, researchers may investigate the moderating roles of knowledge, skills, and abilities (Morgeson & Humphrey, 2008) and orientations toward work as a job versus career versus calling (Wrzesniewski, McCauley, Rozin, & Schwartz, 1997). With respect to work orientations, we may be witnessing a rise of job orientations as free time and leisure activities have increased substantially in the past few decades (e.g., Hunnicutt, 1988; Snir & Harpaz, 2002). Some have even argued that this increase in the importance of leisure time signifies a decrease in work importance (for a review, see Snir & Harpaz, 2002). According to compensation models, employees who experience deprivation at work will compensate in their choice of non–work activities (e.g., Kohn & Schooler, 1982; Snir & Harpaz, 2002; Wilensky, 1960; cf. Judge, Bono, & Locke, 2000). This suggests that employees who lack enriched jobs will seek out enrichment in other life domains, and employees who lack enriched nonwork lives may seek out enriched jobs. These reactions, however, may depend on employees’ work orientations, with calling-oriented employees seeking out greater involvement and identity engagement in work and job-oriented employees preferring to invest their time, energy, and identities in nonwork activities. Researchers may also attend to the impact of gender differences on job design, returning to classic research on orientations toward people versus things and data versus ideas (Fine, 1955; Lippa, 1998; Little, 1972; Morgeson & Campion, 2003; Rousseau, 1982), as well as to debates about whether gender differences are due to evolutionary and biogenetic sources (Buss, 1995) or social roles, expectations, and stereotypes (Eagly & Wood, 1999). With respect to gender, the past few decades have witnessed significant increases in working women and dual-career

Copyright American Psychological Association. Not for further distribution.

Work Matters

families (e.g., Parker et al., 2001). This, in turn, has increased the potential for work–family conflict (Oldham, 1996; Parker et al., 2001). Such conflict, unless being carefully managed, can adversely affect employees’ abilities to function in demanding work environments (Fried et al., 2008). We clearly need more research on the effect of dual-career issues on employees’ reactions to job design, and on what organizational policies and choices can enable dual-career employees to successfully manage high job demands without creating work–family conflict. Finally, research is also needed on the roles that individual values (Grant, 2008a), interests (Holland, 1996), and knowledge, skills, and abilities (Morgeson et al., 2005) play in moderating reactions to job design. Vocational psychology may offer particularly useful contributions in this area (Gustafson & Mumford, 1995). Job design as a decision-making process. At present, we know little about how managers make decisions about jobs (Campion & Stevens, 1991). From a sociological perspective, managers’ decisions may be influenced by institutional norms and mimicry of similar firms (Meyer & Rowan, 1977), as well as the fads and fashions that take the popular press by storm (Abrahamson, 1996). From a psychological perspective, managers may use heuristics to guide decisions about how to design jobs (Heath, Larrick, & Klayman, 1998). For example, research on the false consensus bias suggests that managers may rely on their own preferences and personalities to infer their employees’ preferences (Marks & Miller, 1987; Ross, Greene, & House, 1977). Similarly, as noted earlier, researchers have shown that many decision makers systematically underestimate the importance of enriched job characteristics in motivating employee performance, relying instead on work simplification principles (Campion & Stevens, 1991) and extrinsic rewards (Heath, 1999). Along these lines, recent research has suggested that job perceptions are a mechanism through which transformational leaders may inspire higher task performance and citizenship behavior. More specifically, researchers have found that perceptions of jobs as motivating and meaningful mediate the associations

of transformational leadership with the outcomes of task performance and citizenship behavior (Piccolo & Colquitt, 2006; Purvanova, Bono, & Dzieweczynski, 2006). Additionally, the concept of evocation offered by Buss (1987) implies that managers may base job design decisions in part on the personality traits of employees. Managers may offer task significance and autonomy to conscientious employees, high interpersonal contact to agreeable extraverts, and jobs with strong creative requirements to open-minded employees. Once managers have made these decisions, how do they implement them? For instance, when seeking to enhance an employee’s task significance, do managers share inspiring stories, implement contact with beneficiaries, provide more autonomy and support for job crafting, or even delegate their own significant tasks to employees? Although sparse research has attended to the processes through which managers make and implement job design decisions, we believe that this is a fruitful avenue that could spawn an entire literature. From a different angle, the field would also benefit from research on the political and social processes that affect job crafting when leaders, structures, and climates are not supportive of job changes initiated by individual employees. Interactions among job characteristics. Researchers have largely neglected efforts to systematically investigate how multiple job characteristics interact to influence attitudes and performance (Dodd & Ganster, 1996). Hackman and Oldham (1976, 1980) proposed that autonomy would enhance the motivational effects of meaningrelated job characteristics such as task significance and task identity, such that task significance and task identity would produce more favorable effects on attitudes and performance when employees had autonomy. This synergistic effect has received little support (Dodd & Ganster, 1996; Oldham & Hackman, 2005). Perhaps it is time for researchers to abandon the synergistic hypothesis in favor of a compensatory hypothesis. For example, high-reliability organizations (HROs), such as air traffic control systems and nuclear power plants, place high priority on preventing errors (Hofmann & Stetzer, 1998; Weick & Roberts, 439

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

1993; Zohar & Luria, 2003). The expected growth of these organizations is consistent with the projected increase in importance of public safety and security needs, as well as the increased complexity of technology and its impact on society. HROs often use restrictive rules and procedures to reduce individual error (Weick, Sutcliffe, & Obstfeld, 1999). The potential motivational costs of this lack of autonomy may be offset by the high levels of task significance inherent in the mission of protecting public safety and human well-being. We hope that researchers will examine whether high task significance compensates for low autonomy in HROs, and explore other new patterns of interactions between job characteristics (see also Morgeson, Johnson, Campion, Medsker, & Mumford, 2006). On a related note, researchers have paid little attention to possible interactions between job feedback and interpersonal feedback. It may be the case that when one source of feedback is lacking, the other source of feedback may serve a compensatory function. For example, knowledge workers responsible for abstract ideas and ambiguous projects are unlikely to receive direct feedback from the job itself, which may increase their reliance on interpersonal feedback. The direction of the interactive effects may depend on contextual factors. For instance, knowledge workers responsible for well-structured tasks—such as fixing bugs in computer programs—may be able to use feedback from the task itself regardless of feedback from other people. There is a need to develop a more systematic theoretical integration between the constructs of job feedback and interpersonal feedback. Curvilinear effects of job characteristics. The majority of job design theory and research has focused on linear, monotonic associations between job characteristics and attitudinal and behavioral outcomes. However, several studies have revealed curvilinear relationships between several job characteristics and outcomes. Much like vitamins, in high doses, “enriched” job characteristics may actually have detrimental effects (Warr, 2007). For instance, Xie and Johns (1995) found a U-shaped relationship between objective ratings of job complexity and self-reports of emotional exhaustion. Similar costs of highly complex or enlarged jobs 440

have been noted by other researchers (e.g., Campion & McClelland, 1993; Elsbach & Hargadon, 2006). Social psychologists have even begun to identify boundaries on autonomy, returning evidence that high levels of choice can lead to dissatisfaction, regret, and indecision (Chua & Iyengar, 2006; Schwartz, 2000). Such effects may be explained by theories of person– environment fit, which suggest that job characteristics are most likely to engender negative effects when they are supplied at levels that exceed employees’ preferences and abilities (e.g., Cable & Edwards, 2004; Ostroff & Judge, 2007). We hope to see researchers answer calls from Warr (2007) to address these types of curvilinear effects and explain their mechanisms and boundary conditions. Units of analysis for understanding the structures of work. Which work structures should we choose as our units of analysis? Should we retain a focus on jobs and tasks, shift to an emphasis on roles, or consider “middle-range” (Weick, 1974) or intermediate units? Such intermediate units may include activities or duties (Morgeson & Campion, 2002), projects (Grant, Little, & Phillips, 2006; Weick, 1999, 2003), and workdays (Elsbach & Hargadon, 2006). Researchers have yet to achieve consensus on the meaning and potential utility of these more molecular versus more global conceptualizations of work structures. Multidimensionality of job characteristics. Multidimensionality is an issue that warrants greater consideration in ongoing research. Researchers have increasingly recognized that specific job characteristics are multifaceted. For example, researchers have identified autonomy as varying in terms of decision making, scheduling, and methods dimensions (Breaugh, 1985; Morgeson & Humphrey, 2006; Wall et al., 1992), task significance as varying in terms of magnitude, scope, frequency, focus, beneficiary, and wellbeing domain dimensions (Grant, 2007), and interpersonal contact as varying in terms of duration, frequency, intensity or depth, directness or proximity, and breadth (Cordes & Dougherty,

Copyright American Psychological Association. Not for further distribution.

Work Matters

1993; Grant, 2007). It is puzzling that other job characteristics have not been seen as multidimensional when related literatures have highlighted multiple facets. For instance, psychologists and sociologists typically differentiate between emotional and instrumental forms of social support (e.g., Carver, Scheier, & Weintraub, 1989; House, 1981), or between more specific forms such as relieving emotional distress, giving advice, teaching skills, and providing material aid (e.g., Duffy, Ganster, & Pagon, 2002). Similarly, although job feedback and interpersonal feedback are seen by job design researchers as unidimensional characteristics (Hackman & Oldham, 1980; Morgeson & Humphrey, 2006), the feedback literature suggests that feedback can vary in terms of sign/ valence (positive vs. negative), focus of attention (learning, motivation, meta-task), and medium (verbal vs. written), specificity, credibility, and timeliness (Kluger & DeNisi, 1996), and the performance monitoring literature suggests that feedback can also vary in terms of purpose and perceived intensity (Holman, Chissick, & Totterdell, 2002). As a third example, although many job design researchers focus on the initiated versus received dimension of task interdependence (Kiggundu, 1981, 1983; Morgeson & Humphrey, 2006), researchers have highlighted a number of other dimensions of interdependence. Wong and Campion (1991) divided interdependence into three broad dimensions, each with multiple facets: task inputs (materials or supplies, information, product or service), task processes (input–output relationship, method, scheduling, supervision, sequencing, time sharing, support service, tools), and task outputs (goal, performance, quality). Others have distinguished between means or task interdependence and resource interdependence (Johnson & Johnson, 1999; Wageman, 1995) and pooled versus sequential versus reciprocal interdependence (Thompson, 1967). From a pragmatic standpoint, whether researchers study single or multiple dimensions of job characteristics may involve trade-offs between respondent burden and potential redundancy with comprehensiveness. From a theoretical standpoint, however, we believe that our understanding of job design can be enhanced by considering the multiple dimensions along which key job characteristics

may vary. The more dimensions that we can generate, the more opportunities we can identify for redesigning jobs. HOW SHOULD WE GET THERE? THEORY-BUILDING AND METHODS IN RESEARCH AND PRACTICE Thus far, we have focused our attention primarily on past, present, and possible future theoretical perspectives and empirical findings. In this section, we consider the different theory-building and methodological approaches that have been used in the past, and may help to advance the future, of job design research.

Theory-Building Approaches Researchers have taken different approaches to building job design theories. Some researchers have adopted a theory-focused approach (Weick, 1992), generating conceptual models with the goal of contributing to knowledge by filling gaps or resolving tensions in the literature. For example, Campion and Thayer (1985) developed their perspective on interdisciplinary job design to compare, reconcile, and synthesize different approaches recommended in organizational psychology, industrial engineering, physiology and ergonomics, and cognitive psychology. Similarly, Clegg and Spencer’s (2007) dynamic model of job design was guided by the observation that job design theorists had not yet integrated key insights that challenged several assumptions of the dominant existing models. In contrast, other researchers have adopted a problem-focused approach (Lawrence, 1992), recognizing problems or challenges in the field and then generating theories to solve these problems. For instance, Wrzesniewski and Dutton (2001) noticed that hospital cleaners were taking initiative to alter their tasks and relationships in ways that were not part of their job descriptions. They developed their theoretical perspective on job crafting to describe and explain these behaviors. Likewise, Grant and colleagues noticed in field research that many employees were doing jobs high in task significance but were left disconnected from seeing their impact on beneficiaries. This observation fueled the theoretical development and empirical test 441

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

of a model of relational job design that could describe, explain, and resolve this problem (Grant, 2007, 2008a; Grant et al., 2007). We see value in both theory-focused and problemfocused approaches to building job design theory. However, we expect that new theoretical perspectives on job design will be increasingly problem driven. Since no theory of social behavior can be simultaneously simple, general, and accurate (Thorngate, 1976), it is unlikely that any single theoretical model will be able to capture all of the important dimensions, antecedents, consequences, mechanisms, and boundary conditions of job design. Instead, we anticipate that researchers will generate novel “middlerange theories” (Weick, 1974) to describe, explain, and resolve job design challenges that emerge in practice. Such problem-driven approaches will require researchers to pay close attention to context (Johns, 2006) to capture the organizational, occupational, social, environmental, and technological opportunities and constraints that affect how jobs are designed, enacted, and experienced.

Methodological Approaches The job design literature is an exemplar in I/O psychology and organizational behavior research for its methodological diversity. In many cases, job design researchers have followed advice from methodologists to allow their research questions to guide their choices of methods (McGrath, 1981), which results in excellent fit between the theoretical question being posed and the suitability of the method for addressing it. When seeking to inductively identify the dimensions along which incumbents perceive job characteristics, researchers have used multidimensional scaling methods (e.g., Stone & Gueutal, 1985). When seeking to test complex models with multiple antecedent, mediating, moderating, and outcome variables, researchers have used surveys of broad cross-sections of jobs (e.g., Hackman & Oldham, 1976). When seeking to cumulate knowledge, researchers have used meta-analyses to draw broad conclusions about relationships among dimensions of job characteristics (Fried, 1991; Taber & Taylor, 1990) and between these dimensions and work related outcomes (Fried & Ferris, 1987; Morgeson & Humphrey, 2006). When seeking to determine how 442

job characteristics influence intraindividual changes in daily well-being, researchers have used experiencesampling studies to capture micro-level experiences (Sonnentag & Zijlstra, 2006; Totterdell, Wood, & Wall, 2006). When seeking to address questions of causality and internal validity that are difficult to control in the field, researchers have used laboratory experiments (e.g., Dodd & Ganster, 1996; Grant et al., 2007; Griffin et al., 1987; White & Mitchell, 1979). When seeking to achieve both internal and external validity, researchers have used field experiments and quasi-experiments, randomly assigning different groups of employees to controlled manipulation and treatment conditions (e.g., Grant, 2008a) or capitalizing on naturally occurring interventions that allow for the comparison of nonequivalent treatment groups (e.g., Campion & McClelland, 1991, 1993; Griffin, 1991; Lieberman, 1956; Morgeson & Campion, 2002; Morgeson et al., 2006; Oldham et al., 1995; Parker, 2003; Wall et al., 1986). However, our assessment is that the job design literature features too many cross-sectional or singlemethod, single-source survey studies in which it is difficult to rule out alternative explanations such as reverse causality, omitted variables, and selection threats. Such studies hamper not only the conclusions drawn by individual authors, but also the ability of the broader community of scholars to draw generalizable conclusions from meta-analyses: “garbage in, garbage out.” As is true in many areas of applied psychology and organizational behavior, the strongest study designs also tend to be the most invasive and time-sensitive designs. However, we believe that in the coming years, the job design literature is most likely to be advanced by four types of studies: field experiments and quasi-experiments, longitudinal survey and experience-sampling studies, qualitative studies, and multimethod and multisource designs. Next, we elaborate on the potential contributions of each methodological approach. Field experiments and quasi-experiments: Combining internal and external validity and supporting job redesign. Many researchers see field experiments and quasi-experiments as the gold standard for studying job design and redesign. Such experiments allow researchers to

Copyright American Psychological Association. Not for further distribution.

Work Matters

support causal inferences by ruling out alternative explanations, and also facilitate generalizability to the field settings that we are ultimately studying. Furthermore, these experiments make it possible for researchers to achieve applied goals of diagnosing, implementing, and evaluating job redesign interventions. As highlighted in our two introductory vignettes, in studying job redesign, researchers typically begin by conducting interviews with managers and observations or surveys of employees. These interviews and surveys make it possible to identify job characteristics that may be constraining and undermining outcomes such as satisfaction, motivation, performance, initiative, and health, as well as job changes that might help to enhance and enable these outcomes. Researchers then collect pretest data on perceptions of job characteristics, the outcomes of interest, and the mediators and moderators expected to carry and bound the effects of an intervention. Interventions are then designed and implemented by researchers or practitioners, dividing employees into different treatment and control groups, and researchers follow up with measures of perceived job characteristics, outcome variables, and mediators to examine and evaluate the effects of the intervention on each group. In this process, researchers can contribute to theory by achieving high levels of both internal and external validity, and also contribute to practice by helping to diagnose, implement, and evaluate job redesign interventions (for further advice, see Grant & Wall, 2009). Longitudinal survey and experience-sampling studies: Supporting internal and external validity when experiments are not possible or not ethical. The job design literature also features surprisingly few longitudinal studies. Most of the longitudinal studies in this literature take the form of long-term evaluations of the effects of field experiments and quasi-experiments (e.g., Campion & McClelland, 1993; Griffin, 1991; Lieberman, 1956; Morgeson & Campion, 2002; Morgeson et al., 2006; Parker, 2003; Wall et al., 1986). When it is not possible or ethical for researchers to conduct experiments, we recommend more longitudinal survey and experience-

sampling studies. Such studies allow for much stronger causal inferences than cross-sectional studies while maintaining greater fidelity to external validity than lab experiments allow. Experience-sampling studies may also help researchers capture daily and weekly effects of task-level experiences (e.g., Sonnentag & Zijlstra, 2006; Totterdell et al., 2006). Qualitative studies: Identifying new job characteristics and mechanisms. In addition, we hope to see more qualitative studies in the job design literature. Job design researchers, being trained primarily in I/O psychology and organizational behavior, have predominantly used quantitative methods to deductively test hypotheses. However, Barley and Kunda (2001) called for more detailed, in-depth studies of work to enrich our understanding of how work is changing in its methods and meaning. Accordingly, we believe that qualitative methods will help researchers to inductively build theory about new job characteristics and mechanisms. For example, qualitative studies have facilitated the discovery of the importance of informal social interaction in job experiences (Roy, 1959), the phenomenon of job crafting (Wrzesniewski & Dutton, 2001), and new explanations for how job control facilitates performance (Wall et al., 1992). Along these lines, although they have received little attention in the job design literature, organizational scholars have recently conducted a number of qualitative studies that have implications for issues of interest to job design researchers. For example, researchers have investigated the work conditions that enable psychological engagement (Kahn, 1990), the strategies that managers use to help employees doing “dirty work” cope with and counter occupational stigma (Ashforth, Kreiner, Clark, & Fugate, 2007), how medical residents resolve work-identity violations when they find that their actions do not match their identities (Pratt, Rockmann, & Kaufmann, 2006), and how managers, doctors, police officers, and addiction counselors express compassion when their tasks require them to harm others in the interest of a greater good (Margolis & Molinsky, 2008). We need more qualitative 443

Grant, Fried, and Juillerat

Copyright American Psychological Association. Not for further distribution.

research of this kind, using a combination of case study, interview, and observational–ethnographic methods, to identify new job characteristics and fresh mechanisms through which these job characteristics may influence employees’ attitudes and behaviors. Multimethod, multisource designs: Triangulating results. Historically, the Job Diagnostic Survey (Hackman & Oldham, 1975, 1980) and the Job Characteristics Inventory (Sims et al., 1976) have been the two most popular scales for measuring job design (for a comparison, see Fried, 1991). Recently, Morgeson and Humphrey (2006) developed the more comprehensive Work Design Questionnaire to allow researchers to measure a much broader set of task, knowledge, physical, and social characteristics of work. These scales rely primarily on Likert-type scales in which respondents indicate the extent of agreement versus disagreement with statements about a job. Most often, respondents are providing self-reports on their own jobs, but some researchers have supplemented these self-reports with ratings from observers such as coworkers, supervisors, and spouses, as well as with independent job classification data from the Dictionary of Occupational Titles or the O*NET (e.g., Morgeson & Humphrey, 2006; Xie & Johns, 1995). Ideally, rather than relying primarily on a single approach, researchers will use multiple methods and sources to strengthen their findings and interpretations. Incorporating sophisticated research designs that combine quantitative and qualitative data, field and laboratory studies, and measures of psychological, behavioral and physiological outcomes obviously requires substantial investments of time, resources, and energy. However, as Campbell and Fiske (1959) articulated, the internal and external validity of our conclusions is ultimately dependent on our ability to triangulate results across different methods and sources of data. CONCLUSION Job design is a topic that continues to fascinate (and sometimes frustrate) both scholars and practitioners. Given its theoretical and practical importance, we are 444

confident that research on job design will continue to flourish in the coming decades and centuries. We hope that in addition to dutifully testing existing theories, researchers will keep their eyes open to new phenomena that help us gain a deeper understanding of job design. As Einstein once quipped, “If we knew what we were doing, it wouldn’t be called research, would it?”

References Abrahamson, E. (1996). Management fashion. Academy of Management Review, 21, 254–285. Ambrose, M. L., & Kulik, C. T. (1999). Old friends, new faces: Motivation in the 1990s. Journal of Management, 25, 231–292. Ariely, D., Kamenica, E., & Prelec, D. (2008). Man’s search for meaning: The case of Legos. Journal of Economic Behavior and Organization, 67, 671–677. Ashford, S. J., Blatt, R., & VandeWalle, D. (2003). Reflections on the looking glass: A review of research on feedback-seeking behavior in organizations. Journal of Management, 29, 769–799. Ashforth, B. E., & Kreiner, G. E. (1999). ‘How can you do it?’: Dirty work and the challenge of constructing a positive identity. Academy of Management Review, 24, 413–434. Ashforth, B. E., Kreiner, G. E., Clark, M. A., & Fugate, M. (2007). Normalizing dirty work: Managerial tactics for countering occupational taint. Academy of Management Journal, 50, 149–174. Axtell, C. M., & Parker, S. K. (2003). Promoting role breadth self-efficacy through involvement, work design, and training. Human Relations, 56, 112–131. Babbage, C. (1835). On the economy of machinery and manufacturing. London: Knight. Bakker, A. B., & Demerouti, E. (2007). The job demands– resources model: State of the art. Journal of Managerial Psychology, 22, 309–328. Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3, 193–209. Barley, S. R., & Kunda, G. (2001). Bringing work back in. Organization Science, 12, 76–95. Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A metaanalysis. Personnel Psychology, 44, 1–26. Black, J. S., & Ashford, S. J. (1995). Fitting in or making jobs fit: Factors affecting mode of adjustment for new hires. Human Relations, 48, 421–437. Blau, G., & Katerberg, R. (1982). Toward enhancing research with the social information processing

Work Matters

approach to job design. Academy of Management Review, 7, 543–550. Bond, F. W., Flaxman, P. E., & Bunce, D. (2008). The influence of psychological flexibility on work redesign: Mediated moderation of a work reorganization intervention. Journal of Applied Psychology, 93, 645–654. Brass, D. J. (1981). Structural relationships, job characteristics, and worker satisfaction and performance. Administrative Science Quarterly, 26, 331–348.

Copyright American Psychological Association. Not for further distribution.

Breaugh, J. A. (1985). The measurement of work autonomy. Human Relations, 38, 551–570. Breaugh, J. A., & Becker, A. S. (1987). Further examination of the work autonomy scales: Three studies. Human Relations, 40, 381–399. Buss, D. M. (1987). Selection, evocation, and manipulation. Journal of Personality and Social Psychology, 53, 1214–1221.

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56, 267–283. Chirkov, V., Ryan, R. M., Kim, Y., & Kaplan, U. (2003). Differentiating autonomy from individualism and independence: A self-determination theory perspective on internalization of cultural orientations and well-being. Journal of Personality and Social Psychology, 84, 97–109. Chua, R. Y.-J., & Iyengar, S. S. (2006). Empowerment through choice? A critical analysis of the effects of choice in organizations. In B. Staw & R. M. Kramer (Eds.), Research in Organizational Behavior, 27 (pp. 41–79). Greenwich, CT: Elsevier. Clegg, C., & Spencer, C. (2007). A circular and dynamic model of the process of job design. Journal of Occupational and Organizational Psychology, 80, 321–339.

Buss, D. M. (1995). Psychological sex differences: Origins through sexual selection. American Psychologist, 50, 164–168.

Conlon, D. E., Van Dyne, L., Milner, M., & Ng, K. Y. (2004). The effects of physical and social context on evaluations of captive, intensive service relationships. Academy of Management Journal, 47, 433–445.

Cable, D. M., & Edwards, J. R. (2004). Complementary and supplementary fit: A theoretical and empirical integration. Journal of Applied Psychology, 89, 822–834.

Cordery, J. L., & Sevastos, P. P. (1993). Responses to the original and revised Job Diagnostic Survey: Is education a factor in responses to negatively worded items? Journal of Applied Psychology, 78, 141–143.

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81–105.

Cordes, C. L., & Dougherty, T. W. (1993). A review and integration of research on job burnout. Academy of Management Review, 18, 621–656.

Campion, M. A. (1988). Interdisciplinary approaches to job design: A constructive replication with extensions. Journal of Applied Psychology, 73, 467–481.

Cummings, T. G. (1986). A concluding note: Future directions of sociotechnical theory and research. Journal of Applied Behavioral Science, 22, 355–360.

Campion, M. A., & McClelland, C. L. (1991). Interdisciplinary examination of the costs and benefits of enlarged jobs: A job design quasi-experiment. Journal of Applied Psychology, 76, 186–198.

Dean, J. W., & Snell, S. A. (1991). Integrated manufacturing and job design: Moderating effects of organizational inertia. Academy of Management Journal, 34, 776–804.

Campion, M. A., & McClelland, C. L. (1993). Follow-up and extension of the interdisciplinary costs and benefits of enlarged jobs. Journal of Applied Psychology, 78, 339–351.

Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demand–resources model of burnout. Journal of Applied Psychology, 86, 499–512.

Campion, M. A., Mumford, T. V., Morgeson, F. P., & Nahrgang, J. D. (2005). Work redesign: Eight obstacles and opportunities. Human Resource Management, 44, 367–390. Campion, M. A., & Stevens, M. J. (1991). Neglected questions in job design: How people design jobs, task-job predictability, and influence of training. Journal of Business and Psychology, 6, 169–191.

Dierdorff, E. C., & Morgeson, F. P. (2007). Consensus in work role requirements: The influence of discrete occupational context on role expectations. Journal of Applied Psychology, 92, 1228–1241. Dodd, N. G., & Ganster, D. C. (1996). The interactive effects of task variety, autonomy, and feedback on attitudes and performance. Journal of Organizational Behavior, 17, 329–347.

Campion, M. A., & Thayer, P. W. (1985). Development and field evaluation of an interdisciplinary measure of job design. Journal of Applied Psychology, 70, 29–43.

Drach-Zahavy, A. (2004). The proficiency trap: How to balance enriched job designs and the team’s need for support. Journal of Organizational Behavior, 25, 979–996.

Carlopio, J. R. (1996). Construct validity of a physical work environment satisfaction questionnaire. Journal of Occupational Health Psychology, 1, 330–344.

Duffy, M. K., Ganster, D. C., & Pagon, M. (2002). Social undermining in the workplace. Academy of Management Journal, 45, 331–351. 445

Grant, Fried, and Juillerat

Dutton, J. E., & Ashford, S. J. (1993). Selling issues to top management. Academy of Management Review, 18, 397–428. Eagly, A. H., & Wood, W. (1999). The origins of sex differences in human behavior: Evolved dispositions versus social roles. American Psychologist, 54, 408–423. Edwards, J. R., Scully, J. A., & Brtek, M. D. (1999). The measurement of work: Hierarchical representation of the Multimethod Job Design Questionnaire. Personnel Psychology, 52, 305–334.

Copyright American Psychological Association. Not for further distribution.

Edwards, J. R., Scully, J. A., & Brtek, M. D. (2000). The nature and outcomes of work: A replication and extension of interdisciplinary work-design research. Journal of Applied Psychology, 85, 860–868. Elsbach, K. D., & Hargadon, A. B. (2006). Enhancing creativity through “mindless” work: A framework of workday design. Organization Science, 17, 470–483. Fine, S. A. (1955). A structure of worker functions. Personnel and Guidance Journal, 34, 66–73. Fortunato, V. J., & Stone-Romero, E. F. (2001). Positive affectivity as a moderator of the objective-task characteristics/perceived-task characteristics relationship. Journal of Applied Social Psychology, 31, 1248–1278. Fost, D. (2008, February 20). They’re working on their own, just side by side. New York Times. Retrieved February 27, 2008, from http://www.nytimes.com Frese, M., & Fay, D. (2001). Personal initiative: An active performance concept for work in the 21st century. In B. M. Staw & R. I. Sutton (Eds.), Research in organizational behavior, 23 (pp. 133–187).

Fried, Y., Levi, A. S., & Laurence, G. (2007). The Job Characteristics Model and LMX-MMX leadership. In J. B. Graen (Ed.), LMX leadership: The series, 5 (pp. 157–197). Charlotte, NC: Information Age Publishing. Fried, Y., Levi, A. S., & Laurence, G. (2008). Motivation and job design in the new world of work. In C. Cooper & C. Cartwright (Eds.), The Oxford handbook of personnel psychology. Oxford, England: Oxford University Press. Fried, Y., Melamed, S., & Ben-David, H. A. (2002). The joint effects of noise, job complexity, and gender on employee sickness absence: An exploratory study across 21 organizations—The CORDIS study. Journal of Occupational and Organizational Psychology, 75, 131–144. Fried, Y., Slowik, L. H., Ben-David, H. A., & Tiegs, R. B. (2001). Exploring the relationship between workspace density and employee attitudinal reactions: An integrative model. Journal of Occupational and Organizational Psychology, 74, 359–372. Gelfand, M. J., Erez, M., & Aycan, Z. (2007). Crosscultural organizational behavior. Annual Review of Psychology, 58, 479–514. Graen, G. B. (1976). Role-making processes within complex organizations. In M. D. Dunnette (Ed.), Handbook of industrial/organizational psychology (pp. 1201–1245). Chicago: Rand McNally. Grant, A. M. (2007). Relational job design and the motivation to make a prosocial difference. Academy of Management Review, 32, 393–417.

Frese, M., Garst, H., & Fay, D. (2007). Making things happen: Reciprocal relationships between work characteristics and personal initiative in a four-wave longitudinal structural equation model. Journal of Applied Psychology, 92, 1084–1102.

Grant, A. M. (2008a). The significance of task significance: Job performance effects, relational mechanisms, and boundary conditions. Journal of Applied Psychology, 93, 108–124.

Frese, M., Kring, W., Soose, A., & Zempel, J. (1996). Personal initiative at work: Differences between East and West Germany. Academy of Management Journal, 39, 37–63.

Grant, A. M. (2008b). Employees without a cause: The motivational effects of prosocial impact in public service. International Public Management Journal, 11, 48–66.

Fried, Y. (1990). Workspace characteristics, behavioral interferences, and screening ability as joint predictors of employee reactions: An examination of the intensification approach. Journal of Organizational Behavior, 11, 267–280. Fried, Y. (1991). Meta-analytic comparison of the job diagnostic survey and job characteristics inventory as correlates of work satisfaction and performance. Journal of Applied Psychology, 76, 690–697. Fried, Y., & Ferris, G. R. (1987). The validity of the job characteristics model: A review and meta-analysis. Personnel Psychology, 40, 287–322. Fried, Y., Grant, A. M., Levi, A. S., Hadani, M., & Slowik, L. H. (2007). Job design in temporal context: A career dynamics perspective. Journal of Organizational Behavior, 28, 911–927. 446

Grant, A. M. (2008c). Does intrinsic motivation fuel the prosocial fire? Motivational synergy in predicting persistence, performance, and productivity. Journal of Applied Psychology, 93, 48–58. Grant, A. M., & Ashford, S. J. (2008). The dynamics of proactivity at work. Research in Organizational Behavior, 28, 3–34. Grant, A. M., & Campbell, E. M. (2007). Doing good, doing harm, being well and burning out: The interactions of perceived prosocial and antisocial impact in service work. Journal of Occupational and Organizational Psychology, 80, 665–691. Grant, A. M., Campbell, E. M., Chen, G., Cottone, K., Lapedis, D., & Lee, K. (2007). Impact and the art of motivation maintenance: The effects of contact with

Work Matters

beneficiaries on persistence behavior. Organizational Behavior and Human Decision Processes, 103, 53–67. Grant, A. M., Little, B. R., & Phillips, S. D. (2006). Personal projects and organizational lives: When personal projects are not merely personal. In B. R. Little, K. Salmela-Aro, & S. D. Phillips (Eds.), Personal project pursuit: Goals, action, and human flourishing (pp. 221–246). Mahwah, NJ: Erlbaum.

Copyright American Psychological Association. Not for further distribution.

Grant, A. M., & Parker, S. K. (2009). Redesigning work design theories: The rise of relational and proactive perspectives. Academy of Management Annals, 3, 317–375. Grant, A. M., & Wall, T. D. (2009). The neglected science and art of quasi-experimentation: Why-to, when-to, and how-to advice for organizational researchers. Organizational Research Methods 12, 653–686.

Hambrick, D. C., Finkelstein, S., & Mooney, A. C. (2005). Executive job demands: New insights for explaining strategic decisions and leader behaviors. Academy of Management Review, 30, 472–481. Heath, C. (1999). On the social psychology of agency relationships: Lay theories of motivation overemphasize extrinsic incentives. Organizational Behavior and Human Decision Processes, 78, 25–62. Heath, C., Larrick, R. P., & Klayman, J. (1998). Cognitive repairs: How organizations compensate for the shortcoming of individual learners. Research in Organizational Behavior, 20, 1–37. Herzberg, F. (1966). Work and the nature of man. Cleveland, OH: World. Herzberg, F., Mausner, B., & Snyderman, B. B. (1967). The motivation to work (2nd Ed.). New York: Wiley.

Griffin, M. A., Neal, A., & Parker, S. K. (2007). A new model of work role performance: Positive behavior in uncertain and interdependent contexts. Academy of Management Journal, 50, 327–347.

Hofmann, D. A., & Stetzer, A. (1998). The role of safety climate and communication in accident interpretation: Implications for learning from negative events. Academy of Management Journal, 41, 644–657.

Griffin, R. W. (1983). Objective and social sources of information in task redesign: A field experiment. Administrative Science Quarterly, 28, 184–200.

Holland, J. L. (1996). Exploring careers with a typology: What we have learned and some new directions. American Psychologist, 51, 397–406.

Griffin, R. W. (1987). Toward an integrated theory of task design. Research in Organizational Behavior, 9, 79–120.

Holman, D., Chissick, C., & Totterdell, P. (2002). The effects of performance monitoring on emotional labor and well-being in call centers. Motivation and Emotion, 26, 57–81.

Griffin, R. W. (1991). Effects of work redesign on employee perceptions, attitudes, and behaviors: A long-term investigation. Academy of Management Journal, 34, 425–435. Griffin, R. W., Bateman, T. S., Wayne, S. J., & Head, T. C. (1987). Objective and social factors as determinants of task perceptions and responses: An integrated perspective and empirical investigation. Academy of Management Journal, 30, 501–523. Gustafson, S. B., & Mumford, M. D. (1995). Personal style and person environment fit: A pattern approach. Journal of Vocational Behavior, 46, 163–188. Hackman, J. R., & Lawler, E. E. (1971). Employee reactions to job characteristics. Journal of Applied Psychology, 55, 259–286. Hackman, J. R., & Oldham, G. R. (1975). Development of the Job Diagnostic Survey. Journal of Applied Psychology, 60, 159–170. Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16, 250–279.

Holman, D. J., Clegg, C., & Waterson, P. (2002). Navigating the territory of job design. Applied Ergonomics, 33, 197–205. Holman, D. J., & Wall, T. D. (2002). Work characteristics, learning-related outcomes, and strain: A test of competing direct effects, mediated, and moderated models. Journal of Occupational Health Psychology, 7, 283–301. House, J. S. (1981). Work, stress, and social support. Reading, MA: Addison-Wesley. Hsueh, Y. (2002). The Hawthorne experiments and the introduction of Jean Piaget in American industrial psychology, 1929–1932. History of Psychology, 5, 163–189. Huang, X., & Van De Vliert, E. (2003). Where intrinsic job satisfaction fails to work: National moderators of intrinsic motivation. Journal of Organizational Behavior, 24, 159–179.

Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, MA: Addison-Wesley.

Humphrey, S. E., Nahrgang, J. D., & Morgeson, F. P. (2007). Integrating motivational, social, and contextual work design features: A meta-analytic summary and theoretical extension of the work design literature. Journal of Applied Psychology, 92, 1332–1356.

Halbesleben, J. R. B., & Buckley, R. M. (2004). Burnout in organizational life. Journal of Management, 30, 859–879.

Hunnicutt, B. K. (1988). Work, leisure, and labor supply: An analysis of the 1980 U.S. census data. International Review of Modern Sociology, 18, 31–55. 447

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

Ilgen, D., & Hollenbeck, J. (1991). The structure of work: Job design and roles. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2, pp. 165–208). Palo Alto, CA: Consulting Psychologists Press.

Kirkman, B. L., Rosen, B., Tesluk, P. E., & Gibson, C. B. (2004). The impact of team empowerment on virtual team performance: The moderating role of face-to-face interaction. Academy of Management Journal, 47, 175–192.

Jackson, P. R., Wall, T. D., Martin, R., & Davids, K. (1993). New measures of job control, cognitive demand, and production responsibility. Journal of Applied Psychology, 78, 753–762.

Kluger, A. N., & DeNisi, A. (1996). Effects of feedback intervention on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin, 119, 254–284.

Janz, B. D. & Prasarnphanich, P. (2003). Understanding the antecedents of effective knowledge management: The importance of knowledge-centered culture. Decision Sciences, 34, 351–384.

Kohn, M. L., & Schooler, C. (1982). Job conditions and personality: A longitudinal assessment of their reciprocal effects. American Journal of Sociology, 87, 1257–1286.

Jermier, J. M., Gaines, J., & McIntosh, N. J. (1989). Reactions to physically dangerous work: A conceptual and empirical analysis. Journal of Organizational Behavior, 10, 15–33.

Kulik, C. T., Oldham, G. R., & Langner, P. H. (1988). Measurement of job characteristics: Comparison of the original and the revised Job Diagnostic Survey. Journal of Applied Psychology, 73, 462–466.

Jex, S. M., & Spector, P. E. (1989). The generalizability of Social Information Processing to organizational settings: A summary of two field experiments. Perceptual and Motor Skills, 69, 883–893.

Langfred, C. W. (2007). The downside of self-management: A longitudinal study of the effects of conflict on trust, autonomy, and task interdependence in self-managing teams. Academy of Management Journal, 50, 885–900.

Johns, G. (2006). The essential impact of context on organizational behavior. Academy of Management Review, 31, 386–408.

Langfred, C. W., & Moye, N. A. (2004). Effects of task autonomy on performance: An extended model considering motivational, informational, and structural mechanisms. Journal of Applied Psychology, 89, 934–945.

Johns, G., Xie, J. L., & Fang, Y. (1992). Mediating and moderating effects in job design. Journal of Management, 18, 657–676. Johnson, D. W., & Johnson, R. T. (1989). Cooperation and competition: Theory and research. Edina, MN: Interaction Book Company. Judge, T. A., Bono, J. E., & Locke, E. A. (2000). Personality and job satisfaction: The mediating role of job characteristics. Journal of Applied Psychology, 85, 237–249. Kahn, W. A. (1990). Psychological conditions of personal engagement and disengagement at work. Academy of Management Journal, 33, 692–724. Kanfer, R., & Ackerman, P. L. (2004). Aging, adult development, and work motivation. Academy of Management Review, 29, 440–458. Karasek, R. A. (1979). Job demands, job decision latitude, and mental strain: Implications for job redesign. Administration Science Quarterly, 24, 285–307.

Latham, G. P., & Pinder, C. C. (2005). Work motivation theory and research at the dawn of the twenty-first century. Annual Review of Psychology, 56, 495–516. Lawrence, P. R. (1992). The challenge of problem-oriented research. Journal of Management Inquiry, 1, 139–142. Leung, K. (2001). Different carrots for different rabbits: Effects of individualism–collectivism and power distance on work motivation. In M. Erez, U. Kleinbeck, & H. Thierry (Eds.), Work motivation in the context of a globalizing economy (pp. 329–339). Mahwah, NJ: Erlbaum. Liden, R. C., Wayne, S. J., & Sparrowe, R. T. (2000). An examination of the mediating role of psychological empowerment on the relations between the job, interpersonal relationships, and work outcomes. Journal of Applied Psychology, 85, 407–416.

Karasek, R. A., & Theorell, T. (1990). Healthy work. New York: Basic Books.

Lieberman, S. (1956). The effects of changes in roles on the attitudes of role occupants. Human Relations, 9, 385–402.

Kiggundu, M. N. (1981). Task interdependence and the theory of job design. Academy of Management Review, 6, 499–508.

Likert, R. (1961). New patterns of management. New York: McGraw-Hill.

Kiggundu, M. N. (1983). Task interdependence and job design: Test of a theory. Organizational Behavior and Human Performance, 31, 145–172. Kilduff, M., & Regan, D. T. (1988). What people say and what they do: The differential effects of informational cues and task design. Organizational Behavior and Human Decision Processes, 41, 83–97. 448

Likert, R. (1967). The human organization. New York: McGraw-Hill. Lindsley, D. H., Brass, D. J., & Thomas, J. B. (1995). Efficacy-performance spirals: A multilevel perspective. Academy of Management Review, 20, 645–678. Lippa, R. (1998). Gender-related individual differences and the structure of vocational interests: The impor-

Work Matters

tance of the people–things dimension. Journal of Personality and Social Psychology, 74, 996–1009. Little, B. R. (1972). Psychological man as scientist, humanist and specialist. Journal of Experimental Research in Personality, 6, 95–118. Little. B. R. (1989). Personal projects analysis: Trivial pursuits, magnificent obsessions and the search for coherence. In D. M. Buss & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 15–31). New York: Springer-Verlag.

Copyright American Psychological Association. Not for further distribution.

Locke, E. A., & Henne, D. (1986) Work motivation theories. In C. Cooper & I. Robertson (Eds.), Internal review of industrial and organizational psychology (pp. 1–35). Chichester, England: Wiley. Loher, B. T., Noe, R. A., Moeller, N. L., & Fitzgerald, M. P. (1985). A meta-analysis of the relation of job characteristics to job satisfaction. Journal of Applied Psychology, 70, 280–289. Margolis, J. D., & Molinsky, A. L. (2008). Navigating the bind of necessary evils: Psychological engagement and the production of interpersonally sensitive behavior. Academy of Management Journal, 51, 847–872. Marks, G., & Miller, N. (1987). Ten years of research on the false-consensus effect: An empirical and theoretical review. Psychological Bulletin, 102, 72–90. Martin, R., & Wall, T. D. (1989). Attentional demand and cost responsibility as stressors in shopfloor jobs. Academy of Management Journal, 32, 69–86.

Mayo, E. (1945). The social problems of an industrial civilization. Cambridge, MA: Harvard University Press. McGrath, J. E. (1981). Dilemmatics: The study of research choices and dilemmas. American Behavioral Scientist, 25, 179–210. McGregor, D. (1960). The human side of the enterprise. New York: McGraw-Hill. Melamed, S., Fried, Y., & Froom, P. (2001). The interactive effect of chronic exposure to noise and job complexity on changes in blood pressure and job satisfaction: A longitudinal study of industrial employees. Journal of Occupational Health Psychology, 6, 182–195. Meyer, J. W., & Rowan, B. (1977). Institutional organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83, 340–363. Miner, J. B. (1984). The validity and usefulness of theories in an emerging organizational science. Academy of Management Review, 9, 296–306. Miner, J. B. (2003). The rated importance, scientific validity, and practical usefulness of organizational behavior theories: A quantitative review. Academy of Management Learning & Education, 2, 250–268. Mintzberg, H. (1973). The nature of managerial work. New York: Harper & Row. Molinsky, A., & Margolis, J. (2005). Necessary evils and interpersonal sensitivity in organizations. Academy of Management Review, 30, 245–268.

Mathieu, J. E., Hofmann, D. A., & Farr, J. L. (1993). Job perception—job satisfaction relations: An empirical comparison of three competing theories. Organizational Behavior and Human Decision Processes, 56, 370–387.

Morgeson, F. P., & Campion, M. A. (2002). Avoiding tradeoffs when redesigning work: Evidence from a longitudinal quasi-experiment. Personnel Psychology, 55, 589–612.

May, D. R., Oldham, G. R., & Rathert, C. (2005). Employee affective and behavioral reactions to the spatial density of physical work environments. Human Resource Management, 44, 21–33.

Morgeson, F. P., & Campion, M. A. (2003). Work design. In W. Borman, R. Klimoski, & D. Ilgen (Eds.), Handbook of psychology, volume twelve: Industrial and organizational psychology (pp. 423–452). New York: Wiley.

May, D. R., Reed, K., Schwoerer, C. E., & Potter, P. (2004). Ergonomic office design and aging: A quasiexperimental field study of employee reactions to an ergonomics intervention program. Journal of Occupational Health Psychology, 9, 123–135. May, D. R., & Schwoerer, C. E. (1994). Employee health by design: Using employee involvement teams in ergonomic job redesign. Personnel Psychology, 47, 861–876. May, D. R., Schwoerer, C. E., Reed, K., & Potter, P. (1997). Employee reactions to ergonomic job design: The moderating effects of health locus of control and selfefficacy. Journal of Occupational Health Psychology, 2, 11–24. Mayo, E. (1933). The human problems of an industrial civilization. New York: MacMillan.

Morgeson, F. P., Delaney-Klinger, K., & Hemingway, M. A. (2005). The importance of job autonomy, cognitive ability, and job-related skill for predicting role breadth and job performance. Journal of Applied Psychology, 90, 399–406. Morgeson, F. P., & Humphrey, S. E. (2006). The work design questionnaire (WDQ): Developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91, 1321–1339. Morgeson, F. P., & Humphrey, S. E. (2008). Job and team design: Toward a more integrative conceptualization of work design. In J. Martocchio (Ed.), Research in personnel and human resource management (Vol. 27, pp. 39–92). Bingley, England: Emerald Group Publishing. 449

Copyright American Psychological Association. Not for further distribution.

Grant, Fried, and Juillerat

Morgeson, F. P., Johnson, M. D., Campion, M. A., Medsker, G. J., & Mumford, T. V. (2006). Understanding reactions to job redesign: A quasiexperimental investigation of the moderating effects of organizational context on perceptions of performance behavior. Personnel Psychology, 59, 333–363.

Oldham, G. R., & Rotchford, N. L. (1983). Relationships between office characteristics and employee reactions: A study of the physical environment. Administrative Science Quarterly, 28, 542–556.

Morrison, E. W. (1994). Role definitions and organizational citizenship behavior: The importance of employee’s perspective. Academy of Management Journal, 37, 1543–1567.

Parker, S. K. (1998). Enhancing role breadth self-efficacy: The roles of job enrichment and other organizational interventions. Journal of Applied Psychology, 83, 835–852.

Ohly, S., Sonnentag, S., & Pluntke, F. (2006). Routinization, work characteristics and their relationships with creative and proactive behaviors. Journal of Organizational Behavior, 27, 257–279.

Parker, S. K. (2000). From passive to proactive motivation: The importance of flexible role orientations and role breadth self-efficacy. Applied Psychology: An International Review, 49, 447–469.

Oldham, G. R. (1988). Effects of changes in workspace partitions and spatial density on employee reactions: A quasi-experiment. Journal of Applied Psychology, 73, 253–258.

Parker, S. K. (2003). Longitudinal effects of lean production on employee outcomes and the mediating role of work characteristics. Journal of Applied Psychology, 88, 620–634.

Oldham, G. R. (1996). Job design. International Review of Industrial and Organizational Psychology, 11, 33–60.

Parker, S. K. (2007). “That is my job”: How employees’ role orientation affects their job performance. Human Relations, 60, 403–434.

Oldham, G. R., & Cummings, A. (1996). Employee creativity: Personal and contextual factors at work. Academy of Management Journal, 39, 607–634. Oldham, G. R., Cummings, A., Mischel, L. J., Schmidtke, J. M., & Zhou, J. (1995). Listen while you work? Quasi-experimental relations between personalstereo headset use and employee work responses. Journal of Applied Psychology, 80, 547–564. Oldham, G. R., & Hackman, J. R. (1981). Relationships between organizational structure and employee reactions: Comparing alternative frameworks. Administrative Science Quarterly, 26, 66–83. Oldham, G. R., & Hackman, J. R. (2005). How job characteristics theory happened. In K. G. Smith & M. A. Hitt (Eds.), The Oxford handbook of management theory: The process of theory development. Oxford, England: Oxford University Press. Oldham, G. R., Hackman, J. R., & Pearce, J. L. (1976). Conditions under which employees respond positively to enriched work. Journal of Applied Psychology, 61, 395–403. Oldham, G. R., Kulik, C. T., & Stepina, L. P. (1991). Physical environments and employee reactions: Effects of stimulus-screening skills and job complexity. Academy of Management Journal, 34, 929–938. Oldham, G. R., Kulik, C. T., Ambrose, M. L., Stepina, L. P., & Brand, J. F. (1986). Relations between job facet comparisons and employee reactions. Organizational Behavior and Human Decision Processes, 38, 28–47. Oldham, G. R., Kulik, C. T., Stepina, L. P., & Ambrose, M. L. (1986). Relations between situational factors and the comparative referents used by employees. Academy of Management Journal, 29, 599–608. 450

Ostroff, C., & Judge, T. A. (Eds.). (2007). Perspectives on organizational fit. Mahwah, NJ: Erlbaum.

Parker, S. K., & Ohly, S. (2008). Designing motivating jobs. In R. Kanfer, G. Chen, & R. Pritchard (Eds.), Work motivation: Past, present, and future. Parker, S. K., & Sprigg, C. A. 1999. Minimizing strain and maximizing learning: The role of job demands, job control, and proactive personality. Journal of Applied Psychology, 84, 925–939. Parker, S. K., & Wall, T. D. (1998). Job and work design: Organizing work to promote well-being and effectiveness. London: Sage. Parker, S. K., Wall, T. D., & Cordery, J. L. (2001). Future work design research and practice: Towards an elaborated model of work design. Journal of Occupational and Organizational Psychology, 74, 413–440. Parker, S. K., Wall, T. D., & Jackson, P. R. (1997). ‘That’s not my job’: Developing flexible employee work orientations. Academy of Management Journal, 40, 899–929. Parker, S. K., Williams, H. M., & Turner, N. (2006). Modeling the antecedents of proactive behavior at work. Journal of Applied Psychology, 91, 636–652. Pasmore, W. A., Francis, C., Haldeman, J., & Shani, A. (1982). Sociotechnical systems: A North American reflection on empirical studies of the seventies. Human Relations, 35, 1179–1204. Perlow, L., & Weeks, J. (2002). Who’s helping whom? Layers of culture and workplace behavior. Journal of Organizational Behavior, 23, 345–361. Pfeffer, J. (1994). Competitive advantage through people: Unleashing the power of the workforce. Boston: Harvard Business School Press. Piccolo, R. F., & Colquitt, J. A. (2006). Transformational leadership and job behaviors: The mediating role of

Work Matters

core job characteristics. Academy of Management Journal, 49, 327–340.

Schwartz, B. (2000). Self determination: The tyranny of freedom. American Psychologist, 55, 79–88.

Porter, L. W., & Lawler, E. E. (1968). Managerial attitudes and performance. Homewood, IL: Dorsey Press.

Shamir, B. (1992). Home: The perfect workplace? In S. Zedeck (Ed.), Work, families, and organizations (pp. 272–311). San Francisco: Jossey-Bass.

Pratt, M. G., Rockmann, K. W., & Kaufmann, J. B. (2006). Constructing professional identity: The role of work and identity learning cycles in the customization of identity among medical residents. Academy of Management Journal, 49, 235–262.

Copyright American Psychological Association. Not for further distribution.

Purvanova, R. K., Bono, J., & Dzieweczynski, J. (2006). Transformational leadership, job characteristics, and organizational citizenship performance. Human Performance, 19, 1–22. Robert, C., Probst, T. M., Martocchio, J. J., Drasgow, F., & Lawler, J. J. (2000). Empowerment and continuous improvement in the United States, Mexico, Poland, and India: Predicting fit on the basis of the dimension of power distance and individualism. Journal of Applied Psychology, 85, 643–658. Roberts, K. H., & Glick, W. (1981). The job characteristics approach to task design: A critical review. Journal of Applied Psychology, 66, 193–217. Roe, R. A., Zinovieva, I. L., Dienes, E., & Ten Horn, L. A. (2000). A comparison of work motivation in Bulgaria, Hungary, and the Netherlands: Test of a model. Applied Psychology: An International Review, 49, 658–687. Roethlisberger, F. J., & Dickson, W. J. (1939). Management and the worker. Boston: Harvard University Press. Ross, L., Greene, D., & House, P. (1977). The “false consensus effect”: An egocentric bias in social perception and attribution processes. Journal of Experimental Social Psychology, 13, 279–301. Rousseau, D. M. (1977). Technological differences in job characteristics, employee satisfaction, and motivation: A synthesis of job design research and sociotechnical systems theory. Organizational Behavior & Human Performance, 19, 18–42. Rousseau, D. M. (1982). Job perceptions when working with data, people and things. Journal of Occupational Psychology, 55, 43–52. Rousseau, D. M., & Fried, Y. (2001). Location, location, location: Contextualizing organizational behavior. Journal of Organizational Behavior, 22, 1–15. Rousseau, D. M., Ho, V. T., & Greenberg, J. (2006). I-deals: Idiosyncratic terms in employment relationships. Academy of Management Review, 31, 977–994. Roy, D. F. (1959). Banana time: Job satisfaction and informal interaction. Human Organization, 18, 158–168. Salancik, G. R., & Pfeffer, J. (1978). A social information processing approach to job attitudes and task design. Administrative Science Quarterly, 23, 224–253.

Shamir, B., & Salomon, I. (1985). Work-at-home and the quality of working life. Academy of Management Review, 10, 455–464. Shaw, J. B., & Weekley, J. A. (1981). The effects of socially provided task information on task perceptions, satisfaction and performance. Paper presented at the Southern Management Association Conference. Sims, H. P., Szilagyi, A. D., & Keller, R. T. (1976). The measurement of job characteristics. Academy of Management Journal, 19, 195–212. Sinha, K. K., & Van de Ven, A. H. (2005). Designing work within and between organizations. Organization Science, 16, 389–408. Smith, A. (1776). An inquiry into the nature and causes of the wealth of nations. London: W. Strahan and T. Cadell. Snir, R., & Harpaz, I. (2002). Work-leisure relations: Leisure orientation and the meaning of work. Journal of Leisure Research, 34, 178–201. Sonnentag, S., & Zijlstra, F. R. H. (2006). Job characteristics and off-job activities as predictors of need for recovery, well-being, and fatigue. Journal of Applied Psychology, 91, 330–350. Spector, P. E. (1985). Higher-order need strength as a moderator of the job scope–employee outcome relationship: A meta-analysis. Journal of Occupational Psychology, 58, 119–127. Spreitzer, G. M. (1996). Social structural characteristics of psychological empowerment. Academy of Management Journal, 39, 483–504. Staw, B. M., & Boettger, R. D. (1990). Task revision: A neglected form of work performance. Academy of Management Journal, 33, 534–559. Stone, E. F., & Gueutal, H. G. (1985). An empirical derivation of the dimensions along which characteristics of jobs are perceived. Academy of Management Journal, 28, 376–396. Taber, T. D., & Taylor, E. (1990). A review and evaluation of the psychometric properties of the Job Diagnostic Survey. Personnel Psychology, 43, 467–500. Taber, T. D., & Alliger, G. M. (1995). A task-level assessment of job satisfaction. Journal of Organizational Behavior, 16, 101–121. Taylor, F. W. (1911). The principles of scientific management. New York: W. W. Norton. Terkel, S. (1972). Working. New York: Pantheon. Theorell, T., & Karasek, R. A. (1996). Current issues relating to psychosocial job strain and cardiovascular 451

Grant, Fried, and Juillerat

disease research. Journal of Occupational Health Psychology, 1, 9–26. Thomas, J., & Griffin, R. (1983). The social information processing model of task design: A review of the literature. Academy of Management Review, 8, 672–682. Thompson, J. D. (1967). Organizations in action. New York: McGraw-Hill.

Copyright American Psychological Association. Not for further distribution.

Thorngate, W. (1976). Possible limits on a science of social behavior. In J. H. Strickland, F. E. Aboud, & K. J. Gergen (Eds.), Social psychology in transition (pp. 121–139). New York: Plenum. Tiegs, R. B., Tetrick, L. E., & Fried, Y. (1992). Growth need strength and context satisfactions as moderators of the relations of the job characteristics model. Journal of Management, 18, 575–593. Totterdell, P., Wood, S., & Wall, T. (2006). An intraindividual test of the demands-control model: A weekly diary study of psychological strain in portfolio workers. Journal of Occupational and Organizational Psychology, 79, 63–84. Trist, E. L. (1981). The sociotechnical perspective. In A. H. Van de Ven & W. F. Joyce (Eds.), Perspectives on organization design and behavior (pp. 19–75). New York: Wiley. Trist, E. L., & Bamforth, K. M. (1951). Some social and psychological consequences of the longwall method of coal-getting. Human Relations, 4, 3–38. Turner, A. N., & Lawrence, P. R. (1965). Industrial jobs and the worker. Boston: Harvard University Press. Unsworth, K. L., Wall, T. D., & Carter, A. (2005). Creative requirement: A neglected construct in the study of employee creativity? Group & Organization Management, 30, 541–560. van der Doef, M., & Maes, S. (1999). The job demandcontrol (-support) model and psychological wellbeing: A review of 20 years of empirical research. Work & Stress, 13, 87–114. Van Yperen, N. W., & Hagedoorn, M. (2003). Do high job demands increase intrinsic motivation or fatigue or both? The role of job control and job social support. Academy of Management Journal, 46, 339–348. Vroom, V. H. (1964). Work and motivation. New York: Wiley. Wageman, R. (1995). Interdependence and group effectiveness. Administrative Science Quarterly, 40, 145–181. Wagner-Tsukamoto, S. (2007). An institutional economics reconstruction of scientific management: On the lost theoretical logic of Taylorism. Academy of Management Review, 32, 105–117. Walker, C. R., & Guest, R. H. (1952). The man on the assembly line. Cambridge: Harvard University Press. 452

Wall, T. D., Corbett, J. M., Martin, R., Clegg, C. W., & Jackson, P. R. (1990). Advanced manufacturing technology, work design, and performance: A change study. Journal of Applied Psychology, 75, 691–697. Wall, T. D., & Jackson, P. R. (1995). Changes in manufacturing and shopfloor job design. In A. Howard (Ed.), The changing nature of work (pp. 164–211). San Francisco: Jossey-Bass. Wall, T. D., Jackson, P. R., & Davids, K. (1992). Operator work design and robotics system performance: A serendipitous field study. Journal of Applied Psychology, 77, 353–362. Wall, T. D., Jackson, P. R., & Mullarkey, S. (1995). Further evidence on some new measures of job control, cognitive demand and production responsibility. Journal of Organizational Behavior, 16, 31–455. Wall, T. D., Kemp, N. J., Jackson, P. R., & Clegg, C. W. (1986). Outcomes of autonomous workgroups: A long-term field experiment. Academy of Management Journal, 29, 280–304. Wall, T. D., & Martin, R. (1987). Job and work design. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (pp. 61–91). Oxford, England: Wiley. Warr, P. (2007). Work, happiness, and unhappiness. Mahwah, NJ: Erlbaum. Weick, K. E. (1974). Middle range theories of social systems. Behavioral Science, 19, 357–367. Weick, K. E. (1979). The social psychology of organizing, 2nd ed. Reading, MA: Addison-Wesley. Weick, K. E. (1992). Agenda setting in organizational behavior: A theory-focused approach. Journal of Management Inquiry, 1, 171–182. Weick, K. E. (1999). That’s moving: Theories that matter. Journal of Management Inquiry, 8, 134–142. Weick, K. E. (2003). How projects lose meaning: The dynamics of renewal. In R. Stablein & P. Frost (Eds.), Renewing research practice. Newbury Park, CA: Sage. Weick, K. E., & Roberts, K. (1993). Collective mind in organizations: Heedful interrelating on flight decks. Administrative Science Quarterly, 38, 357–381. Weick, K. E., Sutcliffe, K. M., & Obstfeld, D. (1999). Organizing for high reliability: Processes of collective mindfulness. In R. I. Sutton & B. M. Staw (Eds.), Research in organizational behavior, 21 (pp. 81–123). New York: Elsevier Science/JAI Press. White, S. E., & Mitchell, T. R. (1979). Job enrichment versus social cues: A comparison and competitive test. Journal of Applied Psychology, 64, 1–9. White, S. E., Mitchell, T. R., & Bell, C. H. (1977). Goal setting, evaluation apprehension, and social cues as

Work Matters

determinants of job performance and job satisfaction in a simulated organization. Journal of Applied Psychology, 62, 665–673. Wilensky, H. L. (1960). Work, careers and social integration. International Social Science Journal, 12, 543–560. Wong, C., & Campion, M. A. (1991). Development and test of a task level model of motivational job design. Journal of Applied Psychology, 76, 825–837.

Copyright American Psychological Association. Not for further distribution.

Wright, B. M., & Cordery, J. L. (1999). Production uncertainty as a contextual moderator of employee reactions to job design. Journal of Applied Psychology, 84, 456–463. Wrzesniewski, A., & Dutton, J. E. (2001). Crafting a job: Revisioning employees as active crafters of their work. Academy of Management Review, 26, 179–201. Wrzesniewski, A., McCauley, C. R., Rozin, P., Schwartz, B. (1997). Jobs, careers, and callings: People’s relations

to their work. Journal of Research in Personality, 31, 21–33. Xie, J. L., & Johns, G. (1995). Job scope and stress: Can job scope be too high? Academy of Management Journal, 38, 1288–1309. Zalesny, M. D., & Ford, J. K. (1990). Extending the social information processing perspective: New links to attitudes, behaviors, and perceptions. Organizational Behavior and Human Decision Processes, 47, 205–246. Zhou, J., Oldham, G. R., & Cummings, A. (1998). Employee reactions to the physical work environment: The role of childhood residential attributes. Journal of Applied Social Psychology, 28, 2213–2238. Zohar, D. & Luria, G. (2003). Organizational metascripts as a source of high reliability: The case of an Army armored Brigade. Journal of Organizational Behavior, 24, 837–859.

453

CHAPTER 14

WORKPLACE SAFETY AND ACCIDENTS: AN INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY PERSPECTIVE Copyright American Psychological Association. Not for further distribution.

Seth Kaplan and Lois E. Tetrick

The prevalence of workplace accidents is rather staggering. For instance, in each of the past 15 years, between 5,500 and 6,700 American workers have died from injuries suffered in the workplace (Bureau of Labor Statistics, 2007); that works out to an average of about 16 workers dying each day. Also alarming is the prevalence of nonfatal injuries and illness, with 4.1 million such incidents reported in the private sector in 2006, which translates into 4.4 incidents for every 100 workers (Bureau of Labor Statistics, 2007). The victims and witnesses of these incidents experience various hardships: They miss work; incur resultant financial difficulties; and are prone to depression, anxiety, alcohol and drug use, and increased family strife (Asmundson, Norton, Allerdings, Norton, & Larsen, 1998). Although notoriously difficult to estimate, the financial cost of workplace injuries is also enormous. The National Safety Council (NSC) estimated that in 2001 the total cost of work-related injuries in the United States was $132.1 billion and that organizations lost 130 million workdays because of injuries (NSC, 2002). That is equivalent to about $490 per capita or about $1,280 per household (NSC, 2002). More recently, the Liberty Mutual Institute for Research Safety (2008) estimated direct U.S. workers compensation costs for the 10 most disabling workplace injuries and illnesses in 2005 to be $48.3 billion. Furthermore, the Institute reported that these costs had risen since its last report in 1998, despite a drop in the number of overall incidents. This financial burden

to the U.S. economy outweighs those associated with diseases such as AIDS, Alzheimer’s, and arthritis and approaches those from cancer and heart disease (Leigh, Markowitz, Fahs, & Landrigan, 2000). Moreover, figures like this do not include other, less obvious organizational costs, such as decreases in workplace morale and tarnished organizational reputations. Industrial and organizational (I/O) psychology has a long history of examining and trying to improve workplace safety and reduce such accidents, dating back at least to Münsterberg’s (1913) efforts to reduce trolley car accidents by designing a simulation to select workers. Over the past decade or so, I/O psychology research on safety and accidents has mushroomed, as evidenced by the Society for Industrial and Organizational Psychology (SIOP) publishing the edited volume Health and Safety in Organizations: A Multilevel Perspective (Hofmann & Tetrick, 2003); the American Psychological Association (APA) publishing an edited volume titled The Psychology of Workplace Safety (Barling & Frone, 2004); the introduction of the Journal of Occupational and Health Psychology; and the formation of the Society for Occupational Health Psychology, which specifically addresses issues of safety and well-being in the workplace. The purpose of this chapter is to provide an overview of the I/O psychology research on occupational safety and accidents. To this end, the chapter unfolds as follows. First, we briefly provide descriptive information regarding the prevalence of various types of workplace accidents. Next, we outline

http://dx.doi.org/10.1037/12169-014 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

455

Kaplan and Tetrick

recent conceptual work attempting to describe and model the nature of safety performance. After this, we spend much of the chapter reviewing evidence on different antecedents of safety behavior and accidents. Finally, the chapter concludes with recommendations for future scholarly research in this area.

Copyright American Psychological Association. Not for further distribution.

PREVALENCE OF WORKPLACE ACCIDENTS The Occupational Safety and Health Act of 1970 established two federal bodies to oversee workplace health and safety. These bodies are the Occupational Safety and Health Administration (OSHA), which is housed in the Department of Labor, and the National Institute for Occupational Safety and Health, which is part of the Centers for Disease Control and Prevention in the U.S. Department of Health and Human Services. In addition to their other functions (e.g., investigating and documenting accidents), both bodies also collect comprehensive data on workplace safety and accidents. We review some key statistics below. Although the number of fatal and nonfatal accidents remains high (as noted earlier), there has been a significant downward trend in incidents over time. In particular, since the inception of these federal agencies, workplace fatalities have decreased by more than 60% and occupational injury and illness rates have fallen by 40% (Bureau of Labor Statistics, 2007). Although one cannot be certain of the causes of such trends, these agencies, through various mechanisms (e.g., research, investigating and assigning penalties for infractions, education) likely have played a significant role. It is not surprising that incident rates vary considerably across industries and occupations. In terms of fatalities, construction, transportation and warehousing, agriculture, fishing, forestry, and mining consistently rank among the most dangerous industries. Specifically, transportation and material moving consistently is the most dangerous in terms of the number of fatalities, although farming, fishing, and agriculture is often the highest in terms of rate of fatalities. As one might infer from these figures, transportation incidents (e.g., driving accidents, being struck by a vehicle) far outweigh other types of accidents in causing fatalities, with highway 456

incidents alone accounting for approximately 20% to 25% of all work-related deaths each year (Bureau of Labor Statistics, 2007). Workplace homicides were the second leading cause of fatality during the early 1990s but have decreased significantly beginning in about 1995. The other two major causes are falls and being struck by an object. The statistics regarding nonfatal injuries and illnesses largely parallel those for fatalities, with some additional industries associated with significant injury and illness but less so with fatality (e.g., retail and food service). In terms of the causes of nonfatal injuries and illnesses, overexertion, which includes pulling, pushing, carrying, and the like, consistently emerges as the number one type of incident, with falls reliably being the second most common (Liberty Mutual Research Institute for Safety, 2008). These statistics reflect accidents in the United States only. The worldwide figures are especially troubling. In 2005, the International Labour Office (ILO) issued a report providing an overview of the most recent estimates of worldwide occupational and work-related accidents and diseases (ILO, 2005). According to this report, approximately 2.2 million people die worldwide each year because of occupational injuries and illnesses. Of this total, fatal accidents are the third most common cause of death, accounting for approximately 350,000, or roughly 16%, of all fatalities. Communicable diseases, circulatory diseases, and cancer are other major causes. Of the fatal occupational injuries, about 26% occur in China, 22% occur in other Asian countries, 15% occur in sub-Saharan Africa, and 11% occur in India. Estimates from Western countries are far lower. Approximately 22,000 of these accident fatalities are working children. These figures do not take into account the approximately 270 million serious nonfatal occupational accidents that occur each year. Although these numbers are disheartening, research and regulatory efforts have the potential to greatly reduce their frequency and impact. Specifically, the ILO reports that if all of their member nations used the best accident prevention strategies and practices, about 300,000 of the 350,000 deaths and 200 million of the 270 million accidents could be prevented.

Workplace Safety and Accidents

Copyright American Psychological Association. Not for further distribution.

CONCEPTUALIZING THE CRITERION DOMAIN: ACCIDENTS AND SAFETY PERFORMANCE Multiple fields and disciplines in addition to I/O psychology study occupational safety, including public health, safety engineering, and human factors and ergonomics. Typically, safety studies in these other areas focus on specific occupational hazards (e.g., a given chemical or chemical spills in general) and/or type of injury or illness (e.g., chemical burns), often in a given job or occupation. A title such as “Safety Goggles: Are They Adequate to Prevent Eye Injuries Caused by Rotating Wire Brushes?” (Paul & Lewis, 2008) typifies the types of studies often conducted. Instead of concentrating on specific hazards or injury types, I/O psychology research tries to provide more general frameworks for conceptualizing safety behavior. In particular, a main contribution I/O psychology has brought to this field is conceptual clarity with regard to the criterion domain of safety behaviors and outcomes (e.g., Burke, Sarpy, Tesluk, & Smith-Crowe, 2002; Kaminski, 2001). Two specific advances with respect to conceptualizing the criterion domain are particularly noteworthy. First, I/O psychology research has highlighted and brought attention to the significant, but largely disregarded, distinction between safety outcomes (e.g., accidents, injuries) and safety performance. Traditionally, studies have treated outcomes, such as fatalities, injuries, or “near-misses” (Zohar, 2000) as synonymous with or proxies for workplace safety. However, outcomes are not synonymous with their antecedent behaviors (Campbell, 1990). Injuries, illnesses, and so on, are outcomes that only in part reflect safety behavior (i.e., performance), such as wearing the correct protective gear (Burke et al., 2002). This distinction has several significant implications. First, it emphasizes that outcomes, unlike behaviors, are not fully under the person’s control but also are partially a function of contextual and higher level influences (Campbell, 1990). In doing so, this distinction directs research efforts away from simply identifying those employees especially susceptible to accidents (i.e., “accident prone”) and

necessitates consideration of how contextual factors (e.g., safety climate, group norms) interact with individual difference variables in accident occurrence (e.g., Dekker, 2002; see Hofmann & Tetrick, 2003). In addition, this distinction highlights the recognition that there is more to safe work than a lack of accidents and, relatedly, that the “safest” workplaces could actually be those with higher accident frequencies (Christian, Bradley, Wallace, & Burke, 2009). For example, even though a mining organization may report more fatalities than a manufacturing plant, employees in the latter may be more lax about following safety procedures than those in the former. As a result, the manufacturing plant employees still likely will experience more safety accidents than they should because their safety behavior is suboptimal even if their safety outcomes are less severe than those of the miners. The second, and related, contribution in this area is I/O psychology research explicitly delineating the nature of safety performance. Burke et al. (2002) defined safety performance as “actions or behaviors that individuals exhibit in almost all jobs to promote the health and safety of workers, clients, the public, and the environment” (p. 432). Recent efforts have sought to identify the behaviors that constitute such performance. For instance, research suggests that safety performance includes both compliant behaviors (e.g., adhering to stated policies and procedures) and discretionary or participatory behaviors (e.g., making hazards known to others; Griffin & Neal, 2000; Hofmann, Morgesson, & Gerras, 2003). The most comprehensive conceptualization has been offered by Burke et al. (2002). On the basis of their review of relevant literature and of safety training lesson plans from the hazardous waste industry, they proposed a four-dimensional conceptualization of safety performance. The four dimensions are (a) using personal protective equipment, (b) engaging in work practices to reduce risk, (c) communicating health and safety information, and (d) exercising employee rights and responsibilities. The researchers created a measure to assess each of these dimensions, with items corresponding to specific behaviors (e.g., “Dons all personal protective equipment correctly” is a behavior and item for the first dimension). Empirical studies with hazardous 457

Copyright American Psychological Association. Not for further distribution.

Kaplan and Tetrick

waste workers at a nuclear clean-up site demonstrated the reliability of the measure. In addition, the results supported the proposed four-dimensional structure, although the factors were strongly correlated (rs were between .62 and .79), suggesting the existence of a higher order factor. This research provides a useful starting point in developing the understanding of safety performance. Burke et al.’s (2002) framework potentially allows one to compare results from individual studies and across industries and occupations. In addition, it affords the opportunity to systematically identify which antecedents and human resources practices most strongly relate to the different performance dimensions. For instance, to the degree that some of these dimensions are more knowledge based and others are more skill based, organizations could design training programs accordingly (e.g., in terms of the content and design of the training). At present, though, empirical efforts examining the nature and dimensionality of safety performance are still lacking, making it difficult to address propositions like those in the paragraph above. Future work adopting a more conceptual and theoretically grounded approach to safety performance would be useful, as would efforts to examine the generalizability of this four-dimensional conceptualization across industries, countries, and so on. With regard to other research on safety performance, also enlightening would be efforts to assess the relationships between safety performance and task- or in-role performance and contextual or extrarole performance (Borman & Motowidlo, 1993). Hofmann et al. (2003) generated six categories of safety citizenship behaviors: helping, voice, stewardship, whistle-blowing, civic virtue, and initiating safety-related change. However, our suspicion is that the degree to which workers conceive of specific behaviors as in-role versus extrarole is contingent on various factors. For instance, some behaviors, such as donning personal equipment, may seem like job requirements (e.g., for firefighters), whereas others, such as communicating health and safety information, may appear more discretionary and outside the formal bounds of one’s job duties. Also, there may be some behaviors that leaders regard as part of workers’ jobs but that the workers see as 458

more discretionary. Studies examining when workers see safety behaviors as within, versus beyond, their formal duties would be informative. ANTECEDENTS OF SAFETY PERFORMANCE AND ACCIDENTS I/O psychology research on safety and accidents historically has paralleled the broader safety literature in largely focusing on predictors of unsafe behaviors and outcomes. Accordingly, we devote much of this chapter to these antecedent variables. In the final section of the chapter, we also discuss the ways in which more recent I/O psychology research has become more sophisticated conceptually, especially in exploring the psychological mechanisms that mediate the predictor–criterion relationships (Griffin & Neal, 2000).

Individual Differences and Personnel Selection Traditionally, I/O psychology’s focus with respect to antecedents of workplace safety and accidents has been on examining personal characteristics, seeking to identify the individual differences that make one especially susceptible to accidents, or accident prone (e.g., Shaw & Sichel, 1971). This idea of accident proneness dates to an early study by Greenwood and Woods (1919), who found that accident occurrence in a British munitions factory was unevenly distributed among workers, with a relatively small proportion of workers accounting for most of the accidents. That I/O psychology has followed this individual difference approach is consistent with the field’s historical concentration on personnel selection and, correspondingly, on the identification and assessment of stable personal characteristics related to performance (e.g., Murphy, 1996). A major premise underlying selection is that job success is a function of the fit between the characteristics of the individual and those of the job (Murphy, 1996). Supportive of the notion that some people are especially likely to be involved in accidents (i.e., are accident prone), evidence does suggest that some individuals are particularly susceptible to repeated accidents (Visser, Pijl, Stolk, Neeleman, & Rosmalen, 2007), implying that organizations per-

Copyright American Psychological Association. Not for further distribution.

Workplace Safety and Accidents

haps should try to identify such individuals in the selection process. We discuss here three sets of individual differences—demographic variables, personality traits, and ability factors—that may be related to involvement in accidents, and we then discuss implications with regard to personnel selection. In terms of demographics, some groups tend to experience more accidents than others. In general, men, younger workers, less experienced workers, and those with less education and lower socioeconomic status are at a higher risk of workplace injury, illness, and fatalities (e.g., Leigh, 1986; Strong & Zimmerman, 2005). Certainly, much if not most of these effects are due to differences in job type. Specifically, these groups hold a disproportionately large number of physically demanding jobs as well as those with more noxious environmental characteristics (e.g., noise, vibration, lack of autonomy). However, there also exists a myriad of other psychological factors that partially may explain these results. For instance, because men are higher in sensation seeking and in risk taking than women (Byrnes, Miller, & Schafer, 1999), they may also take more chances or fewer precautions at work, thereby contributing to this differential. The same logic also may explain age differences (Westaby & Lowe, 2005). Future research exploring the predictive validity of these demographic characteristics within job type would be informative. For instance, results showing that men experience more accidents than do their female counterparts, performing the same job in the same organization, would provide indirect support for the role of these additional psychological factors. A second set of individual differences that relates to workplace safety is personality traits. (See also Vol. 2, chap. 5, this handbook.) I/O psychology researchers almost exclusively evaluate these personality characteristics using self-report measures, which then are correlated with indices of safety or accidents (Clarke & Robertson, 2008). There has been a considerable amount of research on this topic (e.g., Hansen, 1989), much of which derives from the idea of accident proneness, which was introduced above (Shaw & Sichel, 1971). Over the last 2 decades, most of this work has focused on the importance of the Big Five personality factors (i.e., traits of the five-factor model; Costa

& McCrae, 1985). According to proponents of the Big Five framework, individual differences in personality can be reasonably summarized by five broad traits, namely Neuroticism (or its inverse, Emotional Stability), Extraversion, Openness to Experience, Agreeableness, and Conscientiousness. Two recent meta-analyses have been conducted examining personality in relation to safety and accidents. Clarke and Robertson (2008) found that that all of the five factors, with the exception of Extraversion, had moderate to fairly strong relationships with accident involvement (which was based primarily on archival organizational records). Specifically, the corrected correlations for the five traits were Neuroticism = .30, Extraversion = .02, Openness to Experience = .50, Agreeableness = −.44, and Conscientiousness = −.31. In a comprehensive meta-analysis of several predictors of safety criteria, Christian et al. (2009) also examined some of the Big Five traits and occupational accidents and injuries, finding largely similar results: Extraversion = −.07, Neuroticism = .17, and Conscientiousness = −.26. They also looked separately at Conscientiousness and safety performance, finding a corrected correlation of .17. Other personality factors that predict accident occurrence include internal locus of control (r = .26; Christian et al., 2009), propensity for risk taking (see Turner, McClure, & Pirozzo, 2004, for a review), social maladjustment (r = .28; Hansen, 1989), positive affectivity (r = −.22; Iverson & Erwin, 1997), and negative affectivity (r = .20; Kaplan, Bradley, Luchman, & Haynes, 2009). To date, there is limited research exploring the psychological mechanisms underlying these relationships. Some relationships seem fairly intuitive. For instance, the positive effects of conscientiousness likely result from specific facets of this trait (e.g., dutifulness, self-discipline) promoting adherence to safety precautions. Similarly, a prevention focus, which is associated with an avoidance mindset, may increase attention to safety concerns during task activity (Wallace & Chen, 2006). However, the manner through which some of the other traits impact workplace safety is less clear. The reasons that being more agreeable or less neurotic would lead to fewer accidents, for example, are unknown. Future work examining these questions, such as 459

Copyright American Psychological Association. Not for further distribution.

Kaplan and Tetrick

research determining which facets of these broad traits predict safety criteria, would be useful. A third set of individual difference variables that relate to occupational safety are ability factors. (See also Vol. 2, chap. 4, this handbook.) With regard to safety, I/O psychology generally has not devoted as much attention to ability antecedents as to personality traits (but see Hansen, 1989). This relative lack of research may seem surprising, given that intelligence is consistently related to task performance (Schmidt & Hunter, 1998). This dearth of attention is consistent, however, with findings that accidents are more likely to stem from cognitive factors such as absentmindedness and lapses of attention than from a lack of complex reasoning skills (Lawton & Parker, 1998). In a recent I/O psychology study, Wallace and Vodanovich (2003) found direct support for the role of such factors, demonstrating that cognitive failure (e.g., distractibility, forgetfulness; Broadbent, Cooper, Fitzgerald, & Parkes, 1982) predicted unsafe behavior and occupational accidents, as did the interaction between cognitive failure and Conscientiousness. The above findings suggest that organizations may benefit from explicitly considering safety performance when selecting and placing workers. However, evidence indicates that most managers do not consider safety concerns when making workplace decisions. For instance, when Adams-Roy, Knap, and Barling (1995) asked groups of MBA and executive MBA students to list five issues they believed to be important for managerial performance, none of the 89 individuals listed occupational safety issues. This neglect of safety issues likely carries over to personnel selection decisions also, as leaders in most industries presumably focus instead on selecting for general task performance. However, such an approach seems shortsighted, given the significant individual, organizational, and societal consequences of workplace injuries and fatalities (Barling, Kelloway, & Iverson, 2003; Leigh et al., 2000). Further underscoring the importance of these individual difference variables is the generally neglected consideration that one “bad apple” can propagate unsafe practices among other employees. For instance, one lax manager could make decisions or behave in such a manner to undermine the safety climate of the group or organization (e.g., Zohar, 460

2000). Moreover, employees in the same group can influence each others’ safety behavior by establishing safety norms and influencing one another’s trust in management (Salancik & Pfeffer, 1978). Such considerations suggest that individual selection decisions can have a more widespread impact on safety outcomes than usually recognized. However, we also would note that there remain important and largely unexamined issues in terms of selecting for safety performance. First, although safety performance may be important in all jobs and at all job levels, the nature of such performance, and therefore the factors predictive of it, likely vary (Campbell, 1990). The sort of safety performance required of managers (e.g., making and enforcing safety-relevant policies), for instance, probably is not identical to that required of operative employees. More generally, contextual factors likely moderate several of these individual-differences–performance relationships. In fact, many of the credibility intervals from the meta-analyses cited above are large in magnitude and include zero, suggesting the presence of moderators. An important future direction for I/O psychology scholars therefore is documenting when and why these relationships do and do not generalize and which particular factors moderate these findings. Promising candidates to examine as moderators include factors such as nature of industry and occupation, strength of safety climate, amount of control or autonomy over safety-related behavior, and the degree to which the compensation system penalizes or rewards safe behavior. Worth noting is that examining the interplay of these higher level contextual factors and individual-level characteristics generally necessitates a multilevel approach. To date, only a few studies have explicitly called attention to the multilevel nature of workplace safety and to the corresponding conceptual and methodological considerations (e.g., Hofmann & Stetzer, 1996; Hofmann & Tetrick, 2003; Simard & Marchand, 1995; Wallace & Chen, 2006; Zohar & Luria, 2005). Also worthy of investigation are questions regarding the utility of selecting for safety performance, especially in terms of its relationship to task performance. Studies examining whether safety performance can be predicted more precisely than task performance, whether prediction of the two is addi-

Workplace Safety and Accidents

tive in nature, and whether some characteristics (personality traits, cognitive ability) predict one better than the other would all be useful and currently are needed. Related to this utility issue, we also note that, given the modest size of the correlations involving these individual differences, organizations also must consider the ethical and legal implications of selecting on the basis of these characteristics.

Copyright American Psychological Association. Not for further distribution.

Work Characteristics and Job Design A second factor related to workplace outcomes includes the characteristics and design of the job and of the workplace. (See also Vol. 2, chap. 1, this handbook.) Reflecting its roots in scientific management (Taylor, 1911) and the human relations movement (Mayo, 1933), I/O psychology has a long history of examining the impact of workplace features on outcomes such as motivation and productivity. Notably, much less I/O psychology research has focused explicitly on the role of work characteristics on safety. Several factors make summarizing the literature on work characteristics and safety and accidents somewhat challenging. First, because some researchers treat workplace features as aspects or indicators of safety climate (Brown & Holmes, 1986; Zohar, 2008) and others treat the two as conceptually separate (Neal & Griffin, 2004), one must carefully attend to each study’s measurement and analytical scheme. For the purposes of this chapter, we treat the two separately, discussing safety climate below. Similarly, researchers regularly use terms such as stressors, hazards, and risks in dissimilar ways, making cross-study comparisons challenging. Despite these limitations and with these caveats in mind, we review here the influence of two related categories on safety—those pertaining to task characteristics and the work environment and those related to the degree to which the work is demanding. With regard to the nature of job tasks (e.g., requiring sustained vigilance) and the physical working environment (e.g., excessive noise, presence of job hazards), studies consistently link these factors to increased accidents and injuries (Frone, 1998; Leigh, 1986). We follow the general trend in the literature in discussing these two factors together, as they tend to be highly correlated (see Melamed, Luz, Najenson, Jucha, & Green, 1989).

In an effort to quantify the influence of these factors, Melamed et al. (1989) developed an instrument to measure ergonomic stress level that assesses the combined exposure to 17 potential risk factors, including the presence of hazards, poor control and safety guards, physical effort, and vibration (Melamed et al., 1989). It is not surprising that higher scores on this index predict greater likelihood of occupational accidents and injuries (e.g., Melamed, Yekutieli, Froom, Kristal-Boneh, & Ribak, 1999). Studies suggest that the existence (or at least perception) of certain hazards (e.g., danger of getting trapped, use of faulty ladders) is especially predictive of safety outcomes (Melamed et al., 1999; Tomás, Meliá, & Oliver, 1999), whereas other factors such as poor lighting and vibration are relatively less important (Melamed et al., 1999), at least in certain jobs. The role of these factors and their high multicollinearity partially explain why workers in certain industries and occupations are at much greater risk than others (Bureau of Labor Statistics, 2007) and why a small percentage of the workforce experiences a large proportion of workplace accidents (see McKenna, 1983). In addition to impacting safety outcomes, features of the physical work environment also have a relatively large influence on safety performance. In an older study of 44 automotive plant departments, for instance, Keenan, Kerr, and Sherman (1951) found that a clean and comfortable working environment was the strongest predictor of good safety performance. More recently, Tomás et al. (1999), across three different samples, found that perceived job hazards positively predicted safety behavior and that such behavior in turn led to less perceived risk of accidents. In addition to such hazards, the degree to which the job is demanding and entails a higher workload is also potentially relevant to occupational safety. Much work in this domain implicitly or explicitly focuses on the role of management in promoting unsafe behavior by stressing productivity and efficiency at the expense of safety (Hofmann & Tetrick, 2003; Wright, 1986). It is surprising that studies examining the relationships between workload and related variables and safety criteria, however, have produced inconsistent results. For instance, 461

Copyright American Psychological Association. Not for further distribution.

Kaplan and Tetrick

Hofmann and Stetzer (1996) showed that perceived role overload significantly predicted unsafe work behavior (e.g., inappropriate tool use) in a chemical processing plant, but Parker, Axtell, and Turner (2001) failed to find support for this relationship in a sample of manufacturing workers. Studies investigating safety outcomes (e.g., accidents, injuries) have produced equally inconsistent findings. Although some studies have found significant relationships (Frone, 1998), others have not (Iverson & Erwin, 1997). Also worth noting is that role underload and boredom may be as deleterious, if not more so, than overload (Frone, 1998), perhaps because workers seek stimulation from other sources (Fisher, 1993). Thus, one cannot draw firm conclusions regarding the influence of workload or role demands on workplace safety. One explanation for these findings is that overload does not result in higher accident rates unless it fosters or co-occurs with more proximal predictors of accidents such as lack of control and perceived pressure (e.g., to take shortcuts; Wright, 1986). Supportive of this explanation, studies have revealed the benefits of autonomy, “safety control,” and participative decision making on safety behavior (e.g., Parker et al., 2001) and accident rates (see Zacharatos, Barling, & Iverson, 2005). As Simard and Marchand (2005) demonstrated, this greater control and participation encourages employees to be more proactive in their own safety, rather than just complying with stated policies and procedures. Another stream of studies has focused not on perceived overload or demands but on the number of hours worked. Findings from these investigations consistently reveal that working more hours over a given time period (e.g., per day or per week) predicts higher accidents rates (Lusa, Häkkänen, Luukkonen, & Viikari-Juntura, 2002). For instance, Leigh (1986) reported that individuals who worked overtime had a roughly 50% greater chance of experiencing an accident than those who did not work overtime. In addition, Lusa et al. (2002) found that those working 50 or more hours per week were substantially more likely than those working fewer than 50 hours to experience an accident. Rather than fostering unsafe behavior, as perceived demands and pressure do (Wright, 1986), working more hours likely results in 462

accidents through mechanisms such as fatigue and resultant cognitive failure (Reason, 1997). In sum, the effects of objective job conditions and characteristics (e.g., a large physical exertion component) on safety performance and accidents are not as strong as intuition might suggest. In fact, Christian et al. (2009), in their meta-analysis, found perceived risk to be a weaker correlate of these outcomes than were most of the personal characteristics (e.g., personality traits, job attitudes) they examined. Working in unsafe conditions seems to be most problematic when it co-occurs with a sense of perceived pressure and a lack of control. Another job characteristic that is consistently related to safety performance and accidents is the scheduling of work and, in particular, the effects of shift work.

Shift Work Several researchers, and most notably Folkard, Smith, and colleagues, have demonstrated the pernicious effects of shift work (i.e., nontraditional working arrangements that do not follow a standard 9-to-5 schedule) on safety and accidents (Folkard, Åkerstedt, & Macdonald, 2000; see also Smith, Sulksy, & Ormond, 2003, for a review). Shift work is prevalent in manufacturing organizations (e.g., factories), in “high-reliability” organizations (e.g., police, nuclear power plant control room operators), and also in the service sector (e.g., health care settings, restaurants). (See also chap. 17, this volume.) The noxious influence of shift work on workplace injuries and accidents is well documented (e.g., Rajaratnam & Arendt, 2001). For instance, Folkard et al. (2000) showed that accidents are more common during the night shift, even after controlling for other risk factors as did Smith, Folkard, and Poole (1994). Moreover Glazner (1996) found that firefighters were disproportionately more likely to get injured during nighttime versus daytime fires and also that the injuries were more severe at night. Furthermore, Williamson and Feyer (1995) found that the proportion of fatal accidents for workers at night was double that of the proportion during the day. Several factors underlie these findings. First, night work is inconsistent with human circadian rhythms (e.g., related to body temperature, the sleep–wake cycle; see Monk, Folkard, & Wedderburn, 1996). In

Copyright American Psychological Association. Not for further distribution.

Workplace Safety and Accidents

addition, having to work during the night attenuates the total amount of sleep one obtains (Folkard & Barton, 1993), especially given the disruptions (e.g., family obligations) that make sleeping during the day challenging. Owing to these factors, shift workers experience increased sleepiness, fatigue, and moodiness and decreased alertness and vigilance (Dinges, 1995), ultimately leading to more errors, accidents, and injuries. Compounding these factors, a reduced nighttime supervisory or managerial presence can further contribute to this increase in accident frequency and severity. Because shift work cannot be wholly eliminated, researchers have devoted considerable attention to minimizing its damaging effects (Czeisler, Moore-Ede, & Coleman, 1982; Monk, 2000). A seemingly intuitive solution is to have some individuals work night shifts only, thereby presumably allowing their circadian rhythms to readjust in alignment with their schedule. However, this solution is impractical both because few workers are willing to always work the night shift and because complete circadian adjustment is difficult even among night workers because of the difficulty of sleeping during the day (Folkard & Barton, 1993). Another approach is to determine the optimal shift and rotation schedule (Monk, 2000). However, this strategy also has drawbacks. There does not seem to be one perfect system, as each rotation system (e.g., rotating shifts every 2 weeks vs. every few days) has advantages and drawbacks (Monk et al., 1996). The optimal solution is to reduce night work (Knauth, 1993). When doing so is not feasible, the plusses and minuses of the different schedules must be considered in selecting the best option (Monk et al., 1996). Another method is to select workers who can best handle such work. For instance, hiring younger workers and those without family obligations to interfere with their daytime sleep are feasible options (given legal considerations obviously). In addition, measures such as the Circadian Type Inventory (Di Milia, Smith, & Folkard, 1984) could potentially be used to select individuals for shift work. Although none of these strategies will eliminate the accidents that shift work and night work promote, collectively their implementation could reduce the amount and

magnitude of such accidents. A more general construct that incorporates these and other practices reflecting management’s commitment to occupational safety is safety climate.

Safety Climate Safety climate centers on the policies, procedures, and practices in an organization relative to the importance and priority given to safety and reflects the shared perceptions of individuals within an organization or a specific organizational unit with respect to these policies, procedures, and practices (Zohar, in press; see also chap. 12, this volume). Because safety climate refers to policies and procedures relative to safety established by the organization and management, it is not surprising that management commitment to safety has been consistently identified as a major dimension of safety climate (Brown & Holmes, 1986; Coyle, Sleeman, & Adams, 1995; Dedobbeleer & Beland, 1991; Zohar, 1980). Within the organizational climate and the safety climate literatures, there has been a distinction between safety climate and perceptions of safety climate along the lines identified by James and Jones (1974), with safety climate being a group- or organizationallevel construct and perceptions of safety climate being an individual-level construct. To the extent that individuals within a group or organization agree, then one can aggregate the individuals’ perceptions to the group or organization level (Chan, 1998; Kozlowski & Klein, 2000; Zohar, in press). Several studies adopting a multilevel approach have demonstrated the importance of safety climate in safety performance and accidents. For instance, Hofmann and Stetzer (1996) documented high levels of within-team agreement in safety climate perceptions in a chemical processing plant. In turn, aggregated (i.e., team-level) safety climate predicted both unsafe behaviors and actual accidents. Neal and Griffin (2006) also found substantial within-unit agreement in safety climate ratings and further demonstrated that unit-level safety climate predicted accidents 3 years later. Three recent meta-analyses (Christian et al., 2009; Clarke, 2006; Nahrgang, Morgeson, & Hofmann, 2008) documented the paramount role that safety climate plays in safety outcomes. Corrected 463

Copyright American Psychological Association. Not for further distribution.

Kaplan and Tetrick

correlations between overall safety climate and safety behavior or performance were quite strong, ranging from .41 to .62 across the studies. The corrected correlations involving accidents and injury rates were weaker but still of considerable magnitude, ranging from −.22 to −.39. According to Neal and Griffin (2004), safety climate is hierarchically structured. At the broader level is general, or overall, safety climate, which reflects the degree to which individuals see safety as being valued in the workplace. These broader perceptions derive from judgments about more specific safety facets, such as safety-relevant human resource management (HRM) practices, the amount of risk or danger in the job, and supervisory support for safe behaviors, among others. Christian et al. (2009) adopted this framework in their meta-analysis and found that certain facets were especially strong predictors of safety outcomes (e.g., HRM practices) and others were relatively less strong predictors (e.g., perceived job risk). In addition to examining the level of climate dimensions, I/O psychology researchers also have investigated the strength of climate (Schneider, Salvaggio, & Subirats, 2002; Zohar, in press). Climate strength reflects the degree of agreement among members of a unit relative to (safety) climate, with greater agreement representing a stronger climate. The notion of climate strength has only recently been incorporated into the safety climate literature (Luria, 2008; Zohar & Tenne-Gazit, 2008). In Zohar and Luria (2004), safety climate strength was not related to injury rate although safety climate level was. However, Zohar and Luria (2005) found support for the strength of organizational-level climate affecting group-level climate strength, which was related to safety performance at the group level, although the effects for climate strength were smaller than the effects for the level of climate.

Leadership One of the main factors contributing to safety climate is leadership. (See also chap. 7, this volume.) Accordingly, recent meta-analytic evidence strongly supports the role of leadership in safety performance and accident occurrence (Christian et al., 2009; Nahrgang et al., 2008). In fact, in their comprehen464

sive meta-analysis of the predictors of safety criteria which included 90 studies and 1744 effect size, Christian et al. (2009) found that leadership and supervisory support were among the strongest predictors of these outcomes, especially with regard to safety performance. In particular, they reported a meta-analytic correlation of .24 between leadership and safety compliance behavior (e.g., following safety procedures) and of .35 between leadership and safety participation behavior (e.g., initiating safety changes). Furthermore, Nahrgang et al. (2008), in regression analyses, found that only leadership and job demands remained significant predictors when other organizational variables, including safety climate, were controlled for. The researchers suggested that safety climate may be related to leadership, as well as to some of the other variables included in the regression analyses. Empirical evidence supports this interpretation. Several findings indicate a relationship between leadership and safety climate, with safety climate then predicting safety performance and outcomes (Luria, 2008; Zohar, 2002). Although these studies suggest that leadership is an antecedent of safety climate, Hofmann et al. (2003) found that safety climate acted as a moderator of the relation between leadership and positive safety behaviors. This study suggests that leadership has a direct effect on safety performance. In addition, Zohar (2002) and Zohar and Luria (2005) provided evidence that leadership practices, especially the consistency of those practices, affect safety performance. In other work, Zohar and Luria (2003) focused on supervisory monitoring and rewarding of employees’ safety behavior as a means to improve safety behavior. Additional research supports the effectiveness of transformational leadership in creating a positive safety climate as well as potentially having direct effects on safety performance (e.g., Barling, Loughlin, & Kelloway, 2002; Mullen, 2005; Zohar & Tenne-Gazit, 2008). Another way leaders potentially may affect safety performance is by using compensation. Although it is perhaps intuitive that financial incentives will promote safe behavior and therefore fewer accidents, this approach has met with mixed success (Sinclair & Tetrick, 2004). McAfee and Winn (1989) concluded

Copyright American Psychological Association. Not for further distribution.

Workplace Safety and Accidents

in their review of the literature that incentives or feedback reduce accidents, yet Kaminski (2001) found that pay-for-performance plans were actually related to higher injury rates, and Pransky, Snyder, Dembe, and Himmelstein (1999) suggested that incentive systems can lead to under-reporting of injuries and accidents. One explanation for these mixed findings is that they reflect different components of the incentive programs. For example, some safety incentive programs link individual behavior to individual incentives and others link individual behavior to workgroup performance. The latter designs may generate social pressures not to report minor injuries (Gallagher & Myers, 1996). An alternative explanation for the mixed findings is that the success of these programs depends on what is being rewarded (Kerr, 1995). If one wants to increase safety performance, then why reward lack of reported accidents? Clearly, systematic reviews of the incentive-program–safety-performance relation are needed, as is more research conducted to establish the mechanisms by which incentive programs, and perhaps the broader realm of compensation systems, can enhance safety performance and for whom. As evident in the preceding text, most of the research focusing on the relation between leadership and safety performance has taken a positive perspective and has not examined the potential negative consequences of passive and abusive leadership. As Mullen and Kelloway (in press) summarized, there is a growing literature that suggests that passive leadership, as well as abusive supervision, may actually be negatively related to safety performance (see Kelloway, Mullen, & Francis, 2006; Teed, Kelloway, & Mullen, 2008). It is clear that further theoretical development of the role of leadership in safety performance and strong empirical tests of the subsequent theoretical propositions are needed, considering that leaders are the primary agents of the organization and are generally considered to be more proximal in influencing employees’ behavior. Next, we focus on one of the ways that leaders can exert such influence by implementing safety training programs.

Safety Training Safety training represents another HRM practice that organizations can use to increase safe behavior and reduce injuries and fatalities. (See also Vol. 2, chap. 16, this handbook.) Although organizations implement some of this training voluntarily, much safety training is the result of OSHA directives. Much of what researchers know about safety training and the content on which it should focus comes from analysis of operations in higher high-reliability industries such as aviation, nuclear power plant operations, and surface transport and from examination of the catastrophes that have occurred in those industries (e.g., Helmreich & Foushee, 1993; Weick & Roberts, 1993). Burke and Sarpy (2003) identified four major types of safety training programs. One category focuses on increasing workers’ fundamental knowledge regarding topics such as company policies and procedures and proper operating procedures. In general, these programs are designed to increase declarative knowledge and therefore are often presented in a lecture or computerized format (Burke et al., 2006). A second category of programs includes those designed to increase workers’ ability to recognize and appropriately report workplace hazards or safety violations. A third category focuses on developing problem-solving and analytical skills to address safety problems. Finally, the fourth category of training programs is meant to improve decision-making skills and to take systemic steps to reduce future incidents. Evidence generally supports the effectiveness of safety training programs in terms of improving safety-related behavior and in reducing accidents, injuries, and fatalities (see Burke & Sarpy, 2003; Burke et al., 2006; Colligan & Cohen, 2004). Such benefits accrue through the accumulation of safetyrelevant knowledge and skills that enable workers to recognize hazards and operate consistently with stated organizational policies and procedures. In addition, training can demonstrate management’s commitment and establish norms regarding safe behavior. However, as Colligan and Cohen (2004) pointed out, establishing trust among workers and a positive safety climate requires a holistic and systemic approach. An organization that offers safety training but has old, unsafe equipment, for example, 465

Copyright American Psychological Association. Not for further distribution.

Kaplan and Tetrick

can foster cynicism and resentment, not improved safety performance as workers come to regard the training as “window dressing.” The effectiveness of safety training depends on several factors, many of which are relevant to training in general. First, safety training ideally should begin with a needs analysis to identify what knowledge or skills the program should emphasize. In reality, however, OSHA requirements, not organizational needs, likely often drive the training intervention. Second, organizations must carefully consider the modality of the training. Consistent with results from the broader training literature (Frese &, Zapf, 1999), Burke et al. (2006) provided meta-analytic evidence that more engaging training methods (e.g., behavioral modeling, use of dialogue between the trainer and trainee) were more effective than more passive methods (e.g., lecture, video) in terms of knowledge acquisition, as well as decreased accidents and injuries. In fact, they reported that, on the basis of their results, the most engaging methods are, on average, about three times more effective than the least engaging methods with regard to facilitating knowledge and skill acquisition. This finding is especially significant given that passive methods, such as lectures, videos, and more recently, computerized training, likely constitute the vast majority of safety training programs. Finally, safety training will only be effective to the degree that the organization also has removed barriers to the implementation of the knowledge and skills acquired during training. Colligan and Cohen (2004) made this point by describing a study by Carlton (1987) in which kitchen workers who received training in safely lifting kitchen trays to reduce back strain later could not implement the learned technique because of physical obstacles in their actual job site and nonmodeled time constraints. Another potential barrier that can limit the degree to which learned behavior actually translates to onthe-job performance is a lack of managerial support for implementation of the newly acquired knowledge and skills. This is especially true in terms of training workers to communicate safety hazards and concerns. Calling attention to safety issues potentially endangers the organization’s reputation and can also lead to work stoppages or closures and, in turn, potential 466

layoffs or rebukes. Managerial support, norms for open communication, and a positive safety climate make it much more likely that workers will report their safety observations and concerns (Mullen, 2005; Tucker, Chmiel, Turner, Hershcovis, & Stride, 2008). Research examining other factors that predict reporting, such as personality factors and values, is currently lacking, though, and would be useful. SUMMARY, CONCLUSIONS, AND FUTURE RESEARCH DIRECTIONS I/O psychology research on occupational safety and accidents has burgeoned over the past decade. In addition to increasing dramatically in magnitude, this research also has begun to change markedly in terms of approach and content. We view several advances in the recent safety literature as especially significant. First, by attending to recent workforce trends, researchers have broadened their investigation of potential antecedents of workplace safety and accidents. For instance, scholars have begun examining safety issues in specific populations who increasingly compose the workforce, such as older workers (e.g., Kowalski-Trakofler, Steiner, & Scwerha, 2005) and temporary and part-time workers (e.g., Quinlan & Bohle, 2004). Such studies are important given that these populations may face unique safety issues and challenges and, therefore, may require different interventions as well. In addition to examining these demographic trends, I/O psychology scholars have started to appreciate the manifestations of broader societal changes, including increasing job insecurity (e.g., Probst, 2004) and increasing work–family conflict (Cullen & Hammer, 2007) in occupational safety. These advances are significant yet currently few in number. The research on shift work highlights the largely ignored fact that safety performance does not begin and end in the workplace. Future work must incorporate people’s nonwork lives and their larger life circumstances to more fully understand workplace safety. A second significant advance in the safety literature, including the relevant work in I/O psychology,

Copyright American Psychological Association. Not for further distribution.

Workplace Safety and Accidents

is that it has begun taking a noticeably more scholarly and conceptual turn. Instead of simply correlating antecedents (e.g., personality traits, environmental risks) with safety outcomes, studies increasingly adopt a more sophisticated approach, both in terms of theoretical and methodological rigor. Such is evident in the increasing number of studies (a) examining the interactions between various antecedents (e.g., Wallace & Vodanovich, 2003), (b) examining safety in a multilevel framework (e.g., Hofmann & Stetzer, 1996), and (c) using longitudinal designs (e.g., Neal & Griffin, 2006). In addition, measurement advances, including those with respect to conceptualizing and assessing safety performance (Burke et al., 2002; Neal & Griffin, 2004; Hoffman et al., 2003) and with regard to capturing higher level factors (e.g., safety climate; Zohar, in press), are further evidence of this more sophisticated approach. Perhaps most noteworthy is the increasing focus on the psychological mechanisms mediating predictor–outcome safety relationships (Griffin & Neal, 2000; Wallace, Popp, & Mondore, 2006). I/O psychologists are uniquely positioned to examine these mechanisms given their training in, and knowledge of, more general psychological processes. Focusing on the mediating processes is also important given that these are the more proximal mechanisms through which the various antecedents (e.g., demographic variables) ultimately operate. Accordingly, these mediating variables are more strongly related to safety performance than are more distal predictors (Griffin & Neal, 2000; Hofmann & Stetzer, 1996). This rationale also suggests that organizational interventions should aim at affecting these mediators. That is, the underlying processes responsible for safety and accidents should drive the choice and nature of human resources interventions, not the converse. Given this recognition, we present in Table 14.1 a general summary listing some of the psychological process variables reliably found to be predictive of occupational safety. Also, we list specific organizational interventions that research suggests would be most effective in addressing (e.g., combating or enhancing) these processes. This catalog is not meant to be definitive or exhaustive, as we regard the I/O

psychology research as too nascent to offer conclusive safety recommendations. In addition, the efficacy of any HR practice is contingent on specific organizational, contextual, and personal factors. Thus, rather than purporting “best practices,” our objective here is simply to provide a framework that others might borrow from and enhance through scholarly and applied efforts. Perhaps an even more laudable and ambitious undertaking that researchers might work toward is the development of a model linking specific predictors to specific dimensions of safety performance through these mediators. Such a model, in turn, would drive subsequent theoretical development, model testing, and applied practices. As evident in this chapter, I/O psychology research is moving toward this goal, especially by clarifying the nature of the criterion space (Burke et al., 2002), expanding examination of relevant predictors (see above), and constructing models of safety performance (Griffin & Neal, 2000). These advances are encouraging and likely will increase in number in the future as I/O psychology safety research continues to expand. In our view, there are several other exciting and important safety questions that I/O psychology researchers can (and, in some cases, already have) take the lead in addressing. For instance, given the importance of safety climate in predicting accidents (see above), scholars should continue examining the relative importance of different safety climate facets (Christian et al., 2009), the predictors of safety climate (Wallace et al., 2006), and the mechanisms through which it impacts safety outcomes (Zohar, in press). An even more fundamental issue, from our perspective, is that of further refining the understanding of the “safety climate” concept. If one considers safety climate as reflecting shared perceptions (e.g., of organizational practices; Zohar, 2008), a reasonable next question would be, Where do those shared perceptions originate? Although they certainly derive in part from objective work characteristics (e.g., HRM policies), they likely also reflect patterns of interaction and general job attitudes. Determining the source of these shared perceptions and whether the perceptions predict beyond the objective workplace factors would be extremely valuable. 467

Kaplan and Tetrick

TABLE 14.1

Copyright American Psychological Association. Not for further distribution.

Preliminary Framework Linking Recommended Organizational Interventions to Specific Psychological Predictors of Safety and Accidents Psychological predictor

Recommended organizational interventions

Fatigue and decrements in vigilance

1. To combat the effects of shift work, eliminate night work to the degree possible. When considering rotating schedules, consider the advantages and drawbacks of each (see Shift Work section). Also consider hiring individuals who would prefer to work the night shift only. 2. Incorporate frequent rest breaks. Note that such breaks do not, however, negate the effects of shift work (Tucker et al., 2006).

Stress and anxiety

1. Recognize that safety behavior does not occur in isolation. Attempt to reduce stress from all sources that would interfere with safety behavior. 2. Eliminate role stressors and, especially, role overload. 3. Develop cohesion and supportive leaders. 4. Redesign jobs to reduce the influence of physical stressors (e.g., noise, vibration).

Perceived lack of control in job execution and safety issues

1. Restructure jobs to provide more autonomy and decision latitude. 2. Use decentralized team-based structures. 3. Reduce status distinctions; encourage employees to talk with management about safety. 4. Implement a health safety committee.

Lack of safety motivation

1. Use compensation to reward safe behavior. 2. Establish a strong safety climate. Such is accomplished through means such as effective leader behavior, implementation of safety-oriented organizational policies and reward structures, and provision of necessary training and equipment.

Lack of safety knowledge and skills

1. Use safety training. In general, more engaging training is more beneficial than passive training methods. Training must be supplemented with a genuine concern by management to support and enforce the behaviors developed during training. See the “Safety Training” section for other considerations.

Another useful direction for I/O psychology scholars to follow is to be more integrative in their research. Given that I/O psychology is at the intersection of science and practice, I/O psychology scholars are well-placed to foster cross-fertilization and collaboration with other fields such as human factors and industrial engineering. Other collaborators, such as those in public health and sociology, also could inform, and be informed by, I/O psychology research. Each of these fields offers unique theoretical and methodological tools that would provide for a more comprehensive analysis of workplace accidents. For instance, sociology research could provide insights into the role of community factors in organizations’ climates, and economists could speak to the importance of labor market forces on organizations’ safety practices. These are just two examples of the types of knowledge that other fields can contribute and that I/O psychology can help bring together. 468

I/O psychology researchers also can be more integrative with respect to examining safety issues in other countries and cultures. There is ongoing work between U.S. (NIOSH) and European (European Agency for Safety and Health at Work) governmental agencies, but more international collaboration among researchers and practitioners is needed. Perhaps most important is extending I/O psychology research to developing countries and to those countries and organizations who report a disproportionately high number of accidents and workplace fatalities (see above). If this research is to have the most significant impact, researchers must begin applying findings in the countries in most need and also addressing the specific boundary conditions that prevent current findings from generalizing to these countries. Although we recognize that there are practical challenges in conducting such work, we call on I/O psychology scholars to begin undertaking these endeavors, as they represent a very

Workplace Safety and Accidents

Copyright American Psychological Association. Not for further distribution.

practical way to make a tremendous difference in people’s lives. In sum, I/O psychology research has made significant advances in understanding and improving workplace safety behavior and in reducing workplace accidents. However, much more needs to be done. Our hope is that this chapter might help in this expansion by highlighting some theoretical and conceptual gaps in the current literature and perhaps also by providing some starting points for these subsequent efforts.

& D. Hoffman (Eds.), Health and safety in organizations: A multilevel perspective (pp. 56–90). San Francisco: Jossey-Bass. Burke, M. J., Sarpy, S. A., Smith-Crowe, K., Chan-Serafin, S., Salvador, R. O., & Islam, G. (2006). Relative effectiveness of worker safety and health training methods. American Journal of Public Health, 96, 315–324. Burke, M. J., Sarpy, S. A., Tesluk, P. E., & Smith-Crowe, K. (2002). General safety performance: A test of a grounded theoretical model. Personnel Psychology, 55, 429–457.

References

Byrnes, J. P., Miller, D. C., & Schafer, W. D. (1999). Gender differences in risk taking: A meta-analysis. Psychological Bulletin, 125, 367–383.

Adams-Roy, J. E., Knap, M. A., & Barling, J. (1995). Commitment to occupational health and safety in management training. In L. E. Tetrick & J. Barling (Eds.), Changing employment relations: Behavioral and social perspectives (pp. 133–143). Washington, DC: American Psychological Association.

Campbell, J. P. (1990). Modeling the performance prediction problem in industrial and organizational psychology. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd, ed., pp. 687–732). Palo Alto, CA: Consulting Psychologists Press.

Asmundson, G. J., Norton, G. R., Allerdings, M. D., Norton, P. J., & Larsen, D. K. (1998). Posttraumatic stress disorder and work-related injury. Journal of Anxiety Disorders, 12, 57–69.

Carlton, R. S. (1987). The effects of body mechanics instruction on work performance. American Journal of Occupational Therapy, 41, 16–20.

Barling, J., & Frone, M. R. (Eds.). (2004). The psychology of workplace safety. Washington, DC: American Psychological Association. Barling, J., Kelloway, E. K., & Iverson, R. D. (2003). High-quality work, job satisfaction, and occupational injuries. Journal of Applied Psychology, 88, 276–283. Barling, J., Loughlin, C., & Kelloway, E. K. (2002). Development and test of a model linking safetyspecific transformational leadership and occupational safety. Journal of Applied Psychology, 87, 488–496. Borman, W. C., & Motowidlo, S. J. (1993). Expanding the criterion domain to include elements of contextual performance. In N. Schmitt & W. C. Borman, (Eds.), Personnel selection in organizations (pp. 71–98). San Francisco: Jossey-Bass. Broadbent, D. E., Cooper, P. F., Fitzgerald, P., & Parkes, K. R. (1982). The Cognitive Failures Questionnaire (CFQ) and its correlates. British Journal of Clinical Psychology, 21, 1–16. Brown, R. L., & Holmes, H. (1986). The use of factoranalytic procedures for assessing the validity of an employee safety climate model. Accident Analysis and Prevention, 18, 455–470. Bureau of Labor Statistics, U.S. Department of Labor. (2007). [Industry injury and illness data]. Available at http://www.bls.gov/iif/oshsum.htm#06Summary%20 News%20Release Burke, M. J., & Sarpy, S. A. (2003). Improving worker safety and health through interventions. In L. Tetrick

Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. Christian, M. S., Bradley, J. C., Wallace, J. C., & Burke, M. J. (2009) Antecedents of occupational safety performance and outcomes: A meta-analysis. Journal of Applied Psychology, 94, 1103–1127. Clarke, S. (2006). The relationship between safety climate and safety performance: A meta-analytic review. Journal of Occupational Health Psychology, 11, 31–327. Clarke, S., & Robertson, I. (2008). An examination of the role of personality in work accidents using metaanalysis. Applied Psychology: An International Review, 57, 94–108. Colligan, M. J., & Cohen A. (2004). The role of training in promoting workplace safety and health. In J. Barling & M. R. Frone (Eds.), The psychology of workplace safety (pp. 223–248). Washington, DC: American Psychological Association. Costa, P. T., & McCrae, R. R. (1985). The NEO Personality Inventory manual. Odessa, FL: Psychological Assessment Resources. Coyle, I. R., Sleeman, S. D., & Adams, N. (1995). Safety climate. Journal of Safety Research, 26, 247–254. Cullen, J. C., & Hammer, L. B. (2007). Developing and testing a theoretical model linking work–family conflict to employee safety. Journal of Occupational Health Psychology, 12, 266–278. 469

Kaplan and Tetrick

Czeisler, C. A., Moore-Ede, M. C., & Coleman, R. C. (1982, July 30). Rotating shift work schedules that disrupt sleep are improved by applying circadian principles. Science, 217, 460–463.

Copyright American Psychological Association. Not for further distribution.

Dedobbeleer, N., & Beland, F. (1991). A safety climate measure for construction sites. Journal of Safety Research, 22, 97–103.

Helmreich, R. L., & Foushee, H. C. (1993). Why crew resource management? Empirical and theoretical bases of human factors training in aviation. In E. Weiner, B. Kanki, & R. Helmreich (Eds.), Cockpit resource management (pp. 3–45). San Diego, CA: Academic Press.

Dekker, S. (2002). Reconstructing human contributions to accidents: The new view on error and performance. Journal of Safety Research, 33, 371–385.

Hofmann, D. A., Morgeson, F. P., & Gerras, S. J. (2003). Climate as a moderator of the relationship between leader-member exchange and content specific citizenship: Safety climate as an exemplar. Journal of Applied Psychology, 88, 170–178.

Di Milia, L., Smith, P. A., & Folkard, S. (2004). Refining the psychometric properties of the circadian type inventory. Personality and Individual Differences, 36, 1953–1964.

Hofmann, D. A., & Stetzer, A., (1996). A cross-level investigation of factors influencing unsafe behaviors and accidents. Personnel Psychology, 49, 307–339.

Dinges D. F. (1995). An overview of sleepiness and accidents. Journal of Sleep Research, 4, 4–14.

Hofmann, D. A., & Tetrick, L. E. (Eds.). (2003). Health and safety in organizations: A multilevel perspective. San Francisco: Jossey-Bass.

Fisher, C. D. (1993). Boredom at work: A neglected concept. Human Relations, 46, 395–417. Folkard, S., Akerstedt, T., & Macdonald, I. (2000) Refinement of the three-process model of alertness to account for trend in accident risk. In S. Hornberger, P. Knauth, G. Costa, & S. Folkard (Eds.), Shiftwork in the 21st century (pp. 49–54). Frankfurt, Germany: Peter Lang. Folkard, S., & Barton, J. (1993). Does the “forbidden zone” for sleep onset influence morning shift sleep duration? Ergonomics, 36, 85–91. Frese, M., & Zapf, D. (1999). Action as the core of work psychology: A German approach. In H. C. Triandis, M. D. Dunnette, & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (pp. 271–340). Palo Alto, CA: Consulting Psychologists Press. Frone, M. R. (1998). Predictors of work injuries among employed adolescents. Journal of Applied Psychology, 83, 565–576. Gallagher, R. M., & Myers, P. (1996). Referral delay in back pain patients on worker’s compensation: Costs and policy implications. Psychosomatics, 37, 270–284. Glazner, L. K. (1996). Factors related to injury of shiftworking firefighters in the Northeastern United States. Safety Science, 21, 255–263. Greenwood, M., &Woods, H. M. (1919). The incidence of industrial accidents upon individuals with special reference to multiple accidents (Industrial Fatigue Research Board, Medical Research Committee, Report No. 4). London: Her Majesty’s Stationery Office. Griffin, M. A., & Neal, A. (2000). Perceptions of safety at work: A framework for linking safety climate to safety performance, knowledge, and motivation. Journal of Occupational Health Psychology, 3, 347–358. Hansen, C. P. (1989). A causal model of the relationship among accidents, biodata, personality, and cognitive factors. Journal of Applied Psychology, 74, 81–90. 470

International Labour Organization. (2005). Decent work–safe work: Introductory report to the XVII World Congress on Safety and Health at Work, 2005. Available at http://www.ilo.org/public/english/ protection/safework/wdcongrs17/intrep.pdf Iverson, R. D., & Erwin, P. J. (1997). Predicting occupational injury: The role of affectivity. Journal of Occupational and Organizational Psychology, 70, 113–128. James, L. R., & Jones, A. P. (1974). Organizational climate: A review of theory and research. Psychological Bulletin, 81, 1096–1112. Kaminski, M. (2001). Unintended consequences: Organizational practices and their impact on workplace safety and productivity. Journal of Occupational Health Psychology, 6, 127–138. Kaplan, S., Bradley, J., Luchman, J., & Haynes, D. (2009). Dispositional affect and job behaviors: A metaanalytic investigation. Journal of Applied Psychology, 94, 162–176. Keenan, V., Kerr, W., & Sherman, W. (1951). Psychological climate and accidents in an automotive plant. Journal of Applied Psychology, 35, 108–111. Kelloway, E. K., Mullen, J. E., & Francis, L. (2006). Divergent effects of passive and transformational leadership on safety outcomes. Journal of Occupational Health Psychology, 11, 76–86. Kerr, S. (1995). On the folly of rewarding A, while hoping for B. Academy of Management Executive, 9, 7–14. Knauth, P. (1993). The design of shift systems. Ergonomics, 36, 15–28. Kowalski-Trakofler, K., Steiner, L., Schwerha, D. (2005). Safety considerations for the aging workforce. Journal of Safety Science, 43, 779–793. Kozlowski, S. W., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations:

Workplace Safety and Accidents

Contextual, temporal, and emergent processes. In K. J. Kline & S. W. Kozlowski (Eds.), Multilevel theory, research, and methods in organization (pp. 3–90). San Francisco: Jossey-Bass. Lawton, R., & Parker, D. (1998). Individual differences in accident liability: A review and integrative approach. Human Factors, 40, 655–671. Leigh, J. P. (1986). Individual and job characteristics as predictors of industrial accidents. Accident Analysis and Prevention, 18, 209–216.

Copyright American Psychological Association. Not for further distribution.

Leigh, J. P., Markowitz, S., Fahs, M., & Landrigan, P. (2000). Costs of occupational injuries and illnesses. Ann Arbor: University of Michigan Press. Liberty Mutual Research Institute for Safety. (2008). 2008 Workplace safety index. Available from http://www. libertymutualgroup.com/omapps/ContentServer? cid=1138365240689&pagename=LMGResearch Institute%2Fcms_document%2FShowDoc&c=cms_ document Luria, G. (2008). Climate strength—How leaders form consensus. Leadership Quarterly, 19, 42–53. Lusa, S., Häkkänen, M., Luukkonen, R., & ViikariJuntura, E. (2002). Perceived physical work capacity, stress, sleep disturbance and occupational accidents among firefighters working during a strike. Work & Stress, 16, 264–274. Mayo, E. (1933). The human problems of an industrial civilization. New York: Macmillan. McAfee, R. B., & Winn, A. R. (1989). The use of incentives/feedback to enhance workplace safety: A critique of the literature. Journal of Safety Research, 20, 7–19. McKenna, F. P. (1983). Accident proneness: A conceptual analysis. Accident Analysis & Prevention, 15, 65–71. Melamed, S., Luz, J., Najenson, T., Jucha, E., & Green, M. S. (1989). Ergonomic stress levels, personal characteristics, accident occurrence, and sickness absence among factory workers. Ergonomics, 32, 1101–1110. Melamed, S., Yekutieli, D., Froom, P., Kristal-Boneh, E., & Ribak, J. (1999). Adverse work and environmental conditions predict accident involvement in industrial workers—The Israeli CORDIS Study. American Journal of Epidemiology, 150, 18–26. Monk, T. H. (2000). What can the chronobiologist do to help the shift worker? Journal of Biological Rhythms, 15, 86–94. Monk, T. H., Folkard, S., & Wedderburn, A. I. (1996) Maintaining safety and high performance on shift work. Applied Ergonomics, 27, 17–23. Mullen, J. E. (2005). Testing a model of employee willingness to raise safety issues. Canadian Journal of Behavioral Sciences, 37, 273–282. Mullen, J. E., & Kelloway, E. K. (in press). Leading to occupational health and safety. In J. C. Quick & L. E. Tetrick (Eds.), Handbook of occupational health psy-

chology (2nd ed.). Washington, DC: American Psychological Association. Münsterberg, H. (1913). Psychology and industrial efficiency. Boston: Houghton Mifflin. Murphy, K. R. (1996). Individual differences and behavior in organizations. San Francisco: Jossey Bass. Nahrgang, J. D., Morgeson, F. P., & Hofmann, D. A. (2008, April). Predicting safety performance: A metaanalysis of safety and organizational constructs. Paper presented at the 23rd Annual Conference of the Society for Industrial and Organizational Psychology, San Francisco. Neal, A., & Griffin, M. A. (2004). Safety climate and safety work. In J. Barling & M. R. Frone (Eds.), The psychology of workplace safety (pp. 15–34). Washington, DC: American Psychological Association. Neal, A., & Griffin, M. A. (2006). A study of the lagged relationships among safety climate, safety motivation, safety behavior, and accidents at the individual and group levels. Journal of Applied Psychology, 91, 946–953. National Safety Council. (2002). Injury facts. Itasca, IL: Author. Occupational Safety and Health Act of 1970, 29 U.S.C. § 651 (1970). Parker, S. K., Axtell, C. M., & Turner, N. (2001). Designing a safer workplace: Importance of job autonomy, communication quality, and supportive supervisors. Journal of Occupational Health Psychology, 6, 211–228. Paul, A., & Lewis, A. (2008). Safety goggles: Are they adequate to prevent eye injuries caused by rotating wire brushes? Emergency Medicine Journal, 25, 385. Pransky, G., Snyder, T., Dembe, A., & Himmelstein, J. (1999). Underreporting work-related disorders in the workplace: A case study and review of the literature. Ergonomics, 42, 171–182. Probst, T. M. (2004). Job insecurity: Exploring a new threat to employee safety. In J. Barling & M. R. Frone (Eds.), The psychology of workplace safety (pp. 63–80). Washington, DC: American Psychological Association. Quinlan, M., & Bohle, P. (2004). Contingent work and occupational safety. In J. Barling & M. R. Frone (Eds.), The psychology of workplace safety (pp. 81–105). Washington, DC: American Psychological Association. Rajaratnam, S. M., & Arendt, J. (2001). Health in a 24-h society. Lancet, 358, 999–1005. Reason, J. T. (1997). Managing the risks of organizational accidents. Aldershot, England: Ashgate. Salancik, G. R, & Pfeffer, J. (1978). A social information processing approach to job attitudes and task design. Administrative Science Quarterly, 23, 224–253. 471

Kaplan and Tetrick

Schneider, G., Salvaggio, A. N., & Subirats, M. (2002). Climate strength: A new direction for climate research. Journal of Applied Psychology, 87, 220–229. Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124, 262–274. Shaw, I., & Sichel, H. (1971). Accident proneness. Elmsford, NY: Pergamon Press.

Copyright American Psychological Association. Not for further distribution.

Simard, M., & Marchand, A. (1995). A multilevel analysis of organizational factors related to the taking of safety initiative by work groups. Safety Science, 21, 113–129. Sinclair, R. R., & Tetrick, L. E. (2004). Pay and benefits: The role of compensation systems in workplace safety. In J. Barling & M. R. Frone (Eds.), The psychology of workplace safety (pp. 181–201). Washington, DC: American Psychological Association. Smith, L., Folkard, S., & Poole, C. J. (1994). Increased injuries on night shift. Lancet, 334, 1137–1139. Strong, L. L., & Zimmerman, F. J. (2005). Occupational injury and absence from work among African American, Hispanic, and non-Hispanic White workers in the National Longitudinal Survey of Youth. American Journal of Public Health 95, 1226–1232. Taylor, F. W. (1911). The principles of scientific management. New York: Harper & Bros. Teed, M., Kelloway, E. K., & Mullen, J. E. (2008, March). Young workers’ safety: The impact of inconsistent leadership. Paper presented at the Seventh International Conference on Occupational Stress and Health: Work, Stress, and Health, Washington, DC. Tomás, J. M., Meliá, J. L., & Oliver, A. (1999). A crossvalidation of a structural equation model of accidents: Organizational and psychological variables as predictors of work safety. Work & Stress, 13, 49–58. Tucker, S., Chmiel, N., Turner, N., Hershcovis, M. S., & Stride, C. B. (2008). Perceived organizational support for safety and employee safety voice: The mediating role of coworker support for safety. Journal of Occupational Health Psychology, 13, 319–330. Turner, C., McClure, R., & Pirozzo, S. (2004). Injury and risk-taking behavior—A systematic review. Accident Analysis & Prevention, 36, 93–101. Visser, E., Pijl, Y. L., Stolk, R. P., Neeleman, J., & Rosmalen, J. G. M. (2007). Accident proneness, does it exist? A review and meta-analysis. Accident Analysis and Prevention, 39, 556–564.

Wallace, J. C., & Vodanovich, S. J. (2003). Workplace safety performance: Conscientiousness, cognitive failure, and their interaction. Journal of Occupational Health Psychology, 8, 316–327. Weick, K. E., & Roberts, K. H. (1993). Collective mind in organizations: Heedful interrelating on flight decks. Administrative Science Quarterly, 38, 357–381. Westaby, J. D., & Lowe, J. K. (2005). Risk-taking orientation and injury among youth workers: Examining the social influence of supervisors, coworkers, and parents. Journal of Applied Psychology, 90, 1027–1035. Williamson, A. M., & Feyer, A. M. (1995). Causes of accidents and time of day. Work & Stress, 9, 158–164. Wright, C. (1986). Routine deaths: Fatal accidents in the oil industry. Sociological Review, 4, 265–289. Zacharatos, A., Barling, J., & Iverson, D. I. (2005). High performance work systems and occupational safety. Journal of Applied Psychology, 90, 77–93. Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and applied implications. Journal of Applied Psychology, 65, 96–102. Zohar, D. (2000). A group-level model of safety climate: Testing the effect of group climate on microaccidents in manufacturing jobs. Journal of Applied Psychology, 85, 587–596. Zohar, D. (2002). Modifying supervisory practices to improve subunit safety: A leadership-based intervention model. Journal of Applied Psychology, 87, 156–163. Zohar, D. (2008). Safety climate and beyond: A multilevel multi-climate framework. Safety Science, 46, 376–387. Zohar, D. (in press). Safety climate: Conceptual and measurement issues. In J. C. Quick & L. E. Tetrick (Eds.), Handbook of occupational health psychology (2nd ed.). Washington, DC: American Psychological Association. Zohar, D., & Luria, G. (2003). The use of supervisory practices as leverage to improve safety behavior: A cross-level intervention model. Journal of Safety Research, 34, 567–577. Zohar, D., & Luria, G. (2004). Climate as a social– cognitive construction of supervisory safety practices: Scripts as proxy of behavior patterns. Journal of Applied Psychology, 89, 322–333.

Wallace, J. C., & Chen, G. (2006). A multilevel integration of personality, climate, self-regulation, and performance. Personnel Psychology, 59, 529–557.

Zohar, D., & Luria, G. (2005). A multilevel model of safety climate: Cross-level relationships between organization and group-level climates. Journal of Applied Psychology, 90, 616–628.

Wallace, J. C., Popp, E., & Mondore, S. (2006). Safety climate as a mediator between foundation climates and occupational accidents: A group-level investigation. Journal of Applied Psychology, 91, 681–688.

Zohar, D., & Tenne-Gazit, O. (2008). Transformational leadership and group interaction as climate antecedents: A social network analysis. Journal of Applied Psychology, 93, 744–757.

472

CHAPTER 15

DISABILITY AND EMPLOYMENT: NEW DIRECTIONS FOR INDUSTRIAL AND ORGANIZATIONAL PSYCHOLOGY Copyright American Psychological Association. Not for further distribution.

Adrienne J. Colella and Susanne M. Bruyère

In the past 2 decades, the issue of disability in the workplace has garnered a great deal of attention. Concern exists over discrimination against and accommodation of people with disabilities, to name just a few issues. In the United States, the passage of the Americans With Disabilities Act of 1990 (ADA) broadened the concern over disability and work to the fields of law, sociology, economics, and rehabilitation psychology. Similar legislation in other countries (e.g., England, Australia, Canada) has also fueled applied concern and scholarly activity.1 Research on disability and employment has been less developed in the field of industrial and organizational (I/O) psychology, despite the fact that many issues pertaining to the selection, development, and integration of people with disabilities into the workforce are relevant to the domain of I/O psychology. Indeed, since 1990, very few articles related to disability issues have appeared in the Journal of Applied Psychology and Personnel Psychology, compared with scores of articles focusing on gender and race issues. Despite the relative dearth of research on disability employment issues in the field of I/O psychology, there is a plethora of

theory and research from other fields that addresses these issues. The purpose of this chapter is to provide an overview of this literature and to engage I/O scholars and practitioners in the ongoing work related to the integration of people with disabilities into the workforce. We begin this chapter by discussing the definition of disability. Disability is defined in many ways for many purposes. Not only does the definition of disability provide fodder for the courts (e.g., Sutton v. United Airlines, 1999), but it also has methodological implications for research. We then focus on three areas related to disability and employment: accommodation, selection, and workplace integration. These three areas were chosen because they are areas in which significant research has been conducted, they are areas that can benefit from the application of I/O psychology, and they are issues that are important in the employment of people with disabilities. Before moving on, we need to state some boundary conditions. First of all, we do not review case law on the ADA, nor do we discuss the ADA in detail, except where necessary. Much has been written on

Susanne M. Bruyère has been supported in part by U.S. Department of Education, National Institute on Disability and Rehabilitation Research Grants H133G040265 and H133B040013 to Cornell University. The contents of this chapter do not necessarily represent the policy of the U.S. Department of Education or any other federal agency, and endorsement by the federal government should not be assumed. Susanne M. Bruyère would like to acknowledge the contribution of Sara VanLooy, Cornell University Employment and Disability Institute Administrative/Research Assistant, in the preparation of this chapter. 1

In the United Kingdom, the Disability Discrimination Act of 1995 protects and promotes civil rights for people with disabilities. The text of the act as currently in force can be read at http://www.statutelaw.gov.uk/ content.aspx?activeTextDocId=3330327; also see http://www.direct. gov.uk/en/ DisabledPeople/RightsAndObligations/DisabilityRights/DG_4001068 for information and guidance for the public and for employers. For Australia, more information on the Australian Disability Discrimination Act of 1992 can be found at http://www.hreoc.gov.au/ disability_rights/. In Canada, the province of Ontario has passed the Accessibility for Ontarians With Disabilities Act of 2005, and readers are directed to http://www.mcss.gov.on.ca/ mcss/english/pillars/ accessibilityOntario/ for information and guidance.

http://dx.doi.org/10.1037/12169-015 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

473

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

this (e.g., Lee, 2003), and although legal issues are certainly relevant in examining issues such as selection, the intent of our chapter is to focus on psychological issues. Second, there is a great deal of literature on disability insurance (e.g., Chen & van der Klaauw, 2008; Golosov & Tsyvinski, 2006) and return to work (e.g., Pransky, Gatchel, Linton, & Loisel, 2005; Young et al., 2005), both of which are beyond the scope of this chapter, unless related to the selection, retention, and integration of workers with disabilities. Third, there is an entire field of work that addresses disability employment issues from a rehabilitation framework. This literature addresses questions such as, “What are the predictors of job success?” (e.g., Chan, Cheing, Chan, Rosenthal, & Chronister, 2006), “What influences the readiness to work of people with severe disabilities?” (e.g., Roberts & Pratt, 2007), and “What are effective ways of providing job coaches?” (Becker et al., 2007). Again, this vast body of literature is beyond the scope of the chapter but is referred to when relevant. Finally, there is a significant body of work focusing on psychologically based theoretical reasons behind why people with disabilities are treated differently than people without disabilities (e.g., Colella & Stone, 2005; D. L. Stone & Colella, 1996). Theoretical reasons abound for why one may expect there to be negative bias against minority groups, including persons with disabilities, and also for why there may be positive bias. Because of space constraints, these arguments are presented in Table 15.1. For a thorough review of these theories, see D. L. Stone and Colella (1996) and Colella (1996). Theoretical arguments are expounded upon when relevant throughout the text. DEFINITION OF DISABILITY A discussion of the definition of disability is included here because it is important that employers understand their obligations under related laws and also that applicants and employees understand whether they are protected by the various laws related to people with disabilities. Disability is a complex and multidimensional concept. Several theoretical frameworks have been developed to characterize the various dimensions of disability and to model the process 474

of disablement (Livermore & She, 2007). There are a variety of sources and considerations that one can include when considering the definition of disability—legal and public policy, economic, conceptual–philosophical, and applied–practical. In the discussion that follows, we attempt to provide an appreciation of these varying definitions while narrowing this discussion specifically to the realm of employment. There is no single, universally accepted definition of disability. Mashaw and Reno (1996) documented over 20 definitions of disability used by programs, government agencies, or researchers and argued that the appropriateness of any definition can be judged only in the context in which it is used. In general, there are two prominent conceptualizations of disability (Burkhauser, Houtenville, & Wittenburg, 2003). One views disability as an impairment that limits a person’s capacity to function at work, in society, or in daily life. Another defines certain conditions as disabling and counts people with those conditions as disabled. For public policy purposes, national surveys have often been used to assist in identifying the population of people with disabilities. The definitions of disability that result from these surveys vary (Houtenville, 2003). Each survey asks questions of respondents that are used to determine an individual’s disability status. For example, the 2000 U.S. Census asked, “Does this person have any of the following long-lasting conditions: Blindness, deafness, or a severe vision or hearing impairment?” (Erickson & Houtenville, 2005). Respondents who answered affirmatively became the population that has sensory disabilities. Other questions in the 2000 Census determined whether people had conditions that “substantially limit” one or more basic physical activities, such as walking, climbing stairs, lifting, or carrying (physical disability; Erickson & Houtenville, 2005). The definition of mental disability when using 2000 Census data is drawn from the question, “Because of a physical, mental, or emotional condition lasting 6 months or more, does this person have any difficulty: Learning, remembering, or concentrating?” Other surveys result in a work-limitation definition of disability. The Current Population Survey

Disability and Employment

TABLE 15.1 Psychologically Based Theoretical Reasons for Biased Treatment of People With Disabilities

Copyright American Psychological Association. Not for further distribution.

Reason

Explanation

Stereotyping

People hold negative stereotypes about people with disabilities that influence how they process and recall information about them (e.g., D. L. Stone & Colella, 1996). Stereotypes include helpless, benevolent, hypersensitive, inferior, depressed, distant, shy, unappealing, unsociable, bitter, insecure, nonaggressive, unhappy, submissive (Fichten & Amsel, 1986); saintly (Colella, 1996); and less capable of competing (Makas, 1988). Specific stereotypes vary with the type of disability.

Just world hypothesis

People with disabilities are blamed for the disability or viewed as deserving of the disability so that others may maintain a belief in a just world and thus can ease their anxiety that they could one day become disabled (Lerner, 1980). Negative attributes are ascribed to people with disabilities to make the disability seem justified.

Existential anxiety

Nondisabled people identify with people with disabilities, which creates anxiety over the possibility that they, too, may become disabled (Hahn, 1988; Livneh, 1982). Anxiety leads to avoidance.

Aesthetic anxiety

People experience discomfort due to interacting with someone who deviates from society’s norms about what is considered physically attractive (Hahn, 1988; Livneh, 1982). Anxiety leads to avoidance.

Norm to be kind or sympathy effects

People wish to view themselves as good people, which means that they follow normative beliefs about treating those less fortunate in a kind manner (Carver, Glass, Snyder, & Katz, 1977). Effect can lead to avoidance of negative feedback (Hastorf, Northcraft, & Piciotto, 1979), overly positive performance evaluations (Czajka & DeNisi, 1988), and paternalism (Colella & Stone, 2005).

Ambivalence response theory

People hold ambivalent affect toward people with disabilities, feeling both disgust and pity or nurturance at the same time. They respond by displaying heightened affect cued by the social situation (Katz, 1979, 1981).

Social adaptation theory

People choose the most adaptive response in a given situation or context (Piner & Kahle, 1984). Sympathy effects should therefore occur in situations of little importance (i.e., there is nothing to lose), whereas negative stigmatizing reactions should occur in situations the nondisabled actor believes are important or have serious consequences.

Stigmatization

Disability is viewed as a social stigma, which guides the cognitive processing of information about people with disabilities and the scripts that people engage to interact with them (Jones et al., 1984). Common scripts include “child among adults,” “illness and disease,” and “moral deviate” (Jones et al., 1984).

(CPS) asks, “Does anyone in this household have a health problem or disability which prevents them from working or which limits the kind or amount of work they can do?” (Burkhauser & Houtenville, 2006). A common denominator of all these definitions is that they rely on the self-report of the persons being interviewed, thus requiring these persons to decide for themselves whether their condition prevents them from working or limits the kinds of work available to them. Because some people with

disabilities who are in the workforce would not consider themselves work limited, these surveys do not always accurately identify or define people with disabilities (Burkhauser, Daly, & Houtenville, 2002). Currently, the U.S. Census and three major noninstitutional surveys (the American Community Survey, the National Health Interview Survey, and the Survey of Income Program participation) all contain information on sensory disability, functional limitations, mental disabilities, limitations in 475

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

activities of daily living, limitations in instrumental activities of daily living, and work disabilities, although definitions vary (Burkhauser, Houtenville, & Wittenburg, 2003; She & Stapleton, 2006). Over the past 10 years or more, an international effort led by the World Health Organization (WHO) has been reframing how disability is defined, making it more consistent with a social rather than a medical model of disability and impairment. The result has been the International Classification of Functioning, Disability, and Health (ICF), which includes measures of activities and participation and environmental factors, in addition to health and disability classifications (WHO, 2001). In these very early stages of the ICF implementation, there have been a significant number of efforts both in the United States and abroad to examine the conceptual utility of this classification framework to the fields of rehabilitation, health, and other clinical practices. In addition, serious work has begun to examine how this classification system relates to specific national and international surveys where measures relevant to the interests of people with disabilities are reflected, as well as in a wide variety of clinical and research settings (Bruyère, VanLooy, & Peterson, 2005). One of the fertile areas for further work by I/O and rehabilitation psychologists is applications of the ICF conceptual model to employment-related issues for people with disabilities (Bruyère, 2005). Within the “Activities and Participation” section of the ICF is Chapter 8, “Major Life Areas,” which is about carrying out the tasks and actions required to engage in education, work, and employment and to conduct economic transactions. Some of the specific relevant components for consideration include apprenticeship (work preparation); vocational training; acquiring, keeping, and terminating a job; and remunerative and nonremunerative employment. The “Activities and Participation” section of the ICF includes indicators of the individual’s capacity and performance ability with and without assistance; the employment-related items can therefore be further coded using qualifiers for activity limitations and participation restrictions (WHO, 2001). Ongoing work is needed to prove and support the utility of the ICF, especially an 476

examination of employment and workplace-related applications. With many definitions possible, legislators and federal agencies define disability differently, depending on the particular protection or eligibility criteria for services that they are attempting to afford. The ADA definition of disability is rooted in the impairment model of disability. The ADA defines disability as (with respect to an individual) (a) a physical or mental impairment that substantially limits one or more of the major life activities of the individual, (b) having a record of such impairment, or (c) being regarded as having such an impairment (Equal Employment Opportunity Commission [EEOC], 1992, 2002). The Social Security Administration (SSA) has an extensive approach to defining the impairment of an individual for purposes of determining eligibility for benefits (SSA, 2006). When Congress added disability assistance to Social Security in 1950, it limited eligibility to those who were “totally and permanently disabled” (Social Security Advisory Board, 2003). Today, the SSA defines disability as the inability to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairments which can be expected to result in death or which has lasted or can be expected to last for not less than 10 months. (Social Security Online, 2008, ¶ 4) Although the core definition of disability has not changed, SSA continues to examine their definition of disability to make certain that it stays current with advances in rehabilitation knowledge and technology that impact people’s ability to work (Social Security Advisory Board, 2003). However, the ADA does not require that someone be considered disabled by the SSA to be protected. What matters is that there is some impairment that substantially affects a major life activity. Generally, the existence of an impairment is relatively clear. What is less clear is whether this impairment rises to the level of a disability for the purposes of the ADA. To be defined as a disability and to be eligible for accommodations under the

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

ADA, a physical or mental impairment must also meet the “substantially limits” test (Duston, 2001). For example, someone with a learning disability affecting auditory perception might have difficulty gaining information from a staff meeting. This would likely be an ADA disability (Tominey, 2001). There is no specific listing of names of ADA-eligible impairments. It is the effect of an impairment or condition on the life of the particular person that defines a disability (EEOC, 1992). Some impairments (e.g., blindness, deafness, AIDS) are by their nature substantially limiting, whereas others may be limiting for some individuals and not for others (EEOC, 1992). Historically, people with physical impairments have had an easier time making the case to employers, the EEOC, and the courts that they have a substantial limitation to major life activities. The sporadic nature of some psychiatric impairments, the subjective nature of some of the impacts on a person’s ability to work (e.g., difficulty concentrating is hard to measure), and the specific exclusion in the ADA of personality traits as impairments all make mental impairments more difficult to substantiate (Paetzold, 2005). It is also important to be aware of changes in interpretation of these definitions that occur over time on the basis of court rulings. For instance, in Sutton v. United Airlines (1999), the U.S. Supreme Court ruled that factors that mitigate the disabling nature of an impairment may be considered when determining whether a person is covered under the ADA. More recently, on the basis of the assessment that this Supreme Court decision has resulted in some individuals with substantial impairments incorrectly being found as not having disabilities, amendments to the ADA that attempt to reinstate a broad interpretation of protections available under the ADA have been passed (e.g., the ADA Amendments Act of 2008, or ADAAA). During the past few years since the Sutton v. United Airlines decision, individuals with disabilities appear to have been filing for claims of employment discrimination under state laws that provided greater protections than the ADA affords (Bruyère et al., 2007). This illustrates the importance of knowledge of state laws, as well as federal statutes, that provide

employment disability nondiscrimination protections for people with disabilities. The ADA definition of disability, intended by legislators to be broad and open ended, was so contentious and so limited by subsequent court rulings that the ADAAA of 2008 was passed. The intent of the ADAAA was to reject the increasingly narrow judicial constructions of the ADA, constructions that often had a disproportionate impact on people with psychiatric disabilities, in favor of broad coverage (Center, in press; EEOC, 2008d). The ADAAA specifically overturned the Sutton decision by making explicit the legislative intent that mitigating measures not be considered in deciding whether a person has an ADA disability (Center, in press). It retained the basic definition of disability but provided a clearer definition of “substantially limits,” expanded on the definition of “major life activities” by providing nonexhaustive lists of examples, and clarified that a condition that is episodic or in remission (such as bipolar disorder or depression) can still be a disability if it would be considered a disability when active (EEOC, 2008d). To be protected by the ADA, an individual must also be qualified for the job in question. The regulations define this as a person who “satisfies the requisite skill, experience, education and other jobrelated requirements of the employment position such individual holds or desires, and who, with or without reasonable accommodation, can perform the essential functions of such position” (EEOC, 1992, ¶ 2, Section 2.3). Essential functions are job duties that have been required of employees in the position because the position exists to perform that function, because the job is highly specialized and the person is hired for that special expertise, or because there are a limited number of other employees to perform the function. For example, the ability to proofread accurately would be an essential function of a copy editor’s job. The ability to work nights would be an essential function of a floating supervisor. The ability to communicate with coworkers via e-mail might be an essential part of a job; however, doing this via reading and typing might not be essential if the information could be listened to via speech synthesis software and entered via voice recognition dictation software. 477

Colella and Bruyère

Copyright American Psychological Association. Not for further distribution.

ACCOMMODATION The ADA differs from most other civil rights legislation in that it requires employers to provide reasonable accommodation.2 Accommodations are modifications in the job, work environment, work process, or conditions of work that reduce physical and social barriers so that people with disabilities experience equal opportunity in a competitive work environment (EEOC, 1992). Reasonable accommodations are those that do not cause undue hardship, which is defined as causing significant expense or difficulty, not minimal cost or difficulty (Knapp, Faley, & Long, 2006). Costs must be shown to be excessive in relation to the benefits of the accommodation or the employer’s financial survival (Vande Zande v. Wisconsin Dept. of Admin., 1995). Soon after the ADA was passed, employers were worried about the financial costs of accommodation. However, data soon demonstrated that monetary costs were quite low (Braddock & Bachelder, 1994; Schartz, Hendricks, & Blanck, 2006). Yet, lawsuits over the failure to provide accommodation to disabled employees constitute approximately 25% of ADA discrimination charges (Bjelland, Bruyère, Houtenville, Ruiz-Quintanilla, & Webber, 2008), the second largest category of discrimination allegations brought up under the ADA (McMahon, Edwards, Rumrill, & Hursh, 2005; McMahon & Shaw, 2005); discharge is the number one category. Case law has shed light on two important issues (Knapp et al., 2006) that can and should be addressed by I/O psychologists: the accommodation process and the costs and benefits of accommodation.

The Accommodation Process Employees, as the ones with the most knowledge of their needs for reasonable accommodation, are generally responsible for requesting accommodation from their employers (EEOC, 2002; Knapp et al., 2006) and must explain to the employer what the impairments are that require an accommodation (Taylor v. Principal Financial Group, Inc., 1996). After the request is made, it is the employer’s responsibility to engage in an interactive dialogue with the requesting 2

employee to determine which accommodations are appropriate (EEOC, 2002; Knapp, et al., 2006). Most commonly, the immediate supervisor, with or without consultation of his or her supervisor, makes the final decision (Bruyère, Erickson, & Horne, 2002). The requesting individual must establish that he or she has a disability and that there are reasonable accommodations that would allow him or her to perform the essential functions of the job. If the employer determines that the organization is unable to provide the requested accommodation, then it is the employer’s responsibility to provide specific evidence demonstrating that an accommodation would cause undue hardship. It should also be noted that employers do not have to provide the specific requested accommodation if an alternative accommodation is available. The employer should give primary consideration to the employee’s preference, but, ultimately, the choice between accommodations rests with the employer, as long as the accommodation provided is effective— meaning that it allows the employee to perform the essential functions of the job and offers equal opportunities to enjoy the full rights and privileges of employment (EEOC, 2002; Knapp et al., 2006). Although a great deal of legal and some public policy research has been conducted about the accommodation process, very little research has been done from a psychological perspective. One issue that deserves attention is the willingness of disabled employees to request an accommodation. Individuals with disabilities who are entitled to accommodation and who would benefit from accommodation are often reluctant to request an accommodation (Anderson & Williams, 1996; Lee, 1997). Research that asks people with disabilities whether they have requested an accommodation shows that fewer than those who need accommodations ask for them. For example, Allaire, Li, and LaValley (2003) found that although 98% of their sample of people with arthritis and rheumatoid disease faced barriers at work because of their disability, only 38% requested an accommodation. Balser (2007), in her study of people with mobility-related disorders, found that 72% had requested an accommodation.

Reasonable accommodation is also required for employees’ sincerely held religious practices by Title VII of the Civil Rights Act of 1964. See http://www.eeoc.gov/types/religion.html for EEOC guidance on accommodations for religious practice.

478

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

Baldridge and Veiga (2001) developed a model to explain the likelihood that a person with a disability would request an accommodation at work. According to this model, situational variables likely to influence accommodation requests include the accommodation culture (the extent to which the organization supports people with disabilities), the magnitude of the accommodation (e.g., cost, disruption of regular routines), and the onset controllability of the disability. These situational characteristics are proposed to influence the requester’s beliefs about asking for an accommodation, such as the perceived usefulness of the accommodation, the anticipated image cost, the perceived fairness of the accommodation, and anticipated compliance. Normative assessments also occur, including the perceived help-seeking appropriateness and the perceived social obligation toward others with disabilities. To our knowledge, the only empirical test of this model was conducted by Baldridge and Veiga (2006). They surveyed hearing-impaired individuals to determine the extent to which perceptions of supervisory compliance, personal cost of requesting an accommodation, and normative appropriateness mediate the impact of the monetary cost and imposition (as rated by experts) of various accommodations on the likelihood of asking for that accommodation. They found support for their model, suggesting that hearing-impaired individuals strongly considered the social cost of asking for an accommodation. For example, they found that respondents’ perceptions of the personal cost and the normative appropriateness of an accommodation accounted for 10% and 11%, respectively, of the variance in the decision to request an accommodation, over and above the perceived monetary cost and imposition of the accommodation. Clearly, more research is needed to test this model to determine whether these findings generalize across other types of disabilities. This was a hearing-impaired sample, whose coworkers were likely to have known about their disabilities and who were more likely to have a political identity associated with their disability than people with other types of disabilities. Others have suggested additional factors that influence the likelihood of requesting an accommodation. For example, severity of need (Cleveland,

Barnes-Farrell, & Ratz, 1997), perceptions of selfcompetence (Cleveland et al., 1997), and fear of disclosure of a hidden disability (McLaughlin & Gray, 1998) have also been proposed to influence requesting behavior. We know of no research that has examined these other factors. Thus, a great deal of research is needed concerning what influences people with disabilities to ask for an accommodation. Another avenue for future research is to examine what actually happens to people with disabilities who request accommodations and whether their expectations are justified. Are these concerns on the part of people with disabilities justified? Recent research by Paetzold et al. (2009) found that when accommodations were viewed as increasing the competitive performance of a person with disabilities, others thought that the accommodation was unfair. Another important question related to the accommodation process is what influences employers to either grant or deny an accommodation request, an issue that has been approached from several different perspectives. First of all, it should be noted that it is a relatively common event for a supervisor to be asked for an accommodation. A recent survey of federal agency supervisors (Bruyère et al., 2002) reported that half of the supervisor respondents had received at least one accommodation request in the past 5 years, about one third had received one to three requests, 11% had received 4 to 10 requests, and 4% had received more than 10 requests. It should be noted that the employment rate of people with disabilities in the federal government is not particularly high (Bruyère et al., 2002) and that these findings coincide with a study of 131 privatesector managers (Florey & Harrison, 2000), which found that 45% had received an accommodation request. To our knowledge, there are no data indicating the percentage of requests that are accepted and the percentage that are refused. One way of addressing the question of what affects managers’ likelihood of granting an accommodation is to ask them. Lee (1997) conducted a survey of a well-derived sample of 500 employers from various industries. These employers agreed that the cost of accommodations and excessive structural modifications were major barriers to accommodating individuals. They agreed to a lesser 479

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

extent that higher insurance costs, resistance from coworkers and managers, union contracts, and safety risks were major barriers. However, these data were collected only immediately after the ADA was passed, when employer fears of the cost of accommodation were high (McFarlin, Song, & Sonntag, 1991). A study of supervisors relating to workplace accommodation practices was also conducted in the federal sector with over 1,000 white- and blue-collar employees across 16 agencies (Bruyère et al., 2002). Costs were seen as less of a barrier to employment for people with disabilities than were supervisor knowledge of which accommodation to make; needs, attitudes, and stereotypes toward people with disabilities; and the lack of related experience and requisite skills in the person with a disability. Subsequent research has shown that the cost of accommodating individuals with disabilities is quite low (Blanck, 1996; Lee & Newman, 1995) and not necessarily perceived as a barrier by employers. Bruyère (2000) reported that less than 10% of private-sector respondents felt that cost was a barrier to providing accommodations. In Lee’s (1996) sample, 62% of respondents reported that the accommodations provided cost less than $500 and only 11% reported that accommodations cost more than $5,000. This research needs to be conducted again, after employers have had almost 20 years of experience with the ADA. Another way of examining the question has been to conduct laboratory research experiments where extraneous variables are held constant and internal validity is high. T. L. Mitchell and Kovera (2006) conducted two scenario studies and found, as hypothesized, that respondents were more likely to grant accommodations (and more expensive ones) when the disability was viewed as externally caused compared with self-caused (η2 = .65 and η2 = .61, respectively). Contrary to predictions, they found mixed results across studies regarding the effect of work history on accommodation granting (η2 = .30 and η2 = .58). Florey and Harrison (2000) conducted two scenario studies in which they examined the impact of origin of the disability (or onset controllability) and request magnitude on attitudes toward providing a specific accommodation to a hearingimpaired employee. As in T. L. Mitchell and Kovera’s 480

study, they found that respondents intended to give more accommodations to an employee with an externally caused disability than to those whose impairment was self-caused (ω2 = .06 and ω2 = .02, respectively)—for example, a person who uses a wheelchair after a drunk-driving car accident (external cause) or a person with health issues resulting from alcohol or drug abuse (internal cause). They also found support for a positive relationship between past performance and intentions to accommodate (ω2 = .05). They found that the magnitude of the accommodation was only related to attitudes toward accommodation and felt obligation to accommodate but not to actual behavioral intentions. In addition to the experimentally manipulated variables, they also found that performance instrumentality, previous contact with people with disabilities, and perceived fairness of the accommodation influenced intentions to accommodate. A final way to determine what influences whether accommodations are granted has been to conduct large-scale correlational studies examining the correlates of the types of accommodations provided (e.g., Allaire et al., 2003; Balser, 2007; Campolieti, 2004). In a study of workers with various injuries returning to work, Campolieti (2004) found that only previous vocational training, returning back to the previous employer, and having sprains versus other types of disabilities increased one’s chances of receiving an accommodation. Returning to one’s previous employer had the strongest impact, accounting for about an 8% increase in the probability that an employee would receive an accommodation. Allaire et al. (2003), in a study of persons with rheumatic diseases and arthritis, found that educational attainment, functional limitation, physical demands of the job, having a professional–managerial occupation, and believing that requesting an accommodation would result in receiving one were positively related to the likelihood of receiving an accommodation. Functional limitation had the strongest effect, with a 1-standard-deviation increase in limitation increasing the likelihood of accommodation by 60%. McDonald-Wilson, Rogers, and Massaro (2003) found that for persons with psychiatric disabilities, the strongest predictor of the number of accommodations they received was the

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

number of functional limitations (e.g., difficulty conversing with others, displaying relevant affect, following instructions). Those with one limitation had a mean number of 1.2 accommodations, whereas those with four limitations had a mean of 2.5 accommodations. The most theoretically developed study in this area, by Balser (2007), examined the likelihood that people with mobility-related disabilities would receive various accommodations. She found that being an employee in the nonprofit sector, unionization, race, and employee input into the accommodation process were related to receiving various accommodations. She found that different factors predicted the likelihood of receiving different types of accommodation. For example, union membership was not related to scheduling flexibility, but union members were 95% less likely to receive workat-home accommodations. Others have found that employers are more likely to grant accommodations to employees of higher socioeconomic status (Burkhauser & Daly, 1996; Collignon, 1986). Taken as a whole, it is difficult to draw conclusions from this research because separate disabilities and accommodations have been studied, different lists of factors have been excluded and included, and different questions have been asked. Furthermore, the results of this research do not tell us whether those who do not receive accommodations do not get them because (a) they do not ask for them, (b) they were denied them, or (c) they do not need them. One general conclusion that can be drawn here is that factors reflecting human capital (education, training, occupational level, performance history) are the ones most strongly related to accommodation. Another general conclusion is that factors associated with other types of discrimination (cause of disability or race) also influence the granting of accommodations. One way in which I/O psychologists can contribute to the literature on the accommodation process is to examine more closely the dialogue and negotiation that takes place between the requester and the grantor (most often the immediate supervisor; McLaughlin & Gray, 1998). Rousseau’s (2001, 2004) and Rousseau, Ho, and Greenberg’s (2006) work on idiosyncratic treatment of employees could help shed light on this issue. By describing how idiosyn-

cratic deals and flexible work arrangements arise (Rousseau, 2001) and by discussing when such customized work arrangements are likely to be functional and dysfunctional (Rousseau et al., 2006), this new area of research could serve as a basis of a framework for when accommodations for disability reasons will most likely benefit accommodated workers and be viewed as fair by others. Another aspect of the accommodation process that is in the purview of I/O psychologists is to study the decisionmaking process of those deciding to grant the accommodation. What we have now is very little research outlining the factors related to when accommodations are granted and/or received. However, we know very little about what accommodation grantors think about when making those decisions. Research addressing this issue would make it easier to generalize across disabilities and accommodations. One question to examine is how managers interpret the meaning of “reasonable accommodation.” The ADA stated that reasonable accommodation is any change or adjustment to a job or work environment that permits a qualified applicant or employee with a disability to participate in the job application process, to perform the essential functions of a job, or to enjoy benefits and privileges of employment equal to those enjoyed by employees without disabilities. (EEOC, 2008a, “What Are My Obligations to Provide Reasonable Accommodations?” section) With continuing questions about what is “reasonable,” ongoing guidance has been needed to bring more precision to both employers’ and applicants’– employees’ understanding of what is reasonable. Are accommodations to be provided to just enable an employee to perform the essential job functions at an acceptable level? Are they to be provided to make the employee more competitive or perform at a higher level? Are they to be provided to make the workplace more tolerable for employees with disabilities? In an attempt to respond to early feedback about a need for further clarity in defining reasonable 481

Colella and Bruyère

Copyright American Psychological Association. Not for further distribution.

accommodation, the EEOC has continued to issue additional guidance on this topic. Enforcement guidance issued in the fall of 2002 addresses what constitutes a request for reasonable accommodation, the form and substance of the request, and an employer’s ability to ask questions and seek documentation after a request has been made (EEOC, 2002). For example, the EEOC (2002, ¶ 4; see http://www.eeoc. gov/policy/docs/accommodation.html) stated, A modification or adjustment is “reasonable” if it “seems reasonable on its face, i.e., ordinarily or in the run of cases;” this means it is “reasonable” if it appears to be “feasible” or “plausible.” An accommodation also must be effective in meeting the needs of the individual. In the context of job performance, this means that a reasonable accommodation enables the individual to perform the essential functions of the position. Similarly, a reasonable accommodation enables an applicant with a disability to have an equal opportunity to participate in the application process and to be considered for a job. Finally, a reasonable accommodation allows an employee with a disability an equal opportunity to enjoy the benefits and privileges of employment that employees without disabilities enjoy. Still, to many managers, the law and its accommodation provisions remain unclear in their implementation. Thus, two important research questions are, How do managers interpret the intended purpose of accommodations? and How do these interpretations influence organizational willingness to provide accommodations? Another question that can be more systematically addressed is, How do supervisors decide on what is an appropriate accommodation? We do know that supervisors tend to have little information about how and when to grant accommodation (Bruyère, Erikson, & VanLooy, 2000; Bruyère et al., 2007), despite such helpful resources as the Job Accommodation Network (see http://www.jan. wvu.edu/). The research cited above suggests that 482

human capital is related to the granting and receiving of accommodations and that grantors may engage in some type of cost–benefit analysis. The finding that stigmatizing factors are associated with the granting of accommodations suggests that this decision, in part, is also influenced by prejudice. Neither of these decision heuristics is acceptable in considering whether to grant an accommodation. Thus, further research at the behavioral level is warranted. Further complicating this issue is the sheer breadth of what constitutes an accommodation. The Job Accommodation Network provides employers a wealth of information about what accommodations may be suitable for a great variety of disabilities and the ensuing functional limitations. Different researchers use different ways of grouping and classifying accommodations. For example, from the field of rehabilitation medicine, a taxonomy of assistive technology device outcomes has been developed for use by practitioners in the field (Jutai, Fuhrer, Demers, Scherer, & DeRuyter, 2005), and from the field of rehabilitation engineering, workplace accommodation and computer use classification schemes are beginning to emerge (Bruce & Sanford, in press; Milchus & Bruce, in press). However, to date, no overarching definitive taxonomy of workplace accommodations exists. To provide an example of the variety of possible accommodations, Table 15.2 lists various accommodations that frequently appear in the literature. Such a taxonomy of accommodations connected to a taxonomy of functional limitations would be useful in developing tools for managers to use in deciding what accommodations are appropriate and effective in addressing various impairments and for research assessing the efficacy of classes of accommodations. The Job Accommodation Network does provide employers with typical functional limitations for a myriad of disabilities and describes accompanying accommodations. However, no taxonomy of potential accommodations is provided. This is one area in which I/O-based research would be useful.

Costs and Benefits of Accommodation Evidence suggests that the passage of the ADA has not led to better employment figures for people

Disability and Employment

TABLE 15.2 Common Types of Accommodations

Copyright American Psychological Association. Not for further distribution.

Assistive technologies Screen reader Adapted keyboard Hearing aid TTY communicator Braille output device Telephone headset Wheelchair Wrist splints Adapted mouse Amplified telephone Mouth stick

Environmental accessibility

Personal assistance

Job restructuring

Ramps Automatic doors Ergonomic table Ergonomic chair Modified restrooms Special parking Removal of toxins

Job coach Assistant Personal assistance services Readers Interpreters Service animal Coworker training (e.g., in use of ASL)

Reassignment to different job Reassign tasks to coworkers Flextime Modified work schedule Additional supervision Excused absences Short breaks

Note. TTY = teletypewriter; ASL = American Sign Language.

with disabilities (see Bjelland, Burkhauser, & Houtenville, 2008; Houtenville, Stapleton, Weathers, & Burkhauser, in press; Stapleton & Burkhauser, 2003, for an in-depth discussion of this issue), with some economists arguing that the ADA has caused a decline in employment for people with disabilities (Acemoglu & Angrist, 2001; DeLeire, 2003) and others taking a more positive approach and arguing that the employment statistics are not nearly as bad when one accounts for other factors, such as different definitions of disability or employment (Houtenville & Burkhauser, 2004; Kruse & Schur, 2003). Some also point to the impact of accommodation on the retention of employees. Burkhauser, Butler, and Weathers (2001/2002) found that men and women who are accommodated by their employers are significantly less likely to apply for Social Security Disability benefits in each of the first few years after their conditions begin to impact their ability to work than are those who are not accommodated. In any event, no one argues that the ADA has had a strong positive impact on the employment prospects for people with disabilities. One rationale given for why the ADA may not have had the intended effect (there are others; see, e.g., Lee, 2003) is presented by economists, who suggest that because of the accommodation stipulation of the ADA, people with disabilities are more expensive for employers to employ (DeLeire, 2000;

Donohue, 1994; Weaver, 1991) and thus employers are less likely to hire people with disabilities. This argument states that because employers are legally prohibited from paying people with disabilities less, their mandate to provide accommodation makes a person with a disability more expensive to employ in contrast to another employee at the same productivity level who does not require accommodation. There are several economic counterarguments to this point of view (Blanck, Schur, Kruse, Schwochau, & Song, 2003; Blanck, Schwochau, & Song, 2003; Jolls, 2000). One such argument questions the assumption that people with disabilities are less productive without accommodation than are those without disabilities (Blanck, Schur, et al., 2003) and that accommodations are always provided at a financial loss (Schartz et al., 2006). Another argues that providing accommodations lowers everyone’s wages and consequently is beneficial to organizations (Jolls, 2000). Most of these arguments are in the realm of economic theory and analysis. However, one question emerges that is in the domain of I/O psychologists: What are the costs and value to the organization of providing employees with accommodations? Economists and legal scholars have addressed the issue of cost–benefit analyses of accommodations (e.g., Blanck, 1996; Schartz et al., 2006; Stein, 2000). A study conducted by the Job Accommodation Network (1997) found that the median cost of an 483

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

accommodation was $200. A large-scale longitudinal study conducted at Sears & Roebuck (Blanck, 1996) found that 72% of accommodations cost nothing, 17% cost less than $100, 10% cost less than $599, and only 1% cost more than $500. In a largescale study of all private employers in New Jersey, Lee (1997) found that when asked about the most expensive accommodation that had been provided, 38% reported that it cost nothing, 24% reported it cost less than $500, 26% reported that it cost between $500 and $5,000, and 11% said it cost more than $5,000. In dollar terms, the accommodations do not cost very much. Schartz et al. (2006), in a survey of 890 employers who consulted the Job Accommodation Network, found that almost half reported that there was zero cost associated with an accommodation, with a median 1-year cost of $600. Only 10% of private-sector employers selected the costs of accommodation or additional supervision as potential barriers to the employment or advancement of people with disabilities in Bruyère’s (2000) employer survey. There are other indirect costs associated with accommodations. Many disabilities require accommodations involving help and time from supervisors and coworkers (Colella, 2001; McLaughlin & Gray, 1998). For example, McDonald-Wilson, Rogers, Massaro, Lyass, and Crean (2002), in a study of accommodations for individuals with psychiatric disorders, found that supervisors spent an average of 5 hr extra per month and coworkers spent an extra 9 hours per month in activities for the accommodation of a coworker with psychiatric disabilities. Another indirect cost of accommodation is the potential negative reaction on the part of coworkers (Colella, 2001; Colella, Paetzold, & Belliveau, 2004). Bruyère, Erickson, and Ferrentino (2003) reported that employers found changing coworker and supervisor attitudes to be the most difficult change to make when accommodating employees, and Paetzold et al. (2009) found that coworkers perceive accommodations as unfair when they are viewed as providing a competitive advantage. Cleveland et al. (1997) found that reactions to accommodation varied with the rationale provided for the accommodation, such that other employees may see the presence of a disabled coworker as coercive if the accommodation is pre484

sented as being done for legal reasons. Another study found that coworkers resented accommodations that involved modified job duties for people with disabilities and reassignment of certain tasks to coworkers (Stoddard, 2006). Indeed, coworker reactions have been taken into consideration by courts of law in determining undue hardship (Benson v. Northwest Airlines, 1995; Kralik v. Durbin, 1997; Wooten v. Farmland Foods, 1995). Systematic research that empirically assesses the benefits of accommodation for the organization and the individual being accommodated is relatively scarce. Most studies on the impact of accommodations are case studies (Butterfield & Ramseur, 2004); examine rehabilitation-related variables, such as client satisfaction; do not use reliable and valid assessment methodology; and do not focus on criteria that organizations would find important. One exception involves studies that demonstrate that accommodating people with disabilities is related to their return to work and staying at work (Burkhauser, Butler, & Kim, 1995; Burkhauser et al., 2001/2002; Butterfield & Ramseur, 2004; Fabian, Waterworth, & Ripke, 1993). Another potential benefit of providing accommodations is that accommodations initially provided for disability purposes may be applied to all employees and improve everyone’s performance and/or morale. Flextime would be one example of such an accommodation. Given that employers are still concerned about the costs of an accommodation and believe that it is unlikely that benefits exceed costs (Schartz et al., 2006), research providing a more balanced, systematic, and in-depth cost analysis of accommodations is needed. Parallels can be drawn from similar such attempts to create a balanced perspective on the costs of managing diversity for other protected populations (i.e., race–ethnicity, gender, age). Robinson and Deschant (1997) used higher turnover rates, higher absenteeism rates, and lawsuits on sexual, race, and age discrimination, which can occur when people become more aware of discriminatory practices, as reasons for employers to attend to the need for creating a nondiscriminatory work environment where difference is valued. A recent study by Schartz et al. (2006) demonstrates what such an analysis should look like. She

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

and her colleagues surveyed employers who used the Job Accommodation Network to obtain employer estimates of the direct costs, indirect costs, direct benefits, and indirect benefits of providing specific accommodations in their companies. Direct cost was the 1-year estimate of out-of-pocket expenses for providing the accommodation. Indirect costs included lost time due to training, supervisor’s time, coworkers’ time, and loss of production. Direct benefits included allowing the company to hire, retain, and promote a qualified employee; eliminating the cost of training new employees; savings on workers’ compensation and other insurance costs; improved employee attendance and productivity; and increased diversity. Indirect benefits included increased overall company productivity, attendance, morale, profitability, and workplace safety; increased customer base; and improved interactions with coworkers and customers. They found that the mean calendar year net benefit for providing an accommodation was $11,335. Almost 60% of respondents reported a positive net benefit, whereas 18% reported a negative net benefit. Schartz et al. (2006) asked employers to estimate these costs and benefits. Given the criticism of employer estimates of the dollar value of various personnel interventions that surfaced in the utility analysis literature (Latham & Whyte, 1994), research that empirically connects accommodations to various cost and benefit outcomes is needed. For example, do accommodations actually increase productivity? What types of accommodations are likely to result in the highest indirect costs? In parallel research for other protected populations, Barrington and Troske (2001) empirically assessed the relationship between workforce diversity and the economic performance of an establishment using the New Worker– Establishment Characteristics Database, a nationwide employer–employee matched data set, to estimate the association between productivity and workforce diversity. In this research, their main finding was that diversity either was positively associated with productivity or had no significant relationship with productivity. In summary, the issue of accommodation needs and offers multiple research opportunities for I/O psychologists. Accommodations may result in direct

and indirect costs, such as monetary expense and coworker resentment, respectively. In contrast, there may be direct benefits, such as the increased performance, satisfaction, and commitment of the accommodated person, as well as indirect benefits, such as increased performance and morale of coworkers. What we lack is a systematic manner in which to assess these outcomes. The evaluation of accommodation outcomes lags behind the methodology for assessing the value of other personnel procedures (cf. Cascio & Boudreau, 2008). As the workforce ages, the issue of accommodation will become ever more important both to individuals and organizations. As of now, there is scant advice we can offer to organizations on how to best implement disability accommodation procedures and to individuals on how to secure the accommodations they may need. Hopefully, future research will move us toward this goal. SELECTION AND ENTRY One thing we do know is that people with disabilities have a harder time entering into employment than do those without disabilities, even when they are qualified, want to work, and are able to work. Statistics derived from the 2006 American Community Survey (Rehabilitation Research and Training Center on Disability Demographics and Statistics, 2007) indicate that only 37.7% of people with disabilities are employed overall and that only 87.1% of those who report that they are able to work actually do, compared with 79.7% and 95.0% of the general population, respectively. Results from the CPS coincide with these findings, revealing that people with disabilities had an unemployment rate of 10.5% compared with 4.7% of the general population (McMenamin, Miller, & Polivka, 2006). Significant numbers of people with disabilities are not finding jobs even though they want to work and are able to work. As a result of these statistics, there is a significant amount of research examining the selection, hiring, and entry of persons with disabilities into the workforce. D. L. Stone and Williams (1997) laid out a research agenda for the impact of the ADA on organizational selection that provides an excellent 485

Colella and Bruyère

organizing framework for a review of the literature in the past 10 years since their review. We begin by examining disability issues related to job analysis, followed by recruiting, the application process, and testing and assessment. We complete this discussion by addressing the final selection decision.

Copyright American Psychological Association. Not for further distribution.

Job Analysis Early on in the ADA’s history, there was much discussion about the need for organizations to redo job analyses and ensuing position descriptions to comply with the law. (See Vol. 2, chap. 1, this handbook.) The ADA stipulates that a qualified person with a disability is an individual with a disability “who . . . with or without reasonable accommodation, can perform the essential functions of such a position” (EEOC, 1992, p. 11). Although the ADA does not require that new job analyses be performed (K. E. Mitchell, Alliger, & Morfopolous, 1997), it does bring up two issues that were not typically addressed by previous job analysis methods: identifying essential job functions and the role of accommodation in task performance. According to the ADA (EEOC, 1992, pp. 13–18), a job function is essential if (a) the position exists to perform the function; or (b) there are a limited number of other employees who could perform the function, or among whom the function can be distributed; or (c) a function is highly specialized, and the person is hired for special expertise or ability to perform it. The EEOC (1992, pp. 13–18) went on to state what information can be used to determine the above criteria. This information includes the employer’s judgment, a written position description, the amount of time spent on the function, the consequences of not performing the function, the terms of a collective bargaining agreement, work experience of those performing similar jobs, and other relevant factors. The implication of the accommodation stipulation suggests that employers focus on outcomes rather than processes when conducting job analyses (K. E. Mitchell et al., 1997), acknowledging that employees with disabilities may be able to get the job done but in a different way than those without disabilities. K. E. Mitchell et al. (1997) conducted a survey to determine the extent to which people with disabilities performed their jobs in a different man486

ner, and, in accordance with the accommodation literature, they found that 54% performed their jobs differently. Thirty percent performed their jobs differently because they used different tools and technological devices (e.g., special software), 5% changed the amount of time worked by taking short naps, and 16% actually switched tasks with other employees. There has been very little research on job analysis and disability issues. There are many articles and resources written for practitioners that give specific advice. A visit to the Occupational Information Network (http://online.onetcenter.org/) takes the interested employer to the Job Accommodation Network site for help determining what accommodations exist for certain aspects of performance. There are also more in-depth accounts of how job analyses were conducted to meet ADA requirements. For example, Brannick, Brannick, and Levine (1992) conducted a job analysis for pharmacists. Unique aspects of this job analysis were that incumbents and supervisors were asked to rate how essential each task was in getting the job done and how many other people were available to perform the task in question. They came up with four criteria to judge whether a task was essential: the pharmacists thought it was essential, the task had serious consequences if performed incorrectly, no one was available to cover the task, and/or over 70% of incumbents did the task or reported spending an above average amount of time on it. The job analysts then determined the minimum level of sensory and motor requirements for performing each task and had incumbents and supervisors estimate the importance of various knowledge, skills, and abilities for performing the task. Instead of creating a new procedure, K. E. Mitchell et al. (1997) reviewed common current job analysis methods (task analysis, e.g., the Comprehensive Occupational Data Analysis Program; Positional Analysis Questionnaire; critical incidence techniques; threshold traits analysis; and functional job analysis) to determine which met ADA stipulations. They concluded that functional job analysis (Fine, 1989) was the most satisfactory method because it describes jobs in terms of outcomes and provides information that should help in determining appropriate accommodations (such as the tools and technology used for the job).

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

In summary, there are plenty of resources and guidance available to practitioners on how to do a job analysis that would be defensible under the ADA. Thus, we see little need for more research on this issue. However, one issue that does relate to job analysis is the way in which position descriptions are written and the effect that has on the likelihood that people with disabilities apply for jobs and are considered suitable applicants. We address this issue in the next session on recruiting. Also, research using people with disabilities as subject-matter experts in job analysis may result in job descriptions that are more flexible and amenable to people who would benefit from such flexibility.

Recruiting People With Disabilities Little research has been conducted on how to recruit people with disabilities into the organization, perhaps because disability is rarely considered an aspect of diversity programs and because there is no affirmative action requirement under the ADA. (See also Vol. 2, chap. 2, this handbook.) Various states and municipalities offer tax incentives for organizations to hire people with disabilities (see Legnick-Hall, 2007, for a listing). Many sources exist for employers to locate people with disabilities, such as rehabilitation agencies, the Employer Assistance and Recruiting Network (http://www.dol.gov/ odep/programs/earn. htm), independent living centers, state employment agencies, special Web sites, and disabled person associations (D. L. Stone & Williams, 1997). Organizations that are acknowledged as being “disability friendly” and that have a good record of hiring and retaining people with disabilities tend to be very proactive in recruiting people with disabilities (see Legnick-Hall, 2007, for a series of case studies). For example, Hewlett-Packard engages in multiple partnerships with rehabilitation organizations and schools to recruit people with disabilities (Gaunt, 2007). However, there is evidence that employers are generally unaware of these incentives and sources for hiring people with disabilities. A recent Society for Human Resources Poll (Schramm, 2007) found that 87% of employers were aware of welfare to work tax credits, but only 34% and 27%, respectively, were aware of the disabled tax credit and the SSA employment network cash provision.

The growth of online recruiting on both corporate Web sites and job boards, such as Monster.com, has raised concerns about new barriers to the workplace for people with disabilities. A Cornell University survey of 10 job boards and 31 corporate recruitment sites found that because of poor accessibility features and usability characteristics, it was not possible to complete an application and submit a resume on more than two thirds of the sites (Bruyère, Erickson, & VanLooy, 2005). As nearly 90% of employers surveyed in 2002 were using online job postings, and less than 20% reported any familiarity with existing guidelines for accessible Web design, this is an area of concern (Bruyère, Erickson, & VanLooy, 2005). These technologies may improve access to employment opportunities for some people with disabilities but can preclude people with other types of disabilities from finding available positions. An important research question is, why do so few employers actively recruit people with disabilities? That is, why are the majority of employers unwilling or unknowledgeable about how to go about recruiting from this demographic? How can employers be motivated to recruit people with disabilities? In September of 2006 and June of 2008, the Interagency Committee on Disability Research held national summit meetings focused on the future direction of disability employment research. The general theme was that this research should move away from a rehabilitation, supply side perspective and take on an employer-centered, demand-side perspective. The above research question is one example of a demandside research question and one in which I/O psychologists are uniquely qualified to address. Another recruiting issue of interest to I/O psychologists is how certain recruiting methods and content influence the likelihood that people with disabilities will apply. Such research has already been conducted on how recruiting ads impact the recruitment of Black (e.g., Highhouse, Zickar, Thorsteinson, Stierwalt, & Slaughter, 1999; McKay & Avery, 2006) and female (Avery & McKay, 2006) employees. D. L. Stone and Williams (1997) suggested that factors such as whether recruiting messages emphasized essential job functions, focused on ideal applicants, or used “disability” in the affirmative action 487

Colella and Bruyère

statement could influence the attractiveness of jobs for people with disabilities. To our knowledge, no published research has examined this issue.

Copyright American Psychological Association. Not for further distribution.

Application Process, Testing, and Assessment of People With Disabilities One implication of the ADA is that all prehire assessments must not unfairly discriminate on the basis of disability. Reasonable accommodations must be provided so that job applicants with disabilities have access to all application procedures, and prehire disability-related questions are prohibited (EEOC, 1992). For example, accommodations must be made so that applicants in wheelchairs may attend a job interview, and job notices should be provided in an accessible format (e.g., large print) for applicants with visual impairments. Some accommodations at application time will be unrelated to actual testing, such as holding interviews in accessible locations. However, other accommodations may require that the form of the assessment be changed. For example, if a job does not require vision, accommodations should be made so that paper-and-pencil tests are read to visually impaired applicants. Prior to hiring, employers may not ask applicants about the “existence, nature, or severity” of a disability (EEOC, 2009). Instead, employers may only ask about employees’ ability to perform essential job functions and to describe or demonstrate how they would perform these functions, with or without reasonable accommodation. Employers also may not require applicants to take medical exams prior to the job offer. After the offer, receiving the job may be made contingent upon the results of a medical exam only if all employees in the job category are required to take the exam and exclusion based on exam results is related to ability to perform the job (EEOC, 2009). One natural research question for I/O psychologists is whether the validity of tests that have been modified because of accommodation (e.g., presented verbally rather than visually) generalizes for disabled applicants. The Educational Testing Service and the U.S. Office of Personnel Management (reported in Brannick et al., 1992) have conducted tests to see whether validity generalizes across various forms (Braille, large print, audiocassette) of cog488

nitive ability tests. Psychometric characteristics of the tests remained stable across different test versions. However, there were performance-level differences for people with varying degrees of disability. Thus, one fruitful avenue for future research is the extent to which accommodating people with disabilities during the applicant assessment phase may actually result in unfair testing or adverse impact. There has been no systematic research or concern about the effectiveness of different selection techniques for people with disabilities, with two exceptions: personality testing and job interviews. (See also Vol. 2, chaps. 5 and 6, this handbook.) Legally, prehire personality testing has been a contentious issue under the ADA (Knapp et al., 2006) because some personality tests are designed to detect mental illness (a disability that can be protected under the ADA). In Karraker et al. v. Rent-a-Center, Inc. (2005), the Seventh Circuit Court of Appeals ruled that the Minnesota Multiphasic Personality Inventory cannot be used in preemployment testing (without showing job relatedness or business necessity) because it can be considered a medical exam. The Minnesota Multiphasic Personality Inventory also screens out people with certain mental disabilities (i.e., mental illnesses, such as anxiety and depression). One interesting line of research would be to determine whether other more commonly used employment personality tests (not designed to assess mental illness), such as those assessing the Big Five, result in adverse impact against people with disabilities. If commonly assessed personality traits such as Conscientiousness or Neuroticism correlate with mental illness, then a case could be made that the use of these personality tests are creating adverse impact (cf. Stone-Romero, 2005). Several studies examined whether employment interviews of people with disabilities resulted in unfair evaluations (Cesare, Tannenbaum, & Dalessio, 1990; Farina & Felner, 1973; Gouvier, Steiner, Jackson, Schlater, & Rain, 1991; Jasper & Klassen, 1990; Johnson & Heal, 1976; Krefting & Brief, 1976; Marchioro & Bartels, 1994; Miceli, Harvey, & Buckley, 2001; Nordstrom, Huffaker, & Williams, 1998; Reilly, Bocketti, Maser, & Wennet, 2006; Rose & Brief, 1979; C. I. Stone & Sawatzki, 1980; Tagalakis, Amsel, & Fichten, 1988). Although these individual studies tended to have mixed results, Ren,

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

Paetzold, and Colella (2008) conducted a metaanalysis of these studies (and others examining reactions to paper people) and found that there was a significant effect of disability on hiring decisions (mean d = −.09, 95% confidence interval [CI] = −.14 to −.04). They also found that the negative effect was worse for mental disabilities (mean d = −.58) than physical disabilities (mean d = −.08). Recent work not included in this meta-analysis found supporting results. Reilly et al. (2006) found stronger bias against those with depression and substance abuse problems compared with those with a history of cancer. Reilly et al. and Brecher, Bragger, and Kutcher (2006) found that structured interviews helped mitigate the negative bias toward people with histories of disability. Thus, there is pretty strong evidence that bias does occur in interviews and the strength of that bias depends on the type of disability. The next step in this line of research is to understand why this bias may occur and how to mitigate it. Hebl and her colleagues (Hebl & Kleck, 2002; Hebl & Skorinko, 2005) have conducted a series of experiments examining whether acknowledging one’s disability during a job interview can mitigate bias, based on a line of social psychological research demonstrating that if a stigmatized person acknowledges the stigma in social interaction, that person is viewed more favorably (Belgrave & Mills, 1981; Davis, 1961; Farina, Sherman, & Allen, 1968; Hastorf, Wildfogel, & Cassman, 1979; Mills, Belgrave, & Boyer, 1984). The reasons for this effect range from making the nonstigmatized person more at ease to leading to perceptions that the stigmatized person is better adjusted. Hebl and Kleck (2002) found that an interviewee in a wheelchair was viewed more positively, more likely to be hired, better liked, and considered to have more positive traits and skills than was an interviewee who did not acknowledge the disability. Hebl and Skorinko (2005) found similar results using the same paradigm and also found that the positive effects of acknowledgement only occurred when the acknowledgement was made early in the interview, as compared with later or not at all. The research on the bias-mitigating effects of acknowledging one’s disability during interviews is important because it leads to implications about how

people with disabilities can combat negative reactions. However, most of this research concerns visible disabilities, and the manipulation used in the interview was a wheelchair. Evidence suggests that personnel decision bias is weaker for physical disabilities (Gouvier et al., 1991; Ren et al., 2008), thus conducting such research using more heavily stigmatized and nonvisible disabilities (e.g., depression, AIDS) is warranted. Any research on mental disabilities tends to appear in the rehabilitation literature and focuses on issues such as having a supportive employment assistant present at interviews (e.g., Gervey & Kowal, 2005). A dilemma faced by people with hidden disabilities is whether to acknowledge the disability to gain an accommodation at the risk of experiencing discrimination (Duckett, 2000). This issue presents a future research need.

Final Selection Decisions In this section, we examine research that generally is aimed at examining why employers do not hire people with disabilities. Research examining why people with disabilities are underemployed has taken two general tracks. One track is to focus on the characteristics of people with disabilities (e.g., age, severity of disability, skill level) that correlate with employment. Most of this research is from a vocational rehabilitation or economics perspective. The other track focuses on decision makers’ attitudes and bias toward people with disabilities. This research is divided into either large-scale surveys of employer attitudes and beliefs about hiring people with disabilities or experimental studies, mostly conducted in the laboratory, which attempt to discern hiring discrimination. Employer surveys generally show that employers have somewhat positive attitudes toward hiring people with disabilities. Hernandez, Keys, and Balcazar (2000), Unger (2002), and Ainspan (2006) reviewed the myriad of employer surveys that occurred within the decade after the 1990 ADA. Hernandez et al. reviewed 37 studies of various employer attitudes and beliefs toward people with disabilities and the ADA. They found 7 studies in which employers from a variety of settings reported positive global attitudes toward people with a variety of disabilities at all severity levels. They found 11 studies in which 489

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

employers reported negative attitudes; however, these studies assessed more specific attitudes or intentions (e.g., Would you hire?) and were more likely to focus on a specific disability (intellectual disabilities, epilepsy, psychiatric disabilities, learning disabilities, back pain). Unger (2002) reviewed 24 studies focusing solely on employer attitudes. Like Hernandez et al. (2000), she obtained mixed results. She found that employers were more positive when the disability in question was physical as opposed to mental or emotional, they had previous experience with employing people with disabilities, and they had a sense of corporate social responsibility. Common concerns expressed about employees with disabilities included work performance, productivity, safety, dependability, attendance, coworker acceptance, lack of skills, costs (workers compensation, insurance, accommodation), and extra training and supervision. Ainspan (2006) qualitatively reviewed the studies in the previous reviews plus others and came to the same conclusions. This literature has been criticized on several grounds, including being subject to social desirability (Colella & Stone, 2005; Unger, 2002); focusing on human resource managers who may not be the people making the decisions (Ainspan, 2006; Colella & Stone, 2005); using a diversity of measures, many of which are not validated (Hernandez et al., 2000); and using methodology that makes it difficult to compare studies (Unger, 2002). Also, such surveys were quite popular during the 1990s, but there have been very few since. Furthermore, there has been little research that directly links these attitudes to actual hiring behavior. This leads to the next set of studies that attempt to get more directly at hiring decisions. In a recent meta-analysis, Ren et al. (2008) located 37 studies that experimentally examined hiring decisions concerning people with disabilities. Most of these studies were conducted in a laboratory setting, using paper or videotaped applicants with a disability and comparing them with a control without a disability. They found a mean effect size (d) of −.09 for disability, with a 95% CI ranging from −.14 to −.04, indicating a small negative bias toward people with disabilities. This finding is corroborated by a study by Schur (2002), who found that 63.2% of 490

people with disabilities who felt they had been discriminated against felt that the discrimination had occurred at hiring, compared with no individuals (0%) in a nondisabled sample who felt they had been discriminated against. The nondisabled sample was more likely to feel like they had been discriminated against in terms of losing jobs (75%, compared with 32.5% of the disabled sample). Ren et al. (2008) also found that type of disability moderated the effect size, which was mean d = −.58 for people with mental disabilities (k = 6) compared with −.08 for people with physical disabilities (k = 28). This finding is corroborated by a study by Chan, McMahon, Cheing, Rosenthal, and Bezyak (2005), who studied EEOC case statistics and found that perceived discrimination (allegations filed with the EEOC) was greater for controllable but unstable disabilities (depression, schizophrenia, alcohol and drug abuse, HIV–AIDS) than for uncontrollable but stable disabilities (visual impairment, cancer, cardiovascular disease, spinal cord injuries). It is interesting to note that Chan et al. found that actual discrimination (EEOC merit resolutions) occurred at higher levels for uncontrollable but stable disabilities. This brings up the possibility not only that people with mental disabilities face more discrimination at hiring time but also that it is more difficult for them to prove it. A major criticism of research on discrimination conducted in the laboratory using paper people is that it overestimates the impact of bias and does not generalize to the field (Landy, in press). Ren et al. (2008) were able to compare laboratory and field study effects. They did find that negative bias effects were stronger in the laboratory (mean d = −.12, k = 25, 95% CI = −.19 to −.04) versus the field (mean d = −.07, k = 12, 95% CI = −.14 to −.01) but that negative bias also remained in the field setting. Thus, it appears that discrimination does appear in the “real world.”

Implications for Needed Research Labor statistics that highlight the gap in employment levels between those with and without disabilities make it quite clear that people with disabilities who are able and willing to work have a more difficult time finding and maintaining employment than those who are not disabled. Research has approached

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

the “why” behind this issue from several angles, focusing on psychological–cognitive explanations (e.g., employer attitudes), more economically based explanations (e.g., increased cost because of accommodation and insurance), inaccessibility issues (lack of accommodation in selection procedures), or inaccurate job analyses. What future research needs to do is to simultaneously consider these different explanations to truly understand the obstacles facing people with disabilities trying to enter into employment. Different explanations lead to different remedies. For example, if the reason for not hiring employees with disabilities rests on stereotypes and erroneous assumptions, then education efforts aimed at hiring managers would help. In contrast, if actual costs are the issue, then such things as tax incentives and grants should help. Rehabilitation professionals could focus on such solutions as better training of the person with the disability. What is interesting is that there is no one answer to the question of how we can improve hiring rates for people with disabilities and that the most effective solutions may vary by disability. There may be a frame of reference problem here, where people from different disciplines are more inclined to focus on problems relevant to their own field and ignore other potential solutions. Cross-disciplinary research would help address this issue. INTEGRATION INTO THE WORKPLACE A great deal of research has examined the entry of persons with disabilities into the workplace, but relatively less research has focused on what happens to them once they get there. More recent research on disability and employment has considered what happens within organizations regarding disability policy and how people with disabilities are treated. We focus on three lines of work here: performance evaluations and promotion, inclusion, and organization culture. (See also chap. 12, this volume; Vol. 2, chap. 9, and Vol. 3, chap. 2, this handbook.)

Performance Evaluations and Promotion A fair amount of research has examined whether people with disabilities are evaluated unfairly in terms of present or past performance and contribu-

tions (for a review of disability and the performance appraisal process, see Colella, DeNisi, & Varma, 1997). Ren et al. (2008) located 13 studies that were controlled experiments examining the impact of disability on performance evaluations. In line with both the norm to be kind and sympathy effects (Carver, Glass, Snyder, & Katz, 1977; Czajka & DeNisi, 1988; Hastorf, Northcraft, & Piciotto, 1979), they found that the average effect size of the difference between performance evaluations of targets with disabilities and control targets without disabilities was .25, indicating evaluations favored people with disabilities. The 95% CI ranged from .14 to .36. They also examined, but did not find, moderating effects for the sex of the target and type of disability. All of these studies were conducted in the laboratory. Individual studies have posited and found moderators of performance evaluation effects. One of the most studied (in terms of all types of discrimination) is the controllability or the perceived cause of the impairment, that is, the degree to which the cause of a person’s disability is internally attributed (Bordieri & Drehmer, 1987; Chan et al., 2005; Weiner, Perry, & Magnusson, 1988). When the cause of a person’s disability is blamed on some aspect of their own behavior or character, evaluations are more negative than when it is perceived that they incurred the disability through no fault of their own. D. L. Stone and Colella (1996) also posited that the following characteristics of disabilities are likely to lead to more discriminatory responses: unattractive aesthetic qualities, progressive course, visibility, disruptiveness, and danger or peril. Colella and her colleagues (Colella, DeNisi, & Varma, 1998; Colella & Varma, 1999) posited and found that stereotypes about how well a given disability fit with a specific job moderated the impact that disability had on performance evaluations in that job. Given the experimental evidence above and employer reports that people with disabilities perform as well as anyone else (Legnick-Hall, 2007), it does not seem that bias in performance evaluations is a big problem for people with disabilities. This, however, does not translate into better employment opportunities or bias elsewhere in the integration process. It appears that even though raters are fair or even overly generous in their assessments of the 491

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

performance of people with disabilities, they still hold lower performance expectations for people with disabilities (D. L. Stone & Colella, 1996). Ren et al. (2008) meta-analyzed the results of 14 studies examining future performance expectations that might lead to lessened opportunity. They found that performance expectations were lower for people with disabilities than for control groups of people without disabilities, holding past performance constant (d = –.14). They found this negative effect to be greater for people with mental disabilities than physical disabilities, for men compared with women, and in field studies compared with laboratory studies. What is most interesting about these results is that the field effects were so much larger than the laboratory effects, contrary to criticisms that laboratory discrimination studies may inflate bias (Landy, in press). Colella (1996) posited that such low performance expectations would lead to unrealistic feedback, fewer growth opportunities, and poorer promotion rates for people with disabilities. We know of no research that directly examines the impact of low performance expectations on these outcomes; however, there is research that does suggest that people with disabilities experience these problems at work. In a much-cited study, Hastorf, Northcraft, and Piciotto (1979) found that nondisabled participants gave disabled performers unrealistically positive feedback. They attributed their results to the norm to be kind; however, lower performance expectations could also result in inflated feedback. Furthermore, paternalistic and patronizing behavior has often been cited as a complaint of people with disabilities (Ainsley, 1988; Colella & Stone, 2005) when others are overzealous in their attempts to protect them from unpleasantness. Some data do suggest that people with disabilities experience fewer growth opportunities, in terms of training and challenging job assignments, than people without disabilities. In a controlled laboratory study, Colella and Varma (1999) found that raters were less likely to assign a telemarketer with a disability (visual and hearing impairments) to training opportunities than a telemarketer without a disability. Furthermore, studies that examine the quality of work of people with disabilities report that people with disabilities tend to experience jobs 492

with less autonomy and decision making as well as jobs that require less education than comparable nondisabled samples (Grzywacz & Dooley, 2003; Schur, Kruse, Blasi, & Blanck, 2009; Yelin & Trupin, 2003). The third highest reason for filing a discrimination claim under the ADA is unequal terms and conditions of employment (28,528 allegations from 1992 through 2003), with another 4,516 allegations of unequal job assignments and 1,675 allegations of refusing training (McMahon et al., 2005; McMahon & Shaw, 2005). Finally, people with disabilities also appear to be at a disadvantage in terms of promotions. A largescale study of disability employment in the federal government (EEOC, 2008c) found that between fiscal years 2002 and 2006 the number and rate of promotions decreased by 25.2% for people with disabilities, compared with a decrease of 3.99% for federal employees with no disabilities. This lack of promotional opportunities was posited as one reason why people with disabilities leave federal government positions at a much greater rate than other federal employees. A large-scale study of all EEOC ADA allegations (see Bruyère et al., 2007; McMahon et al., 2005; McMahon & Shaw, 2005) found promotion to be one of the top 10 reasons why people filed discrimination charges under the ADA. The EEOC has attempted to clarify issues around performance standards, releasing guidance on the subject of performance evaluations and reasonable accommodations (EEOC, 2008b). In summary, unfair evaluations of actual work performance do not seem to be a major problem for people with disabilities. However, despite good past performance, evaluators tend to hold unfairly low expectations about future performance. These low performance expectations may be one reason why there is evidence that people with disabilities are prone to receive unrealistic feedback, unsatisfactory growth opportunities, and a disadvantage in promotion tournaments.

Inclusion Into the Organization In this section, we address the scant research that focuses on how well people with disabilities are included into the organization by its members. By inclusion, we mean the extent to which people with

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

disabilities are accepted, helped, and treated as others by their coworkers. Ample theoretical reasoning exists for why people with disabilities may suffer from exclusion (Colella, 1996; D. L. Stone & Colella, 1996). Reasons range from communication difficulties, anxiety, stigmatization, stereotyping, and distrust (see Table 15.1). However, aside from studies of people with severe cognitive disabilities in supported employment settings (e.g., Wehman, 2003), this is the least examined area in the field of disability and employment. Evidence for exclusion comes from social psychological laboratory experiments showing that nondisabled people avoid interacting with people with disabilities (Farina & Ring, 1965; Snyder, Kleck, Strenta, & Mentzer, 1979). More recent studies that frame the issue in a work context have found mixed results (Colella et al., 1998; D. L. Stone & Michaels, 1994). It is interesting to note that Colella et al. (1998) found that nondisabled people were less willing to work with a person with a disability only when their own rewards depended on the disabled person’s performance. The results of these lab studies do coincide with survey research of employers and people with disabilities. Bruyère et al. (2003) surveyed 865 private employers and 403 federal employers about the barriers facing employees with disabilities in their organizations. One of the most significant barriers was coworkers’ and supervisors’ attitudes and stereotypes (43% of federal respondents and 22% of private-sector respondents). Results of surveys asking people with disabilities about their interactions with and inclusion by supervisors and coworkers are mixed, with some finding that interactions with coworkers and supervisors are not a problem for people with disabilities (e.g., Schur, 2002) and others finding that these interpersonal relationships are more problematic than they are for people without disabilities (e.g., Crudden & McBroom, 1999; Uppal, 2005). A recent large-scale survey (N = 29,897) comparing people with disabilities with those without found that people with disabilities are less likely to be included on teams and receive informal training from coworkers, suggesting that there is a lack of inclusion (Schur, Kruse, Blasi, & Blanck, 2009).

Colella and Varma (2001) examined the extent to which people with disabilities form meaningful relationships with their supervisors compared with those without disabilities from a leader–member exchange framework (Graen, 1976). In both a laboratory experiment and a field survey, they found that when people with disabilities engaged in impression management behavior, their relationships with their supervisors were just as good as those without disabilities. However, in the absence of impression management behavior, people with disabilities had worse relationships with their supervisors than those without disabilities. It appears that engaging in impression management behavior mitigated the negative bias effects on supervisor– subordinate relationships. In summary, not much is known about the experiences of people with disabilities at work. Self-reports about inclusion are mixed, as are the results of laboratory studies focused on work contexts. However, research in this area has not progressed to systematically examining moderators of the effects of disability on workplace inclusion. One exception is Colella and Varma (1999), who found that employee impression management mitigated negative disability effects. It may be that any behavior that people engage in to enhance their inclusion and success in the workplace is even more important for people with disabilities than for those without them. Furthermore, there are a number of other factors that can influence the inclusion of people with disabilities, such as disability type, job type, reward structure, personality of coworkers, and so forth (cf. D. L. Stone & Colella, 1996). Clearly, this is one area of research in which I/O psychologists can make a significant contribution. Recently, scholars (Legnick-Hall, 2007; Schur, Kruse, & Blanck, 2005; Schur et al., 2009; Stensrud & Gilbride, 2006) have begun to focus on organizational culture as a moderator of the workplace experiences of people with disabilities. The WHO’s ICF has the potential to assist disability service administrators, policymakers, and practitioners with creating a transferable conceptual framework for defining indicators of successful outcomes in the integration of persons with disabilities into the workforce and community (Bruyère, 2005). 493

Colella and Bruyère

Copyright American Psychological Association. Not for further distribution.

Culture Given the findings reported above about the comparably unfavorable work experiences of people with disabilities and the recent focus on the demand side of disability employment issues, several scholars have begun to examine the role that organizational culture plays in determining the work experiences of people with disabilities. Cultural factors include an inclusive, diversity-friendly climate; policies and practices that allow for flexible work arrangements; top management support of the inclusion of people with disabilities; and shared values concerning the worth of all employees. This issue has been studied using a variety of methodologies. Legnick-Hall (2007) conducted case studies of seven companies recognized as being exemplary employers of people with disabilities. On the basis of these case studies, Legnick-Hall came to several conclusions about the aspects of an organizational culture that is disability-friendly. These are presented in the Appendix. Schur et al. (2009) surveyed almost 30,000 employees in 14 companies and determined that people with disabilities were paid less and had less desirable job experiences than employees without disabilities. However, they also found that these results did not hold across work sites (n = 175). They found that in work sites where employees (including those without disabilities) reported a fair and responsive climate, differences among the experiences between employees with disabilities and those without disappeared. As workplace climate became less fair and responsive, the gap between employees with disabilities and those without increased. The results of Schur et al.’s (2009) study are intriguing. However, it should be noted that the data came from a survey that was not designed to specifically test the effects of culture on the workplace welfare of people with disabilities. Stensrud and Gilbride (2006) developed the Employment Opportunity Survey to help rehabilitation specialists find organizations that are most responsive to hiring and retaining people with disabilities. This instrument could possibly be used to assess disability culture in future empirical studies. In conclusion, most past research has focused on micro factors and processes that result in disability 494

discrimination (e.g., stereotypes, fear of costs) or more macro-economic factors (e.g., effects of insurance). By focusing on organizational-level factors, such as culture, current, and hopefully future, research is expanding what we know about disability effects in the workplace. To some extent, what goes on in the organization has been ignored by disability and employment researchers. More recent research is addressing this problem. Furthermore, most research on disability has taken place at one level (either micro or macro). This research suggests that future research on disability and employment should be addressed from a multilevel analysis perspective. CONCLUSION We opened this chapter by stating that not much disability research has been conducted in the field of I/O psychology. However, this review makes it clear that I/O psychologists have a great deal to add to our knowledge about how to integrate people with disabilities into the workforce and ensure that they enjoy the same privileges as everyone else. Disability employment research is taking a new turn by focusing on the employer’s perspective, rather than on rehabilitation, legal, or economic perspectives. It is from this demand-side perspective that I/O psychologists can make the greatest contribution because of the field’s knowledge of hiring, workplace integration, discrimination, justice, and diversity. In contrast, if the field of I/O psychology focuses more on disability and employment, then much can be added to our literature in terms of new content and impact. Three particular areas stand out: accommodation, the aging workforce, and the impact on human welfare. This issue of accommodation was discussed in detail earlier in the chapter. Suffice it to say that accommodation brings a relatively new phenomenon into consideration, particularly when we usually define justice as being treated equitably or equally. Furthermore, the issue of accommodation is not limited to people with disabilities. Others, such as parents or older workers, may need accommodations. The aging workforce is likely to result in increasing numbers of workers with disabilities who may have difficulty staying employed. The U.S. Census Bureau

Copyright American Psychological Association. Not for further distribution.

Disability and Employment

projects that the 45–54- and 55–64-year-old population in the United States will have grown by nearly 44.2 million (17%) and 35 million (39%) between 2000 and 2010 (U.S. Census Bureau, 2004). By the year 2010, this group will account for nearly half (44%) of the working-age population (20–64 years), and the number of people with disabilities between the ages of 50 and 65 will almost double (Weathers, 2006). Accommodation policies and practices and increased understanding of workplace diversity and disability employment nondiscrimination readily lend themselves to addressing the challenges that employers will face with an aging workforce and the increasing prevalence of disability that these demographics bring. Proactive workplace education about employment accommodation and ways to maximize the productivity of an aging workforce can significantly contribute to aging-worker retention. Related research can also demonstrate which accommodations, workplace modifications, and policy and practice changes may positively impact the retention and productivity of an aging workforce. I/O psychology researchers are in a unique position to provide the design conceptualization, metrics, and analyses to test the array of interventions the United States will be exploring to keep our aging workforce healthy and fully engaged in the employment environment. I/O psychologists may, in addition, assist with implementation of these interventions. A focus globally on these issues by the I/O psychology field is also imperative. One example for the need of an international perspective on accommodation and analysis of disability and aging-related workplace factors is the increasing effort in the disability field to find globally acceptable measures for the experience of people with disabilities. Of particular interest is the need to extend the paradigm beyond the medical model to an analysis of how environmental factors impact people with disabilities. There is a growing consensus within the disability research community that environmental factors have a profound effect on the lives of people with disabilities. Elements of the environment are cited for increasing impairments, adding activity limitations, presenting participation restrictions, and contributing to a variety of secondary conditions. However, few methods exist for

measuring its effect. The WHO’s ICF highlights this environmental impact and affords us the beginnings of a schema upon which we may be able to build a globally recognized classification of the experience of people with disabilities (Brooke, Whiteneck, & Terrill, 2004). I/O psychologists can certainly contribute to these efforts, particularly in assisting in finding ways to better quantify the employment experience of people with disabilities, such as workplace inclusion, job satisfaction, and equitable access to the terms and conditions of employment. Finally, with the recent passing of the United Nations Convention on the Rights of Persons With Disabilities (UNCRPD), there is now even greater international contextual support for pursuing the interests of persons with disabilities. The purpose of this Convention is to “promote, protect and ensure the full and equal enjoyment of all human rights by persons with disabilities” (see http://www. un.org/disabilities/convention/conventionfullshtml, Article 1, ¶ 1). It covers a number of relevant key areas, such as equality and nondiscrimination (Article 5), equal recognition before the law (Article 12), employment (Article 27), and participation in political life (Article 29). The Convention marks a shift in thinking about disability from a social welfare concern to a human rights issue, which acknowledges that societal barriers and prejudices are themselves disabling. Most of the research cited in this chapter is based on North American samples and organizations or other English-speaking countries (England, Australia). What is missing is work that examines how disability may be viewed differently in different cultural contexts and what impact such differences may have on adopting an international perspective. The UNCRPD should serve as a catalyst for such work. Future research and implementation efforts made by I/O psychology in more effective employment accommodation, inclusion, and advancement for people with disabilities can now ultimately have a significant international influence. The UNCRPD affords a global platform upon which to transfer the knowledge gained in increasing workplace participation for applicants and employees with disabilities in corporate America to the ever-increasing multinational organizational environment and perhaps 495

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

even ultimately into the informal economies of developing countries. Many years of opportunity for significant national and international contribution in this area await.

Americans With Disabilities Act Amendments Act of 2008, Pub. L. No. 110-325, (S. 3406).

APPENDIX: ASPECTS OF DISABILITYFRIENDLY CULTURES BASED ON LEGNICK-HALL (2007) ■ Top management support for the employment of persons with disabilities. ■ Emphasize a business case for employing people with disabilities. ■ Align disability initiatives with corporate strategy. ■ Develop a disability philosophy that focuses on abilities rather than on disability. ■ Have systematic accommodation policies and procedures. ■ Address supervisors’ and coworkers’ negative attitudes and concerns. ■ Partner with community resources and schools to recruit people with disabilities. ■ Create affinity groups, task forces, and information clearinghouses related to disability. ■ Address disability issues in employee diversity training and orientations. ■ Monitor how people with disabilities fare in the organization. ■ Publicize achievements of people with disabilities. ■ Examine personnel policies that could lead to more hiring of people with disabilities (e.g., flextime).

Avery, D. R., & McKay, P. F. (2006). Target practice: An organizational impression management approach to attracting minority and female job applicants. Personnel Psychology, 59, 157–187.

References Acemoglu, D., & Angrist, J. (2001). Consequences of employment protection? The case of the Americans With Disabilities Act. Journal of Political Economy, 109, 915–957. Ainsley, S. C. (1988). Aging and new vision loss: Disruptions here and now. Journal of Social Issues, 44, 79–94. Ainspan, N. D. (2006, May). Employers’ opinions and attitudes of employing people with disabilities. Paper presented at the annual meeting of the Society for Industrial and Organizational Psychology, Dallas, TX. Allaire, S. H., Li, W., & LaValley, M. P. (2003). Work barriers experienced and job accommodation used by persons with arthritis and other rheumatic diseases. Rehabilitation Counseling Bulletin, 46, 147–156. Americans With Disabilities Act of 1990, 42 U.S.C.A. § 12101 et seq. (West, 1993). 496

Anderson, S. E., & Williams, L. L. (1996). Interpersonal, job, and individual factors related to help seeking processes at work. Journal of Applied Psychology, 81, 282–296.

Baldridge, D. C., & Veiga, J. F. (2001). Toward greater understanding of the willingness to request an accommodation: Can requesters’ beliefs disable the ADA? Academy of Management Review, 26, 85–99. Baldridge, D. C., & Veiga, J. F. (2006). The impact of anticipated social consequences on recurring disability accommodation requests. Journal of Management, 32, 158–179. Balser, D. B. (2007). Predictors of workplace accommodations for employees with mobility-related disabilities. Administration and Society, 39, 656–683. Barrington, L., & Troske, K. (2001). Workforce diversity and productivity: An analysis of employer–employee matched data. Washington, DC: U.S. Census Bureau. Becker, D. R., Baker, S. R., Carlson, L., Flint, L., Howell, R., Lindsay, S., et al. (2007). Critical strategies for implementing supported employment. Journal of Vocational Rehabilitation, 27, 13–20. Belgrave, F. Z., & Mills, J. (1981). Effect upon desire for social interaction with a physically disabled person of mentioning the disability in different contexts. Journal of Applied Social Psychology, 11, 44–57. Benson v. Northwest Airlines, 62 F.3d 1108 1114 (8th Cir. 1995). Bjelland, M. J., Bruyère, S. M., Houtenville, A. J., RuizQuintanilla, S. A., & Webber, D. A. (2008). Trends in disability employment discrimination claims: Implications for rehabilitation counseling practice, administration, training, and research. Unpublished manuscript, Cornell University, Employment and Disability Institute, Ithaca, NY. Bjelland, M. J., Burkhauser, R. V., & Houtenville, A. J. (2008). 2008 progress report on the economic wellbeing of working-age people with disabilities. Ithaca, NY: Cornell University, Employment and Disability Institute, Rehabilitation Research and Training Center for Economic Research on Employment Policy for Persons With Disabilities. Blanck, P. (1996). Empirical study of the Americans With Disabilities Act: Employment issues from 1990–1994. Behavioral Sciences and the Law, 14, 5–27. Blanck, P., Schur, L., Kruse, D., Schwochau, S., & Song, C. (2003). Calibrating the impact of the ADA’s employ-

Disability and Employment

ment provisions. Stanford Law & Policy Review, 14.2, 267–290. Blanck, P., Schwochau, S., & Song, C. (2003). Is it time to declare the ADA a failed law? In D. C. Stapleton & R. W. Burkhauser (Eds.), The decline in employment of people with disabilities: A policy puzzle (pp. 301–337). Kalamazoo, MI: Upjohn Institute for Employment Research.

Copyright American Psychological Association. Not for further distribution.

Bordieri, J. E., & Drehmer, D. E. (1987). Attribution of responsibility and predicted social acceptance of disabled workers. Rehabilitation Counseling Bulletin, 30, 218–226. Braddock, D., & Bachelder, L. (1994). The glass ceiling and persons with disabilities. Washington, DC: U.S. Department of Labor, Glass Ceiling Commission. Brannick, M. T., Brannick, J. P., & Levine, E. L. (1992). Job analysis, personnel selection, and the ADA. Human Resource Management Review, 3, 171–182. Brecher, E., Bragger, J., & Kutcher, E. (2006). The structured interview: Reducing biases toward job applicants with physical disabilities. Employee Responsibilities and Rights Journal, 18, 155–170. Brooks, C. A., Whiteneck, G. L., & Terrill, M. (2004, June). The relationship of environmental characteristics to participation by people with disabilities in Colorado: A county level analysis of secondary data. Poster session presented at the 10th North American Collaborating Center Conference on ICF, Halifax, Nova Scotia, Canada. Bruce, C., & Sanford, J. (in press). Assessment for workplace accommodations. In E. Mpofu & T. Oakland (Eds.), Assessment in rehabilitation and health. Upper Saddle River, NJ: Allyn & Bacon. Bruyère, S. M. (2000). Disability employment policies and practices in private and federal sector organizations. Ithaca, NY: Cornell University, School of Industrial and Labor Relations. Retrieved September 16, 2008, from http://digitalcommons.ilr.cornell.edu/edicollect/63/ Bruyère, S. M. (2005). Using the International Classification of Functioning, Disability and Health (ICF) to promote employment and community integration in rehabilitation. Rehabilitation Education, 19, 105–117. Bruyère, S. M., Erickson, W., & Ferrentino, J. (2003). Identity and disability in the workplace. William and Mary Law Review, 44, 1173–1196. Retrieved September 17, 2008, from http://digitalcommons.ilr. cornell.edu/edicollect/125 Bruyère, S. M., Erickson, W., & Horne, R. (2002). Disability employment policies and practices in U.S. federal government agencies: EEO/HR and supervisor perspectives. Ithaca, NY: Cornell University, School of Industrial and Labor Relations—Extension Division, Program on Employment and Disability. Retrieved September 15, 2008, from http:// digitalcommons .ilr.cornell.edu/edicollect/66

Bruyère, S. M., Erickson, W., & VanLooy, S. (2000, Autumn). HR’s role in managing disability in the workplace. Employment Relations Today, 47–66. Bruyère, S. M., Erickson, W., & VanLooy, S. (2005). Information technology and the workplace: Implications for persons with disabilities. Disability Studies Quarterly, 25(2). Retrieved September 17, 2008, from http://digitalcommons.ilr.cornell.edu/edicollect/98 Bruyère, S. M., Houtenville, A. J., Ruiz-Quintanilla, S. A., Schwab, S. J., Daly-Rooney, R., Taylor, B., et al. (2007, June). Comparison of employment disability discrimination claims with other statutes across U.S. Equal Opportunity Commission and Fair Employment Practice Agencies nationally. Unpublished manuscript, Cornell University, Ithaca, NY. Bruyère, S. M., VanLooy, S., & Peterson, D. (2005). The International Classification of Functioning, Disability and Health (ICF): Contemporary literature overview. Rehabilitation Psychology, 50, 113–121. Burkhauser, R. V., Butler, J. S., & Kim, Y. (1995). The importance of employer accommodation on the job duration of workers with disabilities: A hazard model approach. Labour Economics, 2, 109–130. Burkhauser, R. V., Butler, J. S., & Weathers, R. R., II. (2001/2002). How policy variables influence the timing of applications for Social Security Disability. Social Security Bulletin, 64, 52–83. Burkhauser, R. V., & Daly, M. C. (1996). Employment and economic well-being following the onset of disability. In J. L. Mashaw, V. Reno, R. Burkhauser, & M. Berkowitz (Eds.), Disability, work and cash benefits (pp. 59–102). Kalamazoo, MI: Upjohn Institute for Employment Research. Burkhauser, R. V., Daly, M. C., & Houtenville, A. J. (2002). Self-reported work-limitation data: What they can and cannot tell us. Demography, 39, 541–555. Burkhauser, R. V., & Houtenville, A. J. (2006). A guide to disability statistics from the Current Population Survey— Annual social and economic supplement (March CPS). Ithaca, NY: Cornell University, Rehabilitation Research and Training Center on Disability Demographics and Statistics. Available from http://digitalcommons.ilr. cornell.edu/edicollect/1233/ Burkhauser, R. V., Houtenville, A. J., & Wittenburg, D. C. (2003). A user’s guide to current statistics on the employment of people with disabilities. In D. C. Stapleton & R. V. Burkhauser (Eds.), The decline in employment of people with disabilities (pp. 26–86). Kalamazoo, MI: Upjohn Institute for Employment Research. Butterfield, T. M., & Ramseur, J. H. (2004). Research and case study findings in the area of workplace accommodations provisions for assistive technology: A literature review. Technology and Disability, 16, 201–210. 497

Colella and Bruyère

Campolieti, M. (2004). The correlates of accommodation for permanently disabled workers. Industrial Relations, 43, 546–572.

Copyright American Psychological Association. Not for further distribution.

Carver, C. S., Glass, D. C., Snyder, M. L., & Katz, I. (1977). Favorable evaluation of stigmatized others. Personality and Social Psychology Bulletin, 3, 97–106.

Colella, A., & Stone, D. L. (2005). Workplace discrimination toward persons with disabilities: A call for some new research directions. In R. Dipboye & A. Colella (Eds.), Discrimination at work: The psychological and organizational bases (pp. 227–253). Mahwah, NJ: Erlbaum.

Cascio, W., & Boudreau, J. (2008). Investing in people: Financial impact of human resource initiatives. Upper Saddle River, NJ: Pearson Education, Inc.

Colella, A., & Varma, A. (1999). Disability–job fit stereotypes and the evaluation of persons with disabilities at work. Journal of Occupational Rehabilitation, 9(2), 79–95.

Center, C. (in press). Law and job accommodations in mental health disability. In I. Schultz & S. Rogers (Eds.), Handbook of job accommodations in mental health. New York: Springer.

Colella, A., & Varma, A. (2001). The impact of subordinate disability on leader–member exchange relationships. Academy of Management Journal, 44, 304–315.

Cesare, S. J., Tannenbaum, R. J., & Dalessio, A. (1990). Interviewers’ decisions related to applicant handicap type and rater empathy. Human Performance, 3, 157–171.

Collignon, F. C. (1986). The role of reasonable accommodation in employing disabled persons in private industry. In M. Berkowitz & M. A. Hill (Eds.), Disability and the labor market (pp. 196–241). Ithaca, NY: ILR Press.

Chan, F., Cheing, G., Chan, J. Y. C., Rosenthal, D. A., & Chronister, J. (2006). Predicting employment outcomes of rehabilitation clients with orthopedic disabilities: A CHAID analysis. Disability and Rehabilitation, 28, 257–270.

Crudden, A., & McBroom, L. W. (1999). Barriers to employment: A survey of employed persons who are visually impaired. Journal of Visual Impairment and Blindness, 93, 341–350.

Chan, F., McMahon, B. T., Cheing, G., Rosenthal, D. A., & Bezyak, J. (2005). Drivers of workplace discrimination against people with disabilities: The utility of attribution theory. Work, 25, 77–88.

Czajka, J. M., & DeNisi, A. S. (1988). Effects of emotional disability and clear performance standards on performance ratings. Academy of Management Journal, 31, 394–404.

Chen, S., & van der Klaauw, W. (2008). The work disincentive effects of the disability insurance program in the 1990s. Journal of Econometrics, 142, 757–784.

Davis, F. (1961). Deviance disavowal: The management of strained interaction by the visually handicapped. Social Problems, 9, 120–132.

Cleveland, J. N., Barnes-Farrell, J. L., & Ratz, J. M. (1997). Accommodation in the workplace. Human Resource Management Review, 7, 77–107.

DeLeire, T. (2000). The wage and employment effects of the Americans With Disabilities Act. Journal of Human Resources, 35, 693–715.

Colella, A. (1996). The organizational socialization of employees with disabilities: Theory and research. In G. R. Ferris (Ed.), Research in personnel and human resources management (Vol. 14, pp. 351–417). Greenwich, CT: JAI Press.

DeLeire, T. (2003). The Americans With Disabilities Act and the employment of people with disabilities. In D. C. Stapleton & R. W. Burkhauser (Eds.), The decline in employment of people with disabilities: A policy puzzle (pp. 259–378). Kalamazoo, MI: Upjohn Institute for Employment Research.

Colella, A. (2001). Coworker distributive fairness judgments of workplace accommodations of employees with disabilities. Academy of Management Review, 26, 100–117.

Donohue, J. J., III. (1994). Employment discrimination law in perspective: Three concepts of equality. Michigan Law Review, 92, 2583–2612.

Colella, A., DeNisi, A., & Varma, A. (1997). Appraising the performance of employees with disabilities: A review and model. Human Resource Management Review, 7, 27–53. Colella, A., DeNisi, A. S., & Varma, A. (1998). The impact of ratee’s disability on performance judgments and choice as partner: The role of disability– job fit stereotypes and interdependence of rewards. Journal of Applied Psychology, 83, 102–111. Colella, A., Paetzold, R., & Belliveau, M. (2004). Factors affecting coworkers’ procedural justice inferences of the workplace accommodation of employees with disabilities. Personnel Psychology, 57, 1–23. 498

Duckett, P. S. (2000). Disabling employment interviews: Warfare to work. Disability and Society, 15, 1019–1039. Duston, S. (2001). Definition of disability under the ADA: A practical overview and update. Ithaca, NY: Cornell University. Retrieved May 15, 2009, from http://www. ilr.cornell.edu/edi/hr_tips/article_1.cfm?b_id=27 Equal Employment Opportunity Commission. (1992). A technical assistance manual on the employment provisions (Title I) of the American With Disabilities Act. Washington, DC: Author. Equal Employment Opportunity Commission. (2002). ADA technical assistance manual addendum. Washington, DC: Author. Retrieved September 20,

Disability and Employment

2008, from http://www.eeoc.gov/policy/docs/ adamanual_add.html Equal Employment Opportunity Commission. (2008a). The ADA: Your responsibilities as an employer. Washington, DC: Author. Retrieved July 22, 2008, from http://www.eeoc.gov/facts/ada17.html

Copyright American Psychological Association. Not for further distribution.

Equal Employment Opportunity Commission. (2008b). The Americans With Disabilities Act: Applying performance and conduct standards to employees with disabilities. Washington, DC: Author. Retrieved September 17, 2008, from http://www.eeoc.gov/ facts/performance-conduct.html Equal Employment Opportunity Commission. (2008c). Improving the participation rate of people with targeted disabilities in the federal work force. Washington, DC: Author. Retrieved May 30, 1998. from http://www. eeoc.gov/federal/report/pwtd.html Equal Employment Opportunity Commission. (2008d). Notice concerning the Americans With Disabilities Act (ADA) Amendments Act of 2008. Retrieved May 15, 2009, from http://www.eeoc.gov/ada/amendments_ notice.html Equal Employment Opportunity Commission. (2009). The ADA: Your responsibilities as an employer. Washington, DC: Author. Retrieved May 19, 2009, from http://www.eeoc.gov/facts/ada17.html Erickson, W. A., & Houtenville, A. J. (2005). A guide to disability statistics from the 2000 decennial census. Ithaca, NY: Cornell University, Rehabilitation Research and Training Center on Disability Demographics and Statistics. Available from http://digitalcommons.ilr. cornell.edu/edicollect/187/ Fabian, E. S., Waterworth, A., & Ripke, B. (1993). Reasonable accommodations for workers with serious mental illness: Type, frequency, and associated outcomes. Psychosocial Rehabilitation Journal, 17, 163–172. Farina, A., & Felner, R. D. (1973). Employment interviewer reactions to former mental patients. Journal of Abnormal Psychology, 82, 268–272. Farina, A., & Ring, K. (1965). The influence of perceived mental illness on interpersonal relations. Journal of Abnormal Psychology, 70, 47–51. Farina, A., Sherman, M., & Allen, J. G. (1968). Role of physical abnormality in interpersonal perception and behavior. Journal of Abnormal Psychology, 6, 590–593. Fichten, C. S., & Amsel, R. (1986). Trait attributions about college students with a physical disability: Circumplex analyses and methodological issues. Journal of Applied Social Psychology, 16, 410–427. Fine, S. A. (1989). Functional job analysis scales: A desk manual. Milwaukee, WI: Author.

Florey, A. T., & Harrison, D. A. (2000). Responses to informal accommodation requests from employees with disabilities: Multistudy evidence on the willingness to comply. Academy of Management Journal, 43, 224–233. Gaunt, P. (2007). Hewlett-Packard. In M. L. LegnickHall (Ed.), Hidden talent: How leading companies hire, retain, and benefit from people with disabilities (pp. 37–47). Westport, CT: Praeger. Gervey, R., & Kowal, H. (2005). The job developer’s presence in the job interview: Is it helpful or harmful to persons with disabilities seeking employment? Psychiatric Rehabilitation Journal, 29, 128–131. Golosov, M., & Tsyvinski, A. (2006). Designing optimal disability insurance: A case for asset testing. Journal of Political Economy, 114, 257–279. Gouvier, W. D., Steiner, D. D., Jackson, W. T., Schlater, D., & Rain, J. (1991). Employment discrimination against handicapped job applicants: An analog study of the effects of neurological causation, visibility of handicap, and public contact. Rehabilitation Psychology, 36, 121–129. Graen, G. (1976). Role making processes within complex organizations. In M. D. Dunette (Ed.), Handbook of industrial and organizational psychology (pp. 1201–1245). Chicago: Rand McNally. Grzywacz, J., & Dooley, D. (2003, April). “Good jobs” to “bad jobs”: Replicated evidence of employment continuum from two large surveys. Social Science and Medicine, 1749–1760. Hahn, H. (1988). The politics of physical differences: Disability and discrimination. Journal of Social Issues, 44, 39–47. Hastorf, A. H., Northcraft, G. B., & Piciotto, S. R. (1979). Helping the handicapped: How realistic is the performance feedback received by the physically handicapped. Personality and Social Psychology Bulletin, 5, 373–376. Hastorf, A. H., Wildfogel, J., & Cassman, T. (1979). Acknowledgment of handicap as a tactic in social interaction. Journal of Personality and Social Psychology, 37, 1790–1797. Hebl, M. R., & Kleck, R. E. (2002). Acknowledging one’s stigma in the interview setting: Effective strategy or liability? Journal of Applied Social Psychology, 32, 223–249. Hebl, M. R., & Skorinko, J. L. (2005). Acknowledging one’s physical disability in the interview: Does “when” make a difference? Journal of Applied Social Psychology, 35, 2477–2492. Hernandez, B., Keys, C., & Balcazar, F. (2000). Employer attitudes toward workers with disabilities and their ADA employment rights: A literature review. Journal of Rehabilitation Research, 66(4), 4–16. 499

Copyright American Psychological Association. Not for further distribution.

Colella and Bruyère

practitioners. Equal Opportunities International, 25, 354–372.

Highhouse, S., Zickar, M. J., Thorsteinson, T. J., Stierwalt, S. L., & Slaughter, J. E. (1999). Assessing company employment image: An example in the fast food industry. Personnel Psychology, 52, 159–172.

Kralik v. Durbin, No. 97-3089 (U.S. Ct. App., 3rd Cir., 1997).

Houtenville, A. (2003). Frequently asked questions: Disability statistics. Retrieved May 15, 2009, from http://www.ilr.cornell.edu/edi/disabilitystatistics/ faq.cfm?n7

Krefting, L. A., & Brief, A. P. (1976). The impact of applicant disability on evaluative judgments in the selection process. Academy of Management Journal, 19, 675–680.

Houtenville, A., & Burkhauser, R. (2004). Did the employment of people with disabilities decline in the 1990s, and was the ADA responsible: A replication and robustness check of Acemoglu and Angrist (2001). Ithaca, NY: Cornell University, Employment and Disability Institute, Rehabilitation Research and Training Center for Economic Research on Employment Policy for People With Disabilities. Retrieved September 18, 2008, from http://digitalcommons. ilr.cornell.edu/edicollect/91/

Kruse, D., & Schur, L. (2003). Does the definition affect the outcome? In D. C. Stapleton & R. W. Burkhauser (Eds.), The decline in employment of people with disabilities: A policy puzzle (pp. 279–300). Kalamazoo, MI: Upjohn Institute for Employment Research.

Houtenville, A., Stapleton, D., Weathers, R., & Burkhauser, R. (in press). The status of the workingage population with disabilities: What current data tell us and options to improve the data. Kalamazoo, MI: Upjohn Institute for Employment Research.

Latham, G. G. P., & Whyte, G. (1994). The futility of utility analysis. Personnel Psychology, 47, 31–46.

Jasper, C. R., & Klassen, M. L. (1990). Perceptions of salespersons’ appearance and evaluation of job performance. Perceptual and Motor Skills, 71, 563–566. Job Accommodation Network. (1997). Accommodation benefit/cost data. Unpublished manuscript, West Virginia University, Morgantown. Johnson, R., & Heal, L. W. (1976). Private employment agency responses to the physically handicapped applicant in a wheelchair. Journal of Applied Rehabilitation Counseling, 7, 12–21. Jolls, C. (2000). Accommodation mandates. Stanford Law Review, 53, 223–306. Jones, E. E., Farina, A., Hastorf, A. H., Markus, H., Miller, D. T., & Scott, R. A. (1984). Social stigma: The psychology of marked relationships. San Francisco: W. H. Freeman. Jutai, J., Fuhrer, M., Demers, L., Scherer, M., & DeRuyter, F. (2005). Toward a taxonomy of assistive technology device outcomes. American Journal of Physical Medicine and Rehabilitation, 84, 294–302. Karraker et al. v. Rent-a-Center, Inc., No. 02-2026, 2005 U.S. App. LEXIS 11142, 411 F.3d 831 (7th Cir.). Katz, I. (1979). Some thoughts about the stigma notion. Personality and Social Psychology Bulletin, 5, 447–460.

Landy, F. J. (in press). Stereotypes, bias, and personnel decisions: Strange and stranger. Industrial and Organizational Psychology: Perspectives on Science and Practice.

Lee, B. (1996). Legal requirements and employer responses to accommodating employees with disabilities. Human Resource Management Review, 6, 231–251. Lee, B. (1997). When the going gets tough, do the tough ask for help? Help seeking and power motivation in organizations. Organizational Behavior and Human Decision Processes, 72, 336–363. Lee, B. (2003). A decade of the Americans With Disabilities Act: Judicial outcomes and unresolved problems. Industrial Relations, 42, 11–30. Lee, B., & Newman, K. (1995). Employer responses to disability: Preliminary evidence and a research agenda. Employee Responsibilities and Rights Journal, 8, 209–229. Legnick-Hall, M. L. (2007). Hidden talent: How leading companies hire, retain, and benefit from people with disabilities. Westport, CT: Praeger. Lerner, M. J. (1980). The belief in a just world: A fundamental delusion. New York: Plenum Press. Livermore, G., & She, P. (2007). Limitations in the national disability data system. Ithaca, NY: Cornell University, Employment and Disability Institute, Rehabilitation Research and Training Center on Disability Demographics and Statistics. Livneh, H. (1982). On the origins of negative attitudes toward people with disabilities. Rehabilitation Literature, 43, 338–347.

Katz, I. (1981). Stigma: A social psychological analysis. Hillsdale, NJ: Erlbaum.

Makas, E. (1988). Positive attitudes towards disabled people: Disabled and nondisabled persons’ perspectives. Journal of Social Issues, 44, 49–61.

Knapp, D. E., Faley, R. H., & Long, L. K. (2006). The Americans With Disabilities Act: A review and synthesis of the legal literature with implications for

Marchioro, C. A., & Bartels, L. K. (1994). Perceptions of a job interviewee with a disability. Journal of Social Behavior and Personality, 9, 383–394.

500

Disability and Employment

Mashaw, J., & Reno, V. P. (Eds.). (1996). Balancing security and opportunity: The challenge of disability income policy. Washington, DC: National Academy of Social Insurance.

Copyright American Psychological Association. Not for further distribution.

McDonald-Wilson, K. L., Rogers, E. S., & Massaro, J. (2003). Identifying relationships between functional limitations, job accommodations, and demographic characteristics of persons with psychiatric disabilities. Journal of Vocational Rehabilitation, 18, 15–24. McDonald-Wilson, K. L., Rogers, E. S., Massaro, J., Lyass, A., & Crean, T. (2002). An investigation of reasonable workplace accommodations for people with psychiatric disabilities: Quantitative findings from a multi-site study. Community Mental Health Journal, 38, 35–50. McFarlin, D. B., Song, J., & Sonntag, M. (1991). Integrating the disabled into the work force: A survey of Fortune 500 company attitudes and practices. Employee Responsibilities and Rights Journal, 4, 107–122. McKay, P. F., & Avery, D. R. (2006). What has race got to do with it? Unraveling the role of racioethnicity in job seekers’ reactions to site visits. Personnel Psychology, 59, 395–429. McLaughlin, M. E., & Gray, D. A. (1998, August). The individual perspective: Disability disclosure and the usefulness of adjustments and accommodations for employees with multiple sclerosis. Paper presented at the annual meeting of the Academy of Management, San Diego, CA. McMahon, B. T., Edwards, R., Rumrill, P. D., & Hursh, N. (2005). An overview of the EEOC ADA research project. Work, 25, 1–7. McMahon, B. T., & Shaw, L. R. (2005). Workplace discrimination and disability. Journal of Vocational Rehabilitation, 23, 137–143. McMenamin, T., Miller, S. M., & Polivka, A. E. (2006). Discussion and presentation of the disability test results from the current population survey. Unpublished manuscript, Bureau of Labor Statistics. Miceli, N. S., Harvey, M., & Buckley, M. R. (2001). Potential discrimination in structured employment interviews. Employee Responsibilities and Rights Journal, 13(1), 15–38. Milchus, K., & Bruce, C. (in press). Computer access in the workplace. In A. Helal, M. Mokhtari, & B. Abdulrazak (Eds.), Smart technology for aging, disability and independence: Computing and engineering design and applications. Hoboken, NJ: Wiley. Mills, J., Belgrave, F. Z., & Boyer, K. M. (1984). Reducing avoidance of social interaction with a physically disabled person by mentioning the disability following a request for aid. Journal of Applied Social Psychology, 14, 1–11.

Mitchell, K. E., Alliger, G. M., & Morfopolous, R. (1997). Toward an ADA-appropriate job analysis. Human Resource Management Review, 7, 5–26. Mitchell, T. L., & Kovera, M. B. (2006). The effects of attribution of responsibility and work history on perceptions of reasonable accommodation. Law and Human Behavior, 30, 733–748. Nordstrom, C. R., Huffaker, B. J., & Williams, K. B. (1998). When physical disabilities are not liabilities: The role of applicant and interviewer characteristics on employment interview outcomes. Journal of Applied Social Psychology, 28, 283–306. Paetzold, R. L. (2005). Law & psychiatry: Mental illness and reasonable accommodations at work—Definition of mental disability under the ADA. Psychiatry Services, 56, 1188–1190. Paetzold, R. L., Garcia, M. F., Colella, A., Ren, L. R., Triana, M. D. C., & Ziebro, M. (2009). Perception of people with disabilities: When is accommodation fair? Basic and Applied Social Psychology, 30, 27–35. Piner, K. E., & Kahle, L. R. (1984). Adapting to the stigmatizing label of mental illness: Forgone but not forgotten. Journal of Personality and Social Psychology, 47, 805–811. Pransky, G., Gatchel, R., Linton, S. J., & Loisel, P. (2005). Improving return to work research. Journal of Occupational Rehabilitation, 15, 453–457. Rehabilitation Research and Training Center on Disability Demographics and Statistics. (2007). 2006 disability status report. Ithaca, NY: Cornell University. Reilly, N. P., Bocketti, S. P., Maser, S. A., & Wennet, C. L. (2006). Benchmarks affect perceptions of prior disability in a structured interview. Journal of Business and Psychology, 20, 489–500. Ren, L. R., Paetzold, R. L., & Colella, A. (2008). A metaanalysis of experimental studies on the effects of disability on human resource judgment. Human Resource Management Review, 18, 191–203. Roberts, M. M., & Pratt, C. W. (2007). Putative evidence of employment readiness. Psychiatric Rehabilitation Journal, 30, 175–181. Robinson, G., & Deschant, K. (1997). Building a business case for diversity. The Academy of Management Executive, 11(3), 21–31. Rose, G. L., & Brief, A. P. (1979). Effects of handicap and job characteristics on selection evaluations. Personnel Psychology, 32, 385–392. Rousseau, D. M. (2001). The idiosyncratic deal: Flexibility versus fairness? Organizational Dynamics, 29, 260–273. Rousseau, D. M. (2004). Idiosyncratic deals: When workers bargain for themselves. New York: Sharpe. 501

Colella and Bruyère

Rousseau, D. M., Ho, V. T., & Greenberg, J. (2006). Ideals: Idiosyncratic terms in employment relationships. Academy of Management Review, 31, 977–994. Schartz, H., Hendricks, D. J., & Blanck, P. (2006). Workplace accommodations: Evidence-based outcomes. Work, 27, 345–354.

Copyright American Psychological Association. Not for further distribution.

Schramm, J. (2007). On-line survey of SHRM members. In Employer perspectives on workers with disabilities: A national summit to develop a research agenda. Washington, DC: U.S. Department of Labor, Interagency Committee on Disability Research. Schur, L. (2002). Discrimination in the workplace: Perceptions and responses of people with disabilities. In Proceedings of the 54th Annual Meeting: January 4–6, 2002, Atlanta (pp. 40–48). Champaign: University of Illinois Press.

Survey (EOS). Paper presented at the annual meeting of the Society for Industrial and Organizational Psychology, Dallas, TX. Stoddard, S. (2006). Personal assistance services as a workplace accommodation. Work, 27, 363–369. Stone, C. I., & Sawatzki, B. (1980). Hiring bias and the disabled interviewee: Effects of manipulating work history and disability information of the disabled job applicant. Journal of Vocational Behavior, 16, 96–104. Stone, D. L., & Colella, A. (1996). A model of factors affecting the treatment of disabled individuals in organizations. Academy of Management Review, 21, 352–401.

Schur, L., Kruse, D., & Blanck, P. (2005). Corporate culture and the employment of people with disabilities. Behavioral Sciences and the Law, 23, 3–20.

Stone, D. L., & Michaels, C. (1994). Effects of nature of the disability and competitiveness of reward system on selection of disabled team members. Paper presented at the annual meeting of the Academy of Management, Dallas, TX.

Schur, L., Kruse, D., Blasi, J., & Blanck, P. (2009). Is disability disabling in all workplaces? Workplace disparities and corporate culture. Industrial Relations: A Journal of Economy and Society, 48, 381–410.

Stone, D. L., & Williams, K. J. (1997). The impact of the ADA on the selection process: Applicant and organizational issues. Human Resource Management Review, 7, 203–231.

She, P., & Stapleton, D. (2006). A review of disability data for the institutional population. Ithaca, NY: Cornell University, Employment and Disability Institute, Rehabilitation Research and Training Center on Disability Demographics and Statistics.

Stone-Romero, E. F. (2005). Personality-based stigmas and unfair discrimination in work organizations. In R. Dipboye & A. Colella (Ed.), Discrimination at work: The psychological and organizational bases (pp. 255–280). Mahwah, NJ: Erlbaum.

Snyder, M. L., Kleck, R. E., Strenta, A., & Mentzer, S. (1979). Avoidance of the handicapped: An attributional ambiguity analysis. Journal of Personality and Social Psychology, 37, 2297–2306.

Sutton v. United Airlines, 527 U.S. 471 (1999).

Social Security Administration. (2006). Disability evaluation under Social Security. Retrieved July 22, 2008, from http://www.ssa.gov/disability/professionals/ bluebook Social Security Advisory Board. (2003). The Social Security definition of disability. Retrieved July 22, 2008, from http://www.ssab.gov/SocialSecurity DefinitionofDisability.pdf Social Security Online. (2008, September). Disability evaluation under Social Security (SSA Publication No. 64-039, ICN 468600). Retrieved July 8, 2009, from http://www.ssa.gov/disability/professionals/ bluebook/general-info.htm Stapleton, D. C., & Burkhauser, R. W. (2003). The decline in employment of people with disabilities: A policy puzzle. Kalamazoo, MI: Upjohn Institute for Employment Research.

Tagalakis, V., Amsel, R., & Fichten, C. S. (1988). Job interview strategies for people with a visible disability. Journal of Applied Social Psychology, 18, 520–532. Taylor v. Principal Financial Group, Inc., 93 F.3d 155 (5th Cir. 1996). Tominey, E. (2001). Working effectively with people with learning disabilities. Ithaca, NY: Cornell University. Retrieved May 15, 2009, from http://www.ilr.cornell. edu/edi/hr_tips/article_1.cfm?b_id=29 Unger, D. D. (2002). Employers’ attitudes toward persons with disabilities in the workforce: Myths or realities? Focus on Autism and Other Developmental Disabilities, 17, 2–10. Uppal, S. (2005). Disability, workplace characteristics, and job satisfaction. International Journal of Manpower, 26, 336–349. U.S. Census Bureau. (2004). U.S. Census Bureau population projects. Retrieved October 11, 2008, from http://www.census.gov/ipc/www/usinterimproj/

Stein, M. A. (2000). Empirical implications of Title I. Iowa Law Review, 85, 1671–1690.

Vande Zande v. Wisconsin Dept. of Admin., 44 F.3d 538 (7th Cir. 1995).

Stensrud, E., & Gilbride, D. (2006, August) Corporate cultures and disabilities: The Employment Opportunity

Weathers, R. R., II. (2006). Disability prevalence rates for an aging workforce. Ithaca, NY: Cornell University,

502

Disability and Employment

Employment and Disability Institute, Rehabilitation Research and Training Center on Employment Policy. Weaver, C. L. (1991). Incentives versus controls in federal disability policy. In C. L. Weaver (Ed.), Disability and work: Incentives, rights, and opportunities (pp. 3–17). Washington, DC: AEI Press. Wehman, P. (2003). Workplace inclusion: Persons with disabilities and coworkers working together. Journal of Vocational Rehabilitation, 18, 131–141.

World Health Organization. (2001). International Classification of Functioning, Disability and Health: ICF. Geneva, Switzerland: Author. Yelin, E. H., & Trupin, L. (2003, May). Disability and the characteristics of employment. Monthly Labor Review, 20–31. Young, A. E., Wasiak, R., Roessler, R. T., McPherson, K. M., Anema, J. R., & van Poppel, M. N. M. (2005). Return-to-work outcomes following work disability: Stakeholder motivations, interests, and concerns. Journal of Occupational Rehabilitation, 15, 543–556.

Copyright American Psychological Association. Not for further distribution.

Weiner, B., Perry, R., & Magnusson, J. (1988). An attributional analysis of reactions to stigmas. Journal of Personality and Social Psychology, 55, 738–748.

Wooten v. Farmland Foods, 58 F.3d 382.386 (8th Cir., 1995).

503

CHAPTER 16

ROLE THEORY IN ORGANIZATIONS: A RELATIONAL PERSPECTIVE

Copyright American Psychological Association. Not for further distribution.

David M. Sluss, Rolf van Dick, and Bryant S. Thompson

The roles that a person plays in life are ubiquitous. Roles such as spouse, parent, charity volunteer, engineering professional, and manager fulfill important functions in a person’s family, community, and workplace. These roles also provide people with a sense of who they are and who they are becoming. Within organizations, many people become, at one time or another, an employee, a subordinate, a manager, a department member, a customer, a supplier, a project team member, and the like. These roles are enacted or played either separately (e.g., giving a performance evaluation as a manager in the morning and receiving a performance evaluation as a subordinate in the afternoon) or simultaneously (e.g., being a member of a product development team representing one’s department, of which one is manager). It seems as though individuals, as well as organizations, cannot function without roles— wherein structured interdependencies organize and create a network of intertwining tasks and responsibilities1 (Ashforth, 2001; Biddle, 1986; Katz & Kahn, 1978; Stryker & Burke, 2000). Roles, as such, become the nexus for how work is designed, communicated, accomplished, evaluated, and experienced (e.g., Welbourne, Johnson, & Erez, 1998). However, this same ubiquity of organizationally bound roles makes it significantly more difficult to understand the influence of roles on individual-level attitude and behavior. For example, organizational research focusing on roles has spanned such wide-

ranging topics as socialization, job transfers, social networks, team functioning, work/family conflict, and work design. One can easily access over 10,000 relevant articles, monographs, books, book chapters, and dissertations by entering role theory (and rolecentered terms) as key words across databases within psychology and sociology. As such, any review of roles in organizations tends to be a daunting task. Indeed, the sprawling and disparate use of the role concept led Biddle (1986) to worry about its use as more metaphorical than theoretical and/or practical. Fortunately, the study of roles within the organizational context remains vibrant and viable. To understand organizational roles and how roles influence attitudes and behaviors, one needs to understand its rooting in and influence on the self-concept. As a result, our review of roles within organizations takes an identity perspective. The concept of identity within organizational research has traditionally focused on how a collective (e.g., organization, team, occupation) or social category (e.g., gender, nationality) influences one’s self-concept and, in turn, behavior (e.g., Ashforth & Mael, 1989; Haslam, 2004; Haslam & Ellemers, 2005). Nevertheless, organizational roles and their subsequent relationships also have a central influence on one’s identity at work (e.g., Ashforth, 2001; Kreiner, Hollensbe, & Sheep, 2006; Sluss & Ashforth, 2007). Both social identity theory (SIT) and role identity theory, via microsociology and social psychology, have focused

We thank Sheldon Zedeck and Sharon Parker for helpful comments on the structure and content of this chapter. 1

Note that (although slight) we distinguish between role and job. Roles include both the tasks and the interdependencies (both relational and structural) subsequent to the tasks. Jobs focus squarely on the tasks. It seems one may hold multiple roles and multiple jobs (as defined here). However, the difference lies in the foci (see Role Definition section).

http://dx.doi.org/10.1037/12169-016 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

505

Copyright American Psychological Association. Not for further distribution.

Sluss, van Dick, and Thompson

consistently on how social structure (i.e., society) influences the self-concept, which, in turn, influences one’s affect, behavior, and cognition (Ashforth, 2001; Biddle, 1986; Mead, 1934; Stryker & Burke, 2000; cf. Katz & Kahn, 1978). As such, identitybased role theory (broadly defined) is multilevel and cross-disciplinary in nature because it “combines . . . [the] psychological (individual contributions) and . . . sociological (organizational framework) perspective[s]” (Welbourne et al., 1998, p. 542; cf. Whetten, 2007). In sum, (a) work is experienced and lived via one’s roles; (b) the subsequent role identities have significant influence on individual affect, behavior, and cognition; and (c) to understand behavior within organizations, we must therefore understand the dynamics and processes of role identity. In this chapter, we review the identity-based role literature. First, we provide a theoretical overview of both role identity and social identity theories— integrating the personalized and relational level of self with that of the depersonalized and collective (cf. Brewer & Gardner, 1996). Second, we turn our attention to three areas of interest2: (a) multiple roles; (b) role crafting3 (i.e., role making/taking, role definition, role innovation, role clarity), as well as a proposed new direction; and (c) role recovery. We delineate promising directions for future research within each of these three major areas. ROLE IDENTITY AND SOCIAL IDENTITY THEORIES A role is traditionally defined as a set of behavioral expectations attached to a position in an organized set of social relationships (Merton, 1957; Stryker, 2007; Stryker & Burke, 2000). In short, these behavioral expectations specify the meaning and character of the role—that is, the role identity. As such, the role is attached to a structural position, whereas the role identity is how the individual (i.e., role occupant) interprets and makes sense of

that role (cf. Ashforth, 2001). A role identity, therefore, is a cognitive schema that organizes and stores the information and meaning attached to the role (via behavioral expectations) and serves as a framework for interpreting in-role as well as extra-role behavior (Stets & Burke, 2000; Tepper, Lockhart, & Hoobler, 2001). These behavioral expectations are subject to social construction and negotiation among role occupants (Mead, 1934; Swann, 1987). Although a structurally positioned role assumes stability in behavioral expectations determined by institutional pressures, microprocesses that create these behavioral expectations exist and, as a result, are more dynamic (e.g., Reay, Golden-Biddle, & Germann, 2006). For example, the role of manager may possess more or less institutionalized behavioral expectations such as allocating resources, providing rewards, and giving performance feedback, but the nuances, content, and focus of these behaviors are still negotiated by those occupying the role (e.g., manager) as well as the counterrole (e.g., subordinate, senior manager, peer manager). As such, role identity theory attempts to integrate both the structural functionalist (Burt, 1982; Merton, 1957; Parsons, 1951) and symbolic interactionist perspectives (Mead, 1934; Serpe, 1987; Stryker & Burke, 2000). Structural functionalism focuses on how the social structure (e.g., the role position such as manager, director, or technician) institutionalizes stable behavioral expectations across situations and, depending upon function, hierarchy, and status, how that position influences the self-concept. Along the same lines, symbolic interactionism focuses on how individuals interrelate across the network of role relationships, creating both meaning for the role occupant (that is, identity) and providing a working template or cognitive schema to interpret in-role and extra-role experience. As such, role identity theory has progressed from simply explaining the shared, institutionalized, and normative expectations given a position in some social structure such as an organi-

2

We integrate findings from both work design research as well as team dynamics, but only as these findings directly relate to role identity theory. However, we do not comprehensively cover these areas as they are covered in other chapters in this volume (see chaps. 13 and 19, this volume).

3

Although role transitions is a widely researched topic within organizational psychology, the role transition literature is mainly under the purview of organizational socialization—newcomers adjusting to and integrating with their new role, job, occupation, and/or organization. We direct the reader’s attention to the chapter in this volume as well as recent reviews (Ashforth, 2001; Ashforth, Sluss, & Harrison, 2007; see also Vol. 3, chap. 2, this handbook).

506

Copyright American Psychological Association. Not for further distribution.

Role Theory in Organizations

zation or community of practice to exploring the processes by which role occupants define themselves and their roles vis-à-vis social interaction with other role occupants (e.g., Reay et al., 2006; Stryker, 2007; cf. Biddle, 1986). As a result, organizational scholars broaden the definition of a role (and its identity) to include not only the structural position but also the goals, values, beliefs, norms, interaction styles, and time horizons that are associated with the particular role (Ashforth, 2001). We assert that roles and their subsequent identities are more dynamic within the organizational context than within other social structures. Traditionally, role identity theory (within microsociology) leans on the supposition that “society” begets role expectations and that society is seen as a “mosaic of relatively durable [italics added] patterned interactions and relationships, embedded in an array of groups” (Stryker & Burke, 2000, p. 285). In contrast, role theory within organizations places society at the organizational level wherein there are myriad potential role identities and subsequent role relationships. Organizational roles are natural outcomes due to a nexus of transactions and contracts (Katz & Kahn, 1978; cf. Williamson, 1985). Roles are born from the negotiation and interactive processes that are inherent in working out the trade of goods, services, information, or whatever is of value to that particular role relationship. These roles are also, as a result, subject to changes inherent in organizational contexts—thus, the organization (as a proxy for society) is less durable than the above definition presupposes. For example, new mergers, alliances, products, or services are all fodder for change in organizational roles and responsibilities (e.g., Chreim, Williams, & Hinings, 2007; Reay et al., 2006). Therefore, role identities within organizations, although influenced by institutionalized pressures for conformity and legitimacy, are under pressure by dynamic situational (both structural and relational) factors resulting in equivalent volatility in role expectations, identity, and behavior. Given the rate of change and multiplicity of roles within organizations, the question remains how individuals organize and manage a seeming mass of role identities. The answer lies (at least partly) in role identity salience.

Role Identity Salience People live out their organizational lives within an ever-expanding, changing, and intertwining network of role relationships. Role identity theory deals directly with how and why individuals make certain role-bound choices in their behavior (Stryker, 2007). In short, role identity theory seeks to answer this “quintessential question: Given situations in which there exist behavioral options aligned with two (or more) sets of role expectations attached to two (or more) positions in networks of social relationships, why do persons choose one particular course of action?” (Stryker & Burke, 2000, p. 286). In other words, role identity theory directly addresses the fact that, within the organizational context, individuals have numerous group memberships and role relationships to enact that, at times, may conflict and cause trade-offs in both attitude and behavior. Each role identity is placed on a salience hierarchy wherein higher salience increases the “readiness to act out [that] identity” (Stryker & Serpe, 1994, p. 17). Salience is determined by two main factors: (a) the number of role relationships tied to the role and (b) the strength or intensity of the relational ties within these role relationships (Hogg, Terry, & White, 1995; Stets & Burke, 2000; Stryker & Burke, 2000; Stryker & Serpe, 1982, 1994). As a result, the higher the number of connected role relationships and the stronger these relational ties, the higher the role identity salience, which, in turn, influences rolechoice behavior (i.e., acting in accordance to one role vs. another). In order to explain (and predict) role-choice behavior, we propose an integrative model in which relational and situational factors facilitate role identity salience. However, role identity salience only activates the identity and does not guarantee rolechoice behavior. We suggest that role identification moderates the relationship between identity salience and role-choice behavior such that role identification will facilitate the association between perceiving the role identity as salient and acting in accordance to that role versus some other salient role (i.e., rolechoice behavior; see Figure 16.1). Although recognizing the influence of these factors on the link between salience and activation, role identity theorists also give room that one’s identity 507

Copyright American Psychological Association. Not for further distribution.

Sluss, van Dick, and Thompson

FIGURE 16.1. Role identity salience process.

salience hierarchy may become a type of “personality variable carried . . . across situations” (Stryker & Serpe, 1994, p. 18)—in effect, allowing one to create situations in which a particular role may be enacted. For example, 1st-year college students decorated their new dorm rooms in a similar fashion to their room at home, thus reaffirming their identity during this transition (Serpe & Stryker, 1987). As such, identities highest on the salience hierarchy tend to be enacted independent of situational and relational cues—due to chronic accessibility (cf. Ashforth & Johnson, 2001)—whereas the remainder varies on the basis of the myriad cues that increase and/or decrease role identity salience (see Figure 16.1). Relational determinants of role identity salience. Role identity theory proposes two major relational factors that influence role identity salience: the number of role relationships (i.e., relational breadth and depth) and the relational tie strength (i.e., relational identification) connected to the particular role. Sluss and Ashforth (2007), building on role identity theory and the personal relationship literature, further delineated these two relational determinants of identity salience. Relational breadth and depth. Thus far, we have discussed role identities as a generalized network of roles and counterroles (e.g., how Jim sees himself as a colleague to other civil engineers, how Jim sees himself as a team leader to team members). 508

Just as one organizational member may have multiple roles (e.g., manager, engineer, subordinate, project team member), the same organizational member will have multiple role relationships attached to each of these roles (Fletcher, Simpson, Thomas, & Giles, 1999; Sluss & Ashforth, 2007)— again, at a generalized level. For example, the role of manager inherently includes role relationships with subordinates, upper management, peer managers, customers, and other organizational stakeholders. As such, role identities are inherently relational in nature (e.g., Brewer & Gardner, 1996; Brickson, 2000; Sluss & Ashforth, 2007). A role is not lived or maintained “solo” but relies on the relational incumbents to give it meaning and substance. For example, the subordinate is only a subordinate when there is a manager; a client is only a client when there is a consultant; and a student is only a student when there is a teacher, other students, or other pertinent relational incumbents. Therefore, role identities may span from generalized understandings of a role (e.g., being an accountant) to particularized (and negotiated) role relationships with specific others (e.g., Sophia the accountant’s relationship with Audrey the external auditor vs. Sophia’s relationship with Jim the coworker; Sluss & Ashforth, 2007). Similarly, there are multiple midrange role identities as well. For example, Sophia the accountant may have a different understanding of role expectations (i.e., role identity) when interacting with external auditors as a whole than when interacting with a particular external auditor, Audrey. Therefore, identity salience may be extended from a simple frequency count of relationships to role relationship “depth” (i.e., number of abstraction levels from generalized to particularized) as well as “breadth” (i.e., number of relational others at each level of abstraction; see Figure 16.2). We argue that the more the depth and breadth of role relationships attached to a role, the higher the role identity salience—that is, the more the individual will be ready to enact that particular role identity. For example, Jim’s role identity as the plant production manager may be much more salient for his work identity than his role identity as a part-time

Role Theory in Organizations

Generalized

Manager Role Identity

Copyright American Psychological Association. Not for further distribution.

Mgr-Customer role relationship

Specific Others

Mgr-Senior Mgr role relationship

Mgr-Employee role relationship

Employees (1st shift)

Specific Others

Employees (Night shift)

Specific Others

Specific Others Particularized

FIGURE 16.2. Role identity hierarchy: relational depth and breadth. Mgr = manager.

member of a corporate-wide quality improvement team. Indeed, Jim may enact his manager identity across these two situations because of overwhelming relational depth and breadth attached to the managerial role. Nevertheless, it is possible that Jim’s managerial role (again, with an overwhelming relational depth and breadth) will subsume to his qualityimprovement team role because of the strength of the particularized relationships existing on the team (e.g., Jane, the vice president of production and quality-improvement team member has been Jim’s career-long mentor)—the topic of our next discussion. Relational identification. Relational tie strength (i.e., relational identification) is the second driver of role identity salience. As above, a relational view of identity provides an expanded definition of relational tie strength via relational identification. Sluss and Ashforth (2007) defined relational identification as the “partial definition of oneself in terms of a given role-relationship” (p. 15). In other words, the self expands to include the role relationship (and, in part, the relational other) as central to the role incumbent’s identity (e.g., Aron

& Aron, 2000). Just as role identity may be generalized or particularized, relational identification can also function at the generalized and particularized level. For example, Dinah, a software engineer, may have a generalized sense of relational identification with project managers as well as a particularized relational identification with David, one particular project manager. Additionally, relational identification may take on both a positive or negative valence, resulting in three types of identification: (a) relational identification (positively viewed relationship); (b) relational disidentification (negatively viewed relationship); and (c) ambivalent relational identification (mixed perceptions concerning the role relationship; Sluss & Ashforth, 2007; cf. Elsbach, 1999). As a result, a role identity may become salient because of a strong relational identification with either a particularized role relationship or a more generalized role relationship. For example, Jane’s role identity as a salesperson to her favorite and most fruitful client may outweigh her role identity as a sales team leader with a large number of salespeople. 509

Sluss, van Dick, and Thompson

Copyright American Psychological Association. Not for further distribution.

In sum, relational identity and identification further delineate role theory’s view of salience by proposing two relational factors: (a) the depth and breadth of relational ties and (b) the strength of the generalized and particularized relational identifications attached to the particular role. Situational determinants of role identity salience. In addition to relational determinants of identity salience, SIT provides insight into situational determinants of identity salience as well as specifies role identification (Hogg et al., 1995). From the perspective of SIT, roles may be viewed as social categories or collectives in which one may be a member (Ashforth, 2001; Pratt, Rockmann, & Kaufmann, 2006). Similar to other social categories or collectives (e.g., organization, occupation, and nationality), a person can define him- or herself in terms of a role resulting in role identification. Situational factors. Both SIT (Tajfel & Turner, 1979, 1986) and self-categorization theory (SCT; J. C. Turner, Hogg, Oakes, Reicher, & Whetherell, 1987) provide additional insight into the role identity salience hierarchy. SIT and SCT provide complementary drivers to identity salience (J. C. Turner et al., 1987). Social identity scholars also recognize that individuals are members of a large variety of collectives and, as a result, are simultaneously incumbents in a plethora of roles (e.g., colleague, supervisor, subordinate). Thus, similar to role identity theory, identity salience becomes an important component in explaining identityrelevant attitudes and behaviors. Within SIT and SCT, salience is considered as a continuous variable formed by an interaction of accessibility and fit within a given situation (J. C. Turner, 1999). A category is more likely to be salient if the individual is predisposed to perceive that category as relevant (accessibility) and if both the category and the situation match the individual’s expectations and reality matches these expectations (see Oakes, Turner, & Haslam, 1991). Accessibility. Accessibility describes the perceiver’s readiness to accept a category (i.e., the extent to which it has prior meaning and significance for the individual). The higher this prior meaning of a category, the less input that is nec510

essary to activate this specific category (Oakes, 1987). Accessibility in SCT has some links with salience in identity theory, in as much as identity theorists also propose a higher probability of a role being activated when the person’s commitment to the role is chronically higher, because of the relational depth and breadth of the role as well as the strength of relational identification (see previous discussion). Nevertheless, according to Hogg et al. (1995), role identity theory tends to view salience as more stable than dynamic, whereas SCT conceptualizes identity salience as flexible and situationally determined. We tend to agree with SCT. Indeed, our relational view of role identity salience (i.e., relational depth and breadth; relational identification, accessibility) allows for change in both number and strength of role-based relational ties, as does the concept of fit. Normative and comparative fit. In addition to accessibility, SCT specifies additional situational determinants—normative and comparative fit. Fit refers to the match between an activated category (e.g., a manager’s role) and the stimulus reality (e.g., the context of a weekly meeting between the manager and his or her subordinates; J. C. Turner, 1999). Oakes (1987) differentiated between comparative and normative fit. Normative fit describes a principle “which suggests that a given category is more likely to become salient to the extent that the pattern of observed contentrelated similarities and differences between category members is consistent with the perceiver’s prior expectations” (Haslam, 2001, p. 382). In addition, comparative fit describes a principle “which suggests that a given category is more likely to become salient to the extent that the differences between members of that category are perceived to be smaller than the differences between members of that category and comparison others” (Haslam, 2001, p. 381). In practice, identity salience increases (a) when the category is simply mentioned, for instance, when a manager mentions another (potentially competing) team in a team meeting (in experimental studies, presenting the colors of the university the student participants belonged to was sufficient to increase performance; e.g., James & Greenberg,

Copyright American Psychological Association. Not for further distribution.

Role Theory in Organizations

1989); (b) when an individual encounters other people within the context of relevant other categories such as a manager meeting a subordinate during an appraisal, or when the context suggests that a specific role identity will be expected (van Dick, Wagner, Stellmacher, & Christ, 2005); and (c) when role incumbents perceive that they are in conflict or have incompatible goals (e.g., managers meeting with union representatives; see Wagner & Ward, 1993). Role identification. SIT and SCT were initially developed to understand intergroup hostility, ingroup favoritism, and outgroup discrimination or bias. SIT and SCT were then later applied to organizational contexts (Ashforth & Mael, 1989; Hogg & Terry, 2000). As a result, the social identity perspective has proved a fruitful framework to explore and understand a great range of organizationally relevant issues, such as leadership, productivity, communication and decision making, diversity, and stress (for reviews, see Haslam, 2004; Haslam, Postmes, & Ellemers, 2003; van Dick, 2004). It has been shown, for instance, that strong organizational identification relates to lower turnover and turnover intentions (van Dick, Wagner, Stellmacher, & Christ, 2004), higher willingness to show extra-role behavior (van Dick, Grojean, Christ, & Wieseke, 2006), stronger customer orientation (Wieseke, Ullrich, Christ, & van Dick, 2007), and more creative effort (Hirst, van Dick, & van Knippenberg, in press). Meta-analyses have also revealed reliable relationships between team and organizational identifications and a range of attitudes and behaviors relevant in organizations, such as job satisfaction, team climate, and in-role and extra-role behavior (Riketta, 2005; Riketta & van Dick, 2005). The relationships between these identifications and important attitudes and behaviors are quite substantial; Riketta (2005), for instance, reported relationships between organizational identification and job satisfaction (r = .54), job involvement (r = .61), turnover intentions (r = −.48), in-role performance (r = .17), and extra-role performance (r = .35). Tajfel, Billig, Bundy, and Flament (1971) used the minimal group paradigm to investigate the con-

ditions under which individuals treat their in-group members better than members of other (out)-groups. In a series of experiments, Tajfel and colleagues showed that, even in completely artificial groups, participants consistently allocated more points to in-group members than to out-group members even at the cost of optimum reward. Considering these findings, Tajfel (1978) started distinguishing between personal and social identities. The personal and the collective identity might be seen as some extreme poles of an identity (i.e., self-defining) continuum with role identities (that is, the relational level) as discussed in role identity theory forming anything between those poles. Outside the laboratory and in organization-based field studies, identity-based intergroup relationships have also been investigated. Richter, West, van Dick, and Dawson (2006), for instance, showed that strong team identities relate to more conflicts with members of other teams but only for those employees who are not simultaneously identified with the larger organization. As another example, Terry and Callan (1998) demonstrated intergroup bias driven by identity-based status differential in the context of an organizational merger. Brewer and Gardner (1996) argued that the social self can be further differentiated as relational and collective. The relational level of self is derived from interpersonal relationships with the welfare of the dyad as the focus of motivation. The collective level of self (à la SIT) is derived from impersonal group memberships with the welfare of the group as the focus of motivation. Tajfel (1978) defined social identity as “that part of an individual’s self-concept which derives from his knowledge of his membership of a social group together with the value and emotional significance attached to that membership” (p. 63). It should be noted that identification is somewhat distinct from simple self-categorization as a group or role-set member—wherein the latter is recognition of membership in a collective (Hogg & Terry, 2000). However, the recognition of one’s group membership can produce a psychological connection between the self and the group . . . 511

Sluss, van Dick, and Thompson

Copyright American Psychological Association. Not for further distribution.

there is a considerable amount of variability in the degree to which people feel a subjective sense of interconnectedness with their groups; that is, in the degree to which they include the ingroup in the self.” (Tropp & Wright, 2001, p. 586) In sum, identification is more than just categorizing oneself as a role occupant. Following this, at least three components of identification can be differentiated: (a) an affective component (How much do I value the identity?), (b) an evaluative component (How much do I think others value this identity?), as well as (c) a cognitive component (self-categorization; Do I see myself as a member of the social category?; see van Dick, 2001). In its basic form, SIT predicts that (a) people strive for a positive self-concept and that (b) one’s identity partly consists of one’s memberships in social groups—the roles one holds in organizations in particular (see Hogg & Terry, 2000). Holding the role of a manager, employee, or member of a specific team (e.g., marketer or production worker) partly answers the individual’s question of “Who am I?” and thus contributes to his or her self-definition. SIT thus would predict that organizational members’ identification with their roles will be associated with role-related attitudes and behavior. Indeed, Pratt (1998) elaborated on the point that identification serves the individual’s needs for belonging and safety. Following this, an individual who identifies more strongly with a role will have more of his or her needs satisfied via the role and is more likely to enact the role expectations and role-choice behavior. As our model alludes, we argue that, whereas relational (i.e., relational depth/breadth, relational identification) and structural factors (i.e., accessibility and fit) drive role identity salience, role identification strengthens the relationship between role identity salience (wherein the identity is activated) and role-choice behavior (see Figure 16.1; cf. van Dick, Wagner, et al., 2004). So far, we have provided an in-depth account of the theoretical bases of roles in organizations. We have integrated role identity, relational identity, and SITs to specify the nature, forms, underlying motives 512

and situational dependencies of role identities. We have provided an overall framework for how both relational (i.e., relational breadth and depth, relational identification) and situational (i.e., accessibility, fit) factors influence role identity salience and, in the end, role-choice behavior. We now apply our framework to specific areas of current research— namely, multiple roles, role making/taking, role definition, role innovation, and role clarity—as well as propose a new direction: role recovery. MULTIPLE ROLES As we have indicated throughout this chapter, an employee is not only an organizational member but also someone who is involved in multiple groups and role relationships within and between organizational boundaries. As a result, an individual may have a veritable jungle of role identities that potentially influence attitudes and behaviors. In our discussion of role identity salience, we have already suggested how one role identity may become activated and influence role-choice behavior over another. Here, we focus on research that examines simultaneous or interacting identities. We show that the individual has multiple coexisting identities and that the level of salience determines which of those identities becomes relevant for one’s behavior provided that there is an identity–behavior match.

Current Research Self-theories are the multiple hypotheses a person has about one’s self to guide perceptions, thoughts, and actions. The concept of the self refers to the insider’s view of personality (Markus & Cross, 1990). As a fundamentally social construal (Banaji & Prentice 1994), the “working self-concept is influential in the shaping and controlling of intrapersonal behavior (self-relevant information processing, affect regulation, and motivational processes) and interpersonal processes, which include social perception, social comparison, and social interaction” (Markus & Cross, 1990, p. 578). For example, a person’s theory of what makes a good leader guides that person’s responses to and interactions with his or her subordinates and even his or her own leaders (see Dweck, 1991).

Copyright American Psychological Association. Not for further distribution.

Role Theory in Organizations

Certainly, current thinking holds that a person may possess many possible selves (or role identities; Ibarra, 1999; Leonard, Beauvais, & Scholl, 1999; Lord, Brown, & Freidberg, 1999; Markus & Nurius, 1986). Although consensus has yet to be reached within the literature, scholars agree on three points— namely, that the self-concept (a) contains a stable aspect; (b) may change in differing situations; and (c) is closely linked to traits, values, and behavior (e.g. Donahue, 1994; Higgins, Klein, & Strauman, 1987; Roberts & Donahue, 1994). The multiple and differing possible self-concepts that individuals possess may be considered as separate role identities. These identities carry information about what are considered proper behaviors in a particular situation, what values are salient in that situation, and what attitudes are relevant for a particular situation (Leonard et al., 1999). Indeed, although the stable aspect of (global) identity provides a starting point for role identities, it is the specific role identities that provide the relevant information that influence behavior within that role (e.g. traits, competencies, values). We obviously do not have room here for discussing the many role identities conceivable in organizational contexts. These possibilities include organizational members who also have a strong professional or occupational role identity, or a manager who, in addition to his role identity, may still be a team member (as a primus inter pares) with a strong associated identity to that role. Johnson, Morgeson, Ilgen, Meyer, and Lloyd (2006), for instance, explored veterinarians’ identities concerning their profession, their organization, and their work group and found that patterns of those identities varied according to the veterinarians’ status (e.g., as working in a veterinary medicine vs. nonveterinary medicine organization, or being an associate vs. an owner of the organization). Pratt and Forman (2000) suggested that those and other multiple identities in organizations can be managed by changing the number of and/or relationships among the identities through managerial responses of compartmentalization, deletion, integration, or aggregation. To illustrate just one of the many pairs or combinations of multiple role identities, we discuss two of the probably most prominent identities—that is, the role of member of the organization as a whole and

the role of member in smaller organizational subunits (e.g., team, department, workgroup). Ashforth and Johnson (2001) presented a theoretical analysis of these two identities, whereas van Knippenberg and van Schie (2000) were among the first to empirically distinguish between those forms of nested identities. These and several other authors have suggested that the identities and roles associated with smaller entities should be stronger than roles associated with the larger organization. Among other reasons (see Ashforth, Harrison, & Corley, 2008, for an exhaustive list of arguments), this is mainly the case because (a) employees in their everyday workplace interactions are more likely to encounter people from other groups and departments rather than those from different organizations, which renders the smaller category more salient, and (b) the principle of optimal distinctiveness (Brewer, 1991) suggests that employees at the same time want to “stand out” as unique individuals and “fit in” into groups to satisfy their needs for belonging; integrating these two motivational forces should be easier in small, exclusive roles than in large, inclusive ones. That said, those individuals in boundary-spanner roles may experience the opposite (e.g., Bartel, 2001). However, supporting the general notion, Riketta and van Dick (2005) found that team identification is stronger than organizational identification. This does not automatically mean, however, that role identifications attached to more localized roles are more important for the prediction and/or development of the broad range of individual attitudes and behaviors. Instead, although more localized identification is stronger, the identification and the behavior in question must match (see van Dick, Christ, et al., 2004). This has been termed the identity-matching principle by Ullrich, Wieseke, Christ, Schulze, and van Dick (2007; see also Ashforth et al., 2008), who proposed that the respective relevant identity needs to “match” (i.e., address the same domain or level of behavior to be of predictive value). In line with this principle, Ullrich et al. (2007) found that organizational identification predicted organization-level customer-oriented behavior, whereas corporate identification predicted corporate citizenship behavior. Christ, van Dick, Wagner, and Stellmacher (2003) found organizational identification in schools to be 513

Copyright American Psychological Association. Not for further distribution.

Sluss, van Dick, and Thompson

related to teachers’ school-related extra-role behavior, whereas team identification was most closely related to helping behavior directed toward the teachers’ immediate colleagues. In sum, the individual has multiple role identities, and the level of salience determines which of those identities become relevant for his or her behavior provided that there is a match between identity and behavior and requisite role identification. Sometimes, those multiple role identities may be in conflict, for instance, when a manager who identifies with his or her role needs to give negative feedback to a subordinate who is also a member of the manager’s team and therefore an ingroup member. These conflicting aspects of an individual’s roles and attached role identities can be a source of great strain. In the following section, we briefly explain how these conflicts can sometimes also lead to processes of disidentification with negative consequences for the individual’s satisfaction and wellbeing. At other times, different identities complement each other or even positively interact. Van Dick, van Knippenberg, Kerschreiter, Hertel, and Wieseke (2008), for instance, recently found (across two samples of bank and travel agency employees) that in cases of positive overlap (i.e., high work group and organizational identification), identifications are more strongly associated with job satisfaction and extra-role behavior than when only one of the identifications is high. In addition, Sluss and Ashforth (2008) proposed that organizationally nested relational identification may converge with organization identification, given that nested identifications prime and resemble each other, eliciting similar responses.

Future Research on Multiple Roles Future research on multiple role identities may reap rewards by focusing on targets (foci) that have not received much attention in the past, such as individuals whose careers or occupations or roles have only relatively recently been created in larger numbers. One example is the “portfolio worker” who is working for several clients simultaneously, perhaps taking the role of a consultant with one client, a more managerial role with a second client, and a teammember role with a third client. A second example is the individual who is a member of category or 514

who holds a role that is not viewed as particularly positive by society (e.g., a worker who produces land mines), but role identification remains important for the individual’s self-concept and should receive more research attention (e.g., Ashforth, Kreiner, Clark, & Fugate, 2007; Kreiner, Ashforth, & Sluss, 2006). Third, in a related vein, very little research has looked into ambivalent or even negative forms of identification, that is, when individuals actively dissociate from certain aspects of their roles. Kreiner and Ashforth (2004) referred to this as disidentification, which occurs when an individual defines him- or herself as not having the same attributes or principles that he or she believes define the organization. A manager, for instance, might identify with his or her role and particularly with the positive impact he or she potentially can make on followers by providing developmental feedback, vision, and guidance. On the other hand, the same manager might disidentify with other aspects of the role, for instance, the requirement to make promotion decisions or to lay off staff. Kreiner and Ashforth emphasized that disidentification is not just the opposite of identification. They argued that disidentification is a separate variable and a unique psychological state. Whereas identification consists of connecting (typically positive) aspects of the organization (whether at the molar or facet level) to oneself, disidentification consists of disconnecting (typically negative) aspects of the organization (whether at the molar or facet level) from oneself. Although a major goal of both identification and disidentification is preservation of a positive social identity, the paths to that goal and the phenomenology of the experience differ appreciably. (p. 3) Kreiner and Ashforth have suggested items to measure disidentification with the organization, and it would be fruitful to use these scales particularly with respect to role identities; a lot is to be done on this interesting research front.

Role Theory in Organizations

Copyright American Psychological Association. Not for further distribution.

Fourth, it would certainly be of great theoretical and practical value to examine the complex interplay of the various factors that drive role identity salience, which are (a) relational (i.e., relational depth/breadth, relational identification), (b) structural (i.e., comparative and normative and fit), and (c) personal (i.e., accessible or chronic role identification). Finally, in our view, future research should pay more attention to both the identity-matching principle (Ullrich et al., 2007; see also Ashforth et al., 2008) and also effects of role identities that are potentially in conflict. ROLE CRAFTING Just as role identity salience (within organizations) is somewhat dynamic, so are the behavioral expectations attached to one’s organizational roles. As such, the defining, negotiating, and crafting of the role itself is an important topic for role theory within organizations. Unfortunately, the research streams in this area are disparate and incommunicative. We use role crafting as a broad term to integrate and review the research that focuses on the establishment and subsequent change of roles within organizations. The streams include (a) role making/taking (Graen & Cashman, 1975; Jablin, 1987; Katz & Kahn, 1978; Quick, 1979); (b) role definition (Bolino, 1999; Morgeson & Camnion, 1997; Hofmann, Morgeson, & Gerras, 2003; Organ, 1988; Organ, 1990; Parker, 1998; Tepper et al., 2001); (c) (forms of) role innovation (Frese, Fay, Hilburger, Leng, & Tag, 1997; Frese, Kring, Soose, & Zempel, 1996; Morrison & Phelps, 1999; Staw & Boettger, 1990; van Maanen & Schein, 1979; Wrzesniewski & Dutton, 2001); and (d) role clarity (Bush & Busch, 1981; Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964; Teas, Wacker, & Hughes, 1979). We briefly define each area and present current research findings. Then, we suggest future research directions across the streams that encompass role crafting.

Role Making/Taking Role making is the process by which two individuals shape and reinforce desired roles through reciprocation. This reciprocation is aimed at increasing mutual trust and exchange of benefits that often occur

between a leader and a member—leader–member exchange (Graen & Cashman, 1975; Graen & Uhl-Bien, 1995; Schriesheim, Castro, & Cogliser, 1999). Gerstner and Day (1997) meta-analytically found that leader–member exchange is related to leader performance ratings (r = .49), member performance ratings (r = .31), satisfaction with supervision (r = .74), overall satisfaction (r = .50), organizational commitment (r = .38), role conflict (r = −.40), role clarity (r = .34), and turnover intentions (r = −.27). For example, Erik, a project manager, may provide Jason, a project member, with valuable resources and increased responsibilities after Jason has demonstrated that he is a reliable project member. Likewise, Jason may use his acquired resources to help reduce Erik’s workload. Role taking is when organizational members learn and accept roles through organizational socialization, instruction, and feedback (Jablin, 1987; Katz & Kahn, 1978). Role taking involves individuals assuming expected roles to align their actions with social norms—viewing themselves as objects of a desirable social transaction (Heimer & Matsueda, 1994). When role taking is reciprocated, joint activity is facilitated and levels of social control increase (Mead, 1934). Together, role making and role taking influence how a role is defined and crafted, with the source being the role occupant, the counterrole occupant (i.e., role sender), and the organization. As such, there are various interpersonal and institutional pressures that affect the creation, maintenance, and augmentation of roles. Role-taking behavior is driven by five major factors: (a) meaning of the self, (b) attitude about role and role sender, (c) anticipating reactions of the role sender, (d) associating with the role senders, and (e) absence of reflective thought with regard to role-sender suggestions (Heimer & Matsueda, 1994), In short, role taking becomes a somewhat passive process in which the role occupant “takes on” the prescribed role sans any type of proactive augmentation of the role. For example, Jason, the project member, may assume more of a leadership role if he (a) perceives himself as having the potential to be a leader and believes that the role sender also perceives him as being a leader, (b) wishes to be a leader and would welcome the opportunity to work more closely 515

Copyright American Psychological Association. Not for further distribution.

Sluss, van Dick, and Thompson

with the role sender, (c) anticipates that the role sender will embrace him as a leader, (d) associates with the role sender for a sufficient amount of time to know how to assume the role of a leader, and (e) responds habitually to the cues of the role sender via learned routines and adopted scripts rather than deliberate cognition—thus, taking on the given role. In addition to these factors, the cultural backgrounds of both the role taker and the role sender are important, such as family orientation, collectivism/ individualism, power distance, and time orientation (Stone-Romero, Stone, & Salas, 2003). An individual’s culture is important because it has demonstrated an influence over individual beliefs, emotions, and values (Hofstede, 1991; Markus & Kitayama, 1991; Triandis, 1980; Zavalloni, 1980). According to the similarity-attraction paradigm, individuals tend to be more attracted to those whom they perceive as similar to themselves (Byrne, 1961). As such, the frequency of communication, relational identification, and rate of social integration increase when similarity is perceived—often due to a perception of trust and association between similar individuals (Bond, 1983; Bond & Forgas, 1984; Lincoln & Miller, 1979; O’Reilly, Caldwell, & Barnett, 1989; Thomas & Ravlin, 1995; Zenger & Lawrence, 1989). In many interactions, an individual’s expectations are based on stereotypical assumptions. This is especially true when information is incomplete because of a lack of similarity or incompatible cultural schema (Bell, Wicklund, Manko, & Larkin, 1976; Thomas & Ravlin, 1995). As such, prior information about a person (or cultural group) is factored into how one behaves and constructs work roles (cf. Ajzen & Fishbein, 1975; Darley & Fazio, 1980; Ginossar & Trope, 1980). For example, the interaction between a manager and subordinate will be contingent on whether the work is performed in the home country of the manager. When the subordinate has a strong national identification, variation in the manager’s behavior from the subordinate’s behavioral expectations will be attributed to external causes, such as cultural differences, thus compounding the perception of dissimilarity (Pinkley, LaPrelle, Pyszczynski, & Greenberg, 1988). Managers will attempt to mitigate these effects by engaging in culturally adaptive behaviors, accommodating the subordinates’ cultural 516

norms (to the extent that the norms are known to the manager). Otherwise, subordinates are likely to experience decreases in self-esteem while perceiving their manager to be untrustworthy and ineffective (Francis, 1991; Giles & Smith, 1979; Thomas & Ravlin, 1995). Role making has focused mainly on the outcomes of a mutually committed and high quality relationship (via leader–member exchange theory; Gertsner & Day, 1997). Although scholars have criticized the role-making literature as lacking a chronological cohesion and adequate explanation as to the rationale of how role making develops (Schriesheim, Castro, & Cogliser, 1999), Bauer and Green (1996) revealed that as individuals endeavor to engage in role making (via leader–member exchange), both the leader and member attempt to evaluate the behavior and underlying motivations of the relational other and then make a choice as to the nature and degree of relational exchange to pursue. Contributing to whether individuals will pursue and succeed in forging valuable exchange relationships are perceived individual similarities, which are followed by trust (Dienesch & Liden, 1986; Graen & Scandura, 1987; Mayer, Davis, & Schoorman, 1995; Meglino, Ravlin, & Adkins, 1989). Trust is a significant contributor toward the development of the exchange relationship because trust increases behavior predictability by establishing a common system of communication and providing more stable interactions (Blau, 1964; Dienesch & Liden, 1986; Meglino, Ravlin, & Adkins, 1991). As trust increases, greater levels of delegation ensue, freeing the leader to expend less energy on the delegated duties while augmenting the scope of the member’s desired responsibilities (Barber, 1983; Mayer et al., 1995). Thus, leader–member exchange is also a trust-building exercise. As such, trust is both an antecedent and consequence of leader–member exchange over time (Bauer & Green, 1996). Role making can also develop through the use of socialization tactics. Organizations use socialization tactics to facilitate newcomer assimilation from one role to another (van Maanen & Schein, 1979). Van Maanen and Schein list the following six dimensions, referred to as socialization processes: collective (vs. individual), formal (vs. informal), sequential (vs. random), fixed (vs. variable), serial (vs. disjunctive),

Role Theory in Organizations

and investiture (vs. divestiture). Of these, the institutionalized tactics (collective, formal, serial, and investiture) are more likely to lead to individuals passively accepting prescribed roles, whereas the individualized tactics (individual, informal, random, variable, disjunctive, and divestiture) tend to result in individuals being more proactive in shaping (or making) their roles (Ashforth, Sluss, & Harrison, 2007; Ashforth, Sluss, & Saks, 2007; Jones, 1986).

Copyright American Psychological Association. Not for further distribution.

Role Definition Thus far, we have discussed role behavior as unidimensional. However, individuals performing similar activities may define their roles quite differently and assume broader in-role responsibilities than others, whereas others may view their roles narrowly (Hofmann et al., 2003; Morrison, 1994; Salancik & Pfeffer, 1978). Individuals may also define roles according to time horizons or values—whether their role orientation is long term (vs. short term) or collaborative (vs. individual; Ashforth, 2001; Parker, 2007). As such, roles (rather than being defined objectively) are contingent upon individual differences, socialization, and role orientation (Graen, 1976; Hofmann et al., 2003; Korsgaard, Sapienza, & Schweiger, 2002; Morrison, 1994; Parker, 2007; Parker, Wall, & Jackson, 1997; Robinson, Kraatz, & Rousseau, 1994; Saks & Ashforth, 1997; Salancik & Pfeffer, 1978). As such, role behavior may be construed as either falling within the prescribed role (i.e., in-role behavior) or as beneficial for the role but not required (i.e., extra-role behavior). Role definition, a part of the role-crafting process, attempts to categorize these behaviors. Role definition takes shape as individuals form perceptions of their work role on the basis of role preference, role ability, and expectations of others—with an ongoing assessment and subsequent modifications occurring as social cues and individual inclinations converge (or diverge; Graen, 1976; N. Turner, Chmiel, & Walls, 2005). More specifically, the distinction between in-role behavior and extra-role behavior is that in-role behavior consists of activities that are formally required and rewarded, whereas extra-role activities are neither required nor rewarded (Katz & Kahn, 1978; Organ, 1990). In addition, extra-role activities are discre-

tionary activities not directly associated with a job description or formal reward system (Katz & Kahn, 1978). Even though performance of in-role activities is more likely than performance of extra-role behaviors (Morrison, 1994), extra-role behaviors are essential for organizations to thrive and function (Organ, 1988). For example, when task completion is contingent upon helping behavior, helping (i.e., extrarole behavior) becomes an integral component to task interdependence. As such, the extra-role behavior becomes embedded within what may be expected as in-role behavior (Griffin, Neal, & Parker, 2007). The most prominent of the extra-role behaviors is organizational citizenship behavior (OCB), which is a similar construct to contextual performance (Borman & Motowidlo, 1993). OCBs are extra-role activities primarily related to prosocial, contextual performance rather than to task performance (Bolino, 1999; Organ, 1988, 1990). There are two major dimensions of OCBs: altruism (behavior that benefits specific individuals) and conscientiousness (behavior that benefits the organization; Rioux & Penner, 2001). For example, Jennifer may sense that Taylor is having a challenging workday and offer to take Taylor to lunch, with the principal benefit going to Taylor (as an altruistic act by Jennifer). In short, Jennifer’s primary concern is the well-being of Taylor. Likewise, Jennifer may offer to help Taylor practice for an upcoming client presentation—thus benefiting the organization. OCB role perceptions encompass perceived role breadth, perceived instrumentality, perceived role efficacy, and perceived role discretion (McAllister, Morrison, Kamdar, & Turban, 2007). For example, job autonomy and cognitive ability positively related to role breadth, with role breadth in turn mediating the relationship between job autonomy and job performance as well as the relationship between cognitive ability and job performance (Hofmann et al., 2003; Morgeson, Delaney-Klinger, & Hemingway, 2005; Morrison, 1994; Parker, 1998; Parker, Williams, & Turner, 2006). McAllister et al. (2007) demonstrated that these role perceptions provide further precision with regard to our understanding concerning in-role versus extra-role perceptions. For example, higher levels of perceived instrumentality may be concurrent with lower levels of perceived role 517

Copyright American Psychological Association. Not for further distribution.

Sluss, van Dick, and Thompson

discretion, thus confounding the extra-role perception. As such, it is efficacious to recognize these role perceptions as conceptually and empirically different, with relatively independent effects. With regard to perceived role discretion, the discretionary nature of OCBs facilitates the expression of individual employee attitudes (Organ, 1977). Several scholars have argued that OCBs are contingent upon perceived fairness (Moorman, 1991; Moorman, Niehoff, & Organ, 1993; Organ & Konovsky, 1989). As such, in response to fair treatment, individuals will perform OCBs as a compensatory reward. Likewise, individuals will withhold OCBs as a retaliatory response to unfairness. This behavior is consistent with Adams’s (1965) equity theory and Blau’s (1964) social exchange theory in that OCBs may be used to reciprocally restore equity (Tepper et al., 2001). Similarly, Morrison (1994) found that employees with favorable attitudes tend to view OCBs as in-role behavior (vs. extra-role behavior). This may be because viewing OCBs as part of one’s job is compatible with holding favorable attitudes and being a good citizen; however, it is also possible that this correlation is more a product of a socially desirable position driven by impression management and/or social exchange tactics rather than being based on bona-fide beliefs (Bolino, 1999; Morrison, 1994; Organ, 1990). Likewise, there is a link between OCBs and role identity—with OCBs being likely to increase as an individual’s role identity increases. Finkelstein and Penner (2004) found empirical evidence linking role identity with OCBs (r = .38). Because a particular role is internalized and becomes part of the individual’s self-concept, individuals begin to identify with normative expectations when a role identity is redefined by the increasing congruence between the self and the collective over time, with the individual striving to behave consistent with the emerging role (Piliavin & Callero, 1991). Thus, contributions to a collective through OCBs become a part of who one is rather than just what one does; the boundary between in-role and extra-role behavior often remains elusive, obscuring role expectations and incentives. This is especially true when unfixed roles evolve because of negotiation, clarification, and identification. The ramifications for completing a mandatory 518

(vs. discretionary) activity may elicit a broad spectrum of reactions, ranging from reward to rebuke (Morrison, 1994).

Role Innovation Once a role is defined, individuals may find that the role needs augmentation, revision, or, at times, wholesale change. Scholars have explored this need via role innovation. Role innovation is when individuals, leaders, and organizations instigate role modifications aimed at enhancing outcomes (e.g., Wrzesniewski & Dutton, 2001). Role innovation includes taking charge, personal initiative, task revision, and job crafting. Taking charge and personal initiative are ways for individuals to be more assertive in shaping work roles. Taking charge and personal initiative involve individuals making informal, voluntary, and discretionary efforts in order to effectuate organizational improvements, particularly with regard to problem solving as well as during times of change (Frese et al., 1996; Frese et al., 1997; Morrison & Phelps, 1999; Wrzesniewski & Dutton 2001). Individuals are more likely to demonstrate personal initiative when the work is more complex and within their control (Frese, Garst, & Fay, 2007; Hackman & Oldham, 1976; Parker et al., 1997). As such, when individuals are confronted with complex tasks and are granted sufficient control over the task, they tend to engage in proactive behaviors in order to gain the requisite knowledge and skills to overcome potential obstacles to effective task completion, rather than to just react to problems (Parker, Williams, & Turner, 2006). Inasmuch as one’s control orientation is high (i.e., aspires for control, perceives opportunities for control, and has high levels of self-efficacy), there is a strong relationship between control/complexity and personal initiative (Frese et al., 2007). As individuals assume broader job responsibilities through proactive workplace behaviors—being accountable for organizational goals rather than just individual goals—they develop a flexible role orientation (Parker, 2007). As such, there is a positive relationship between a flexible role orientation and performance (r = .37) that increases over time (r = .46). Likewise, this relationship is moderated by high job autonomy (r = .41; Parker, 2007).

Copyright American Psychological Association. Not for further distribution.

Role Theory in Organizations

Taking charge and personal initiative can lead to task revision. Task revision is when action is taken to redirect work from a faulty task (Staw & Boettger, 1990); it occurs when certain norms and processes are rejected in favor of new practices and procedures relative to task performance and relationships, matching actual requirements with current needs, skills, and abilities of the majority of the employees (Nicholson, 1984; Schein. 1971; van Maanen & Schein, 1979; Wrzesniewski & Dutton, 2001). Job crafting is a proactive process by which individuals endeavor to modify the physical and cognitive aspects of their tasks or relationships in the workplace (Wrzesniewski & Dutton, 2001). Individuals are now increasingly usurping crafting license once held only by leaders, proactively guiding work roles and career trajectories (Bridges, 1994; Wrzesniewski & Dutton, 2001). As such, role innovation is being reshaped through an improvisational and evolutionary process crafted by the individual and permitted by the organization. Rather than passively assuming a preconstructed role shaped by proficiency and adaptivity, individuals are claiming latitude in proactively crafting a more desirable work role (Grant, 2007; Griffin et al., 2007). For example, Bill, a project member, may wish to take a more active role with regard to the budgetary component of the project. Bill becomes proactive in learning and completing budgetary tasks. After which, Bill lobbies the project manager to revise Bill’s role to include budgetary responsibilities (possibly replacing other lessdesirable responsibilities). Contemporary role innovation tactics advance a more bottom-up perspective that empowers selfdetermined and competent individuals to play a more active role on the organizational stage through creative identity, forging new parametric contours through proactive crafting (Grant, 2007; Wrzesniewski & Dutton, 2001). This process has led to organizations and individuals entering into idiosyncratic deals (I-deals; Rousseau, 2005). I-deals are informal arrangements between a worker and employer that provide sufficient flexibility to meet the needs of both parties. As such, I-deals are often unique collaborations that facilitate experimentation and innovation in negotiating the parameters of a job—quid pro quo arrangements that yield mutually

beneficial outcomes (Rousseau, 2005). Ilgen and Hollenbeck (1991) broached the issue of role evolution and emergence; however, their view was that the job did not actually change, only the role changed. Wrzesniewski and Dutton (2001) argued that both the job and role change, rejecting the notion that jobs remain objective, irrespective of work role changes. Scholars have argued that individuals engage in role innovation to fulfill three needs: control, positive self-image, and connection to others (Baumeister & Leary, 1995; Braverman, 1974; Frese et al., 2007; Parker et al., 1997; Wrzesniewski & Dutton, 2001). If individuals are satisfying these needs elsewhere, they may choose not to engage in role innovation (Caldwell & O’Reilly, 1990). Individual needs vary based on how individuals view their work—as a job, career, or calling. To the extent possible, individuals will endeavor to craft their work role in accordance with how they view their work. For example, individuals who view their work as a job will craft much different work roles than those who view their work as a calling (Bellah, Madsen, Sullivan, Swidler, & Tipton. 1985; Wrzesniewski, McCauley, Rozin, & Schwartz, 1997). Individuals who feel their organization would not permit or support role innovation may also be deterred from contemporary role innovation activities (Caldwell & O’Reilly, 1990). For example, high control industries such as telemarketing and manufacturing are less likely to welcome role innovation, whereas low-control industries, wherein autonomy and creativity are more prevalent, are more likely to embrace role innovation (Amabile, Tighe, Hill, & Hennessey 1994; Wrzesniewski & Dutton, 2001). Extant research has found that role innovation antecedents are proactivity, openness to negotiation, and work facilitation, whereas potential consequences include job satisfaction and negative role conflict (Ashford, 1986; Graen & Scandura, 1987; Katz & Kahn, 1978; Kristof, 1996; V. D. Miller, Johnson, Hart, & Peterson, 1999; V. D. Miller & Jablin, 1991; K. I. Miller & Monge, 1986; Morrison, 1993; Schein, 1971). Socialization tactics are also associated with role innovation. Researchers have suggested that institutionalized socialization tactics tend to create a negative relationship between socialization tactics 519

Copyright American Psychological Association. Not for further distribution.

Sluss, van Dick, and Thompson

and role innovation, whereas individualized socialization tactics often lead to a positive relationship between socialization tactics and role innovation (N. J. Allen & Meyer, 1990; Ashforth & Saks, 1996; Baker, 1995; Jones, 1986; King & Sethi, 1992; Mignerey, Rubin, & Gorden, 1995). However, Ashforth, Sluss, and Harrison (2007) reminded us that institutionalized socialization tactics only constitute a process, wherein the content may vary. As such, institutionalized tactics may also be used to promote innovation, although contained within the parameters set forth by the organization (Ashforth & Saks, 1996; Ashforth, Sluss, & Harrison, 2007).

Role Clarity Irrespective of whether the role is crafted and defined by means of role making, role taking, or role innovation, it is imperative that role expectations be clear. When role expectations are unambiguous, role clarity ensues. Role clarity refers to whether an individual has certainty with regard to expectations surrounding his or her work role (Bush & Busch, 1981; Kahn et al., 1964; Teas et al., 1979). Conversely, role ambiguity is antonymous with role clarity and refers to one’s degree of uncertainty as to expected behaviors and attitudes (Kahn et al., 1964). Thus, role clarity and role ambiguity are conceptual opposites, with presence or absence of clarity being the sole distinction (Sawyer, 1992). Herein, our focal construct is role clarity. Empirical evidence demonstrates that there are several antecedents of role clarity. These antecedents include newcomer information seeking (r = .17), institutionalized socialization tactics (r = .27), detailed feedback from others, and participating in decision making (Bauer, Bodner, Erdogan, Truxillo, & Tucker, 2007). Newcomer information seeking refers to newcomers being proactive in their search for information pertaining to their work relationships and tasks. Because newcomers are actively and curiously pursuing relevant facts, effective newcomer information seeking facilitates the transmission of knowledge from organizational insiders to new role occupants both in terms of quality as well as frequency, thus increasing role clarity (Berger, 1979; Falcione & Wilson, 1988; van Maanen & Schein, 1979). Also, during the newcomer adjustment process, organiza520

tions tend to use institutionalized socialization tactics to increase newcomer assimilation from one role to another (van Maanen & Schein, 1979). Institutionalized socialization tactics reduce newcomer uncertainty by controlling the type, source, and ease of attaining information (D. G. Allen, 2006; Bauer et al., 2007; Jones, 1986; Saks & Ashforth, 1997). In response to information seeking, feedback from others is the process by which individuals receive guidance regarding previous behavior and can be a means by which clear and direct information is communicated relative to performance of work roles (Hackman & Oldham, 1976). Recognition, praise, or even reprimand provides clarity by signaling either acceptance or disapproval of behavior (Young, Worchel, & Woehr, 1998). Feedback from others is one of the most effective ways to clarify work roles, with the effect size depending on valence, response mode, and directness (Rotheram, LaCour, & Jacobs, 1982; Unzicker, Clow, & Babakus, 2000). Last, participation in decision making also tends to increase role clarity. Individuals who are permitted to participate in the decision-making process achieve increased role clarity through perceived empowerment, control, and legitimacy. This perception helps reconcile ambiguous work roles and augment understanding of expectations. The iterative nature of decision making often helps to elucidate work roles for the participants (Gilmore & Mooreland, 2000; Teas, 1980, 1983; Teas et al., 1979). Role clarity also has a plethora of important workrelated consequences. Role clarity provides a structured and predictable working environment wherein individuals are able to cognitively master the causal structures that affect their lives (DeCarlo, Teas, & McElroy, 1997; Sujan, 1986). Role clarity also increases job performance, both over the short term and over time (Churchill, Ford, Hartley, & Walker, 1985; Fried, et al., 2003). Role clarity is a harbinger of proactivity, confidence, and commitment (V. D. Miller & Jablin, 1991; Saks & Ashforth, 1997; Spreitzer, Kizilos, & Nason, 1997), increasing meaning, competence, and self-determination (Hall, 2008) while reducing negative strain (Abramis, 1994; Behrman & Perreault 1984; Fry, Futrell, Parasuraman, & Chmielewski, 1986; Jackson & Schuler, 1985; Revicki, Whitley, Gallery, & Allison,

Copyright American Psychological Association. Not for further distribution.

Role Theory in Organizations

1993; Teas, 1980; Von Emster & Harrison, 1998) as well as resulting in higher levels of self-efficacy (Chen & Bliese, 2002). Role clarity has also been found to mediate the relationship between information/ feedback seeking and extra-role behavior (e.g., OCB; Whitaker, Dahling, & Levy, 2007). Role clarity also conditions the relationship between job demand and job strain such that higher role clarity reduces the negative impact of job demands on job strain (Lang, Thomas, Bliese, & Adler, 2007). In sum, role clarity positively influences job performance (r = .29), organizational commitment (r = .29), job satisfaction (r = .32), and self-efficacy (r = .45; Babin & Boles, 1996; Bauer et al., 2007; Chen & Bliese, 2002; de Ruyter, Wetzels, & Feinberg, 2001).

Future Research in Role Crafting Given the wide berth of role crafting, we overlay our role identity salience model to suggest several future research directions. First, although much research has been conducted with regard to role making (see Gerstner & Day 1997; Graen & Uhl-Bien, 1995; Schriesheim et al., 1999), we believe much progress could be made in terms of understanding role identity salience through a dyadic process. What happens to role identity salience when the extant dyad has not been properly defined? Do the subordinate (member) and leader follow different role-making or negotiation processes to achieve role clarity? Do both role occupants need to include the role at similar levels of salience in order to successfully negotiate their respective roles? These questions also bring to light a need to focus on the dyad as the locum of action. We suggest scholars use more dynamic and dyadically focused methods such as longitudinal survey techniques testing for the ebbs and flows of consensus/agreement, qualitative observational methods that analyze dyadic processes, or more dynamic approaches (e.g., dynamic interactionism; Hattrup & Jackson, 1996). Second, scholars need to better understand how individuals define what is in-role versus extra-role behavior and how that influences helping behavior. An individual’s definition of his or her work role— whether the employee interprets the helping behavior as being in-role or extra-role—makes a difference (Tepper et al., 2001). We speculate that exploring

the drivers of role identity salience may provide insight into how individuals include extra-role behavior as part of the core role definition. Will relational breadth and depth increase role breadth so as to increase the chance that the extra-role behavior is seen as core to the role? We suggest using relational/ social network methodologies wherein the role is the center of the network and thus produces an operationalization of relational breadth and depth. Research has also shown that helping behaviors increase as they are perceived as in-role expectations, for example, as a result of being rewarded (even implicitly) for the behavior. We speculate that role identification may increase the probability of helping behaviors, regardless of how they are perceived (i.e., in-role vs. extra-role). Third, the level and form of role crafting can vary according to context. Job crafting covers broad degrees of change, whereas job shaping suggests deals only with small scale changes in work roles (Lyons, 2006; Wrzesniewski & Dutton, 2001). Future research should examine contextual variables to determine which factors are more suitable to significant role modification and which conditions permit only slight role changes. These contexts may include the nature of work, organizational culture, and leadership style. For example, virtual work provides a rich context within which to examine what happens when one’s role may be significant although temporary (at best) or fleeting and ill-defined (at worst). Within virtual work as well as the context of telecommuting, does role crafting take on a less significant role in one’s work experience? Or, because of the decreased structure and an increased need for control, does role crafting become even more important? In sum, research that assesses how dynamic working environments affect the degree to which roles are enacted may help clarify contextual implications relative to role crafting. That said, chronic accessibility of the role identity may overpower the context in that the individual attempts to radically change the role so as to better match the chronically accessible role identity. Fourth, drawing from a relational view of role identity theory, future research should focus on deciphering the relationship between role identification and role innovation. For example, when role 521

Copyright American Psychological Association. Not for further distribution.

Sluss, van Dick, and Thompson

innovation takes place, are individuals crafting a role that reflects who they are as opposed to a role that is more indicative of who they should be? What are the conditions under which individuals craft salient roles that are closely associated with their identification, versus crafting roles that are relatively disassociated with who they are? Because of the importance of role identity salience, we speculate that there may be a significant difference between role innovators who shape salient roles that are higher versus lower on the salience hierarchy. Finally, Ashforth, Sluss, and Harrison (2007) suggested the potential for delineated role innovation as either conducted unilaterally or collaboratively. Future research should examine the functional and dysfunctional components of unilateral role innovation (vs. collaborative) as it pertains to role salience. For example, it would be useful to know the conditions and extent to which the relationship between role innovation and role salience is moderated by the level of unilateral (vs. collaborative) innovation. Last, scholars should also explore how role identity salience affects role clarity and its outcomes. For example, what are the limits of role salience when roles are not clear? What happens to the role salience once roles are clarified? Do less salient roles need to be clear? Will clarifying less salient roles make them more salient and more likely to conflict with other roles? Because role clarity provides individuals with additional information about their roles, we know that role clarity can result in decreased role salience when undesirable aspects about a role are revealed. As such, it seems logical to conclude that the level of role salience will be contingent upon whether the clarifying information is viewed as negative or positive. Of course, clarifying one’s role early on in the socialization process may inhibit individuals from crafting a desired role—especially, for example, if it is made clear that an individual’s appropriate role is to be more reactive than proactive. In short, there may be adverse consequences to individuals having a clear understanding about an undesired role. Nevertheless, Wrzesniewski and Dutton (2001) contended that successful role crafting can occur even in jobs that have low levels of autonomy, authority, or complexity because individuals alter the way in which they perform their assigned tasks and frame 522

work relationships. For example, they illustrated how nurses changed their role by paying more attention to patients, even though this additional attention was not formally prescribed. Of course, employers who choose to precisely clarify a role by providing the exact way a task should be performed (and have a means to enforce this stringent requirement) may elicit negative consequences as they provide excessive levels of role clarity. Notwithstanding, we speculate that role salience will be attenuated by role ambiguity irrespective of the nature of the information revealed, because role ambiguity tends to increase uncertainty, perpetuate cognitive dissonance, and decrease relational identification. A NEW DIRECTION: ROLE RECOVERY We have discussed the various ways in which roles can be defined, negotiated, and even innovated (Graen & Cashman, 1975; Grant, 2007; Jablin, 1987; Katz & Kahn, 1978; V. D. Miller et al., 1999; Posner & Butterfield, 1978; Quick, 1979; Wrzesniewski & Dutton, 2001). We also understand that established roles can yield varying levels of clarity and consensus (Biddle, 1979; Bush & Busch, 1981; Kahn et al., 1964; Quick, 1979; Teas et al., 1979). It is surprising, however, that we know less about what happens when expectations for established roles are not met. How do the role occupants recover both the productivity and the familiarity within the relationship? We speculate that the forgiveness literature may provide important insights into how role recovery may happen. Indeed, role recovery is not specifically defined in the extant literature. Role recovery is a term we use to signify the restoration of role expectations through forgiveness—again, restoring both that which is exchanged in the relationship and the relationship itself.

Understanding Role Recovery Individuals are imperfect and will inevitably be perceived to violate accepted relational norms that govern relationships and shape role expectations (Rusbult, Hannon, Stocker, & Finkel, 2005). Role violations can be classified as a breach in the psychological contract. The psychological contract is defined as a set of agreements—albeit unwritten—pertaining to

Copyright American Psychological Association. Not for further distribution.

Role Theory in Organizations

expectations about the giving and receiving of benefits within an organization (Argyris, 1960; Robinson, 1996; Robinson & Morrison, 2000; Schein, 1965; Thomas, Au, & Ravlin, 2003). These expectations may involve workplace promotions, training, job security, and empowerment (Rousseau, 1989; Turnley & Feldman, 2000). The psychological contract emerges as a cognitive representation of perceived mutual obligations between individuals and can be useful in simplifying multifaceted workplace circumstances and relationships (Shore & Tetrick, 1994; Thomas et al., 2003). Scholars have classified the psychological contract as being either transactional or relational, depending on one’s conceptualization of the contract (Rousseau, 1989). Individuals can have large discrepancies with regard to the underpinnings of the psychological contract (Thomas et al., 2003). The transactional psychological contract suggests an exchange-based agreement that is classified as being short term, limited in level of involvement, and viewed as a practical means to an end. The relational psychological contract suggests an identity-based understanding that is categorized as being long term, intensive in level of involvement, and perceived as evoking collective interests (Brown, 1997; Parks & Schmedemann, 1994; Suchman, 1995; Thomas et al., 2003). Individuals perceive a violation to a psychological contract when others act contrary to the perceived parameters of the contract (Rousseau, 1995; Morrison & Robinson, 1997). Psychological contract violations can occur when referent others either intentionally or unintentionally violate the contract, with intentional violations being perceived as more severe violations. When a role violation has occurred and a breach in the psychological contract is perceived, various possible responses are available to affected individuals. Individuals may choose to be angry, bitter, and indignant or even opt for revenge or punitive retribution (Allred, 1999; Bies & Tripp, 1996; Stuckless & Goranson, 1992). Moreover, individuals may choose estrangement (Aquino, Tripp, & Bies, 2006), avoidance, escape, exit (McCullough et al., 1998; Rusbult et al., 2005), cognitive dissonance, denial (Latack, 1986), forgetting (Smedes, 1984, 1996), or feigned forgiveness via impression management tactics (Bies,

1987; Cody & McLaughlin, 1990; Goffman, 1961; Schlenker, 1980; Tomlinson & Mayer, 2009). Conversely, individuals may attempt to recover the relationship through the process of forgiveness. Forgiveness facilitates role recovery because it entails (at least) the removal of negative feelings, cognition, or behavior (Aquino, Grover, Goldman, & Folger, 2003; Horsbrugh, 1974; Lewis, 1980; Murphy, 1988; North, 1987; Richards, 1988) and may also include the return of positive feelings, cognition, and/or behavior (Aquino et al., 2003; Cameron, Dutton, & Quinn, 2003; Tangney, Wagner, Hill-Barlow, Marschall, & Gramzow, 1996). Forgiveness and subsequent role recovery can yield several benefits related to work roles. These benefits include increased self-esteem, creativity, learning, and relational boundaries (Bright, Fry, & Cooperrider, 2006; Maltby, Day, & Barber, 2004; Sandage & Williamson, 2005; Strelan & Covic, 2006; Toussaint, Williams, Musick, & Everson, 2001). However, the perceived role violation should not be forgiven prematurely (Affinito, 1999; Exline, Worthington, Hill, & McCullough, 2003; Lamb & Murphy, 2002; Murphy, 2000; Wiesenthal, 1997). Premature role recovery may result in not adequately contemplating the long-term significance of the role violation (such as legal accountability as well as acceptable and salutary relational conduct; Exline et al., 2003). Because role recovery often requires sacrifice from both parties involved in the violation of role expectations, there needs to be a sufficient motivation from both parties to recover the role (Rusbult, Verette, Whitney, Slovik, & Lipkus, 1991; van Lange et al., 1997). For example, subordinates are more likely to forgive leaders than leaders are to forgive subordinates, especially within a procedurally just environment. This finding is attributed to subordinates having more to lose when roles are not recovered with a leader, particularly when fair processes are in place to evoke positive conflict resolution from leaders and subdue subordinate vigilantism (Aquino et al., 2006). Finally, perceived justice or, in our case, injustice is a critical indicator of how difficult it will be for the role occupant (i.e., the one perceiving a breach) to “recover” and forgive. The nature of the role violation seems to be the largest contributor to how 523

Sluss, van Dick, and Thompson

Copyright American Psychological Association. Not for further distribution.

unjust the violation is perceived. For example, injustice is more likely perceived when the “violator” (a) was responsible, (b) had intentionality, (c) could have avoided the violation, and (d) committed a severe offense (Bradfield & Aquino, 1999; Darby & Schlenker, 1982; Folger & Cropanzano, 1998; Girard & Mullet, 1997; Weiner, 1995). At the same time, injustice is less likely to be perceived when the violator is perceived to express a sincere apology, concedes responsibility, and acts to restore what was violated (Ohbuchi, Kameda, & Agarie, 1989).

Advancing Role Recovery Because human imperfection can cloud understanding as to role clarity and role consensus, individuals may not sufficiently understand role expectations, thus preventing accurate detection of role violations (lack of role clarity; Bush & Busch, 1981; Posner & Butterfield, 1978). Likewise, individuals may not agree or come to consensus regarding each other’s role expectations, thus being unwilling to adopt expected role behavior (lack of role consensus; Biddle, 1979; Quick, 1979). We know that leaders develop different reciprocal relationships with different group members (Graen & Cashman, 1975; Graen & Scandura, 1987), with the boundaries of these relationships being tested as the relationship forms. We suggest that the leader–subordinate relationship is ripe for research concerning role recovery. When there is consensus as to dyadic role expectations, the relationship is likely to be of higher quality in terms of both social exchange and identity (Bauer & Green, 1996; Graen & Uhl-Bien, 1995; Maslyn & Uhl-Bien, 2001; Sluss & Ashforth, 2007; 2008; van Knippenberg, van Dick, & Tavares, 2007). Nevertheless, research has found that subordinate (member) role expectations are based more on social and developmental needs, whereas leader role expectations are based more on work-related issues (Baldwin, 1997; Baldwin & Baccus, 2003; Huang, Wright, Chiu, & Wang, 2008; Kim & Organ, 1982). As such, either party in the leader–subordinate relationship may perceive a role violation, whereas another perceives no breach. We suggest exploring the influence of role clarity (as well as the dyadic role-making process) in conjunction with role recovery. It may also be useful to examine how role recov524

ery works when there are high levels of role clarity versus low levels. Additionally, as there are individuals who internalize work and relational roles as part of who they are, there are also individuals who accept roles externally without embracing them internally. Likewise, there are benevolent forgivers who internalize feelings of empathy and compassion when forgiving, whereas there are pragmatic forgivers who may only forgive externally or for practical purposes (Bright et al., 2006). We recommend future research aimed at examining the effect role identification has on the role-recovery process. For example, is internalized forgiveness that is benevolent and compassionate required when there is strong relational and/or role identification? That is, if an individual has accepted a role as part of who he or she is, are feelings of benevolence and empathy toward the role violator required to restore the role relationship? Likewise, is pragmatic forgiveness sufficient when role identification is weak or of low intensity? The term role recovery seems to suggest that what was sufficient for the original role would be sufficient for the recovered role. However, how does a transgression change role expectations? Furthermore, how does role recovery change role expectations? After role recovery, individuals may feel concerned about transgression recurrence, or they may be emboldened by the fact that their role could be restored. Furthermore, when the dyad is confronted with the need to forgive each other, how does role identity salience influence role recovery? Does forgiving a role expectation violation become more difficult when role identity salience is higher versus lower? Does it make a difference whether the role violator is, for example, a leader (vs. subordinate), or whether the violator views the dyad as more or less salient than the offender views the dyad? These questions converge with symbolic interactionism’s view of role theory wherein emphasis is given to individual participants and the evolution of roles through interaction, whereby the role is the product of the role participant’s cumulative interaction (Biddle, 1986; Mead, 1934). As such, violations, indifference, and conformity are constantly shaping the status of the role. This leads to a potentially controversial question: Are roles stronger, in terms of

Role Theory in Organizations

role salience, after recovering from a role violation than they were prior to the violation? Of course, there are probably various contextual contingencies (moderators) that would influence the answer to this question; nevertheless, answers to this question would shed significant light on role recovery and provide a more dynamic “role” for role theory within organizations.

Copyright American Psychological Association. Not for further distribution.

CONCLUSION Despite an ever faster changing workplace due to flatter hierarchies, virtual work, mergers, alliances, and the like, roles are still important to the human condition and may be more important than ever to help scholars understand what binds the individual and the organization and how the individual is able to cope with stress arising from change. Through our relational perspective of role theory in organizations, we attempted to reconcile the individuals’ natural quest for stability and security with the environmental pressures to change and flexibly respond to new forms of organizing. The individual today both suffers and benefits from the pace of change: suffering from increasing ambiguity and decreasing clarity, but at the same time gaining from having more leeway for developing his or her own definition of what a specific role might entail or even creating new roles. In sum, we hope we have provided an analysis that is helpful for the organizational scholar and practitioner within industrial and organizational psychology as well as management to further develop modern approaches to roles within organizations.

References Abramis, D. J. (1994). Work role ambiguity, job satisfaction, and job performance: Meta-analysis and review. Psychological Reports, 75, 1411–1433. Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 267–299). New York: Academic Press.

Allen, D. G. (2006). Do organizational socialization tactics influence newcomer embeddedness and turnover? Journal of Management, 32, 237–256. Allen, N. J., & Meyer, J. P. (1990). Organizational socialization tactics: A longitudinal analysis of links to newcomers’ commitment and role orientation. Academy of Management Journal, 33, 847–858. Allred, K. G. (1999). Anger driven retaliation: Toward an understanding of impassioned conflict in organizations. In R. J. Lewicki, R. J. Bies, & B. H. Sheppard (Eds.), Research on negotiations in organizations (Vol. 7, pp. 27–58). Greenwich, CT: JAI Press. Amabile, T. M., Tighe, E. M., Hill, K. G., & Hennessey, B. A. (1994). The work preference inventory: Assessing intrinsic and extrinsic motivational orientations. Journal of Personality and Social Psychology, 66, 950–967. Aquino, K., Grover, S. L., Goldman, B., & Folger, R. (2003). When push doesn’t come to shove: Interpersonal forgiveness in workplace relationships. Journal of Management Inquiry, 12, 209–216. Aquino, K., Tripp, T. M., & Bies, R. J. (2006). Getting even or moving on? Power, procedural justice, and types of offense as predictors of revenge, forgiveness, reconciliation, and avoidance in organizations. Journal of Applied Psychology, 91, 653–668. Argyris, C. (1960). Understanding organizational behavior. Homewood, IL: Dorsey Press. Aron, A., & Aron, E. (2000). Self-expansion motivation and including other in the self. In W. Ickes & S. Duck (Eds.), The social psychology of personal relationships (pp. 109–128). Chichester, England: Wiley. Ashford, S. J. (1986). Feedback-seeking in individual adaptation: A resource perspective. Academy of Management Journal, 29, 465–487. Ashforth, B. E. (2001). Role transitions in organizational life: An identity-based perspective. Mahwah, NJ: Erlbaum. Ashforth, B. E., Harrison, S. H., & Corley, K. G. (2008). Identification in organizations: An examination of four fundamental questions. Journal of Management, 34, 325–374. Ashforth, B. E., & Johnson, S. A. (2001). Which hat to wear? The relative salience of multiple identities in organizational contexts. In M. A. Hogg & D. J. Terry (Eds.), Social identity processes in organizational contexts (pp. 31–48). Philadelphia: Psychology Press.

Affinito, M. G. (1999). When to forgive. Oakland, CA: New Harbinger.

Ashforth, B. E., Kreiner, G. E., Clark, M. A., & Fugate, M. (2007). Normalizing dirty work: Managerial tactics for countering occupational taint. Academy of Management Journal, 50, 149–174.

Ajzen, I., & Fishbein, M. (1975). A Bayesian analysis of attribution processes. Psychological Bulletin, 82, 267–277.

Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Journal, 14, 20–39. 525

Sluss, van Dick, and Thompson

Ashforth, B. E., & Saks, A. M. (1996). Socialization tactics: Longitudinal effects on newcomer adjustment. Academy of Management Journal, 39, 149–178. Ashforth, B. E., Sluss, D. M., & Harrison, S. H. (2007). Socialization in organizational contexts. G. P. Hodgkinson & J. K. Ford (Eds.), International review of industrial and organizational psychology (Vol. 22, pp. 1–70). New York: Wiley.

Copyright American Psychological Association. Not for further distribution.

Ashforth, B. E., Sluss, D. M., & Saks, A. M. (2007). Socialization tactics, proactive behavior, and newcomer learning: Integrating socialization models. Journal of Vocational Behavior, 70, 447–462. Babin, B. J., & Boles, J. S. (1996). The effects of perceived co-worker involvement and supervisor support on service provider role stress, performance and job satisfaction. Journal of Retailing, 72(1), 57–75. Baker, W. K. (1995). Allen and Meyer’s 1990 longitudinal study: A reanalysis and reinterpretation using structural equation modeling. Human Relations, 48, 169–186. Baldwin, M. W. (1997). Relational schema as a source of if–then self-inference procedures. Review of General Psychology, 1, 326–335. Baldwin, M. W., & Baccus, J. R. (2003). Relational knowledge and an expectancy-value approach to self-esteem. In S. J. Spencer, S. Fein, M. P. Zanna, & J. M. Olson (Eds.), Motivated social perception: The Ontario symposium (Vol. 9, pp. 167–188). London: Erlbaum. Banaji, M. R., & Prentice, D. A. (1994). The self in social contexts. Annual Review of Psychology, 45(1), 297–332. Barber, B. (1983). The logic and limits of trust. New Brunswick, NJ: Rutgers University Press. Bartel, C. A. (2001). Social comparisons in boundaryspanning work: Effects of community outreach on members’ organizational identity and identification. Administrative Science Quarterly, 46, 379–413. Bauer, T. N., Bodner, T., Erdogan, B., Truxillo, D. M., & Tucker, J. S. (2007). Newcomer adjustment during organizational socialization: A meta-analytic review of antecedents, outcomes, and methods. Journal of Applied Psychology, 92, 707–721. Bauer, T. N., & Green, S. G. (1996). Development of leader–member exchange: A longitudinal test. Academy of Management Journal, 39, 1538–1567. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–529. Behrman, D. N., & Perreault, Jr., W. D. (1984). A role stress model of the performance and satisfaction of industrial salespersons. Journal of Marketing, 48(4), 9–21. 526

Bell, L. G., Wicklund, R. A., Manko, G., & Larkin, C. (1976). When unexpected behavior is attributed to the environment. Journal of Research in Personality, 10, 316–327. Bellah, R. N., Madsen, R., Sullivan, W. M., Swidler, A., & Tipton, S. M. (1985). Habits of the heart: Individualism and commitment in American life. New York: Harper & Row. Berger, C. R. (1979). Beyond initial interaction: Uncertainty, understanding, and the development of interpersonal relationships. In H. Giles & R. N. St. Clair (Eds.), Language and social psychology (pp. 122–144). Baltimore: University Park Press. Biddle, B. J. (1979). Role theory: Expectations, identities, and behaviors. New York: Academic. Biddle, B. J. (1986). Recent developments in role theory. Annual Review of Sociology, 12, 67–92. Bies, R. J. (1987). The predicament of injustice: The management of moral outrage. Research in Organizational Behavior, 9, 289–319. Bies, R. J., & Tripp, T. M. (1996). Beyond distrust: “Getting even” and the need for revenge. In R. M. Kramer & T. R. Tyler (Eds.), Trust in organizations (pp. 246–260). Thousand Oaks, CA: Sage. Blau, P. (1964). Exchange and power in social life. New York: Wiley. Bolino, M. C. (1999). Citizenship and impression management: Good soldiers or good actors? Academy of Management Review, 24(1), 82–98. Bond, M. H. (1983). Linking person perception dimensions to behavioral intention dimensions. Journal of Cross-Cultural Psychology, 14, 41–63. Bond, M. H., & Forgas, J. P. (1984). Linking person perceptions to behavior intention across cultures: The role of cultural collectivism. Journal of Cross-Cultural Psychology, 15, 337–352. Borman, W. C., & Motowidlo, S. J. (1993). Expanding the criterion domain to include elements of contextual performance. In N. Schmitt & W. C. Borman (Eds.), Personality selection (pp. 71–98). San Francisco: Jossey-Bass. Bradfield, M., & Aquino, K. (1999). The effects of blame attributions and offender likableness on revenge and forgiveness in the workplace. Journal of Management, 25, 607–631. Braverman, H. (1974). Labor and monopoly capital: The degradation of work in the twentieth century. New York: Monthly Review Press. Brewer, M. B. (1991). The social self: On being the same and different at the same time. Personality and Social Psychology Bulletin, 17, 475–482. Brewer, M. B., & Gardner, W. (1996). Who is this “we”? Levels of collective identity and self-representations.

Role Theory in Organizations

Journal of Personality and Social Psychology, 71, 83–93. Brickson, S. (2000). The impact of identity orientation on individual and organizational outcomes in demographically diverse settings. Academy of Management Review, 25(1), 82–101. Bridges, W. (1994). Job shift: How to prosper in a workplace without jobs. Reading, MA: Addison-Wesley.

Copyright American Psychological Association. Not for further distribution.

Bright, D. S., Fry, R. E., & Cooperrider, D. L. (2006). Forgiveness from the perspectives of three response modes: Begrudgement, pragmatism, and transcendence. Journal of Management Spirituality and Religion, 2/3(1), 78–103. Brown, A. D. (1997). Narcissism, identity and legitimacy. Academy of Management Review, 22, 643–686. Burt, R. S. (1982). Toward a structural theory of action: Network models of social structure, perception, and action. New York: Academic. Bush, R. F., & Busch, P. (1981). The relationship of tenure and age to role clarity and its consequences in the industrial salesforce. Journal of Personal Selling & Sales Management, 2(1), 17–23. Byrne, D. (1961). Interpersonal attraction and attitude similarity. Journal of Abnormal and Social Psychology, 62, 713–715. Caldwell, D. F., & O’Reilly, III, C. A. (1990). Measuring person-job fit with a profile-comparison process. Journal of Applied Psychology, 75, 648–657. Cameron, K. S., Dutton, J. E., & Quinn, R. E. (2003). Positive organizational scholarship. San Francisco, CA: Berrett-Koehler. Chen, G., & Bliese, P. D. (2002). The role of different of levels of leadership in predicting self- and collective efficacy: Evidence for discontinuity. Journal of Applied Psychology, 87, 549–556. Chreim, S., Williams, B., & Hinings, C. R. (2007). Interlevel influences on the reconstruction of professional role identity. Academy of Management Journal, 50, 1515–1539. Christ, O., van Dick, R., Wagner, U., & Stellmacher, J. (2003). When teachers go the extra-mile: Foci of organizational identification as determinants of different forms of organizational citizenship behavior among schoolteachers. British Journal of Educational Psychology, 73, 329–341. Churchill, Jr., G. A., Ford, N. M., Hartley, S. W., & Walker, O. C. (1985). The determinants of salesperson performance: A meta analysis. Journal of Marketing Research, 22, 103–118. Cody, M. J., & McLaughlin, M. L. (1990). Interpersonal accounting. In H. Giles & W. P. Robinson (Eds.), Handbook of language and psychology (pp. 227–255). Oxford, England: Wiley.

Darby, B. W., & Schlenker, B. R. (1982). Children’s reactions to apologies. Journal of Personality and Social Psychology, 43, 742–753. Darley, J. M., & Fazio, R. H. (1980). Expectancy confirmation processes arising in the social interaction sequence. American Psychologist, 35, 867–881. DeCarlo, T. E., Teas, R. K., & McElroy, J. C. (1997). Salesperson performance attribution processes and the formation of expectancy estimates. Journal of Personal Selling & Sales Management, 17(3), 1–17. de Ruyter, K., Wetzels, M., & Feinberg, R. (2001). Role stress in call centres: Its effects on employee performance and satisfaction. Journal of Interactive Marketing, 15(2), 23–35. Dienesch, R. M., & Liden, R. C. (1986). Leader-member exchange model of leadership: A critique and further development. Academy of Management Review, 11, 618–634. Donahue, E. M. (1994). Do children use the Big Five, too? Content and structural form in personality description. Journal of Personality, 62(11), 45–66. Dweck, C. S. (1991). Self-theories and goals: Their role in motivation, personality, and development. In R. A. Dienstbier (Ed.), Nebraska Symposium on Motivation: Perspectives on motivation (pp. 199–235). Lincoln, NE: University of Nebraska Press. Elsbach, K. D. (1999). An expanded model of organizational identification. Research in Organizational Behavior, 21, 163–200. Exline, J. J., Worthington, Jr., E. L., Hill, P., & McCullough, M. E. (2003). Forgiveness and justice: A research agenda for social and personality psychology. Personality & Social Psychology Review, 7, 337–348. Falcione, R. L., & Wilson, C. E. (1988). Socialization processes in organizations. G. M. Goldhaber & G. A. Barnett (Eds.), Handbook of organizational communication (pp. 301–338). Norwood, NJ: Ablex. Finkelstein, M. A., & Penner, L. A. (2004). Predicting organizational citizenship behavior: Integrating the functional and role identity approaches. Social Behavior and Personality, 32, 383–398. Fletcher, G. J. O., Simpson, J. A., Thomas, G., & Giles, L. (1999). Ideals in intimate relationships. Journal of Personality and Social Psychology, 76(1), 72–89. Folger, R. G., & Cropanzano, R. (1998). Organizational justice and human resource management. Thousand Oaks, CA: Sage. Francis, J. N. P. (1991). When in Rome? The effects of cultural adaption on intercultural business negotiations. Journal of International Business Studies, 22, 403–428. Frese, M., Fay, D., Hilburger, T., Leng, K., & Tag, A. (1997). The concept of personal initiative: 527

Sluss, van Dick, and Thompson

Operationalization, reliability and validity of two German samples. Journal of Occupational & Organizational Psychology, 70, 139–161.

Copyright American Psychological Association. Not for further distribution.

Frese, M., Garst, H. & Fay, D. (2007). Making things happen: Reciprocal relationships between work characteristics and personal initiative in a four-wave longitudinal structural equation model. Journal of Applied Psychology, 92, 1084–1102.

Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of leader– member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. Leadership Quarterly, 6, 219–247. Grant, A. M. (2007). Relational job design and the motivation to make a prosocial difference. Academy of Management Review, 32, 393–417.

Frese, M., Kring, W., Soose, A., & Zempel, J. (1996). Personal initiative at work: Differences between East and West Germany. Academy of Management Journal, 39, 37–63.

Griffin, M. A., Neal, A., & Parker, S. K. (2007). A new model of work role performance: Positive behavior in uncertain and interdependent contexts. Academy of Management Journal, 50, 327–347.

Fried, Y., Slowik, L. H., Shperling, Z., Franz, C., Ben-David, H. A. Avital, N., & Yeverechyahu, U. (2003). The moderating effect of job security on the relation between role clarity and job performance: A longitudinal field study. Human Relations, 56, 787–805.

Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16, 250–279.

Fry, L. W., Futrell, C. M., Parasuraman, A., & Chmielewski, M. A. (1986). An analysis of alternative causal models of salesperson role perceptions and work-related attitudes. Journal of Marketing Research, 23, 153–163. Gerstner, C. R., & Day, D. V. (1997). Meta-analytic review of leader-member exchange theory: Correlates and construct issues. Journal of Applied Psychology, 82, 827–844. Giles, H., & Smith, P. (1979). Accommodation theory: Optimal levels of convergence. In H. Giles & R. N. St. Clair (Eds.), Language and social psychology (pp. 45–63). Baltimore: University Park Press. Gilmore, A., & Mooreland, L. (2000). Call centre: How can service quality be managed? Irish Marketing Review, 13(1), 3–11. Ginossar, Z., & Trope, Y. (1980). The effects of base rates and individuating information on judgment about another person. Journal of Experimental Social Psychology, 16, 228–242. Girard, M., & Mullet, E. (1997). Propensity to forgive in adolescents, young adults, older adults, and elderly people. Journal of Adult Development, 4, 209–220. Goffman, E. (1961). Asylums. Garden City, NY: Anchor Books. Graen, G. (1976). Role making processes within complex organizations. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1201–1245). Chicago: Rand McNally. Graen, G., & Cashman, J. F. (1975). A role-making model of leadership in formal organizations: A developmental approach. In J. G. Hunt & L. L. Larson (Eds.), Leadership frontiers (pp. 143–166). Kent, Ohio: Kent State University Press. Graen, G., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing. Research in Organizational Behavior, 9, 175–208. 528

Hall, M. (2008). The effect of comprehensive performance measurement systems on role clarity, psychological empowerment and managerial performance. Accounting, Organizations & Society, 33, 141–163. Haslam, S. A. (2001). Psychology in organizations: The social identity approach. London: Sage. Haslam, S. A. (2004). Psychology in organizations: The social identity approach (2nd ed.). London: Sage. Haslam, S. A., & Ellemers, N. (2005). Social identity in industrial and organizational psychology: Concepts, controversies, and contributions. In G. P. Hodgkinson & J. Kevin Ford (Eds.), International review of industrial and organizational psychology (Vol. 20, pp. 39–118). England: Wiley. Haslam, S. A., Postmes T., & Ellemers, N. (2003). More than a metaphor: Organizational identity makes organizational life possible. British Journal of Management, 14, 357–369. Hattrup, K., & Jackson, S. E. (1996). Learning about individual differences by taking situations seriously. In K. R. Murphy (Ed.), Individual differences and behavior in organizations (pp. 507–547). San Francisco: Jossey-Bass. Heimer, K., & Matsueda, R. L. (1994). Role-taking, role commitment, and delinquency: A theory of differential social control. American Sociological Review, 59, 365–390. Higgins, E. T., Klein, R. L., & Strauman, T. J. (1987). Selfdiscrepancies: Distinguishing among self-states, selfstate conflicts, and emotional vulnerabilities. In K. M. Yardley & T. M. Honess (Eds.), Self and identity: Psychosocial perspectives (pp. 173–186). New York: Wiley. Hirst, G., van Dick, R., & van Knippenberg, D. (in press). A social identity perspective on leadership and employee creativity. Journal of Organizational Behavior. Hofmann, D. A., Morgeson, F. P., & Gerras, S. J. (2003). Climate as a moderator of the relationship between

Role Theory in Organizations

leader-member exchange and content-specific citizenship: Safety climate as an exemplar. Journal of Applied Psychology, 88, 170–178. Hofstede, G. (1991). Cultures and organizations: Software of the mind. London: McGraw-Hill. Hogg, M. A., & Terry, D. J. (2000). Social identity and self-categorization processes in organizational contexts. Academy of Management Review, 25, 121–140.

Copyright American Psychological Association. Not for further distribution.

Hogg, M. A., Terry, D. J., & White, K. M. (1995). A tale of two theories: A critical comparison of identity theory with social identity theory. Social Psychology Quarterly, 58, 255–269. Horsbrugh, H. J. N. (1974). Forgiveness. Canadian Journal of Philosophy, 4, 269–282. Huang, X., Wright, R. P., Chiu, W. C., & Wang, C. (2008). Relational schemas as sources of evaluation and misevaluation of leader-member exchanges: Some initial evidence. The Leadership Quarterly, 19, 266–282. Ibarra, H. (1999). Provisional selves: Experimenting with image and identity in professional adaptation. Administrative Science Quarterly, 44, 764–791. Ilgen, D. R., & Hollenbeck, J. R. (1991). The structure of work: Job design and roles. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2, pp. 165–207). Palo Alto, CA: Consulting Psychologists Press. Jablin, F. M. (1987). Organizational entry, assimilation, and exit. In F. M. Jablin, L. L. Putnam, K. H. Roberts, & L. W. Porter (Eds.), Handbook of organizational communication: An interdisciplinary perspective (pp. 679–740). Thousand Oaks, CA: Sage. Jackson, S. E., & Schuler, R. S. (1985). A meta-analysis and conceptual critique of research on role ambiguity and role conflict in work settings. Organizational Behavior and Human Decision Processes, 36(1), 16–78. James, K., & Greenberg, J. (1989). In-group salience, intergroup comparison, and individual performance and self-esteem. Personality and Social Psychology Bulletin, 15, 604–616. Johnson, M. J., Morgeson, F. P., Ilgen, D. R., Meyer, C. J., & Lloyd, J. W. (2006). Multiple professional identities: Examining differences in identification across work-related targets. Journal of Applied Psychology, 91, 498–506. Jones, G. R. (1986). Socialization tactics, self-efficacy, and newcomers’ adjustments to organizations. Academy of Management Journal, 29, 262–279. Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964). Occupational stress: Studies in role conflict and ambiguity. New York: Wiley. Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.). New York: Wiley.

Kim, K. I., & Organ, D. W. (1982). Determinants of leader-subordinate exchange relationships. Group and Organization Studies, 7(1), 77–89. King, R. C., & Sethi, V. (1992). Socialization of professionals in high-technology firms. Journal of High Technology Management Research, 3, 147–168. Korsgaard, M. A., Sapienza, H. J., & Schweiger, D. M. (2002). Beaten before begun: The role of procedural justice in planning change. Journal of Management, 28, 497–516. Kreiner, G. E., & Ashforth, B. E. (2004). Evidence toward an expanded model of organizational identification. Journal of Organizational Behavior, 25(1), 1–27. Kreiner, G. E., Ashforth, B. E., & Sluss, D. M. (2006). Identity dynamics in occupational dirty work: Integrating social identity and system justification perspectives. Organization Science, 17, 619–636. Kreiner, G. E., Hollensbe, E. C., & Sheep, M. L. (2006). Where is the me among the we? Identity work and the search for optimal balance. Academy of Management Journal, 49, 1031–1057. Kristof, A. L. (1996). Person-organization fit: An integrative review of its conceptualizations, measurement, and implications. Personnel Psychology, 49(1), 1–49. Lamb, S., & Murphy, J. G. (2002). Before forgiving: Cautionary views of forgiveness in psychotherapy. New York: Oxford University Press. Lang, J., Thomas, J. L., Bliese, P. D., & Adler, A. B. (2007). Job demands and job performance: The mediating effect of psychological and physical strain and the moderating effect of role clarity. Journal of Occupational Health Psychology, 12, 116–124. Latack, J. C. (1986). Coping with job stress: Measures and future decisions for scale development. Journal of Applied Psychology, 71, 377–385. Leonard, N. H., Beauvais, L. L., & Scholl, R. W. (1999). Work motivation: The incorporation of self-conceptbased processes. Human Relations, 52, 969–998. Lewis, M. (1980). On forgiveness. Philosophical Quarterly, 30(120), 236–245. Lincoln, J. R., & Miller, J. (1979). Work and friendship ties in organizations: A comparative analysis of related networks. Administrative Science Quarterly, 24, 181–199. Lord, R. G., Brown, D. J., & Freiberg, S. J. (1999). Understanding the dynamics of leadership: The role of follower self-concepts in the leader/follower relationship. Organizational Behavior & Human Decision Processes, 78, 167–203. Lyons, P. (2006). Individual competitiveness and spontaneous changes in jobs. Advances in Competitiveness Research, 14(1), 90–98. 529

Sluss, van Dick, and Thompson

Maltby, J., Day, L., & Barber, L. (2004). Forgiveness and mental health variables: Interpreting the relationship using an adaptational–continuum model of personality and coping. Personality and Individual Differences, 37, 1629–1641. Markus, H. R., & Cross, S. (1990). The interpersonal self. In L. Pervin (Ed.), Handbook of personality: Theory and research (pp. 576–608). New York: Guilford Press. Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224–253.

Copyright American Psychological Association. Not for further distribution.

Markus, H. R., & Nurius, P. (1986). Possible selves. American Psychologist, 41, 954–969. Maslyn, J. M., & Uhl-Bien, M. (2001). Leader–member exchange and its dimensions: Effects of self-effort and other’s effort on relationship quality. Journal of Applied Psychology, 86(4), 697–708. Mayer, R. C, Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20, 709–734. McAllister, D. J., Morrison, E. W., Kamdar, D., & Turban, D. B. (2007). Disentangling role perceptions: How perceived role breadth, discretion, instrumentality, and efficacy relate to helping and taking charge. Journal of Applied Psychology, 92, 1200–1211. McCullough, M. E., Rachal, K. C., Sandage, S. J., Worthington, Jr., E. L., Brown, S. W., & Hight, T. L. (1998). Interpersonal forgiving in close relationships II: Theoretical elaboration and measurement. Journal of Personality and Social Psychology, 75, 1586–1603. Mead, G. H. (1934). Mind, self, and society. Chicago: University of Chicago Press. Meglino, B. M., Ravlin, E. C., & Adkins, C. L. (1989). A work values approach to corporate culture: A field test of the value congruence process and its relationship to individual outcomes. Journal of Applied Psychology, 74, 424–432. Meglino, B. M., Ravlin, E. C., & Adkins, C. L. (1991). Value congruence and satisfaction with a leader: An examination of the role of interaction. Human Relations, 44, 481–495. Merton, R. K. (1957). Social theory and social structure. Glencoe, IL: Free Press. Mignerey, J. T., Rubin, R. B., & Gorden, W. I. (1995). Organizational entry: An investigation of newcomer communication behavior and uncertainty. Communication Research, 22(1), 54–85. Miller, K. I., & Monge, P. R. (1986). Participation, satisfaction, and productivity: A meta-analytic review. Academy of Management Journal, 29, 727–753. Miller, V. D., & Jablin, F. M. (1991). Information seeking during organizational entry: influences, tactics, and a 530

model of the process. Academy of Management Review, 16, 92–120. Miller, V. D., Johnson, J. R., Hart, Z., & Peterson, D. L. (1999). A test of antecedents and outcomes of employee role negotiation ability. Journal of Applied Communication Research, 27(1), 24–48. Moorman, R. H. (1991). The relationship between organizational justice and organizational citizenship behaviors: Do fairness perceptions influence employee citizenship? Journal of Applied Psychology, 76, 845–855. Moorman, R. H., Niehoff, B. P., & Organ, D. W. (1993). Treating employees fairly and organizational citizenship behaviors: Sorting the effects of job satisfaction, organizational commitment, and procedural justice. Employee Responsibilities & Rights Journal, 6, 209–225. Morgeson, F. P., & Camnion, M. A. (1997). Social and cognitive sources of potential inaccuracy in job analysis. Journal of Applied Psychology, 82, 627–655. Morgeson, F. P., Delaney-Klinger, K. A., & Hemingway, M. A. (2005). The importance of job autonomy, cognitive ability, and job-related skill for predicting role breadth and job performance. Journal of Applied Psychology, 90, 399–406. Morrison, E. W. (1993). Newcomer information seeking: Exploring types, modes, sources, and outcomes. Academy of Management Journal, 36, 557–589. Morrison, E. W. (1994). Role definitions and organizational citizenship behaviors: The importance of the employees’ perspective. Academy of Management Journal, 37, 1543–1567. Morrison, E. W., & Phelps, C. C. (1999). Taking charge at work: Extrarole efforts to initiate workplace change. Academy of Management Journal, 42, 403–419. Morrison, E. W., & Robinson, S. L. (1997). When employees feel betrayed: A model of how psychological contract violation develops. Academy of Management Review, 22, 226–256. Murphy, J. G. (1988). Forgiveness and resentment. In J. G. Murphy & J. Hampton (Eds.), Forgiveness and mercy (pp. 14–33). Cambridge, England: Cambridge University Press. Murphy, J. G. (2000). Two cheers for vindictiveness. Punishment and Society, 2, 131–143. Nicholson, N. (1984). A theory of work role transitions. Administrative Science Quarterly, 29, 172–191. North, J. (1987). Wrongdoing and forgiveness. Philosophy, 62, 499–508. Oakes, P. J. (1987). The salience of social categories. In J. C. Turner, M. A. Hogg, P. J. Oakes, S. D. Reicher, & M. S. Wetherell (Eds.), Rediscovering the social group (pp. 117–141). New York: Blackwell.

Role Theory in Organizations

Oakes, P. J., Turner, J. C., & Haslam, S. A. (1991) Perceiving people as group members: The role of fit in the salience of social categorizations. British Journal of Social Psychology, 30, 125–144. Ohbuchi, K., Kameda, M., & Agarie, N. (1989). Apology as aggression control: Its role in mediating appraisal of and response to harm. Journal of Personality and Social Psychology, 56, 219–227.

Copyright American Psychological Association. Not for further distribution.

O’Reilly, C. A., III, Caldwell, D. F., & Barnett, W. P. (1989). Work group demography, social integration, and turnover. Administrative Science Quarterly, 34(1), 21–37. Organ, D. W. (1977). A reappraisal and reinterpretation of the satisfaction-causes-performance hypothesis. Academy of Management Review, 2, 46–53. Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: Lexington Books. Organ, D. W. (1990). The motivational basis of organizational citizenship behavior. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 12, pp. 43–72). Greenwich, CT: JAI Press. Organ, D. W., & Konovsky, M. A. (1989). Cognitive versus affective determinants of organizational citizenship behavior. Journal of Applied Psychology, 74, 157–164. Parker, S. K. (1998). Enhancing role breadth self-efficacy: The roles of job enrichment and other organizational interventions. Journal of Applied Psychology, 83, 835–852. Parker, S. K. (2007). That is my job: How employees’ role orientation affects their job performance. Human Relations, 60, 403–434. Parker, S. K., Wall, T. D., & Jackson, P. R. (1997). “That’s not my job”: Developing flexible employee work orientations. Academy of Management Journal, 40, 899–929. Parker, S. K., Williams, H. M., & Turner, N. (2006). Modeling the antecedents of proactive behavior at work. Journal of Applied Psychology, 91, 636–652. Parks, J. M., & Schmedemann, D. A. (1994). When promises become contracts: Implied contract and handbook provisions on job security. Human Resource Management, 33, 403–423.

Posner, B. Z., & Butterfield, D. A. (1978). Role clarity and organizational level. Journal of Management, 4(2), 81–90. Pratt, M. G. (1998). To be or not to be? Central questions in organizational identification. In D. A. Whetten & P. C. Godfrey (Eds.), Identity in organizations. Building theory through conversations (pp. 171–207). Thousand Oaks, CA: Sage. Pratt, M. G., & Foreman, P. O. (2000). Classifying managerial responses to multiple organizational identities. Academy of Management Review, 25, 18–42. Pratt, M. G., Rockmann, K., & Kaufmann, J. (2006). Constructing professional identity: The role of work and identity learning cycles in the customization of identity among medical professionals. Academy of Management Journal, 49, 235–262. Quick, J. C. (1979). Dyadic goal setting within organizations: Role-making and motivational considerations. Academy of Management Review, 4, 369–380. Reay, T., Golden-Biddle, K., & Germann, K. (2006). Legitimizing a new role: Small wins and microprocesses of change. Academy of Management Journal, 49, 977–998. Revicki, D. A., Whitley, T. W., Gallery, M. E., & Allison, E. J. (1993). Impact of work environment characteristics on work-related stress and depression in emergency medicine residents: A longitudinal study. Journal of Community & Applied Social Psychology, 3, 273–284. Richards, N. (1988). Forgiveness. Ethics, 99, 77–97. Richter, A. W., West, M. A., van Dick, R., & Dawson, J. F. (2006). Boundary spanners’ identification, intergroup contact and effective intergroup relations. Academy of Management Journal, 49, 1252–1269. Riketta, M. (2005). Organizational identification: A metaanalysis. Journal of Vocational Behavior, 66, 358–384. Riketta, M., & van Dick, R. (2005). Foci of attachment in organizations: A meta-analytic comparison of the strength and correlates of workgroup versus organizational identification and commitment. Journal of Vocational Behavior, 67, 490–510. Rioux, S. M., & Penner, L. A. (2001). The causes of organizational citizenship behavior: A motivational analysis. Journal of Applied Psychology, 86, 1306–1314.

Parsons, T. (1951). The social system. Glencoe, IL: Free Press.

Roberts, B. W., & Donahue, E. M. (1994). One personality, multiple selves: Integrating personality and social roles. Journal of Personality, 62, 199–218.

Piliavin, J. A., & Callero, P. L. (1991). Giving blood: The development of an altruistic identity. Baltimore: Johns Hopkins University Press.

Robinson, S. L. (1996). Trust and breach of the psychological contract. Administrative Science Quarterly, 41, 574–599.

Pinkley, R. L., LaPrelle, J., Pyszczynski, T., & Greenberg, J. (1988). Depression and the self-serving search for consensus after success and failure. Journal of Social and Clinical Psychology, 6, 235–244.

Robinson, S. L., Kraatz, M. S., & Rousseau, D. M. (1994). Changing obligations and the psychological contract: A longitudinal study. Academy of Management Journal, 37, 137–152. 531

Sluss, van Dick, and Thompson

Robinson, S. L., & Morrison, E. W. (2000). The development of psychological contract breach and violations: A longitudinal study. Journal of Organizational Behavior, 21, 525–546. Rotheram, M., LaCour, J., & Jacobs, A. (1982). Variations in group process due to valence, response mode, and directness of feedback. Group & Organization Studies, 7(1), 67–75. Rousseau, D. M. (1989). Psychological and implied contracts in organizations. Employee Responsibilities and Rights Journal, 2, 121–139.

Copyright American Psychological Association. Not for further distribution.

Rousseau, D. M. (1995). Psychological contracts in organizations: Understanding written and unwritten agreements. Thousand Oaks, CA: Sage.

Serpe, R. T. (1987). Stability and change in self: A structural symbolic interactionist explanation. Social Psychology Quarterly, 50(1), 44–55. Serpe, R. T., & Stryker, S. (1987). The construction of self and the reconstruction of social relationships. In E. Lawler & B. Markovsky (Eds.), Advances in group processes: Theory and research (Vol. 4, pp. 41–82). Greenwich, CT: JAI Press. Shore, L. M., & Tetrick, L. E. (1994). The psychological contract as an explanatory framework in the employment relationship. In C. L. Cooper & D. M. Rousseau (Eds.), Trends in organizational behavior (Vol. 1, pp. 91–109). Chichester, England: Wiley.

Rousseau, D. M. (2005). I-deals: Idiosyncratic deals employees bargain for themselves. Armonk, NY: M. E. Sharpe.

Sluss, D. M., & Ashforth, B. E. (2007). Relational identity and identification: Defining ourselves through work relationships. Academy of Management Review, 32, 9–32.

Rusbult, C. E., Hannon, P. A., Stocker, S. L., & Finkel, E. J. (2005). Forgiveness and relationship repair. In E. L. Worthington, Jr. (Ed.), Handbook of forgiveness (pp. 185–206). New York: Brunner-Routledge.

Sluss, D. M., & Ashforth, B. E. (2008). How relational and organizational identification converge: Processes and conditions. Organization Science, 19, 807–823.

Rusbult, C. E., Verette, J., Whitney, G. A., Slovik, L. F., & Lipkus, I. (1991). Accommodation processes in close relationships: Theory and preliminary empirical evidence. Journal of Personality and Social Psychology, 60(1), 53–80.

Smedes, L. B. (1984). Forgive and forget: Healing the hurts we don’t deserve. New York: Harper & Row. Smedes, L. B. (1996). The art of forgiving. Nashville, TN: Moorings.

Saks, A. M., & Ashforth, B. E. (1997). Organizational socialization: Making sense of the past and present as a prologue for the future. Journal of Vocational Behavior, 51, 234–279.

Spreitzer, G. M., Kizilos, M. A., & Nason, S. W. (1997). A dimensional analysis of the relationship between psychological empowerment and effectiveness, satisfaction, and strain. Journal of Management, 23(5), 679–695.

Salancik, G. R., & Pfeffer, J. (1978). A social information processing approach to job attitudes and task design. Administrative Science Quarterly, 23, 224–253.

Staw, B. M., & Boettger, R. D. (1990). Task revision: A neglected form of work performance. Academy of Management Journal, 33(3), 534–559.

Sandage, S. J., & Williamson, I. (2005). Forgiveness in cultural context. In E. L. Worthington, Jr. (Ed.), Handbook of forgiveness (pp. 41–56). New York: Brunner-Routledge.

Stets, J. E., & Burke, P. J. (2000). Identity theory and social identity theory. Social Psychology Quarterly, 63, 224–237.

Sawyer, J. E. (1992). Goal and process clarity: Specification of multiple constructs of role ambiguity and a structural equation model of their antecedents and consequences. Journal of Applied Psychology, 77, 130–142. Schein, E. H. (1965). Organizational psychology. Englewood Cliffs, NJ: Prentice-Hall. Schein, E. H. (1971). Occupational socialization in the professions: The case of role innovation. Journal of Psychiatric Research, 8, 521–530. Schlenker, B. R. (1980). Impression management. Belmont, CA: Brooks/Cole. Schriesheim C. A., Castro S. L., & Cogliser C. C. (1999). Leader-member exchange (LMX) research: A comprehensive review of theory, measurement, and data-analytic practices. Leadership Quarterly, 10(1), 63–113. 532

Stone-Romero, E. F., Stone, D. L., & Salas, E. (2003). The influence of culture on role conceptions and role behavior in organizations. Applied Psychology, 52, 328–362. Strelan, P., & Covic, T. (2006). A review of forgiveness process models and a coping framework to guide future research. Journal of Social & Clinical Psychology, 25, 1059–1085. Stryker, S. (2007). Identity theory and personality theory: Mutual relevance. Journal of Personality, 75, 1083–1102. Stryker, S., & Burke, P. J. (2000). The past, present, and future of an identity theory. Social Psychology Quarterly, 63, 284–297. Stryker, S., & Serpe, R. T. (1982). Commitment, identity salience, and role behavior: Theory and research example. In W. Ickes & E. S. Knowles (Eds.),

Role Theory in Organizations

Personality, roles, and social behaviour (pp. 199–218). New York: Springer-Verlag. Stryker, S., & Serpe, R. T. (1994). Identity salience and psychological centrality: Equivalent, overlapping, or complementary concepts? Social Psychology Quarterly, 57(1), 16–35. Stuckless, N., & Goranson, R. (1992). The vengeance scale: Development of a measure of attitudes toward revenge. Journal of Social Behavior and Personality, 7(1), 25–42.

Copyright American Psychological Association. Not for further distribution.

Suchman, M. C. (1995). Managing legitimacy: Strategic and institutional perspectives. Academy of Management Review, 20, 571–610. Sujan, H. (1986). Smarter versus harder: An exploratory attributional analysis of salespeople’s motivation. Journal of Marketing Research, 23(1), 41–49. Swann, W. B., Jr. (1987). Identity negotiation: Where two roads meet. Journal of Personality and Social Psychology, 53, 1038–1051. Tajfel, H. (1978). Social categorization, social identity, and social comparison. In H. Tajfel (Ed.), Differentiation between social groups. Studies in the social psychology of intergroup relations (pp. 61–76). London: Academic Press.

Terry, D. J., & Callan, V. J. (1998). In-group bias in response to an organizational merger. Group Dynamics: Theory, Research, and Practice, 2(2), 67–81. Thomas, D. C., Au, K., & Ravlin, E. C. (2003). Cultural variation and the psychological contract. Journal of Organizational Behavior, 24, 451–471. Thomas, D. C., & Ravlin, E. C. (1995). Responses of employees to cultural adaptation by a foreign manager. Journal of Applied Psychology, 80(1), 133–146. Tomlinson, E., & Mayer, R. C. (2009). The role of causal attribution dimensions in trust repair. Academy of Management Review, 34, 85–104. Toussaint, L. L., Williams, D. R., Musick, M. A., & Everson, S. A. (2001). Forgiveness and health: Age differences in a U.S. probability sample. Journal of Adult Development, 8, 249–257. Triandis, H. C. (1980). Values, attitudes, and interpersonal behavior. In H. Howe & M. Page (Eds.), Nebraska symposium on motivation (pp. 195–259). Lincoln, NE: University of Nebraska Press. Tropp, L. R., & Wright, S. C. (2001). Ingroup identification as inclusion of ingroup in the self. Personality and Social Psychology Bulletin, 27, 585–600.

Tajfel, H., Billig, M. G., Bundy, R. P., & Flament, C. (1971). Social categorization and intergroup behaviour. European Journal of Social Psychology, 1(2), 149–178.

Turner, J. C. (1999). Social identity and self-categorization. In N. Ellemers, R. Spears, & B. Doosje (Eds.), Social identity: Context, commitment, content (pp. 6–34). Oxford, England: Blackwell.

Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987). Rediscovering the social group: A self-categorization theory. Oxford, England: Blackwell.

Tajfel, H., & Turner, J. C. (1986). The social of identity theory of intergroup behavior. In S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24). Chicago: Nelson-Hall.

Turner, N., Chmiel, N., & Walls, M. (2005). Railing for safety: Job demands, job control, and safety citizenship role definition. Journal of Occupational Health Psychology, 10, 504–512.

Tangney, J. P., Wagner, P. E., Hill-Barlow, D., Marschall, D. E., & Gramzow, R. (1996). Relation of shame and guilt to constructive versus destructive responses to anger across the lifespan. Journal of Personality and Social Psychology, 70, 797–809.

Turnley, W. H., & Feldman, D. C. (2000). Re-examining the effects of psychological contract violations: Unmet expectations and job dissatisfaction as mediators. Journal of Organizational Behavior, 21(1), 25–42.

Teas, R. K. (1980). Churchill and Ford model of sales force motivation and performance. Journal of the Academy of Marketing Science, 8(1/2), 58–72.

Ullrich, J., Wieseke, J., Christ, O., Schulze, M., & van Dick, R. (2007). The identity matching principle: Corporate and organizational identification in a franchising system. British Journal of Management, 18, 29–44.

Teas, R. K. (1983). Supervisory behavior, role stress, and the job satisfaction of industrial salespeople. Journal of Marketing Research, 20, 84–91. Teas, R. K., Wacker, J. G., & Hughes, R. E. (1979). A path analysis of causes and consequences of salesmen’s perceptions of role clarity. Journal of Marketing Research, 16, 335–369. Tepper, B. J., Lockhart, D. L., & Hoobler, J. (2001). Justice, citizenship and role definition effects. Journal of Applied Psychology, 86, 789–796.

Unzicker, D., Clow, K. E., & Babakus, E. (2000). The role of organizational communications on employee perceptions of a firm. Journal of Professional Services Marketing, 21(2), 87–103. van Dick, R. (2001). Identification in organizational contexts: Linking theory and research from social and organizational psychology. International Journal of Management Reviews, 3, 265–283. 533

Sluss, van Dick, and Thompson

Copyright American Psychological Association. Not for further distribution.

van Dick, R. (2004). My job is my castle: Identification in organizational contexts. In C. L. Cooper & I. T. Robertson (Eds.), International Review of Industrial and Organizational Psychology (Vol. 19, pp. 171–204). Chichester, England: Wiley.

Wagner, U., & Ward, P. L. (1993). Variation of out-group presence and evaluation of the in-group. British Journal of Social Psychology, 32, 241–251. Weiner, B. (1995). Judgments of responsibility. New York: Guilford Press.

van Dick, R., Christ, O., Stellmacher, J., Wagner, U., Ahlswede, O., Grubba, C., et al. (2004). Should I stay or should I go? Explaining turnover intentions with organizational identification and job satisfaction. British Journal of Management, 15, 351–360.

Welbourne, T. M., Johnson, D. E., & Erez, A. (1998). The role-based performance scale: Validity analysis of a theory-based measure. Academy of Management Journal, 41, 540–555.

van Dick, R., Grojean, M. W., Christ, O., & Wieseke, J. (2006). Identity and the extra-mile: Relationships between organizational identification and organizational citizenship behaviour. British Journal of Management, 17, 283–301.

Whetten, D. (2007). A critique of organizational identity scholarship: Challenging the uncritical use of social identity theory when social identities are also social actors. In C. A. Bartel, S. Blader, & A. Wrzesniewski (Eds.), Identity and the modern organization (pp. 253–272). Mahwah, NJ: Erlbaum.

van Dick, R., van Knippenberg, D., Kerschreiter, R., Hertel, G., & Wieseke, J. (2008). Interactive effects of work group and organizational identification on job satisfaction and extra-role behavior. Journal of Vocational Behavior, 72, 388–399. van Dick, R., Wagner, U., Stellmacher, J., & Christ, O. (2004). The utility of a broader conceptualization of organizational identification: Which aspects really matter? Journal of Organizational and Occupational Psychology, 77, 171–191. van Dick, R., Wagner, U., Stellmacher, J., & Christ, O. (2005). Category salience and its effects on organizational identification. Journal of Organizational and Occupational Psychology, 78, 273–285. van Knippenberg, D., van Dick, R., & Tavares, S. (2007). Social identity and social exchange: Identification, support, and withdrawal from the job. Journal of Applied Social Psychology, 37, 457–477. van Knippenberg, D., & van Schie, E. C. M. (2000). Foci and correlates of organizational identification. Journal of Occupational and Organizational Psychology, 73, 137–147. van Lange, P. A. M., Drigotas, S. M., Rusbult, C. E., Arriaga, X. B., Witcher, B. S., & Cox, C. L. (1997). Willingness to sacrifice in close relationships. Journal of Personality and Social Psychology, 72, 1373–1395. van Maanen, J., & Schein, E. H. (1979). Toward a theory of organizational socialization. In B. M. Staw (Ed.), Research in Organizational Behavior (Vol. 1, pp. 209–264). Greenwich, GT: JAI Press. von Emster, G. R., & Harrison, A. A. (1998). Role ambiguity, spheres of control, burnout, and work-related attitudes of teleservice professionals. Journal of Social Behavior and Personality, 13, 375–385.

534

Whitaker, B. G., Dahling, J. J., & Levy, P. (2007). The development of a feedback environment and role clarity model of job performance. Journal of Management, 33, 570–591. Wieseke, J., Ullrich, J., Christ, O., & van Dick, R. (2007). Organizational identification as a determinant of customer orientation in service firms. Marketing Letters, 18, 265–278. Wiesenthal, S. (1997). The sunflower: On the possibilities and limits of forgiveness. New York: Schocken. Williamson, O. E. (1985). The economic institutions of capitalism: Firms, markets, relational contracting. New York: Free Press. Wrzesniewski, A., & Dutton, J. E. (2001). Crafting a job: Revisioning employees as active crafters of their work. Academy of Management Review, 26, 179–201. Wrzesniewski, A., McCauley, C., Rozin, P., & Schwartz, B. (1997). Jobs, careers, and callings: People’s relations to their work. Journal of Research in Personality, 31(1), 21–33. Young, B. S., Worchel, S., & Woehr, D. J. (1998). Organizational commitment among public service employees. Public Personnel Management, 27, 339–350. Zavalloni, M. (1980). Values. In H. Triandis, & R. W. Brislin (Eds.), Handbook of cross-cultural psychology (Vol. 5, pp. 73–120). Boston: Allyn & Bacon. Zenger, T. R., & Lawrence, B. S. (1989). Organizational demography: The differential effects of age and tenure on technical communication. Academy of Management Journal, 32, 353–376.

CHAPTER 17

FLEXIBLE WORK SCHEDULES

Copyright American Psychological Association. Not for further distribution.

Ellen Ernst Kossek and Jesse S. Michel

Flextime has made our work force more efficient and more focused while they are working. It is a step backwards to go back to rock solid hours. As long as an employee is getting the job done, they should be treated like an adult. (Hernreich, 2008) Flexible work schedules, such as flextime, telework, or compressed workweeks, are examples of increasing variation in the timing and duration of work hours and the location of work. Although standard work schedules have traditionally been the norm in organizations, growing numbers of employers are experimenting with a wide range of flexible work schedules at the same time as they are transforming employment systems and work processes across time zones and cultures. The increasing proliferation of flexible and more varied work schedules for organizational members is a global employment phenomenon (Jacobs, Gerson, & Gornick, 2004). National country studies from the United States to Australia estimate that only about half of employees work a standard fixed daytime work schedule 5 days a week (Golden, 2001; Watson, Buchanan, Campbell, & Briggs, 2003). As the opening quotation suggests, when implemented with both employer and

employee interests in mind, flexible work schedules can increase efficiency and work focus and empower individuals to self-manage work time (Halpern, 2005; Kossek, 2005). Flexible work schedules are an increasingly important issue for industrial and organizational (I/O) psychology because they reflect the adaptation of human resource practices to the changing nature of work, seen in a labor force increasingly diverse in work time availability and in dramatic changes in the design of work systems in response to a 24–7 global economy. Accordingly, many new challenges are created for I/O psychologists. For example, how can we rigorously assess the benefits of flexible work schedules for individuals and organizations? When and how should flexible work schedules be used to attract and retain an increasingly diverse workforce? What are strategies for managing and socializing talent when people are working many different schedules across different time zones with little face-to-face interaction? What are the best selection tools to identify individuals who will work well in jobs involving global teams with constant technological interaction over a 24–7 period? What is the optimal design of training programs to help supervisors coordinate and motivate employees who have many

We thank Leslie B. Hammer for her input to an earlier version of this chapter and Sheldon Zedeck and anonymous reviewers for their comments. This research was partially supported by the Work, Family and Health Network, which is funded by a cooperative agreement through the National Institutes of Health and the Centers for Disease Control and Prevention: National Institute of Child Health and Human Development (NICHD; Grants U01HD051217, U01HD051218, U01HD051256, U01HD051276), National Institute on Aging (Grant U01AG027669), Office of Behavioral and Social Sciences Research, and National Institute for Occupational Safety and Health (Grant U010H008788). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of these institutes and offices. Special acknowledgment goes to extramural staff science collaborator Rosalind Berkowitz King (NICHD) and Lynne Casper (now of the University of Southern California, Los Angeles) for the design of the original Workplace, Family, Health and Well-Being Network Initiative. Persons interested in learning more about the network should go to http://www.kpchr.org/workplacenetwork/, http://ellenkossek.lir.msu.edu/, and http://www.wfsupport.psy.pdx.edu/.

http://dx.doi.org/10.1037/12169-017 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

535

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

different schedules? How can high performance cultures be created and contributions accurately assessed when employees have less face time at work? What are effective coaching programs to reduce work–life conflicts for virtual workers who have simultaneous access to work and life demands? When are flexible work schedules effective as organizational development interventions to reduce job stress and improve productivity and when do they increase stress? These are just some of the pertinent questions regarding flexible work schedules that pose new issues for the field of I/O psychology to investigate. What we found in our review is that scholars have been more successful in answering the first two research questions on the potential benefits of flexible work schedules, and who desires them, than in clarifying how to ensure successful implementation and adaptation of human resource systems and organizational cultures (Ryan & Kossek, 2008). Our chapter is organized as follows: (a) flexible work schedules overview; (b) relevant theories; (c) measurement challenges and cross-cutting characteristics of what makes a flexible schedule “flexible”; (d) individual and organizational outcomes; and (e) future research and directions. FLEXIBLE WORK SCHEDULES OVERVIEW In this section, we give a brief overview of the history, organizational rationale, and types of flexible work schedules.

History Historically, prior to the U.S. industrialization period of the mid-1800s, most workers were either farmers or self-employed, thus determining their own work schedules (Ronen, 1981). Then standardized employer-set work schedules, with work carried out away from the home or a personal business, started appearing as large factories spread with industrialization. A traditional full-time schedule was assumed to be a 40-hour week during which employees worked an 8-hour day, 5 days a week, with fixed starting and stopping times (Avery & Zabel, 2001). Hunnicutt (1996) described an important historical development that occurred in December 1930. To create jobs for laid-off workers 536

during the Great Depression, the Kellogg Company, the largest manufacturer of cereal in the world, altered the standard of an 8-hour day conducted over three shifts, substituting four 6-hour shifts. Employee morale increased as a result of more leisure time, there were fewer accidents, and the price per unit of production declined as employees worked more productively (Avery & Zabel, 2001). The program was publicized as a national model, supported by many stakeholders from government to labor to business. Although the company briefly went back to offering only 8-hour shifts during the World War II exigencies, both 6-hour and 8-hour shifts were offered in the postwar decades. Hunnicutt (1996, p. 106) noted the “feminization of shorter hours,” as women were the biggest supporters and users opting for the 6-hour day. Except for men near retirement or disabled workers, most men continued to work the 8-hour day. During an economic downturn in the 1980s, in order to reduce headcount and benefits costs, Kellogg ended the 6-hour day, but by then the notion of flexible work schedules had developed as a corporate experiment, primarily serving the needs of women and noncore workers. The 6-hour day initiative provides an important historical remnant for 21st century organizations, as flexible work scheduling has gradually become mainstream, allowing for growing employee discretion over at least some aspects of work scheduling.

Growth From a macro-organizational perspective, labor market, cost, and environmental and technological forces are driving employers to implement flexible work schedules. Labor market demographic shifts reveal a workforce that increasingly needs and values flexibility. Statistics show an explosive growth in the number of individuals who must ensure that family responsibilities are managed while they are at work. Although we cite U.S. statistics here, these trends are mirrored around the world. Since 1975, the labor force participation of U.S. women with children under 18 years age has increased from 47% to 78% (Kossek, 2006). Nearly 40% of all professionals and managers who work at major U.S. companies are now women, many of whom simultaneously juggle caregiving and their jobs (Bond, Thompson, Galinsky, &

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

Prottas, 2003). The U.S. Census Bureau reported that 82% of U.S. families are dual earners or single parents with children under the age of 18 years at home (Bureau of Labor Statistics, 2009). A third of all workers (equally men and women) provide elder care (Bond et al., 2002). Fifty percent of all children will live in a single-parent family before reaching 18 years (Cohen, 2002). Fathers play a greater role in caregiving and value flexibility more than those of previous generations (Pleck, 1997). Millenials, the current generation of workers entering the workforce, take a more balanced approach to work than previous generations (Deal, 2007). Product and labor cost savings are also driving growth of flexible work schedules. The adoption of contingent and part-time work schedules, as well as temporary extra shifts, allows employers to expand and contract workforce size and employment at will in response to variation in product demand, economic uncertainty, and new market developments in the global economy (see also chap. 18, this volume). Globalization and rising consumer demand, as well as the high costs of shutting down continuous processing manufacturing systems, mandate 24–7 operations, with production and service delivery around the clock for many firms. A cross-national sample of firms shows that the information technology sector is at the forefront of having a flexible, mobile, often off-shore workforce, which enables firms to quickly hire staff, form partnerships, and develop a customer base around the globe (Landry, Mahesh, & Hartman, 2005; MacEachen, Polzer, & Clarke, 2008). Contingent work schedules reduce labor costs. Companies typically have a two-tiered workforce: a core group and a noncore group. One group consists of full-time employees who have better health care and pension benefits and some job security. The other is a contingent work group of workers with less favorable benefits and hours, who can be easily laid off to quickly reduce labor costs. This ability to reduce headcount through a contingent contract is especially critical in the European Union (EU), where it is increasingly difficult to lay off regular workers without legally mandated employment severance, which can take months to negotiate (Mery, 2009).

Telework reduces office costs by enabling more efficient facility management and space use (Karnowski & White, 2002). One review summarizing costs savings noted that IBM saved over $75 million in annual real estate costs, whereas the U.S, General Services Administration (GSA) had major reductions in office energy costs (Kurkland & Bailey, 1999). A study by Robèrt and Börjesson (2006) found significant reductions in rental costs from introducing flexible offices and telecommuting at a Swedish telecom. Yet some scholars warn that the employer cost savings may be at worker expense, as shifting operations to workers’ homes increases home office costs (Davenport & Pearlson, 1998). Flexible schedules help employers support the environment and cut workers’ fuel costs at the same time. After gas prices spiked to over $4 a gallon in the United States, Oklahoma and Kentucky adopted state-sponsored telework and flextime programs specifically designed to help workers save on fuel. Utah mandated a 4-day workweek for 17,000 state employees, about 80% of the state workforce (Kossek, 2008). Teleworking and 4-day workweeks or delayed schedule starts reduce traffic congestion, fuel consumption, and air and noise pollution (Balepur, Varma, & Mokhtarian, 1998). Unproductive time spent in traffic is reduced by allowing individuals to commute during off-peak times. Empirical evidence of these effects is mixed. Studies by Bernardino and Ben-Akiva (1996) and Mokhtarian (1998) relying on mathematical models to simulate and estimate the favorable environmental impacts of teleworking found little or no positive effects of teleworking on air pollution reduction. Yet Henderson and Mokhtarian (1996) found that having neighborhood telework centers cut motor vehicle transmissions by half and also increased time spent working, improved performance, and enhanced job satisfaction. Technological changes in the way work is structured due to the growth in use of electronic computer and voice tools have made work more portable, facilitating employees’ abilities to work anywhere, anytime. More employers have become comfortable with flexibility as technological tools enhance the ability to electronically monitor employee productivity (Venkatesh & Johnson (2002). 537

Kossek and Michel

Copyright American Psychological Association. Not for further distribution.

Types of Flexible Work Schedules: When, Where, How Much, and Continuity Current descriptions of flexible work schedules all build on the concept of employee scheduling discretion, thus enabling employees to have some choice to determine how long, when, or where they are engaged in work for various time periods (e.g., days, weeks, seasons). This discretion affects how an individual experiences his or her working time in relation to nonworking time, such as time spent on leisure and domestic activities, from caregiving to household labor to relaxation (Fagan, 2001). Evans, Kunda, and Barley (2004) defined flexible work schedules as allowing employees to determine when they start and stop work hours, how many hours they work, which days or shifts they work, or where they work. Rau (2003) defined flexible work schedules as alternative work options enabling work to be conducted outside the temporal or spatial boundaries of a “standard” workday. Taken together, these definitions provide several organizing criteria. Types of flexible work schedules can be organized into four design criteria: (a) flexibility in when one works, such as the timing of work; (b) flexibility in where one works, such as the location or place of work; (c) flexibility in how much one works, such as the amount of work or workload; and (d) flexibility in the continuity of work, such as short- and long-term breaks in work activity and time off. These design criteria can be overlapping and used in various combinations to create hybrid flexible work arrangements. Drawing on Kossek and Van Dyne (2008), Table 17.1 gives an overview of these schedule types, which are discussed below. Most of the I/O literature focuses on flexible work schedules chosen by employees, which have generally had a positive connotation for employee well-being, particularly when used to reduce work–life conflicts. We also note a related research stream in the sociology and poverty literatures on nonstandard schedules (cf. Presser, 2003), which generally have a negative connotation for worker well-being, particularly when used not by choice by lower wage or hourly workers (e.g., shift work), temporary workers (e.g., contingent work), or professionals feeling compelled to overwork (e.g., work excessive hours or during leisure time). 538

Flexibility in the Timing of Work Most flexible work schedules relate to the timing of work. Flextime is the most common, followed by the compressed workweek, shift work, and contingent work. Flextime. Flextime originated in Germany in the 1970s, and although it quickly spread across Western and Northern Europe, the United States was slower to adopt it, particularly in the private sector (Avery & Zabel, 2001). Under flextime, employees have the discretion to vary the times they arrive and leave work, within management parameters, to meet their personal needs (Avery & Zabel, 2001). Flextime schedules have a predetermined range of times in which employees can arrive (e.g., 6–10 a.m.) and leave (e.g., 3:00–7:00 p.m.), with a core band between work starting and stopping times when all employee must be present (e.g., 10:00 a.m.– 3:00 p.m.). Having core hours helps managers with the coordination of meetings and supervision (Van Dyne, Kossek, & Lobel, 2007). Flextime policies sometimes incorporate daily carryover, where employees can vary their work schedules with regard to daily time spent at work, as long as they spend a predetermined set amount of weekly time at work (e.g., 40 hours per week). Though estimates vary, about one fourth (Golden, 2001) to nearly two fifths (Bond et al., 2003) of U.S. workers have access to flextime, up considerably from about 1 in 10 workers in 1985 (Golden, 2001). Professional and higher level employees are more likely to have access to flextime than are lower level employees. Direct service and manufacturing jobs offer less access to flextime than do jobs in other industries (Kossek & Distelberg, 2009). Compressed workweek. Under a compressed workweek, an employee works a full-time schedule in fewer than 5 days. The most common compressed 40-hour workweek is a 4-day, 10-hour schedule with a Monday or Friday off (Pierce, Newstrom, Dunham, & Barber, 1989). Another concept is the 9-hour work day, with 1 additional hour added to the 8-hour day. Known as a 9–80 schedule, this compressed workweek occurs over a 2-week period. A key benefit of compressed workweeks is that

Flexible Work Schedules

TABLE 17.1 Types and Examples of Flexible Work Schedules Basic type and definition

Examples

Flexibility in the timing of work Definition: Flexibility in when work occurs

■ ■ ■ ■ ■ ■ ■ ■

Copyright American Psychological Association. Not for further distribution.

■

Flexibility in the location or place of work Definition: Flexibility in the location or place where work occurs

■ ■ ■ ■

Flexibility in amount of work (reduced workload and hours) Definition: Flexibility in the amount of work or workload

■ ■ ■ ■ ■ ■ ■ ■ ■ ■

Flexibility in work continuity (short-term breaks in employment or time off) Definition: Flexibility to allow for employment breaks or time off

■

■ ■ ■ ■ ■ ■

employees can have a 3-day weekend every week (with four 10-hour days) or every other week (with eight 9-hour days). Compressed workweeks are more common in North America (especially Canada) than in other parts of the world (Avery & Zabel, 2001). About 15% of U.S. employees have access to the compressed workweek (Bond et al., 2003). It is more common for lower level than senior employees, and in police and nursing occupations more than in other job families. Shift work. Although shift work is not always thought of as a flexible work schedule, it is a common form of nonstandard working time. It can involve

Flextime Core days Results-based professional work Contingent work Rotating shifts Shift work 4-day workweek Compressed workweek Weekend, evening, night work Telework or flexplace satellite offices, neighborhood work centers Required travel or client office work Split locations Informal telework combined with nonstandard working time Job sharing Reduced load or customized work Part-time work Temporary layoffs Temporary shutdown Required reduced or part-time hours Overtime mandates or limits Reduced hours Phased retirement Work-study or coops Short-term or long-term leaves (e.g., educational, travel, family, maternity, disability, military) Sabbaticals Extended or indefinite paid and unpaid leaves of absence Vacation Sick time or disability time off Part-year work Intermittent leave

evening (e.g., 3 p.m.–11:00 p.m.), night (11:00 p.m.– 7:00 a.m.), or weekend hours; rotating shifts (e.g., evenings one day, nights the next), or double shifts (e.g., 16 hours) when a worker is not relieved from 24–7 operations such as in hospitals, prisons, or factories. Sometimes an employee can have a regular but nonstandard schedule, such as a set 8-hour work schedule that always takes place at night (Barnett & Hall, 2007). Some workers do choose shift work, as it allows them to pursue other life pursuits, such as education or child care, during the day. About 15% of the U.S. labor force works nonstandard or irregular schedules, often in the service and technical industries (see the U.S. 539

Kossek and Michel

Copyright American Psychological Association. Not for further distribution.

Bureau of Labor Statistics [USBLS] National Compensation Survey results, http://www.bls. gov/ncs/ebs/). In France, about 10% of workers work nonstandard hours, compared with 20% in other EU countries, such as Greece and the United Kingdom (Presser, 2003). Contingent work. Contingent work is defined as a flexible work arrangement in which an individual does not explicitly or implicitly contract for longterm employment or works minimum hours that vary irregularly (Polivka & Nardone, 1989). Examples of contingent workers include seasonal, temporary in-house, or freelance workers (Connelly & Gallagher, 2004). Under a contingent work schedule, the hiring of workers is based on a temporary fixed-term contract, unlike a traditional employment agreement, which has an expectation of an ongoing employment relationship. The three commonly used government measures of contingent work are (a) whether an employee does not expect a job to last more than a year; (b) whether the employee is self-employed or an independent contractor; and (c) whether the employee has worked in a job less than a year and is expecting it to end within the year. The USBLS estimated that contingent workers accounted for 1.8% to 4.1% of total U.S. employment in 2005. While usually the exception, some contingent workers prefer temporary work because it allows them to choose employers and work hours and take extended time off (Ashford, George, & Blatt, 2008).

Flexibility in the Location or Place of Work Another common form of flexible work schedule relates to the location or place of work. Common arrangements are telework or flexplace, and informal teleworking often combined with nonstandard working time. Telework or flexplace. Under a telework or flexplace schedule, employees work from a location outside of their physical organizational setting. Telework or flexplace is defined as a flexible work arrangement that allows employees to access labor activities from many varied locations, typically using technologies transmitting communication and information (Pérez, Sánchez, & de Luis Carnicer, 2003). Although there are many forms of telework or flexplace, four defin540

ing types capture most of them: (a) telecommuting, (b) satellite offices, (c) neighborhood work centers, and (d) mobile workers (Kurkland & Bailey, 1999). Telecommuters work from home on a regular basis and may or may not use technology in their work. Employees at satellite and neighborhood work offices work outside the home and organization. However, employees at satellite offices are from a single organization, whereas employees at neighborhood work centers can be from multiple organizations but share office space in a local suburban area rather than commuting to a downtown center. Such opportunities allow employees to engage in regular interactions with work colleagues (e.g., conference calls via video feeds) while reducing the length of the commute and the need to purchase urban office space. Mobile workers are transient and typically work from multiple locations that vary depending on the customer being served. These employees are sometimes referred to as “road warriors.” They generally face more cognitive complexity, fatigue, and mobility than do teleworkers who work virtually from a regular location (Kossek & Lautsch, 2008). Of U.S. employees, 15% telework at least 1 day a week (see USBLS National Compensation Survey results, http://www.bls.gov/ncs/ebs/).

Informal Teleworking Combined With Nonstandard Working Time Besides a growth in use of formal human resource policies supporting flexible work schedules, informal flexible work schedules are a rising trend that needs to be considered when referring to teleworking. The nature of many jobs has changed to be increasingly virtual, flexible, and self-regulated with growing access to portable e-work, defined as electronic work from BlackBerrys, cell phones, or laptops (Kossek & Lautsch, 2008). Work is increasingly being diffused over all hours of the day or week, extending later into the night and starting earlier in the morning and also spreading into vacations and weekends (Hamermesh, 1999). It has also spread from employer locations to our homes and to many third places, as from cybercafés to our cell phones and BlackBerrys while commuting. More and more individuals are casually teleworking in planes, trains, and automobiles or in public places like coffee shops and restaurants.

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

Nevertheless, from a formal human resource policy perspective, casual teleworkers such as these would not necessarily be viewed as working on a flexible schedule. Yet this growth in informal flexible scheduling practice needs to be noted in I/O studies. For example, the expansion of casual telework makes studying the effects of formal telework use challenging. One quasi-experimental study found contamination of a control group, identified by the HR department as non-teleworkers, because many of them were often informally telecommuting before or after work or on weekends to handle rising workloads (Kossek, Lautsch, & Eaton, 2006). Pure telework control groups can be difficult to create because telework and nonstandard work hours often occur together. As Golden (2001) reported, use of telework is positively related to an employee’s acknowledging access to flextime regarding the start or end of work during the day. This exemplifies how different types of flexibility may be used in bundles. It is also important to understand the reasons for informal flexible work schedule use, particularly for boundary blurring practices, as some support nonwork demands while others support work demands. As examples of each, a supervisor may regularly allow an employee to work from home unofficially every Friday to accommodate day care constraints for a newborn infant who sleeps most of the day while the parent works. Or an employee who uses e-mail, texting, or cell phones on his or her job habitually is expected to take work phone calls and check e-mail during unofficial working time from home (sometimes referred to as overwork).

Flexibility in Amount of Work (Workload and Hours) A third form of flexible work schedule, part-time work, relates to the amount of work (lower workload or hours). After describing part-time work generally, we discuss two growing subtypes: job sharing and reduced-load work. Part-time work. Under a part-time work schedule, employees work fewer than 35 hours per week (see USBLS National Compensation Survey results, http://www.bls.gov/ncs/ebs/). One of the most common flexible work schedules in the world, part-time

work grew after World War II to accommodate employers’ needs to cut labor costs and the demographic shifts that had brought more women into the labor force (Tilly, 1996). There are several subtypes of part-time work, such as job sharing, in which two people share a job for a reduced workload, or customized work arrangements by which an individual’s workload is reduced in return for less pay or hours. Sometimes health benefits and pensions are not offered with these arrangements unless workers work a minimum number of hours, usually at least 50% or 75% of full-time hours, and even then benefits may be prorated. Nearly one in five U.S. workers is a permanent part-time employee. In the EU, this figure ranges from 5% in Greece to 39% in the Netherlands, with an average of 16% (Avery & Zabel, 2001). There are two main types of part-time jobs: retention part-time jobs, in which workers negotiate part-time as a retention strategy (such as job sharing and reduced-load work); and secondary labor market part-time jobs, in which employees who prefer full-time work take these jobs as a way to enter the labor force (Tilly, 1996). Job sharing. Under a job sharing schedule, two employees voluntarily share work responsibilities where each works less than full-time (Christensen & Staines, 1990). Sometimes job sharers have complementary skills, with each performing a different aspect of a full-time job, such as one person focusing on the human resource aspects and the other on the financial duties (Kossek & Lee, 2005). In other cases, the job sharers split parts of a single full-time job and operate as one. Here there must be considerable trust and coordination between employees. Sometimes these jobs are designed to have some overlap of a few extra hours or a common day to ensure tradeoffs are done smoothly. In still other cases, the job sharers might perform two completely different part-time jobs, but together their work hours add up to a single full-time employee equivalent of work hours (Pierce, Newstrom, Dunham, & Barber, 1989). Customized or reduced-load work. U.S. companies have tremendous latitude to decide what are expected weekly hours for exempt professionals (Kossek & Distelberg, 2009). The U.S. Fair Labor 541

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

Standards Act (FLSA) regulates only overtime pay for nonexempt workers who work more than 40 hours per week. Consequently, professionals and managers habitually can work much longer hours than what the FLSA considers to be a full-time week for nonexempt workers. With work hours increasing, terms such as part-time and full-time have shifted in meaning to be more loosely linked to actual work hours, particularly for professional exempt workers who can work up to 60 or 70 hours a week with no overtime paid (Williams & Calvert, 2002). One reason for the growing work hours of exempt employees is that professionals are being socialized to work “as long as it takes to get the job done.” Working long hours and spending face time at work is construed as commitment and a performance proxy (Jacobs & Winslow, 2004). With recent layoffs and staffing reductions, professionals face rising workloads and may fear job loss if they work less. Given lengthening time demands for professional work, reduced-load or customized work has arisen as a new variant of part-time work developed for professional and managerial jobs. Growing numbers of individuals want to work in a profession but not the 50- or 60-hour workweek that many full-time professionals are socialized to work (Hill, Märtinson, Ferris & Baker, 2004). Under reducedload schedules, employees undergo a reduction in work hours or load and take a pay cut. For example, if the normal load for a research scientist at a pharmaceutical company is four research projects, an individual working 75% load would be assigned three projects instead of four and take a 25% pay cut (Kossek & Lee, 2008). Most reduced-load work arrangements are unique in design and based on an agreement between a specific supervisor and employee to reduce hours or workload. One study of nearly 80 reduced-load workers found professionals customized their working time to an average of 31.9 hours per week, with a range of 20 to 55 hours (Lee, MacDermid, & Buck, 2000). Even though working 55 hours may seem excessive, for some professional jobs—for example, those at the vice president or director level of a major corporation—it can still be socially and practically viewed as involving a workload reduction. Finally, phased retirement is another example of reduced-load work, in which 542

full-time employees are allowed to gradually reduce their workloads and hours before retirement.

Flexibility to Allow for Short-Term Breaks in Employment or Time Off Receiving considerably less attention than other flexible work schedules are sabbaticals, vacations, leaves, and part-year work. These flexible work arrangements allow for short-term breaks in employment without losing one’s job. These are increasingly important flexible work schedule forms because they enable individuals to maintain their relationships with their employers, yet have a break from work responsibilities. Such breaks help individuals to engage in renewal, undergo new skill development, travel, conduct military service, attend to caregiving or health demands, or prevent burnout. Sabbaticals. Under a flexible work arrangement that allows sabbaticals, employees take a prolonged paid time away from work and expect to return to their same jobs at the end of the sabbatical (Etzion, 2003; University of Illinois Office of the Vice President for Academic Affairs [UIOVPAA], 1996). Sabbaticals are traditionally linked to universities and academic positions as a means to allow for skill enhancement or renewal after heavy teaching loads or administrative work. Although less available in the private sector and often distributed on a caseby-case basis to higher-performing employees, sabbaticals have increasingly been adopted by many Fortune 1000 corporations such as Apple, McDonald’s, Segal, American Express, and Du Pont (UIOVPAA, 1996). Leaves, vacation, and flex-leaves. Under a flexible work arrangement that allows for leaves of absence, employees are allowed to be absent from work or work duty for a set period of time to handle domestic or personal needs. This absence can range from a few minutes (e.g., intermittent leave) or hours off during the workday to several weeks, months, or longer (Ivanovic & Collin, 2006). Leaves can be paid or unpaid and granted for many reasons, including military or religious demands, training for a marathon, adoption, short-term disability, maternity, paternity, foster care, caring for a sick child or relative, or educational purposes (Galinsky et al., 2004).

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

One of the most common leaves is maternity leave. The United States is somewhat unique among industrialized countries in that it does not offer mandated publicly paid leave. Employers have no legal obligation to offer paid leaves, specifically for maternity or child care (Stebbins, 2001). Consequently, less than 50% of employed women in the United States receive paid leave during the first 12 weeks after the birth of a child. Only 7% of employers provide paid paternity leave of any duration (U.S. Office of Personnel Management, 2001). In contrast, in Canada, employees may take job-protected maternity leave with full or partial pay for up to 1 year. In the EU, mothers are provided 14 weeks paid leave, which can be extended with additional partial paid parental leave if fathers also use the leave to share in caregiving (Kelly, 2006). The United States does have the Family and Medical Leave Act (FMLA). The FMLA is defined by the U.S. Department of Labor (1993) as a federally mandated law requiring employers with 50 or more part- or full-time employees to provide unpaid leave and time off from work up to 12 weeks in any 12-month period for the birth or adoption of a child, for an employee’s serious health condition, or to enable the employee to care for a spouse, parent, minor, or disabled child who has a serious health condition. The FMLA requires employers to continue employee health care insurance coverage during the leave and, when the employee returns, to provide the same or an equivalent position that the employee held before the leave. Studies show some employers do not publicize the FMLA very effectively and often resist implementation (Baum, 2006). Increasingly, companies are combining vacation time with leaves and sick time to create a paid-timeoff leave bank, where employees can use the time off in increments in whatever combination of time off they would like. Unfortunately, this approach can sometimes mean that employees use their leave time for domestic and caregiving needs and end up not having time left to take vacation to provide for personal leisure, work recovery, or their own illness. Some employees, particularly professionals with heavy workloads and long hours, typically do not take all of their vacation they could officially take under the policy and lose these days off. Many com-

panies have adopted a “use it or lose it” policy whereby firms deny employees carryover of paid vacation as a way to minimize future labor cost liability, without reducing workloads to allow employees to actually use all their vacation days. Even with a use it or lose it penalty, in a bad economy where layoffs are occurring and time at work is viewed as commitment, workers are reluctant to use all of their vacation. In contrast, in EU countries, at least a month of annual vacation is common. Part-year work. Under a part-year work arrangement, workers are typically employed to fulfill seasonal or short-term needs. This enables organizations to maintain flexible and short-term staffing (Druker, White, Hegewisch, & Mayne, 1996). Some professions attract high-level talent by offering seasonal flexibility in annual scheduling, such as academic, teaching, and tourism jobs. Other industries hire seasonal migrant workers, for example, in construction and agricultural jobs, or offer part-year employment to handle variations in customer seasonal demand (e.g., holiday retail jobs, tax accounting firms, ski resorts).

Section Summary As noted in the preceding review, a flexible work schedule allows employee flexibility in one or more of the design criteria: when, where, how much, or the continuity of work. While these design features of different types of flexibility are a good start, most studies are very descriptive, which makes studying flexible work schedules in an integrative and theoretical manner not as easy as it first appears. THEORIES RELEVANT TO FLEXIBLE WORK SCHEDULES In this section, we review several emerging theoretical perspectives relevant to the study of flexible work schedules. They are psychological control, motivation, and work–family conflict perspectives, of which boundary theory is a subset. A growing body of research has shown that using flexible work schedules leads to greater perceptions of control, lower work–family conflict, and lower turnover or intention to leave. Seminal research is also being done on the motivational and boundary management 543

Kossek and Michel

literature on the effects of flexibility work schedule use. Some of the ideas that follow are speculative where noted.

Copyright American Psychological Association. Not for further distribution.

Psychological Job Control Theory Researchers (e.g., Deci & Ryan, 1995; Karasek & Theorell, 1990) have long pointed to the importance of employees’ having high perceptions of job control and support for their individual well-being. Key constructs pertinent to flexible work schedules based on job control theory include perceptions of job control over work hours and perceived job autonomy. A key assumption of the literature on flexible work schedules is that using them relates positively to employee perceptions of job control over scheduling and increased job autonomy in job design. However, not all studies assess whether use of flexible work schedules does indeed relate to greater perceptions of control and autonomy. Control is a concept from the demand–control model of work stress. It is defined as the decision latitude employees have over their job tasks (Karasek, 1979). The demand–control model posits positive relationships between workers’ job demands and their ability to control how and when they perform a job, such as when and how they carry out tasks (Karasek & Theorell, 1990). It is assumed that a job with high demands and low control will lead to stress; however, an individual in the same high-demand job who perceives high control will experience lower strain (Grönlund, 2007). Flexible work schedules are an intervention that could enable greater control by providing tangible and psychological resources to enhance well-being. Although job control traditionally refers to employees’ perceptions of control over how work is done (Karasek, 1979; Karasek & Theorell, 1990), more recently, Kelly and Moen (2007) and Kossek, Lautsch, and Eaton (2006) extended the notion of job control to refer to control over when and where people work, in addition to control over how work is done. Although Kelly and Moen found that perceptions of increased control over the timing and place of work among professionals who work at a corporate headquarters was related to decreased work–family conflict; Kossek and colleagues’ study of teleworkers did not find that use of flexibility necessarily led to more control or lower work– 544

family conflict. An explanation for the lack of positive results for teleworkers may be that they were stigmatized for working in a different way. An additional explanation is that their workloads and pace of work were excessive, and therefore mere use of flexibility did not lead to greater control. It is likely that the type of flexible schedule used may differentially relate to control perceptions, which in turn may moderate individual and organizational outcomes. Nearly all the studies reviewed in this section measured employee perceptions of schedule control and not actual or nonsame source assessment of control over schedules.

Implications of Flexibility Over Timing of Work for Control Employee use of flextime and compressed workweeks allows workers more control over their ability to integrate personal role demands with work role demands. For example, by being able to control the timing of the starting or stopping of work schedules, an employee on flextime can restructure work hours at the end of each day to deal with nonwork demands, such as a late babysitter or the need to attend a school conference, get a car fixed, or go to the doctor, without having to miss an entire day of work. In the case of a compressed workweek (e.g., 4 days of 10 hours each), control over the timing of nonwork demands is increased because an employee can schedule appointments and other nonwork activities during the regularly scheduled fifth day off. Absenteeism is lowered for users of both flextime and compressed workweek because of this ability to cluster personal appointments during employeecontrolled nonworking time. Reviews of shift work suggest its effects are less positive for control (cf. Presser, 2003). Studies show that working a night shift, and especially rotating shifts or a swing shift, even when by choice, is generally bad for health because it disrupts sleep patterns. Often there is less control over the ability to develop an established sleep schedule. One reason for this is that even if night-shift workers always have a regular time off during the day to (hypothetically) sleep, that time may be when other members of the household (e.g., spouses or children) may be awake. The employee often does not get a full period

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

of rest because he or she may sacrifice sleep in order to be involved in daytime domestic life. It is important to note that shift work, in relation to population census representation, tends to be disproportionately delegated to low-income and minority workers (Presser, 1999). Despite these concerns, particularly for night work, shift work allows some employees to have greater control over their ability to participate in other meaningful nonwork roles, such as child care, attending school events, or volunteering, or earning a pay differential. Regarding control linkages to contingent work, individuals working a contingent schedule have a means to be able to control which days or times of the year they will work to enable them to take breaks from work when needed, such as to attend school or care for a sick child (e.g., a substitute teacher, an on-call per diem nurse). But many contingent workers work a contingent schedule as a first step to garnering full-time employment. In this case, working a contingent schedule may not increase perceptions of control—quite the opposite, as the employee often experiences job insecurity or underemployment. For use to lead to greater control, one must assess whether an individual prefers contingent work.

Implications of Flexibility Over Place of Work for Control In a nationwide sample of several hundred salaried professional workers and managers in the financial services and computer industries, Kossek and Lautsch (2009) found that being a formal user of a corporate telework policy was correlated with significantly higher perceptions of personal job control (r = .31), but higher schedule irregularity (r = .12). They also found that individuals who reported that they had a higher volume of “portable E work,” defined as work that was portable electronic work that could be performed away from the main office, reported significantly higher place mobility (r = .22). These individuals were more likely to be working in multiple places, such as one day on an airplane and with customers the next. Thus, use of telework has the trade-offs of increased control over location, but less control over hours.

Implications of Part-Time Work for Control When individuals use part-time work schedules, they will have increased perceptions of control coupled with decreased demands, since workload is reduced. Karasek and Theorell (1990) would argue that this type of situation leads to the most beneficial outcomes for workers. However, when the opportunity is not presented as an option, such as the case with involuntary part-time work (when the worker prefers full-time hours and pay), the sense of control is diminished. While many high-income workers wish to cut back hours and can often afford to do so, low-income workers may face underemployment or forced part-time work, which they may not desire because they need the income and health care benefits. Research suggests that part-time workers are sometimes less likely to get promoted, while women and older people are more likely to work in part-time jobs that permit caregiving and meeting other life demands (Hammer & Barbera, 1997). In some EU countries, there is a concern that part-time work is leading to lower control because hours of pay are being cut, though workloads are not. This phenomenon is referred to as work-intensification where individuals are working fewer hours yet expected to complete the same amount of work in less time.

Implications of Short-Term Breaks for Control Research on work recovery substantiates the importance of giving workers autonomy to control when they may take breaks from work for mental and physical health (Sonnentag, 2001). Control over time away from work counteracts job stress and helps to maintain a person’s well-being. Totterdell, Spelten, Smith, Barton, and Folkard (1995) demonstrated that worker well-being significantly increased with each additional day off from work. Psychological detachment theory suggests that resources necessary for work can then be regained during off-job time so that recovery can occur (Sonnentag, 2001). In one unpublished cross-national study (Davidson, Eden, & Westman, 2004), 16 faculty members reported their level of job stress prior to and after sabbatical, compared with a larger matched control group who were 545

Kossek and Michel

Copyright American Psychological Association. Not for further distribution.

not on leave during the same time period. Those who were on sabbatical reported very small effect size improvements in perceptions of control, positive affect, and life satisfaction. Regarding the effects of vacations, there is very little rigorous research (e.g., studies using quasiexperimental repeated measures). However, in one meta-analysis of only seven studies, findings suggest that although vacation has positive effects on health and well-being, the effects were modest, d = 0.43, and soon fade out after work resumes (d = −0.38; de Bloom et al., 2009). MOTIVATION AND WORK–FAMILY PERSPECTIVES Vroom’s (1964) expectancy theory holds that individuals are more likely to be motivated to exert effort to perform for valued goals they think they can achieve. Under a motivation theoretical perspective, flexible work schedule users are assumed to be more likely to exhibit higher performance because they would have greater resources (e.g., more time, more support), which would enable them to perceive greater expectancy that they can perform both work and family roles well (Kossek & Misra, 2008; see also Vol. 3, chaps. 3 and 11, this handbook). A key issue to measure is the degree to which perceptions of effort to perform are increased because of use of flexible work schedules. Studies would also measure the degree to which individuals perceive reduced constraints to performing well and increased expectancy to stay in the labor force because of the increased access and use to flexible work schedules. Workers who are able to access and use flexible work schedule supports they value, therefore, may be more likely to have higher effort–performance linkages because they will be more likely to believe they can perform both work and family roles well. Research does indeed show that workers individuals may engage in higher extra-role performance when flexibility is available. Lambert (2000) found that employees with access to work–family benefits were more likely to exhibit higher organizational citizenship behaviors. 546

Work–Family and Boundary Linkages A work–family perspective on flexible work scheduling theorizes that these schedules would reduce work–family conflict, defined as when one role interferes with the performance of another role. Use of flexible schedules could also have the potential to increase work–family enrichment, the degree to which resources or skills or knowledge in one role (e.g., work) enhance the other (e.g., family), since users would have greater involvement in both work and family roles. Regarding the latter, Greenhaus and Powell (2006) suggested that resources in one domain will extend to and impact resources in another domain, leading to positive spillover. They believed that increased flexibility will have a positive impact not only in the work role but also in the family role, via positive spillover. For example, by using a flexible work schedule, a worker will have more positive well-being on the job and at home because he or she will experience fewer conflicts. This increased positive mood in each domain, in turn, will cross-transfer, and enhance the overall quality of accumulated role experiences at work and home. Studies are beginning to investigate how use of different types of flexibility may lead to lower work–family conflict or higher enrichment. For example, use of some types of work schedule flexibility (e.g., part-time work) may lead to lower work–family conflict than others (e.g., telework). The latter flexible work schedule simply reshuffles work tasks in location from work to home but does not reduce workload. For example, although Kossek, Lautsch, and Eaton (2006) reported a positive relationship (r = .31) between being a teleworker and perceptions of flexible job control, they found no relationship between being a teleworker and lower work-to-family conflict. After controlling for marital status, gender, and having dependents, the study found that the more teleworkers perceived higher job control, the lower family-to-work conflict (β = −.27) as long as they engaged in a boundary management strategy that discouraged multitasking or managing family activities while working. This finding leads us to discussion of boundary theory, which relates to work–family spillover theories. Boundary theory is based on the idea that individuals construct mental, physical, and emotional

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

fences between roles, such as work and family (Ashforth, 2001). Some individuals prefer to segment work and family roles, whereas others do not care whether work crosses into home and are integrators (Nippert-Eng, 1996). Flexible work schedules affect employee perceived ability to control boundaries between work and home, such as the degree to which the timing and location of work or family roles are flexible and permeable (Kossek, Lautsch, & Eaton, 2005). They can facilitate a boundary management strategy enabling individuals to manage work–family role synthesis in ways that fit with personal values regarding segmentation and integration and role investments (Kossek, Noe, & DeMarr, 1999). It is important for researchers to consider how variation in preferences for segmentation to integration of work–nonwork boundaries, also known as flexstyles, moderate attraction to and the effects of using different types of flexible schedules (Kossek & Lautsch, 2008). For example, a study by Rothbard, Phillips, and Dumas, (2005) found that key job attitudes were moderated by the degree of congruence between an individual’s values for segmentation and the availability of flextime policies enabling restructuring of work and family roles to support segmentation. Individuals who more strongly valued work–family segmentation were more committed to their jobs to the extent that they had access to flextime, compared with those who more strongly valued role integration, even after controlling for many key demographics (e.g., gender). It is clear that the effect on boundaries of using a flexible work schedule varies by schedule type. Telework, for example, is the flexible schedule that most heavily blurs the physical boundary between work and home. Teleworkers, by definition, are more likely to integrate work–family roles and experience higher work–family conflict than other flexibility forms such as part-time work, which allows for a boundary management strategy characterized by more work–family separation. Learning to separate work and family roles requires new socialization of work and family task enactment for teleworkers. There is often growing job and family creep—seepage of the responsibilities of one role into the other (Kossek & Lautsch, 2008). Golden (2001) reported that individuals using technological tools such as

laptops and cell phones tend to have longer work days. They also report more difficulty with escaping or breaking away from work psychologically. They also may have more role transitions, switching more frequently between work tasks and home tasks; leading to switching cost and process losses. (Kossek & Lautsch, 2008). Consequently, while users of telework may hold positive perceptions of higher psychological control over schedule flexibility, this benefit may be offset by teleworkers’ lesser ability to separate boundaries. Weaker work–family boundary separation leads to a greater propensity to take on additional work (e.g., substituting commuting time for job tasks) or nonwork responsibilities (e.g., trying to do the laundry at the same time as working), resulting in an increase in total life workload and work–family conflict (Kossek & Lautsch, 2008).

Section Summary The preceding section has shown that use and availability of flexible work schedules relate to perceptions of job control, motivation, and perceptions of work–family conflict and boundary-blurring. In the next section, we review key measurement challenges, which we integrate with this discussion of relevant theoretical constructs to identify cross-cutting characteristics across schedules. PERSISTENT MEASUREMENT CHALLENGES Our review has identified the need to (a) differentiate between measures of formal flexibility policies and flexibility in job design and (b) clarify measurement of availability and use and level and degree of diffusion within the firm, to better compare prevalence, take-up, and impact.

Formal Policy or Informal Job Characteristic? The literature ranges from being very fragmented, as in studies that examine flexible work schedules separately and use no common theoretical thread or dimensions comparing their design (Rau, 2003), to very general, as in studies of employees’ or employers’ responses to an index listing a wide number of programs available. For example, researchers in the 547

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

latter class might ask a general question, such as whether one had access to a flexible schedule or workplace flexibility, without specifying the type of flexible schedule. This vagueness is problematic as it is difficult to know if the worker had access to flexibility in, for example, timing or workload reduction or to a formal program versus a flexible job design tailored to individual circumstances. This ambiguity leads to a bifurcation in the flexible work schedule literature. One stream mainly conceives of flexible work schedules as a job design feature that refers to an individual’s perceived level of job autonomy over work schedule flexibility (e.g., Richman, Civian, Shannon, Hill, & Brennan, 2008). Flexibility control is seen as a job characteristic. Respondents typically use Likert scales to assess the degree of perceived flexibility control concerning the timing and time of their work (cf. Anderson, Coffey, & Byerly, 2002). Scholars refer to this construct as perceived flexibility control or control over work time (Kelly et al., 2008; Kossek et al., 2006). The main other research stream views flexible work schedules as involving a formal human resource policy or informal supervisory approved work practice. Typically, measurement is conducted in dichotomous terms. A job was defined as either flexible or not, on the basis of a policy or practice. If an individual was a user of a flexible schedule, it was assumed he or she had autonomy, regardless of whether the use of the policy or practice actually led to increased flexibility control or effectiveness. (For a review, see Kelly et al., 2008.)

Clarifying Flexibility Use and Access Within and Across Firms Many studies also confound policy availability and actual use, whether the flexible schedule is a formal human resource (HR) policy available across the firm or an idiosyncratic and informal supervisor practice. Clarifying these issues is very important in studying relationships between antecedents and outcomes. For example, should a firm be considered as offering flexible work schedules if it is offered to only one employee (any employee) on an exception basis, or only to some employee groups such as high-talent professionals but not lower wage workers? 548

Clarifying levels of analysis in measurement within the firm is also confusing. For example, a firm can be listed in a national survey as having a flexibility policy, but there can be wide variation within the firm at the business-unit or work-unit level in the degree to which a practice is available, depending on an employee’s supervisor or occupation. Given these trends, it is not surprising that reports on flexible schedules significantly differ between employees and employers. Employers typically focus on policy adoption and overstate availability. Employees, on the other hand, often focus on perceived barriers to use, such as lack of communication and cultural and supervisor support (Kossek & Distelberg, 2009). If flexible schedules are available on paper but go underused because the organizational culture does not support them, so that users are afraid that they may not be promoted or, worse, could lose their jobs, does a firm really offer flexible work schedules? We will draw on these gaps in measurement accuracy and the previous review of relevant theoretical constructs in order to identify common characteristics that differentiate offerings of flexible work schedules. We identify key themes that must be assessed when evaluating flexible work schedules to ensure they are flexible more than just in name only, in order to improve measurement of antecedents and outcomes across types.

Definition and Cross-Cutting Themes: What Makes a Flexible Work Schedule “Flexible”? Assuming that the flexible work schedule policy involves flexibility related to one of the four main work schedule types—timing, location, workload amount, or continuity of employment hours—we drew on the preceding review of relevant theory and measurement challenges to identify the following five criteria that should be used in any study to assess flexible work schedules. 1. To what extent does the flexible work schedule involve (a) a recognized human resource policy or practice sanctioning work schedule flexibility; and (b) job design characteristics fostering greater perceived job control over work scheduling? The first criterion is that the flexible work schedule

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

should involve both a human resource policy or practice and some link to job design characteristics fostering high perceptions of increased autonomy over continuity of work and when, where, and how much of it is done. Ideally, a formal flexible policy always would be well-linked and supported by informal supervisory practice. However, policy and practice are not always tightly coupled. If the policy just exists on paper and only in principle, use may be restricted. Under this situation, the schedule will not be experienced by an employee as increasing perceptions of job autonomy or control over the work schedule. To be considered a flexible work schedule, the schedule must enable employees to have some perceived autonomy to control or customize one or more of these schedule criteria to meet personal preferences. 2. To what extent does the culture support use of flexible work schedules, so that there is a relatively low gap between availability and use by those who desire a flexible work schedule? The second point is that the organizational culture must support a majority of workers and managers in perceiving that the schedule as readily available. If the schedule is only an informal practice that individual workers request on a case-by-case basis from supervisors who may vary in support, there may be wide variation in equity in how the schedule is administered and implications for whether positive outcomes occur. We do not believe that a firm should be considered as having adopted flexible work schedules if it is not a recognized practice that many workers can request. In some firms, managers permit access only to select higher-performing workers and try to keep the schedule from being known as a work option. Ad hoc “secret” deal-making between individual employee and employer, or I-Deals, on exceptional basis (cf. Rousseau, 2007) can occur; such arrangements would be outside the scope of this review. 3. To what extent is use employee initiated and perceived as voluntary? The third attribute is that the use of the flexible schedule must be employee initiated and enable the workers to have some choice as to whether to use the schedule. This

distinction is important, because voluntary flexibility may be more likely to be psychologically beneficial for the worker (as in perceptions of increased job control and well-being) than would involuntary flexibility, forced by the employer. Measuring “voluntariness” can be tricky, because many professionals are socialized to work long hours and highly identify with the work role; they may use flexibility to work long hours, even if their employer does not require them to do so. 4. To what extent is use of flexible work schedules determined by mutuality in the employment relationship to benefit both employees and employers? The fourth criterion is that flexible schedule use emanates from some mutuality between employer and employee in the power to influence the scheduling of working hours. This criterion helps distinguish a flexible worker from the growing numbers of self-employed workers. Although self-employed workers could be considered as having a flexible schedule, they are outside the boundaries of this review, which focuses on individuals who regularly work for an employer and are considered to be employees. The assumption that use of the flexible work schedule leads to positive outcomes for both employee and employer is an important indicator of mutuality in the power relationship and of a balance in accrued benefits from flexible work schedules. 5. To what extent is the schedule socially constructed as “psychologically different” from a standard schedule in terms of boundary blurring? In this fifth criterion, the work schedule is viewed as being psychologically different from a standard work schedule, particularly in terms of what are considered “standard” norms for the number of hours spent at work or continuity of employment or “normal” relationships regarding the degree of boundary blurring or separation of work and nonwork relationships. This criterion is based on growing evidence that the definition of the term flexible work schedules has a social construction component. Those working on a flexible work schedule are seen as working something other than a regular schedule that a majority of workers use (Ashford et al., 2008; Cappelli, 1999). 549

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

Given this social or normative aspect of flexibility, the meaning of flexible work schedules may shift over time in societal culture and across firms, as they become more prevalent. What is considered a flexible work schedule may vary by organizational and national culture, type of job, or the prevailing work rules of the employer. For example, teleworking may be the standard for an information technology (IT) firm but unusual in a manufacturing firm. Flexibility may not only refer to the schedule of work hours, but may take on social meaning as an attribute ascribed to an employee working in a nonnormative manner. The individual is labeled a flexible worker. FLEXIBLE WORK SCHEDULE OUTCOMES Any summary of outcomes of flexible schedules must be introduced with the caveat that more research needs to be done to isolate the specific effects of various types of flexible schedules with better measures that address the measurement and definitional issues noted in the previous sections. Drawing on selected studies, Table 17.2 shows a summary of main outcomes with effect sizes for selected citations. We summarize here the general trends shown in Table 17.2. GENERAL EMPLOYER OUTCOMES The I/O literature suggests there are two main benefit categories from flexible work schedules for the organization. The first is increased workforce attraction and retention, effort, quality, and productivity, all of which lead to higher job satisfaction, engagement, extra role effort, commitment, higher workforce quality from a larger applicant pool, and lower turnover of talent. The second main employer benefit is cost savings from the ability to attract and retain a motivated workforce, as well as a lowering of rates of dysfunctional employee behaviors, such as absenteeism, turnover, or accidents (Halpern, 2005; Kelly et al., 2009; Kossek, 2006; Kossek & Hammer, 2008). Employers may also have savings in compensation costs, because some employees may be willing to trade off wages for more leisure time off from work. 550

Given these trends, employers who offer flexible work scheduling to accommodate work–life conflicts may have a competitive advantage in external recruitment and internal retention. Evidence does suggest that having flexibility policies does increase the size and quality of the applicant pool (Clifton & Shepard, 2004). Some workers with unique skills, such as high-talent professionals or workers in jobs with higher turnover (e.g., nursing, service jobs) can exert workforce leverage to entice employers to offer flexible schedules or impose preferred administrative structures (e.g., flexible hours) on their organization (Barringer & Milkovich, 1998). Flexible work schedules also enable the development of internal labor markets to retain workers, by making it more unattractive for employees to leave the firm, as it raises opportunity costs of looking for similar alternative employment (Davis & Kalleberg, 2006). This has potential cost savings for employers, because resources and time are not devoted to constantly recruiting and training new workers, who are not likely to be as productive immediately as experienced workers. From the employer perspective, besides the positive productivity effects noted, there are possible countervailing negative effects that simultaneously must be taken into account in more studies. Such effects may include increased administrative costs and the complexity of having to manage what can be increasingly varied schedules to ensure coverage and coordination for client interactions (Van Dyne et al., 2007). Costs also may be incurred if investments are not made to train supervisors to learn new ways to supervise, communicate with, manage, and measure the performance of a workforce that is more dispersed in time at work. Cross-training and better teamwork also may be needed to encourage workers to learn each others’ jobs and self-coordinate schedules to implement flexible work schedules in ways that consider implications for work group efficiency, as opposed to only individual self-interest. Measuring cost reductions is also tricky, as they may occur indirectly, particularly through variables that are important pathways for employee well-being. For example, the relationship between flexible work schedules, turnover, and absenteeism may be mediated via lower job stress, work–family conflict, or

Flexible Work Schedules

TABLE 17.2 Summary of Prevalence and Sample Outcomes From Flexible Work Arrangements Type of work schedule

Definition

Use and availability

Impact on employee and employers

Flextime

A flexible work schedule that allows employees to vary their work hours, within certain parameters, to better suit their needs (Ronen, 1981)

Used by 29 million workers (28%) in the United States (USBLS)a 56% of employers offer flextime (Burke, 2005)

Decrease in negative affect levels for women caregivers (Chesley & Moen, 2006) Higher productivity (r = .22; Baltes, Briggs, Huff, Wright, & Neuman, 1999; see also Pierce & Newstrom, 1983) Higher satisfaction with schedule (Baltes et al., 1999) Lower absenteeism (r = .42; Baltes et al., 1999; see also Dalton & Mesch, 1990) Lower driver stress and time urgency (Lucas & Heady, 2002) Higher job satisfaction (r = .16; Baltes et al., 1999; see also Orpen, 1981) Decreased turnover (Allenspach, 1975; Ralston, 1989; Ronen, 1981; Stavrou, 2005) Lower work-to-family conflict (ρ = −.30; Byron, 2005) and lower family-to-work conflict (ρ = −.30; Byron, 2005) Greater family supportive organizational perceptions and supervisor support, lower turnover intentions (r = −.11), higher organizational commitment (r = .16) and job satisfaction (r = .13; Allen, 2001)

Compressed workweek

A work schedule that allows an employee to work a full week (e.g., 40 hours) in fewer than 5 days (Pierce, Newstrom, Dunham, & Barber, 1989)

33% of employers offer compressed workweeks (Burke, 2005)

Higher supervisor rated performance (r = .21), higher job satisfaction (r = .28), and higher satisfaction with schedule (r = .19; Baltes, Briggs, Huff, Wright, & Neuman, 1999) Lower absenteeism (Goodale, & Aagaard, 1975; Nord & Costigan, 1973) Lower work–family conflict (Dunham, Pierce, & Castaneda, 1987; Allen, 2001) Greater family supportive organizational perceptions and supervisor support, lower turnover intentions, higher organizational commitment and job satisfaction (Allen, 2001)

Shift work

Any organization of working hours that differs from the traditional diurnal work period: work days, evenings, nights, or some form of rotating schedule (Costa, 2003)

Copyright American Psychological Association. Not for further distribution.

Flexibility in the timing of work

Higher work–family conflict (Jansen, Kant, Nijhuis, Swaen, & Kristensen, 2004)

(continued) 551

Kossek and Michel

TABLE 17.2 (Continued)

Copyright American Psychological Association. Not for further distribution.

Summary of Prevalence and Sample Outcomes From Flexible Work Arrangements Type of work schedule

Definition

Use and availability

Impact on employee and employers

Contingent work

Any job in which an individual does not have an explicit or implicit contract for long-term employment (Polivka & Nardone, 1989, p. 11)

10.7% of workers in the US (consisting of independent contractors, on-call workers, temporary help agency workers, and workers provided by contract firms; USBLS)a

Mixed: Reports of low (Van Dyne & Ang, 1998), neutral (Pearce, 1993), and high organizational commitment (McDonald & Makin, 2000) Mixed: Reports of low (Bergman, 2002) and high job satisfaction (Galup, Saunders, Nelson, & Cerveny, 1997; McDonald & Makin, 2000) Mixed: Low (Van Dyne & Ang, 1998) and high organizational citizenship behaviors (Pearce, 1993) Higher levels of subjective health problems (Martens, Nijhuis, Van Boxtel, & Knottnerus, 1999) Lower job performance (Ang & Slaughter, 2001) Higher job-induced tension (Bernhard-Oettel, Sverke, & De Witte, 2005)

Flexibility in the location or place of work Telework or flexplace

A way of flexible working that enables workers to get access to their labor activities from different locations by the use of information and communication technologies (Pérez, Sánchez, & de Luis Carnicer, 2002, p. 733)

37% of employers offer telecommuting (Burke, 2005) 44.4 million American users who performed any kind of work from home (Dieringer Research Group, 2004) 24.1 million of American users who worked at home during business hours at least 1 day per month (Dieringer, 2004)

Increase in personal growth for male caregivers (Chesley & Moen, 2006; marks, 1998) Lower time-based family-to-work conflict (Lapierre & Allen, 2006) Lower work-to-family conflict and higher family-to-work conflict (Golden, Veiga, & Simsek, 2006) Greater family supportive organizational perceptions and supervisor support, lower work–family conflict and turnover intentions, higher organizational commitment and job satisfaction (Allen, 2001) Lower absenteeism (Stavrou, 2005) Organizational performance (Martínez-Sánchez, A., Pérez-Pérez, M., Vela-Jiménez, M. J., & de-Luis-Carnicer, 2007; Stavrou, 2005)

Flexibility in the amount of work (workload and hours) Part-time work

552

Employees who work fewer than 35 hours per week (USBLS)a

A main type of flexible work arrangement in smaller businesses (Maxwell, Rankine, Bell, & MacVicar, 2007) Unskilled, poor pay, little career possibilities, low security (Barnett, 1998; Kahne, 1985)

Lower role overload and work-to-family conflict (β = .18; Rijswijk, Bekker, Rutte, & Croon, 2004; see also Higgins, Duxbury, & Johnson, 2000) Greater family supportive organizational perceptions and supervisor support, lower work–family conflict and turnover intentions, higher organizational commitment and job satisfaction (Allen, 2001) Lower annual staff turnover (Stavrou, 2005) No difference in job satisfaction (Lee & Johnson, 1991; McGinnis & Morrow, 1990)

Flexible Work Schedules

TABLE 17.2 (Continued) Summary of Prevalence and Sample Outcomes From Flexible Work Arrangements Type of work schedule

Definition

Use and availability

Impact on employee and employers

Copyright American Psychological Association. Not for further distribution.

Higher turnover for part-time workers and those who have temporary position (Cohen & Gadon, 1978; Feldman & Doerpinghaus, 1992; Granrose & Appelbaum, 1986) Greater flexibility in scheduling but less continuity in workflow (Olmsted & Smith, 1989) Lower job-induced tension (Bernhard-Oettel, Sverke, & De Witte, 2005) Job sharing

A work schedule that allows two employees voluntarily share the work responsibilities of one full-time position, where each works less than fulltime (Christensen & Staines, 1990)

19% of employers offer job sharing programs (Burke, 2005)

No difference in job satisfaction (Lee & Johnson, 1991; McGinnis & Morrow, 1990) Higher turnover for part-time workers and those who have temporary position (Cohen & Gadon, 1978; Feldman & Doerpinghaus, 1992; Granrose & Appelbaum, 1986) Greater flexibility in scheduling but less continuity in workflow (Olmsted & Smith, 1989) Lower annual staff turnover (Stavrou, 2005)

Customized or reduced-load work

A work schedule where employees lessen their workloads through the reduction of work hours or tasks and being paid less accordingly (Meiksins & Whalley, 2002)

A main type of flexible work arrangement in smaller businesses (Maxwell, Rankine, Bell, & MacVicar, 2007) Mothers working in professional occupations (Hill, Märtinson, Ferris, & Baker, 2004) Managers and professionals (Lee, MacDermid, Williams, Buck, and Leiba-O’Sullivan, 2002)

Heightened levels of work–family balance (Hill, Märtinson, Ferris, & Baker, 2004; Lee, MacDermid, Williams, Buck, & Leiba-O’Sullivan, 2002) No impact on career opportunity (Hill, Märtinson, Ferris, & Baker, 2004) Increased general well-being, positive effects on children and parent-child relationship, higher job satisfaction and performance, satisfaction with career implications (Lee, MacDermid, Williams, Buck, & Leiba-O’Sullivan, 2002) Managers perceive it as maintaining or enhancing work performance, recruitment, and retention (Lee, MacDermid, Williams, Buck, & Leiba-O’Sullivan, 2002)

Flexibility to allow for short-term breaks in employment or time off Sabbaticals

A work schedule that allows employees take a prolonged paid time away from work and expect to return to their same jobs at the end of the sabbatical (UIOVPAA, 1996)

14–24% of American corporations have established sabbatical programs (UIOVPAA, 1996)

Lower burnout (Duetschman, 1994) Approximately 60% of academic faculty reported new research and skill development after sabbatical (UIOVPAA, 1996) Employees enjoy their sabbaticals and feel better when they are done, some employees improve their skills or perform acts of social worth (Kramer, 2001) Avoid technological obsolescence (Bachler, 1995) Employees return to work with a new viewpoint and with new vigor, hiring substitutes to fill in for those on sabbatical could reduce unemployment, and having a sabbatical policy gives an organization a competitive edge (Kramer, 2001) (continued) 553

Kossek and Michel

TABLE 17.2 (Continued)

Copyright American Psychological Association. Not for further distribution.

Summary of Prevalence and Sample Outcomes From Flexible Work Arrangements Type of work schedule

Definition

Use and availability

Impact on employee and employers

Leaves, vacation, and flex-leaves

A work schedule that allows employees to be absent from work or work duty (Ivanovic & Collin, 2006)

British survey estimated 35% have parental leave (Cully, O’Reilly, & Millward, 1998) British survey estimated 56% have paid leave (Cully et al., 1998)

Greater family supportive organizational perceptions and supervisor support, lower work–family conflict and turnover intentions, higher organizational commitment and job satisfaction (Allen, 2001)

Part-year work

A work arrangement where workers are generally employed to fulfill seasonal or short-term needs

Seasonal work and ad hoc industries (cf. Lockyer & Scholarios, 2007) Developing countries, and increasingly more common in developed countries (Houseman & Osawa, 2003) Predominately public administration, education, and heath workers (Local Government Management Board, 1998)

Disproportionately marginalized groups (i.e., women and minority ethnic groups; Conley, 2003) Devalued treatment and stigmatization (Boyce, Ryan, Imus, & Morgeson, 2007) Recruitment difficulties and skill shortages (Lockyer & Scholarios, 2007) Increased flexibility and reduced costs (Boyce, Ryan, Imus, & Morgeson, 2007)

Note. USBLS = U.S. Bureau of Labor Statistics; UIOVPAA = University of Illinois Office of the Vice President for Academic Affairs. aSee USBLS National Compensation Survey results (http://www.bls.gov/ncs/ebs/).

higher work engagement. Nevertheless, the effect sizes shown in Table 17.2 help one conclude that there is a positive relationship between the availability of flexible work schedules and organizational attachment. For example, Allen (2001) found modest positive relationships with organization commitment, job satisfaction, and turnover intentions. Lowering job stress also can positively affect health care costs by lowering blood pressure and reducing negative health behaviors such as alcohol or drug abuse or overeating (cf. Harris & Fennell, 1988). Yet the effects of flexible work schedule use may be lagged, as it may take several months or years before these effects show up on the bottom line. EMPLOYEE OUTCOMES For the individual, a main benefit of using flexible work schedules or having greater access to schedule flexibility relates to increased well-being, lower 554

stress, and health. A second main benefit is better focus, satisfaction, and role quality experiences both in job and nonwork roles. One likely pathway between flexible schedule use and higher levels of well-being, assuming workload is held constant, is lower work–family conflict, which in turn relates to lower job and life satisfaction (Kossek & Ozeki, 1998). Better person–job fit may also ensue, as restructuring work schedules to better fit nonwork demands allows an individual more time to devote to other roles outside of work, such as exercise, seeing friends, or being involved in the community. Assessing benefits from flexible work schedule use for work and family roles is likely to be reciprocal and iterative and can operate via many complex pathways. An example comes from a study that, although it did not examine formal flexible schedules, did research workload perceptions over time. Higher negative work affect related to negative home

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

affect (β = .15), and higher job workload related to negative home affect (β = .04; Ilies et al., 2007). Thus, more positive relationships between work and family roles also may occur in part because better role quality experiences may ensue at work and at home, making the individual feel that those roles are more complementary and not always at odds. Individuals also may be able to perform better at both work and family roles because of positive capitalization of affect and mood transfer between both domains (Ilies et al., 2007) and increased ability to focus on each role. Just as with employer outcomes, when assessing individual outcomes, some countervailing factors must be considered before concluding that the overall effects of work schedule flexibility are positive. For example, if employees are not able to use flexible work schedules that best meet their personal time demands, or if they experience career penalties from using flexibility, the benefits of these schedules will be lessened at best or, worse yet, could become negative. We have noted that in many firms, although flexibility is officially allowed, users are sometimes seen and stigmatized by the organizational culture as being less committed to the firm, or as not being mainstream workers (Kossek & Lee, 2005). They may face a backlash such as lower raises, fewer promotions, or being first to be laid off in a downturn (Golden, 2008). Few studies have actually quantified these costs and linked them to actual flexible work schedule use. Individuals also may experience increased cognitive complexity from using some flexible work schedule types such as telecommuting. This results in more switching cost from increased frequency of role transitions and in higher process losses (Kossek & Lautsch, 2008). For example, individuals can be constantly moving between work roles (e.g., working on a laptop) to multitasking on domestic roles (e.g., supervising a child’s homework while doing the laundry). Telecommuters also face the temptation of overwork and increased work–family conflict, burnout, and role overload from having work or domestic chores constantly available to them all the time. Telecommuters may then be tempted to simply try to take on more work and home tasks simultaneously. Negative mood transfers from work

may also be brought into the home more easily, as well as the reverse. Many of the outcomes of using flexible work schedules may be moderated by the employee’s demographic, psychological, or job background. For example, if access to flexible work schedules is viewed as varying the place of work, it will most benefit individuals who are most in need of flexible schedules as help or support. Thus, workers with extremely long commutes may be more likely to benefit from teleworking than those who live close to the office. Or employees who have higher work–family conflict, such as those with young children, may receive more benefit from flextime to enable them to take them to doctor or school appointments than would someone with fewer domestic demands to manage during the workweek (cf. Hammer, Kossek, Anger, Bodner, & Zimmerman, 2009). Thus, flexible work schedule studies must identify the relevant population most likely to benefit from a specific flexible work schedule type when assessing outcomes, but few do. Cross-level work group and organizational moderators also may be critical when assessing individual outcomes. For example flexible work schedules used in a company with an unsupportive work group or organizational culture may weaken the positive effects at the individual level.

Overview of Outcomes by Schedule Type Baltes, Briggs, Huff, Wright, and Neuman (1999) conducted a meta-analysis that compared effects across schedule types, summarizing 24 years of research (1973–1997). While it was not always clear whether studies were measuring use or access, they found that access to flexible work schedules positively relates to higher job satisfaction and lower absenteeism. Compressed work schedules resulted in higher supervisor ratings of performance. A later meta-analysis by Byron (2005) found that schedule flexibility was negatively related to perceptions of work-to-family conflict (ρ = −.30) and family-towork conflict (ρ = −.17). These relationships were moderated by sample parental status (work-tofamily conflict, r = −.72) and the percentage of the sample that was female (work-to-family conflict, r = .10; family-to-work conflict, r = −.63). 555

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

Drawing on data from the 1997 National Study of the Changing Workforce, Halpern (2005) found a that the larger the number of time-flexible policies an organization offers, the greater the organizational cost savings from lower absenteeism by less “missed time at work, fewer days late for work or left early, and the failure to meet deadlines” (p. 162). Future research should build on this study to isolate the effect sizes from use of the specific types of flexibility in relation to absenteeism or deadlines missed. Richman and colleagues (2008) drew on a consulting firm’s sample of 15 large corporations that were part of a Work Family Directions study. They found an incremental effect size of 8% of the variance in employee engagement linked to employee perceptions of perceived flexibility and a 9% increase in the variance in engagement explained by the presence of family supportive policies. This study is one of the few we found that included measures of both formal (e.g., policies) and perceptions of informal flexibility. An area for improvement is that both of the measures used were one-item measures, which are less reliable and too general to identify the source of cultural support or type of flexibility. The items were “Do you have flexibility or not?” and “Or have supportive policies or not?”

Symbolic Outcomes: Availability of Flexibility as Perceived Organizational Support Eaton (2003) examined work–family policies of seven biopharmaceutical firms in a single state. Based on a sample of 463 employees, Eaton estimated the availability of workplace flexibility via an index of seven flexibility practices (flextime, parttime jobs, flex place, job sharing, compressed workweek, unpaid personal leave, and sick leave to care for ill children). Eaton found that availability of formal and informal policies, perceptions of one’s ability to use policies, and degree of control over flexibility (R2 = .06) were all significant predictors of perceived productivity and organizational commitment, after controlling for multiple individual employee variables (e.g., age, education, tenure, company size). No gender moderating effects were 556

found, indicating that men and women benefit from flexibility. In a study that investigated perceived cultural support within organizations for the family role and considered many flexible work schedules, Allen (2001) used a sample of 522 employees from a variety of settings (technology firm, utility company, women’s professional business association). Allen found that benefits offered (flextime, compressed workweek, flex place, part-time work, and a variety of dependent care supports) were significant predictors of lower work–family conflict, higher job satisfaction, higher organizational commitment, and lower turnover intentions, after controlling for a number of variables (e.g., salary, race, tenure). However, when adding perceptions of organizational support of the family, each of the multiple regressions resulted in a change in R2 between .15 and .24. This suggests both perceived cultural support for the family role may be more important for favorable work attitudes than mere availability of flexibility. Hammer, Neal, Newsome, Brockwood, and Colton (2005) found a positive relationship (β =.16) between use of alternative work schedules and work–family conflict for women. They argued that a potential reason for this non-favorable outcome was that the schedules enabled the women in their study to engage in more non-work-related responsibilities as opposed to using the increased control and time to lower stress and strain outcomes. A study drawback was that the type of alternative work schedule used was not delineated; use was dummy coded, so it was not clear what type of flexible schedule was being used.

Moderators of Outcomes Comparing Flextime With Flexplace A recent study by Shockley and Allen (2007) employed more measurement precision and examination of moderators than many previous studies on linkages between flextime use and lower work– family conflict. Using a highly educated sample of women with an employed spouse and/or at least one child living at home, results suggest that flextime was more highly related to lower work-to-family conflict than to family-to-work conflict. This relationship was stronger for flextime than telework.

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

When controlling for age, marital status, work hours, and parental status, flextime and family responsibility accounted for 9% of the variance in work-tofamily conflict. Family responsibility moderated the relationship between both access to flextime and work-to-family conflict (β = −1.33, ΔR2 = .05) and family-to-work conflict (β = −1.47, ΔR2 = .06). Also important, when perceptions of family-supportive organizational policy availability were considered, the relationship between flextime and work–family conflict became insignificant. Perceptions of familysupportive organizational policies accounted for over a fourth (26%) of the variance in work-to-family conflict and 14% of the variance in family-to-work conflict, after controlling for the demographics noted above. These findings suggest that it’s not necessarily mere access to schedule flexibility that matters; rather, perceptions of how family-supportive the organization is really drive the direct effects to work–family conflict.

Family Outcomes Related to Shift Work Rarely are family measures related to scheduling assessed in management and I/O studies. An exception is an interesting study on shift work conducted by Barnett, Gareis, and Brennan (2008). While most studies focus on negative outcomes of shift work, Barnett focused on when shift work can be positive when matching workers’ preferences for scheduling. Using a sample of 55 dual-earner families with children between the ages of 8 and 14, Barnett and colleagues examined the within-couple relationships between the wife’s work and the spouse’s work–family conflict, psychological distress, and marital-role quality. The most robust finding of this study was that the wife’s shift work was significantly related to her work–family conflict but not to the husband’s level of work–family-conflict. Those who worked evening shifts reported greater work–family conflict than those who worked day shifts. The wife’s shift work by number of hours also was significantly related to her level of psychological distress. Interestingly, only the interaction between shift work and number of hours was significant, as shift work and hours worked had no direct effects. The authors found that wives who worked day shifts had no variation in psychological distress;

meanwhile, those who worked evenings reported higher distress with fewer work hours.

Outcomes Related to Workload Flexibility Hill, Märtinson, Ferris, and Baker (2004) sought to better understand how reduced-load work affects perceptions of work–family balance. Using survey data from nearly 700 professionals from the 1996 IBM Work and Life Issues Survey in the United States, they compared mothers of preschool children with their full-time counterparts. These parttime or reduced-load employees worked on average 47% fewer hours and reported 41% lower income than the full-time group. Hill and colleagues found that reduced hours were positively related to work– family balance (r = .47) but not career opportunity (r = −.02). Likewise, when controlling for occupational level, family income, age, and job flexibility, reduced hours were again positively related to work–family balance (ΔR2 = .09) but not career opportunity (ΔR2 = .01). The mean annual family salary was $100,568 for reduced-load workers, while the mean for their full-time counterparts was $120,590, suggesting that these were relatively high earners who may be more likely to afford the income loss from working part-time than those in lower paying jobs. In a qualitative examination, Lee, MacDermid, Williams, Buck, and Leiba-O’Sullivan (2002) examined the role of contextual factors in the success of 82 managers and professionals working a reducedload work schedule. Lee et al. (2002) found that HR practices and policies for the reduction of work hours were quite successful, with an average reduction of 18 hours per week. In regard to personal outcomes, they found that 91% of respondents reported being more happy and satisfied with their work– family balance, 86% reported positive effects on their children and parent–child relationship, and most reported greater general well-being, less stress, and feeling less worn out and more relaxed. In regard to job and career outcomes, 85% reported neutral or positive implications toward work performance, 67% liked their jobs and felt they were doing challenging and interesting work, and most were satisfied with career implications of a reduced workload. Seventysix percent of senior managers interviewed believed 557

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

that reduced-load work maintained or improved work performance; most felt that it also enhanced recruitment and retention. Lee and colleagues found that 15 contextual factors were strongly endorsed by managers and professionals as being key factors in the success of reduced-load work. Individual factors included personal characteristics, such as having higher levels of work ethic, commitment, an organized and highly concentrated work style, a unique skill set in high demand, being a self-starter and interpersonally skilled, and having strong and clear personal values. Favorable job-context factors related to kinds of work that allowed for higher individual autonomy or were project-oriented. Favorable work-group factors included having a supportive boss and competent and supportive direct reports. Favorable organizational factors were noted for firms that had an organizational culture that valued employees’ needs, saw a business need for retaining skills, and offered wide publicity of work–life policies and programs. Similar results to the research just cited, showing a positive relationship between working part-time and lower work–family conflict and higher levels of well-being, were found in a study by Rijswijk, Bekker, Rutte, and Croon (2004).

Outcomes of Short-Term Breaks and Time for Work Recovery Collectively, outcomes of flexibility policies that allow breaks from work have received far less empirical and theoretical focus, so our review here is more descriptive. Kramer (2001) discussed the potential benefits of sabbaticals. Kramer compiled an impressive list of stories from individuals who opted for a sabbatical, including a former governor of Tennessee, lawyers, clergy, high-tech industry employees, educators, and even store clerks. All of the sabbatical reports reviewed by Kramer revolved around positive features such as feelings of being reenergized, reinvigorated, and refreshed. Kramer found that (a) employees enjoy their sabbaticals and feel better when they are done; (b) employees return to work with a new viewpoint and with new vigor; (c) some employees improve their skills or perform acts of social worth; (d) hiring substitutes to fill in for those on sabbatical could reduce unemployment; and (e) having 558

a sabbatical policy gives an organization a competitive edge. To address the limited empirical evidence examining work recovery, Totterdell and colleagues (1995) explored recovery duration based on 28 days of self-ratings, cognitive-performance tasks, and sleep diary results from a sample of 28 nurses. The longer the time allowed for recovery from the work shift, the greater the employee satisfaction on subsequent workdays. Satisfaction also was significantly higher at the end of day shifts when that shift was preceded by 2 rest days compared with only 1. Results also showed that a number of measures (sleep, mood, and social satisfaction) were worse on the 1st day of rest compared with subsequent days. These results suggest that recovery from work takes time. Although the findings were not longitudinal, they do indicate that short-term breaks benefit the employee and employer. As an extension of this work, Fritz and Sonnentag (2006) conducted a longitudinal study to explore the effects of vacation on employee performance-related outcomes and well-being. Using a working sample of 221 university employees, they found changes in effort expenditure and well-being between responses before and after vacation. Specifically, they noted vacation effects and partial fade-out effects. Vacation experiences (negative work reflection) contributed to well-being immediately after vacation (β = .27) and 2 weeks later (β = .16), after controlling for negative affect and well-being before vacation. Likewise, vacation experiences of negative work reflection (β = .21), relaxation (β = −.13), and nonwork hassles (β = .15) all significantly predicted self-reported effort expenditure 2 weeks after vacation. These results further suggest that short-term breaks in employment are beneficial to employees. FUTURE RESEARCH AND PRACTICE Overall, this chapter has reviewed the growing diversity in the different schedules used to organize working time from both the individual and the organizational perspectives. We have demonstrated that not all forms of schedule flexibility are the same in impacts on the individual and the organization, and that the views of the individual and the organization

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

on the benefits and drawbacks of flexible work schedules sometimes differ. We have also noted that the literature on flexible work schedules is very descriptive and needs both methodological and theoretical development. We have noted the need for studies to include new trends such as the growth in casual use of flexible work schedules, such as checking e-mail during nonwork hours, a social trend that must be accounted for in formal studies of flexible work schedules. Although many implications for research have already been made throughout this chapter, we close with additional suggestions regarding (a) a research agenda specifically focusing on flexible work schedule implementation, (b) the need for I/O theory to consider how flexible work schedules impact growing heterogeneity in work experiences, (c) improving measurement and theoretical linkages, (d) support and context as moderators, (e) assessment of more varied outcomes, and (f ) increased consideration of “the future of flex” as an organizational effectiveness tool.

Implementation Research Agenda Clearly, research needs to move beyond whether flexible work schedules merely exist and attempt to increase understanding of the conditions under which they are effectively implemented and used. We also need greater insights into the variation in antecedents and outcomes and processes related to different types of flexible work schedules used by employees and organizations with varying characteristics. Toward this end, Table 17.3 provides a list of hypotheses on implementation issues for future research. Many of the hypotheses in this table draw on our review of criteria of what makes a work schedule “flexible.” These design criteria can also be used to create new measures to better assess implementation of flexible work arrangements. We have also noted I/O constructs such as job control, valence and expectancies, work–family spillover, and preferences for boundary management integration or segmentation that could be used to assess the implementation of these schedules. We believe it is critical to discuss the implications of different schedule types for control perceptions as a pathway to understand other outcomes. We also have

pointed out that control over work time does not necessarily involve a formal program but can relate to an aspect of job design. We would like to see more integrated studies on implementation that measure human resource policy use, organizational cultural and supervisor support of flexibility, and worker perceptions of flexible-scheduling autonomy to reconcile the gap between policy and practice. We would also like to see more inclusion of family schedule design and flexible scheduling supports in these studies.

New I/O Theory Needed Related to Growing Heterogeneity in Work Experiences I/O theories need to be reviewed to account for the growing heterogeneity in work schedules and arrangements. They also need to adopt a multiplestakeholder approach to determine differential impacts of flexible schedules on managers and on different types of workers and families, as well as communities. As the use and customization of flexible work schedules continues to grow, an increasing important issue is “When are ‘nonstandard hours’ considered standard?” Heterogeneity in work schedules is likely to grow in bad economies as well as good ones. For example, during growing economic activity, they are ways to attract workers or keep up with rising product demand; meanwhile, during an economic downturn, they are ways to retain workers when raises are limited or as an alternative to layoffs. Many organizations will need to manage blended workforces, with employees working standard work schedules working side by side with those working flexible schedules, which can create challenges for managers in implementation (Lautsch & Kossek, 2009). Many basic theories of work, such as motivation, job satisfaction, culture and leadership, and organizational commitment, among many others, implicitly assume standard or regular work schedule and arrangements, with some homogeneity in employment experiences and motivations (see chaps. 7 and 12, this volume; Vol. 3, chaps. 3 and 4, this handbook). The reality—that more employees are working in many different ways with greater heterogeneity of work schedules—influences how 559

Kossek and Michel

TABLE 17.3

Copyright American Psychological Association. Not for further distribution.

Implementation Issues Related to Moderating Effects of Use of Flexible Work Schedules on Employee’s Work Attitudes and Behavior: Hypotheses for Future Research Flexible work schedule design attribute

Hypothesis based on literature trends for future research

1. Employer or employee initiated. Is use of flexible work schedules employer or employee initiated?

Hypothesis 1: When use of flexible work schedules is initiated by the employee, there is a significant positive relationship between use and employee job and family satisfaction and a significant negative relationship with work turnover.

2. Formal policy or informal practice/job characteristic. Does the experience of using the formal flexible work schedule lead to actual employee perceptions of greater control over work hours and workload?

Hypothesis 2: When use of flexible work schedules leads to employee perceptions of greater control over the timing, time, and amount of work, the employee will experience lower work-to-family conflict.

3. Policy availability compared with use. Does the flexible work schedule have availability on paper but lower use by workers?

Hypothesis 3: Flexible work schedules that are available on paper but have low actual use by workers and will have lower influences on reducing work–family conflict, as well as lower influences on positive employee work attitude and behaviors.

4. Social cultural support and use backlash. Does working in different ways have strong cultural support from management and limited negative backlash from use?

Hypothesis 4: When employees initiate use of a flexible work arrangement, if using flexibility is not positively valued by the organizational culture or leads to stigmatization, the positive effects of flexible work schedules will be ameliorated.

5. Organizational attachment effects. How does use of the flexible work schedule affect organizational attachment and the long-term employment attachment?

Hypothesis 5: Flexible work schedules that are designed in ways that support positive organizational identification and attachment to the employment relationship will be most likely to lead to positive employer benefits.

6. Nature of workforce use. Are flexible work schedules used by many different types of employees in many different functions?

Hypothesis 6: The more that flexible work schedules are used by employees of a wide range of demographics and workforce functions, the more flexibility is seen as a socially normalized way of working.

people experience work attachment, work roles, and work culture. Socialization of new employees and resocialization of existing ones will be increasingly difficult as more and more workers have varied time and work, and work at a geographical distance. Increasingly, high-talent employees may not necessarily look the same in how they work and act, nor will they view the work role as primary and be willing to restructure nonwork demands to enable them to devote primary attention and energy to working time (Kossek & Misra, 2009). (See also Vol. 3, chap. 2, this handbook.) On the practitioner side, HR policies, particularly for the high-potential and high-talent workers, are currently designed to most heavily reward employees who work schedules that meet core hours set by the

employer or are willing to increase hours to place working time above personal or family time. Yet, as noted above, growing numbers of employees simply do not work in this way. This was the way work was done in the 1950s, when organizations consisted primarily of men with homogenous careers and schedules (Whyte, 1956). It is not necessarily the way work schedules are enacted in the 21st century. Organizations have not fully adapted scheduling to meet the labor market, technological, and environmental shifts we discussed in this chapter. HR systems related to performance management, training, socialization, and career development, for example, have not kept up with these changes nor have they been adapted to mesh with the flexible organization of today.

560

Flexible Work Schedules

Copyright American Psychological Association. Not for further distribution.

Improving Measurement and Theoretical Linkages This chapter has shown that one of the major limitations in the current literature on flexible work arrangements is the imprecision with which flexible policies are measured. A key implication for research and practice of this chapter is that it is important for I/O psychologists to improve definition and measurement of flexible work schedules and better link measures to theoretical models. We have noted one area of this imprecision in the tendency for researchers to cluster or combine lists of flexible work arrangements (e.g., Allen, 2001; Casper & Harris, 2008; Stavrou, 2005) in order to create a composite score of adopted policies. This skews results toward rating larger organizations as more flexible simply because they have policies on paper. Researchers must also measure effectiveness, access across organizational groups, mixed consequences from use, and flexibility type. In sum, better reporting of specific flexibility design types is needed, perhaps drawing on the framework in this chapter that looks at types of flexibility practices in clusters, as not all forms of flexibility are similar in processes or outcomes. Here is an example of how flexibility type might differentially relate to outcomes comparing flextime and compressed workweeks. Though flexible work schedule practices are often implemented to benefit an organization’s workforce, various flexible schedule types differentially benefit, and potentially hinder, individual workers, depending on their scheduling needs. Flextime greatly benefits an employee with parental responsibilities because they are better able to respond to such needs (e.g., day care or school drop-off and pick-up schedule), while a compressed workweek hinders this same worker’s ability to respond to these needs, by making it difficult to do pick up or drop off a child over a 10- or 12-hour day. As such, it is important to examine individual flexible work arrangements individually, as well as how combinations suppress or change outcome relationships. Key moderators such as level of caregiving should also be assessed. As an example of how drawing on theory more closely could better inform measures, we use motivation theory as an illustration. Currently, the literature generally does not distinguish motivational

effects of different types of flexible work schedules. Studies drawing on a motivation perspective would measure variation in the degree to which employees with different employee backgrounds value different types of flexible scheduling and regard such schedules as instrumental in enhancing their ability to perform on the job. Studies should also measure the degree to which individuals have high expectancy that using flexible work schedules will accrue positive outcomes (such as low backlash and favorable work and family experiences). More research is needed to assess whether individuals who highly value flexible work schedules as a job characteristic and who use them are likely to have higher performance and a stronger relationship between use and performance linkages. Research drawing on motivation perspectives also would measure the degree to which different workers value different types of flexible work schedules and how different types of flexible schedules help individuals achieve important goals, both personal and work related (Kossek & Misra, 2008). It would also be important to measure the degrees to which individuals have high social identity pertaining to work and family roles and also how much they value integrating these roles, as this may predict increased valence regarding flexible work schedule use. (See Lobel, 1991, for a review of relationships between work–family role allocation and social identity.) Individuals who value work and family roles equally highly are often referred to as dualcentric, where two roles are both primary to social identity, and therefore the individuals put high dual investment in both roles. Dualcentric individuals are more likely to value flexible work schedules, as they enable greater participation in work and family roles simultaneously. Another area of imprecision that we have noted is that many studies confound the measurement of availability and use, often only examining the availability of formal flexible work arrangements. Fortunately, this seems to be a trend that some research is now rectifying by examining the unique effects of both the availability and use of flexible schedules (e.g., Casper & Harris, 2008; Parker & Allen, 2001). However, there still seems to be significant imprecision in regard to both the measurement 561

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

of temporality and intensity, amount, or extent of one’s use of flexible work arrangements. For example, Kossek, Barber, and Winters (1999) used survey data from single-item measures that assessed whether respondents had ever used alternative work schedule options; the data were then coded as users versus nonusers. Casper and Harris (2008) assessed use as “don’t use,” “use occasionally,” and “use frequently,” coded as 0, 1, and 2, respectively. These scores were then summed across a variety of policies to determine the amount of schedule flexibility used. Butler, Gasser, and Smart (2004) assessed a variety of flexible schedules with 5-point single items ranging from “never” to “very often,” coded as 0 to 5, respectively. Collectively, these examples represent the general norm within the literature regarding the measurement of use, each lacking in temporal (e.g., frequency over the course of a year) and intensity (e.g., frequency over the course of a week) information. Clearly, what needs to be clarified is how long and how frequently one has to use a schedule to be considered a user in order to have the schedule affect employee behaviors and attitudes. If someone can telework from home once a month or in bad weather, or have flextime when a child is sick, is that sufficient to have an impact on outcomes? Furthermore, what happens when someone uses more than one schedule at the same time, such as flextime with telework? How does one tease out the effects of each type over time? Overall, we need to move from studies reporting descriptive use of work schedule flexibility to measures of the extent of and effectiveness of implementation such as the hypotheses noted in Table 17.3. We also need to link measurement of use to workers’ perceptions of control and satisfaction. It is important for studies to include measures of actual policy use and measures of the degree to which workers’ experience flexibility on the job in the same study, so that scholars can ascertain whether using a flexible work schedule actually enhances job autonomy perceptions. Kossek, Lautsch, and Eaton (2006) suggested that future work–family research should distinguish between descriptions of flexibility use (formal telecommuting policy user, amount of telecommuting practiced), how the individual experiences flexibility psychologically, and performance on and off the job. 562

Support and Context as Moderators We have noted the importance of measuring not only the availability of flexible schedule policies and practices but also the degree to which individuals’ perceive that the company and supervisors are supportive of actually using flexibility without backlash. More studies need to combine measurement of policy availability and use with examination of cultural support for new ways of working. Several studies reported here showed interactions between formal flexible work schedule availability and use and support in relation to work–family conflict reduction. Another area of concern moves beyond the pure measurement of use of flexible work schedules and focuses more on the implications of using flexible benefits. For example, it has been proposed that when individuals take advantage of flexible work arrangements and overtly demonstrate interest in nonwork life, they may face negative judgments regarding their lack of organizational commitment (Allen & Russell, 1999; Fletcher & Bailyn, 1996; Lobel & Kossek, 1996). Accordingly, it has been suggested that an organizational culture for acceptance and use of flexible work schedules is critical to avoid backlash from not only management but peer nonusers as well. For example, Breaugh and Frye (2008) found that employees who reported their supervisors as being more family supportive were more likely to use flexible work schedules; more research is needed to tease out the ordering of this relationship. In addition, future research needs to further explore cultural support of flexible work schedules at the organizational level and work group level, along with the potential backlash of nonsupportive cultures. More studies also need to examine flexible work schedules in personal contexts. By this we mean that studies should examine not only the individual worker’s schedule but also the worker in the context of other family members’ schedules or the prevailing work group and organizational context and variable schedules. For example, we need to examine work schedules as part of a family system and investigate not only the employees’ schedules but how they mesh with those of family members. Similarly, there is a need to examine the compatibility of individual

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

flexible schedules with coworker, manager, and customer schedules. It is also important for studies to state the reasons for the adoption of flexible work schedules; who controls use—the employer or employee? Research is also needed on the degree to which flexible schedules are viewed as integrated into the business context. For example, flexplace may be standard for many mobile IT workers but very unusual for someone working in another industry. In the latter case, flexplace may engender social backlash from use in one context but not another, and studies need to be clear on workplace norms. Cross-level studies on variation in flexibility norms and preferences should be done. At the individual level, research might examine flexstyles such as psychological preferences for integration and segmentation (Kossek & Lautsch, 2008), which may shape preferences for various flexible schedules. These same proclivities could be aggregated at work group and organizational levels to understand the micro and organizational climates for work schedule flexibility and also to unpack the factors leading to growing scheduling conflicts between workers and managers. Cross-cultural research on flexible work scheduling is needed in which studies examine culture differences in the primacy of work to leisure and the perceived need for managers to control workers’ behaviors. Most cross-cultural research on flexible work schedules has been at the national public policy level, such as the availability of leaves across nation states. Very little research has examined the use of flexible work schedules across national cultures at the level of the firm, and these measurement challenges are discussed next. CLARIFYING PUBLIC POLICY CONTEXTUAL MEASUREMENT INFLUENCES IN NATIONAL SURVEYS Future research should aim to reduce measurement ambiguity currently found in national and international surveys in the United States and EU. For example, a review of three national U.S. surveys on flexible work schedules identified a lack of definitional clarity in the published literature on what is meant by flexible work schedules (Kossek &

Distelberg, 2009). The review compared the National Compensation Survey (USLBS, 1999, 2000, 2003, 2007), The National Study of the Changing Workforce (Bond, Thompson, Galinsky, & Prottas, 2003), and a professional association membership survey (Kossek & Distelberg, 2008). Wide variation was found in definitions, measures, and sampling techniques. The lack of agreement on how to study flexible work schedules is problematic because (a) it makes it difficult to compare these national surveys when one is not sure if the samples or measures are similar and (b) it is likely there is higher measurement error in assessment, given the wide latitude that respondents have to interpret general items, making prevalence levels and empirical linkages more suspect. Similarly, Piotet (1988) cautioned against relying on statistics that assess prevalence across the EU, because common definitions either do not exist or, if they do exist, vary from country to country. It is difficult, therefore, to develop international data on worldwide health effects of flexible work schedule use, or even within-country comparisons within the same firm. As an illustration on a more global scale, there currently is no internationally accepted definition of a “standard” work day or schedule; the definition can vary by national law and culture, organizational culture, and occupation (Cappelli, 1999). Although there is wide variation in culture and legislation on flexibility and work hours, little of this variation has been considered in I/O studies of flexible work schedules. Yet these differences do matter. Take France, for example: The French workweek is officially 35 hours (http://www.triplet.org). Employers can pay a fine to allow lower level workers to work longer hours. In France, most stores are closed on Sundays and many employees take a month-long holiday in August. Although an understanding of these trends certainly makes it possible to study flexibility in France, the fact that France is now part of the EU may make it more difficult to make work-hour comparisons across countries, unless there is some legislative support to discourage workers from working long or irregular hours. Even in the EU variation exists, with full-time work ranging from 35 to 39 hours per week. Yet there is much more public policy support to protect workers from long hours than in the United States. For example, 563

Copyright American Psychological Association. Not for further distribution.

Kossek and Michel

EU legislation has been passed limiting the maximum number of weekly work hours for nonexempt workers to 48 even with overtime payments (Crosby, Williams, & Biernat, 2004). Rarely are such differences considered in studies to put international workhour trends and flexible schedule use in context. We should note that a growing trend in some nations is to actively encourage employers to support flexibility, and more I/O studies should consider institutional effects on flexible work schedule adoption and use. In the UK and Australia, laws regarding equal access to flexible employment have instituted employees’ “right to request” a flexible schedule and the investigation of employer ability to accommodate such requests. International studies of work–life will need to consider this variation in labor standards and legislation context when studying workplace flexibility from an I/O frame.

Need for Expansion of Measurement: More and Different Types of Outcomes More research is needed to clarify outcomes from flexible work schedules, including the amount of flexibility used and the chronology of use, attitudes, and behaviors. With the growing cost of oil, there is renewed interest in the productivity and organizational impacts of varying work schedules, but little quality research exists to inform organizations and society of the costs and benefits of multiple stakeholder perspectives (e.g., employee, employer, family, community) of different flexibility forms. We also need more research on linkages between schedule control and employee health and stress, as this is a growing societal concern. New research suggests linkages between support for flexible work schedules and health, including heart rates, blood pressure, sleep quality, depressive symptoms, and physical pain (Kossek & Hammer, 2008). Such findings suggest flexibility is not just a nice thing to do to attract and retain workers; it also may impact longevity, as well as family and societal well-being. New intermediate measures of work productivity such as engagement, focus, creativity, conflicts over availability, and communication patterns should be included in outcome studies. Certainly studies of outcomes need to be based on longitudinal quasi-experimental work with con564

trol groups. As noted, we found relatively few of such studies in our review. We also believe that one finding—that the favorable effects of using flexibility were higher for individuals with higher work–family conflict—suggests that interventions might be tailored to focus on the members of the workforce who have the greater need and interest in flexible work scheduling. This is the target group most in need of workplace support and who are most immediately likely to benefit from workplace innovation in the short run. In the long run, all workers may benefit from having greater control over where, when, and how long they work over the life course. The fact that more and more employees are spending what used to be personal time for work highlights the need for workers (especially those on a flexible work schedule) to increasingly self-regulate boundaries between work and personal time (Nippert-Eng, 1996). It is also important to determine whether the employer exerts social pressure on employees to restructure personal time as work time, particularly if workloads are too high and there are ambiguous norms about work hours (Kossek & Lee, 2008). Among issues that should be addressed are professional work cultures that socially foster overwork and the tendency to use telework and other flexible work schedule forms to manage rising workloads. This issue is especially important during times when companies may be cutting staff in a bad economy and people may be afraid to request or use flexibility. One may ask, “Is there a minimum or optimal amount of work schedule flexibility to promote well-being?” and “Under what conditions does use of which types of flexible work schedules lead to greater perceptions of schedule control?”

The Future of Flex as a Work Group and Organizational Effectiveness Tool Studies should consider factors influencing acceptance and use, such as the importance of need assessments to make sure policies adopted are congruent with workforce characteristics that may wax and wane over the career and family life cycle. Researchers also should examine the degree to which flexible work schedules are integrated with organizational and business objectives, as well as ensuring the development of managerial support

Copyright American Psychological Association. Not for further distribution.

Flexible Work Schedules

and a favorable organizational culture or climate. We found far more research on the latter topic of support and culture than the former on business strategy or workforce fit or even implementation. Although policies are typically adopted at the organizational level, within firms, there is often wide variation and organizational stratification in which different jobs, work groups, and workforce demographics have access to flexible schedules. Relatively little research has been done at the work-group level of analysis, in particular, which is critical for implementation because most policies are implemented on the basis of supervisory discretion. A review by Van Dyne, Kossek, and Lobel (2007) found that motivation and coordinating effects of flexible schedules were the main implementation challenges at the work-group level. Managers are more likely to experience positive work-group performance impacts if they are able to effectively manage coordination of work schedules and learn how to manage equity within the work group. To facilitate this, it is critical for the employer to allocate resources to train managers and employees to learn how to work in new scheduling forms and to monitor the effectiveness of implementation of work schedules (Kossek & Hammer, 2008; Lautsch & Kossek, 2009). The National Work Family Health Network (see http://www.workfailyhealthnetwork.org) is one example of an effort to train and resocialize supervisors to help work groups and employers learn how to redesign social processes to better support employees’ schedule flexibility. This is a cross-university interdisciplinary initiative, sponsored by the U.S. National Institutes of Health and the Centers for Disease Control and Prevention, that began in 2005. New management training and organizational culture change interventions are being designed to increase employee control over work schedules (cf. Kossek & Hammer, 2008; Kelly & Moen, 2007). The premise is that increasing supervisory and cultural support for workplace flexibility will enable employees to have more control over work schedules, reduce work– family conflicts, and ultimately improve worker health, family well-being, and organizational productivity. Conducted in over 60 work sites nationwide, the study uses a longitudinal quasi-experimental design with repeated waves of measurement of I/O

outcomes, as well as measures of biodata and productivity from workers, coworkers, supervisors, and families. Overall, the study will assess the utility of increased work schedule flexibility as an effective workplace intervention to increase worker health and work productivity. It is an example of the kind of integrative future research that is needed to improve the promise of flexible work schedules to benefit workers, employers, and society (http://www. workfamilyhealthnetwork.org).

References Allen, T. D. (2001). Family-supportive work environments: The role of organizational perceptions. Journal of Vocational Behavior, 58, 414–435. doi:10.1006/jvbe.2000.1774 Allen, T. D., & Russell, J. E. A. (1999). Parental leave of absence: Some not so family-friendly implications. Journal of Applied Social Psychology, 29, 166–191. doi:10.1111/j.1559-1816.1999.tb01380.x Allenspach, H. (1975). Flexible working hours. Geneva: International Labor Office. Anderson, S., Coffey, B. S., & Byerly, R. (2002). Formal organizational initiatives and informal workplace practices: Links to work–family conflict and jobrelated outcomes. Journal of Management, 28, 787–810. doi:10.1016/S0149–2063(02)00190–3 Ang, S., & Slaughter, S. A. (2001). Work outcomes and job design for contract versus permanent information systems professionals on software development teams. MIS Quarterly, 25, 321–350. doi:10.2307/3250920 Ashford, S. J., George, E., & Blatt, R. (2008). Old assumptions, new work: The opportunities and challenges of research on nonstandard employment. In J. P. Walsh & A. Brief (Eds.), Academy of Management annals (pp. 65–117), New York: Erlbaum. Ashforth, B. E. (2001). Role transitions in organizations: An identity-based perspective. Mahwah, NJ: LEA Press. Avery, C., & Zabel, D. (2001). The flexible workplace: A sourcebook of information and research. Westport, CT: Quorum. Bachler, C. J. (1995). Workers take leave of job stress. Personnel Journal, 74, 38–48. Balepur, P. N., Varma, K. V., & Mokhtarian, P. L. (1998). Transport impacts of centre-based telecommuting: Interim findings from the neighborhood telecenters project. Transportation, 25, 287–306. doi:10.1023/A: 1005048329523 Baltes, B. B., Briggs, T. E., Huff, J. W., Wright, J. A., & Neuman, G. A. (1999). Flexible and compressed workweek schedules: A meta-analysis of their 565

Kossek and Michel

effects on work-related criteria. Journal of Applied Psychology, 84, 496–513. Barnett, R. C. (1998). Toward a review and reconceptualization of the work/family literature. Genetic, Social, and General Psychology Monographs, 124, 125–182. Barnett, R. C., Gareis, K. C., & Brennan, R. T. (2008). Wives’ shift work schedules and husbands’ and wives’ well-being in dual-earner couples with children: A within-couple analysis. Journal of Family Issues, 29, 396–422.

Copyright American Psychological Association. Not for further distribution.

Barnett, R. C., & Hall, D, T. (2007). The silver lining in shift work: Can your organization take advantage of it? Organizational Dynamics, 36, 404–417. doi:10.1016/j.orgdyn.2007.06.006 Barringer, M., & Milkovich, G. (1998). A theoretical explanation of the adoption and design of flexible benefit plans: A case in human resource innovation. Academy of Management Review, 23, 305–324. doi:10.2307/259376 Baum, C. L., II. (2006). The effects of governmentmandated family leave on employer family leave policies. Contemporary Economic Policy, 24, 432–445. doi:10.1093/cep/byj025 Bergman, M. E. (2002, August). Psychological and objective contingency as predictors of work attitudes and behavior. Paper presented at the Academy of Management meeting, Denver, Colorado. Bernardino, A., & Ben-Akiva, M. (1996). Modeling the process of adoption of telecommuting: Comprehensive framework. Transportation Research Record, 1552, 161–170. doi: 10.3141/1552-22 Bernhard-Oettel, C., Sverke, M., & De Witte, H. (2005). Comparing three alternative types of employment with permanent full-time work: How do employment contract and perceived job conditions related to health complaints? Work and Stress, 19, 301–318. doi:10.1080/02678370500408723 Bond, T. Thompson, C., Galinsky, E., & Prottas, D. (2003). Highlights of the National Study for the Changing Workforce. New York: Families and Work Institute. Boyce, A. S., Ryan, A. M., Imus, A. L., & Morgeson, F. P. (2007). “Temporary worker, permanent loser?” A model of the stigmatization of temporary workers. Journal of Management, 33, 5–29. doi:10.1177/ 0149206306296575 Breaugh, J. A., & Frye, N. K. (2008). Work–family conflict: The importance of family-friendly employment practices and family-supportive supervisors. Journal of Business and Psychology, 22, 345–353. doi:10. 1007/s10869-008-9081-1 Bureau of Labor Statistics. (2009, May 27). Employment characteristics of families in 2008. Retrieved February 19, 2010, from http://www.bls.gov/news.release/ pdf/famee.pdf 566

Burke, M. E. (2005). Benefits Survey Report: A study by the Society for Human Resource Management. Alexandria, VA: Society for Human Resource Management. Butler, A., Gasser, M., & Smart, L. (2004). A socialcognitive perspective on using family-friendly benefits. Journal of Vocational Behavior, 65, 57–70. doi:10.1016/S0001-8791(03)00097-6 Byron, K. (2005). A meta-analytic review of work–family conflict and its antecedents. Journal of Vocational Behavior, 67, 169–198. Cappelli, P. (1999). Career jobs are dead. California Management Review, 42(1), 146–167. Casper, W. J., & Harris, C. M. (2008). Work-life benefits and organizational attachment: Self-interest utility and signaling theory models. Journal of Vocational Behavior, 72, 95–109. Chesley, N., & Moen, P. (2006). When workers care: Dual-earner couples’ caregiving strategies, benefit use, and psychological well-being. American Behavioral Scientist, 49, 1248–1269. doi:10.1177/ 0002764206286388 Christensen, K. E., & Staines, G. L. (1990). Flextime: A viable solution to work/family conflict? Journal of Family Issues, 11, 455–476. doi:10.1177/ 019251390011004007 Clifton, T. J., & Shepard, E. (2004). Work and family programs and productivity: Estimates applying a production function model. International Journal of Manpower, 25, 714–728. doi:10.1108/ 01437720410570036 Cohen, P. (2002). Cohabitation and the declining marriage premium of men. Work and Occupations, 29(3), 346–363. doi:10.1177/0730888402029003004 Cohen, A. R., & Gadon, H. (1978). Alternative work schedules: Integrating individual and organizational needs. Reading, MA: Addison-Wesley. Conley, H. M. (2003). Temporary work in the public services: Implications for equal opportunities. Gender, Work and Organization, 10, 455–477. doi:10.1111/1468–0432.00206 Connelly, C. E., & Gallagher, D. G. (2004). Emerging trends in contingent work research. Journal of Management, 30, 959–983. doi:10.1016/j.jm.2004.06.008 Contingent and alternative work arrangements. (2005, February). Retrieved August 7, 2008, from http://www.bls.gov/news.release/conemp.nr0.htm Costa, G. (2003). Factors influencing health of workers and tolerance to shift work. Theoretical Issues in Ergonomics Science, 4, 263–288. doi:10.1080/ 14639220210158880 Crosby, F. J., Williams, J. C., & Biernat, M. (2004). The maternal wall. Journal of Social Issues, 60, 675–682. doi:10.1111/j.0022-4537.2004.00379.x

Flexible Work Schedules

Cully, M., O’Reilly, A., & Millward, N. (1998). The 1998 Workplace Employee Relations Survey First findings (URN 98/934). London and Swindon, England: Department of Trade and Industry; Economic and Social Research Council; and Advisory, Conciliation and Arbitration Service. Dalton, D. R., & Mesch, D. J. (1990). The impact of flexible scheduling on employee attendance and turnover. Administrative Science Quarterly, 35, 370–387. doi:10.2307/2393395

Copyright American Psychological Association. Not for further distribution.

Davidson, O. B., Eden, D., & Westman, M. (2004). Sabbatical leave as a “time out” from chronic academic job stress: A test of conservation of resources theory. Unpublished manuscript. Davis, A., & Kalleberg, A. (2006). Family friendly organizations: Work and family programs in the 1990s. Work and Occupations, 33, 191–223. doi:10.1177/ 0730888405280446 Davenport, T., & Pearlson, K. (1998). Two cheers for the virtual office. Sloan Management Review, 39(4), 51–65. Deal, J. (2007). Retiring the generation gap: How employees young and old can find common ground. San Francisco: Jossey-Bass and Center for Creative Leadership. De Bloom, J., Kompier, M., Geurts, S., De Weerth, C., Taria, T., & Sonnentag, S. (2009). Do we recover from vacation? Meta-analysis of vacation effects on health and well-being. Journal of Occupational Health, 51(1), 13–25. doi:10.1539/joh.K8004 Deci, E. L., & Ryan, R. M. (1995). Human autonomy: The basis for true self-esteem. In M. Kemis (Ed.), Efficacy, agency, and self-esteem (pp. 31–49). New York: Plenum Press. Dieringer Research Group. (2004). American interactive consumer survey. Brookfield, WI: Author. Druker, J., White, G., Hegewisch, A., & Mayne, L. (1996). Between hard and soft HRM: Human resource management in the construction industry. Construction Management and Economics, 14, 405–416. doi:10.1080/014461996373278 Duetschman, A. (1994). The managing wisdom of hightech superstars. Fortune, 17, 197–206. Dunham, R. B., Pierce, J. L., & Castaneda, M. B. (1987). Alternative work schedules: Two field quasiexperiments. Personnel Psychology, 40, 215–242. doi:10.1111/j.1744-6570.1987.tb00602.x Eaton, S. (2003). If you can use them: Flexibility policies, organizational commitment and perceived performance. Industrial Relations. 42, 145–167. Etzion, D. (2003). Annual vacation: Duration of relief from job stressors and burnout. Anxiety, Stress & Coping, 16, 213–226.

Evans, J., Kunda, G., & Barley, S. (2004). Beach time, bridge time, and billable hours: The temporal structure of technical contracting. Administrative Science Quarterly, 49, 1–38. Fagan, C. (2001). The temporal reorganization of employment and household rhythm of work schedules. American Behavioral Scientist, 44, 1199–1212. doi:10.1177/00027640121956728 Feldman, D. C., & Doerpinghaus, H. L. (1992). Patterns of part-time employment. Journal of Vocational Behavior, 41, 283–294. doi:10.1016/00018791(92)90030-4 Fletcher, J. K., & Bailyn, L. (1996). Challenging the last boundary: Reconnecting work and family. In M. B. Arthur & D. M. Rousseau (Eds.), The boundaryless career (pp. 256–267). New York: Oxford University Press. Fritz, C., & Sonnentag, S. (2006). Recovery, well-being, and performance-related outcomes: The role of workload and vacation experiences. Journal of Applied Psychology, 91, 936–945. Medline doi:10.1037/0021-9010.91.4.936 Galinsky, E., Bond, J. T., & Hill, E. J. (2004). A status report on workplace flexibility: Who has it? Who wants it? What difference does it make? New York: Families & Work Institute. Galup, S., Saunders, C., Nelson, R. E., & Cerveny, R. (1997). The use of temporary staff and managers in a local government environment. Communication Research, 24, 698–730. doi:10.1177/ 0093650297024006005 Golden, L. (2001, March). Flexible work schedules: What are we trading to get them? Monthly Labor Review, 124(3), 50–67. Golden, L. (2008). Limited access: Disparities in flexible work schedules and work-at-home. Journal of Family and Economic Issues, 29, 86–109. doi:10.1007/ s10834-007-9090-7 Golden, T. D., Veiga, J. F., & Simsek, Z. (2006). Telecommuting’s differential impact on work–family conflict: Is there no place like home? Journal of Applied Psychology, 91, 1340–1350. Medline doi:10. 1037/0021-9010.91.6.1340 Goodale, J. G., & Aagaard, A. K. (1975). Factors relating to varying reactions to the 4-day workweek. Journal of Applied Psychology, 60, 33–38. doi:10.1037/h0076345 Granrose, C. S., & Appelbaum, E. (1986, January). The efficiency of temporary help and part-time employment. Personnel Administrator, 31(1), 71–82. Greenhaus, J. H., & Powell, G. N. (2006). When work and family are allies: A theory of work–family enrichment. Academy of Management Review, 31, 72–92. 567

Kossek and Michel

Copyright American Psychological Association. Not for further distribution.

Grönlund, A. (2007). More control, less conflict? Job demand-control, gender and work–family conflict. Gender, Work and Organization, 14, 476–497. doi:10.1111/j.1468-0432.2007.00361.x

women on work and family. Journal of Family and Economic Issues, 25, 121–136. doi:10.1023/B:JEEI. 0000016726.06264.91

Halpern, D. F. (2005). How time-flexible work policies can reduce stress, improve health, and save money. Stress and Health, 21, 157–168. doi:10.1002/smi.1049

Hornung, S., Rousseau, D. M., & Glaser, J. (2008). Creating flexible work arrangements through idiosyncratic deals. Journal of Applied Psychology, 93, 655–664. doi:10.1037/0021-9010.93.3.655

Hammer, L. B., & Barbera, K. M. (1997). Toward an integration of alternative work schedules and human resource systems. Human Resource Planning, 20(2), 28–36.

Houseman, S., & Osawa, M. (Eds.). (2003). Nonstandard work in developed economies: Causes and consequences. Kalamazoo, MI: W.E. Upjohn Institute for Employment Research.

Hammer, L., Kossek, E., Anger, K., Bodner, T., & Zimmerman, K. (2009, April). A field study to clarify work–family intervention processes: The roles of work–family conflict and family supportive supervisor behaviors (Working Paper, Portland State University, OR). Paper presented at the Annual Meeting of the Society for Industrial Psychology.

Hunnicutt, B. K. (1996). Kellogg’s six-hour day. Philadelphia: Temple University Press.

Hammer, L. B., Kossek, E. E., Zimmerman, K., & Daniels, R. (2007). Clarifying the construct of family supportive supervisory behaviors (FSSB): A multilevel perspective. In P. L. Perrewe & D. C. Ganster (Eds.). Research in occupational stress and well-being (Vol. 6, pp. 171–211). Amsterdam: Elsevier. Hammer, L. B., Neal, M. B., Newsom, J., Brockwood, K. J., & Colton, C. (2005). A longitudinal study of the effects of dual-earner couples’ utilization of family-friendly workplace supports on work and family outcomes. Journal of Applied Psychology, 90, 799–810. doi:10.1037/0021-9010.90.4.799 Hamermesh, D. (1999). The timing of work over time. Economic Journal, 109, 37–66. doi:10.1111/ 1468-0297.00390 Harris, H., & Fennell, H. (1988). A multivariate model of job stress and alcoholic consumption. The Sociological Quarterly, 29, 391–406. doi:10.1111/j.1533-8525.1988. tb01260.x Henderson, D., & Mokhtarian, P. (1996). Impacts of center-based telecommuting on travel and emissions: Analysis of the Puget Sound demonstration project. Transportation Research Part D, Transport and Environment, 1(1), 29–45. doi:10.1016/ S1361–9209(96)00009–0 Hernreich, R. (2008). [Post to Ellen Ernst Kossek, Harvard Business Press blog]. Retrieved August 23, 2008, from http://conversationstarter.hbsp.com/ 2008/04/when_flextime_isnt_working.html Higgins, C., Duxbury, L., & Johnson, K. L. (2000). Parttime work for women: Does it really help balance work and family? Human Resource Management, 39, 17–32. doi:10.1002/(SICI)1099-050X(200021)39: 13.0.CO;2-Y Hill, E. J., Märtinson, V. K., Ferris, M., & Baker, R. Z. (2004). Beyond the mommy track: The influence of new-concept part-time work for professional 568

Ilies, R., Schwind, K., Wagner, D., Johnson, M., DeRue, S., & Ilgen, D. (2007). When can employees have a family life? The effects of daily workload and affect on work–family conflict and social behaviors at home. Journal of Applied Psychology, 92, 1368–1379. doi:10.1037/0021-9010.92.5.1368 Ivanovic, A., & Collin, P. (Eds.). (2006). Dictionary of human resources and personnel management. London: A & C Black. Jansen, N. W. H., Kant, I., Nijhuis, F. J. N., Swaen, G. M. H., & Kristensen, T. S. (2004). Impact of worktime arrangements on work-home interference among Dutch employees. Scandinavian Journal of Work, Environment & Health, 30, 139–148. Jacobs, J., Gerson, K., & Gornick, J. (2004). The time divide: Work, family and gender inequality (pp. 127–147). Cambridge: Harvard University Press. Jacobs, J., & Winslow, S. (2004). Overworked faculty: Job stresses and family demands. Annals of the American Academy of Political and Social Science, 596, 104–129. Kahne, H. (1985). Reconceiving part-time work: New perspectives for older workers and women. Totowa, NJ: Rowman & Allanheld. Karnowski, S., & White, B. (2002). The role of facility managers in the diffusion of organizational telecommuting. Environment and Behavior, 34, 322–334. doi:10.1177/0013916502034003003 Karasek, R. (1979). Job demands, job decision latitude and mental strain. Administrative Science Quarterly, 24, 285–308. doi:10.2307/2392498 Karasek, R. A., & Theorell, T. (1990). Healthy work: Stress, productivity, and the reconstruction of working life. New York: Basic Books. Kelly, E. (2006). Work–family policies: The U.S. in international perspective. In M. Pitt-Catsouphes, E. Kossek, & S. Sweet (Eds.), The work and family handbook: Multi-disciplinary perspectives, methods and approaches (pp. 99–124). Mahwah, NJ: Erlbaum. Kelly, E. L., & Moen, P. (2007). Rethinking the clockwork of work: Why schedule control may pay off at

Flexible Work Schedules

work and at home. Advances in Developing Human Resources, 11, 487–506. Kelly, E., Kossek, E., Hammer, L. Durham, M., Bray, J., Chermack, K., et al. (2008). Getting there from here: Research on the effects of work–family initiatives on work–family conflict and business outcomes. In J. P. Walsh & A. Brief (Eds.), Academy of Management annals (Vol. 2, No. 1, pp. 305–349). Briarcliff Manor, NY: Academy of Management.

Copyright American Psychological Association. Not for further distribution.

Kossek, E. (2008). When flextime isn’t working [post on Harvard Business Publishing blog]. Retrieved October 6, 2009, from http://blogs.harvardbusiness. org/cs/2008/04/when_flextime_isnt_working.html Kossek, E., & Distelberg, B. (2008). Work and family employment policy for a transformed work force: Trends and themes. In N. Crouter & A. Booth (Eds.), Work-life (pp. 1–51). Washington, DC: Urban Institute Press. Kossek, E., & Distelberg, B. (2009). Work and family employment policy for a transformed work force: Trends and themes. In N. Crouter & A. Booth (Eds.), Work–life policies that make a real difference for individuals, families, and organizations (pp. 3–51). Washington, DC: Urban Institute Press. Kossek, E., & Hammer, L. (2008, November). Work/life training for supervisors gets big results. Harvard Business Review, p. 36. Kossek, E., & Lautsch, B. (2008). CEO of me: Creating a life that works in the flexible job age. New York: Wharton School. Kossek, E., & Lautsch, B. (2009). “Good teleworking”: Under what conditions does teleworking enhance employees’ well-being? In Y. Amichai-Hamburger (Ed.), Technology and psychological well-being (pp. 148–173). Cambridge, England: Cambridge University Press. Kossek, E., Lautsch, B., & Eaton, S. (2005). Flexibility enactment theory: Implications of flexibility type, control and boundary management for work–family effectiveness. In E. E. Kossek & S. Lambert (Eds.), Work and life integration: Organizational, cultural and psychological perspectives (pp. 243–262). Mahwah, NJ: Erlbaum.

Kossek, E., & Misra, K. (2008). Nonwork influences on work motivation. In R. Kanfer, G. Chen, & R. Pritchard (Eds.), Work motivation: Past, present, and future (pp. 471–500). Oxford, England: Routledge. Kossek, E. E. (2005). Workplace policies and practices to support work and families. In S. Bianchi, L. Casper, & R. King (Eds.), Work, family, health, and well-being (pp. 97–116). Mahwah, NJ: Erlbaum. Kossek, E. E. (2006). Work and family in America: Growing tensions between employment policy and a changing workforce: A thirty year perspective (Commissioned chapter by SHRM Foundation and University of California Center for Organizational Effectiveness for the 30th anniversary of the State of Work in America). In E. Lawler & J. O’Toole (Eds.), America at work: Choices and challenges (pp. 53–72). New York: Palgrave MacMillan. Kossek, E. E., Barber, A., & Winters, D. (1999). Using flexible schedules in the managerial world: The power of peers. Human Resource Management, 38, 33–46. doi:10.1002/(SICI)1099-050X(199921)38: 13.0.CO;2-H Kossek, E. E., Lautsch, B. A., & Eaton, S. C. (2006). Telecommuting, control, and boundary management: Correlates of policy use and practice, job control, and work-family effectiveness. Journal of Vocational Behavior, 68, 347–367. Kossek, E. E., Noe, R., & DeMarr, B. (1999). Work–family role synthesis: Individual, family and organizational determinants. Journal of Conflict Resolution, 10, 102–129. doi:10.1108/eb022820 Kossek, E. E., & Ozeki, C. (1998). Work–family conflict, policies, and the job–life satisfaction relationship: A review and directions for organizational behavior/ human resources research. Journal of Applied Psychology, 83, 139–149. doi:10.1037/00219010.83.2.139 Kossek, E., & Van Dyne, L. (2008). Face time matters: A cross level model of how work life flexibility influences work performance of individuals and groups. In D. Lero, K. Korabick, & D. Whitehead (Eds.), Work family handbook. Amsterdam: Elsevier.

Kossek, E., & Lee, M. (2005). Making flexibility work: What managers have learned about implementing reduced-load work (Alfred P. Sloan Foundation Study Technical Report). Williamsburg and Toronto, Ontario, Canada: Michigan State University and McGill University. Available at http:ellenkossek. lir.msu.edu/

Kurkland, N. B., & Bailey, D. E. (1999). Telework: The advantages and challenges of working here, there, anywhere, and anytime. Organizational Dynamics, 28, 53–68. doi:10.1016/S0090-2616(00)80016-9

Kossek, E., & Lee, M. (2008). Implementing a reducedworkload arrangement to retain high talent: A case study. Journal of Managerial Psychology, 11(1), 49–64. doi:10.1080/10887150801966995

Lambert, S. (2000). Added benefits: The link between work–life benefits and organizational citizenship behavior. Academy of Management Journal, 43, 801–815. doi:10.2307/1556411

Kramer, D. C. (2001). Workplace sabbaticals: Bonus or entitlement? Westport, CT: Quorum.

569

Kossek and Michel

Landry, B., Mahesh, S., & Hartman, S. (2005). The changing nature of work in the age of e-business. Journal of Organizational Change Management, 18, 132–144. doi:10.1108/09534810510589561 Lapierre, L. M., & Allen, T. D. (2006). Work-supportive family, family-supportive supervision, use of organizational benefits, and problem-focused coping: Implications for work–family conflict and employee well-being. Journal of Occupational Health Psychology, 11, 169–181. doi:10.1037/1076-8998.11.2.169

Copyright American Psychological Association. Not for further distribution.

Lautsch, B., & Kossek, E. (2009). Supervisor paradoxes of managing telecommuting. Human Relations, 62, 795–827. doi:10.1177/0018726709104543 Lee, M. D., MacDermid, S. M., & Buck, M. L. (2000). Organizational paradigms of reduced load work: Accommodation, elaboration, transformation. Academy of Management Journal, 43, 1211–1226. doi:10.2307/1556346 Lee, M. D., MacDermid, S. M., Williams, M. L., Buck, M. L., & Leiba-O’Sullivan, S. (2002). Contextual factors in the success of reduced-load work arrangements among managers and professionals. Human Resource Management, 41, 209–223. doi:10. 1002/hrm.10032 Lee, T. W., & Johnson, D. R. (1991). The effects of work schedule and employment status on the organizational commitment and job satisfaction of full-versus part-time employees. Journal of Vocational Behavior, 38, 208–224. doi:10.1016/0001-8791(91)90028-K

and health through flexibility and resilience. Social Science & Medicine, 66, 1019–1033. doi:10.1016/ j.socscimed.2007.11.013 Marks, N. F. (1998). Does it hurt to care? Caregiving, work–family conflict, and midlife well-being. Journal of Marriage and the Family, 60, 951–966. doi:10.2307/ 353637 Martens, M. F. J., Nijhuis, F. J. N., Van Boxtel, M. P. J., & Knottnerus, J. A. (1999). Flexible work schedules and mental and physical health: A study of a working population with non-traditional working hours. Journal of Organizational Behavior, 20, 35–46. doi:10.1002/(SICI)1099-1379(199901)20: 13.0.CO;2-Z Martínez-Sánchez, A., Pérez-Pérez, M., Vela-Jiménez, M. J., & de-Luis-Carnicer, P. (2008). Telework adoption, change management, and firm performance. Journal of Organizational Change Management, 21, 7–31. doi:10.1108/09534810810847011 Maxwell, G., Rankine, L., Bell, S., & MacVicar, A. (2007). The incidence and impact of flexible working arrangements in smaller businesses. Employee Relations, 29, 138–161. doi:10.1108/ 01425450710719987 McDonald, D. J., & Makin, P. J. (2000). The psychological contact, organizational commitment and job satisfaction of temporary staff. Leadership and Organization Development Journal, 21, 84–91. doi:10.1108/01437730010318174

Local Government Management Board. (1998). Flexible working: Working patterns in local authorities and the wider economy. London: Author.

McGinnis, S. K., & Morrow, P. C. (1990). Job attitudes among full- and part-time employees. Journal of Vocational Behavior, 36, 82–96. doi:10.1016/ 0001-8791(90)90016-U

Lobel, S. A. (1991). Allocation of investment in work and family roles: Alternative Theories and implications for research. Academy of Management Review, 16, 507–521. doi:10.2307/258915

Meiksins, P., & Whalley, P. (2002). Putting work in its place: A quiet revolution. Ithaca, NY: Cornell University Press.

Lobel, S. A., & Kossek, E. E. (1996). Human resource strategies to support diversity in work and personal lifestyles: Beyond the “family-friendly” organization. In E. E. Kossek & S. A. Lobel (Eds.), Managing diversity: Human resource strategies for transforming the workplace (pp. 221–244). Cambridge, MA: Blackwell.

Mokhtarian, P. L. (1998). A synthetic approach to estimating impacts of telecommuting on travel. Urban Studies [Edinburgh, Scotland], 35, 215–241. doi:10.1080/0042098984952

Lockyer, C., & Scholarios, D. (2007). The “rain dance” of selection in construction: Rationality as ritual and the logic of informality. Personnel Review, 36, 528–548. doi:10.1108/00483480710752786 Lucas, J. L., & Heady, R. B. (2002). Flextime commuters and their driver stress, feelings of time urgency, and commute satisfaction. Journal of Business and Psychology, 16, 565–571. doi:10.1023/A: 1015402302281 MacEachen, E., Polzer, J., & Clarke, J. (2008). You are free to set your own hours: Governing productivity 570

Mery, A. (2009, May 15). Personal communication. General Electric site visit. Buc, France.

Nippert-Eng, C. (1996). Home and work: Negotiating boundaries through everyday life. Chicago: University of Chicago Press. Nord, W. R., & Costigan, R. (1973). Workers adjustment to the four-day week: A longitudinal study. Journal of Applied Psychology, 58, 60–66. doi:10.1037/h0035419 Olmsted, B., & Smith, S. (1989). Creating a flexible workplace: How to select and manage alternative work options. New York: AMACOM. Orpen, C. (1981). Effect of flexible working hours on employee satisfaction and performance: A field

Flexible Work Schedules

experiment. Journal of Applied Psychology, 66, 113–115. doi:10.1037/0021-9010.66.1.113 Parker, L., & Allen, T. D. (2001). Work/family benefits: Variable related to employees’ fairness perceptions. Journal of Vocational Behavior, 58, 453–468. doi:10.1006/jvbe.2000.1773

Copyright American Psychological Association. Not for further distribution.

Pearce, J. A. (1993). Toward an organizational behavior of contract laborers: Their psychological involvement and effects on employee coworkers. Academy of Management Journal, 36, 1082–1096. doi:10.2307/256646 Pérez, M. P., Sánchez, A. M., & de-Luis-Carnicer, M. P. (2003). The organizational implications of human resources managers’ perceptions of teleworking. Personnel Review, 32, 733–755. doi:10.1108/ 00483480310498693 Pierce, J. L., & Newstrom, J. W. (1983). The design of flexible work schedules and employee responses: Relationships and process. Journal of Occupational Behavior, 4, 247–262. Pierce, J. L., Newstrom, J. W., Dunham, R. B., & Barber, A. E. (1989). Alternative work schedules. Needham Heights, MA: Allyn & Bacon. Piotet, F. (1988). The changing face of work: Researching and debating the issues. Dublin, Ireland: European Foundation for the Improvement of Living and Working Conditions. Pleck, J. H. (1997). Paternal involvement: Levels, origins, and consequences. In M. E. Lamb (Ed.), The role of the father in child development (3rd ed., pp. 66–103). New York: Wiley. Polivka, A. E., & Nardone, T. (1989). The definition of contingent work. Monthly Labor Review, 112, 9–16. Presser, H. (1999, June 11). Toward a 24-hour economy. Science, 284, 1778–1779. doi:10.1126/science.284. 5421.1778 Presser, H. (2003). Working in a 24/7 economy: Challenges for American families. New York: Russell Sage Foundation. Ralston, D. A. (1989). The benefits of flextime: Real or imagined? Journal of Organizational Behavior, 10, 369–373. doi:10.1002/job.4030100407 Rau, B. (2003). Flexible work arrangements. Work family encyclopedia. Sloan Work–Family Network. Retrieved July 23, 2009, from http://wfnetwork. bc.edu/encyclopedia_entry.php?id=240&area=All Richman, A., Civian, J., Shannon, L., Hill, J., & Brennan, R. (2008). The relationship of perceived flexibility, supportive work life policies, and use of formal flexible arrangements and occasional flexibility to employee engagement and expected retention. Community Work & Family, 11(20), 183–197. doi:10.1080/13668800802050350

Robèrt, M., & Börjesson, M. (2006). Company incentives and tools for promoting telecommuting. Environment and Behavior, 38, 521–549. doi:10.1177/ 0013916505283422 Ronen, S. (1981). Flexible working hours: An innovation in the quality of work life. New York: McGraw-Hill. Rothbard, N., Phillips, K., & Dumas, T. (2005). Managing multiple roles: Work–family policies and individual’s desires for segmentation. Organization Science, 16, 243–258. doi:10.1287/orsc.1050.0124 Rousseau, D. (2007). I-Deals: Idiosyncratic deals that employees bargain for themselves. New York: W. E. Sharpe. Ryan, A., & Kossek, E. (2008). Work-Life Policy Implementation: Breaking Down or Creating Barriers to Inclusiveness. Human Resource Management, 47, 295–310. doi:10.1002/hrm.20213 Shockley, K. M., & Allen, T. D. (2007). When flexibility helps: Another look at the availability of flexible work arrangements and work–family conflict. Journal of Vocational Behavior, 71, 479–493. doi:10.1016/ j.jvb.2007.08.006 Sonnentag, S. (2001). Work, recovery activities, and individual well-being: A diary study. Journal of Occupational Health Psychology, 6, 196–210. doi:10.1037/1076-8998.6.3.196 Stavrou, E. T. (2005). Flexible work bundles and organizational competitiveness: A cross-national study of the European work context. Journal of Organizational Behavior, 26, 923–947. doi:10.1002/job.356 Stebbins, L. F. (2001). Work and family in America: A reference handbook. Santa Barbara, CA: Abc-Clio. Tilly, C. (1996). Half a bad job: Bad and good part time jobs in a changing labor market. Philadelphia: Temple University Press. Totterdell, P., Spelten, E., Smith, L., Barton, J., & Folkard, S. (1995). Recovery from work shifts: How long does it take? The Journal of Applied Psychology, 80(1), 43–57.doi:10.1037/0021-9010.80.1.43 UK Department for Business Innovation & Skills. (1998). Workplace employee relations survey. University of Illinois Office of the Vice President for Academic Affairs. (1996). Sabbatical study. Retrieved July 10, 2008, from http://www.vpaa.uillinois.edu/ reports_retreats/sabbatical.asp U.S. Census Bureau. (2002). Statistical abstract of the United States: 2002, The National Data Book, Section 12, Labor force, employment, and earnings, Table No. 578. Persons doing job-related work at home: 2001. Washington, DC: U.S. Census Bureau. U.S. Office of Personnel Management. (2003). Negotiating flexible and compressed work schedules. Retrieved July 2007 from http://www.opm.gov/lmr/html/flexible.asp 571

Kossek and Michel

Williams, J. C., & Calvert, C. T. (2002). Balanced hours: Effective part-time policies for Washington law firms: The Project for Attorney Retention [Final report, 3rd ed.]. William & Mary Journal of Women and the Law, 8, 357. Available at http://www.triplet.com/50–10_ employment/50–20_workingtime.asp

Copyright American Psychological Association. Not for further distribution.

Van Dyne. L. Kossek, E. & Lobel S. (2008). Less need to be there: Cross level effects of work practices that support work–life flexibility and enhance group processes and group-level OCB. Human Relations, 60, 1123–1153. Van Dyne, L., & Ang, S. (1998). Organizational citizenship behavior of contingent workers in Singapore. Academy of Management Journal, 41, 692–703. Van Echtelt, P., Giebbeek, A. & Lindenberg, S. (2006). The new lumpiness of work: Explaining the mis-

572

match between actual and preferred working hours. Work, Employment and Society, 20, 493–512. van Rijswijk, K., Bekker, M. H. J., Rutte, C. G., & Croon, M. A. (2004). The relationships among part-time work, work–family interference, and well-being. Journal of Occupational Health Psychology, 9, 286–295. Venkatesh, V., & Johnson, P. (2002). Telecommuting technology implementations: A within- and betweensubjects longitudinal field study. Personnel Psychology, 55, 661–687. Vroom, V. (1964). Work and motivation. New York: Wiley. Watson, I., Buchanan, J., Campbell, I., & Briggs, C. (2003). Fragmented futures: New challenges in working life. Sydney, Australia: Federation Press. Whyte, W. (1956). The organization man. New York: Simon & Schuster.

CHAPTER 18

NONSTANDARD WORKERS: WORK ARRANGEMENTS AND OUTCOMES

Copyright American Psychological Association. Not for further distribution.

Elizabeth George and Carmen Kaman Ng

What does a temporary office assistant in the United States have in common with a contract engineer in Nigeria, a textile worker on daily wages in Bangladesh, or a banker who lives in Hong Kong but works in Zurich? Among other similarities, they all occupy jobs that we would describe as being nonstandard. They are all in work arrangements that are different from that of the conventional worker who notionally works a fixed number of hours at a specific organizational location and who can reasonably expect long-term employment in that organization. Despite these similarities that distinguish them from standard workers, nonstandard workers are not a homogenous group. As can be seen from our example, they vary on many dimensions, such as their human capital, their work location, their wages or marketability, or the power that they hold relative to organizations. Irrespective of the type of nonstandard work to which we refer, the past 20 years has seen a significant increase in the number of individuals in nonstandard work arrangements. For instance, Houseman and Osawa (2003b) reported that in France, the proportion of the labor force that can be classified as part-time employees, direct-hired temporaries, and temporary help agency workers increased from 16.6% in 1988 to 27.6% in 1998. Similar trends of growth hold for Japan, Germany, the Netherlands, Spain, and the United Kingdom (Houseman & Osawa, 2003b). In addition, the U.S. Census Bureau (2000) reported a rapid increase (23%) in work from home between 1990 and 2000, which is twice the growth rate of the overall work-

force. Workers who work from remote sites are also gaining popularity globally (MacDuffie, 2007). Researchers and practitioners are both challenged by this phenomenon. Researchers have addressed a number of questions related to nonstandard work, such as the conditions under which it arises (von Hippel, Mangum, Greenberger, Heneman, & Skoglind, 1997), attitudinal differences between workers in standard and nonstandard work arrangements (van Dyne & Ang, 1998), and the relationships between standard and nonstandard workers (Davis-Blake, Broschak, & George, 2003). There have also been several reviews of the literature in this area—the work of Connelly and Gallagher (2004) and Ashford, George, and Blatt (2007) being two notable examples. Managers have been concerned with questions such as those relating to productivity of nonstandard workers (Greenwald & Liss, 1973), trust building in virtual teams (Handy, 1995), and outsourcing contract management (Quinn, 1999; Shi, 2007). In this chapter, we explore the work of individuals in nonstandard work arrangements with a focus on the following areas. We first define nonstandard work, relating this term to other terms such as contingent work, temporary work, and employment externalization. We then examine the reasons why organizations have nonstandard work arrangements and why individuals labor under these work conditions. Next, we examine the psychology of nonstandard work, focusing on the attitudes and behaviors of nonstandard workers and standard workers who work alongside nonstandard workers.

http://dx.doi.org/10.1037/12169-018 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

573

Copyright American Psychological Association. Not for further distribution.

George and Ng

We conclude with some prescriptions for future research on this topic, as well as some suggestions for nonstandard workers and their managers on making this work arrangement productive and enjoyable. We should note that we do recognize that nonstandard work is contextually defined and would necessarily vary on the basis of the conditions of the labor market and the legal institutions within which it is found. Our chapter on nonstandard work and workers is limited by the research to which we had access, most of which was done in North America, Europe, or Australia. We recognize that this is a limitation of what follows. (See also chap. 17, this volume.) DEFINING NONSTANDARD WORK Much of the research on nonstandard work draws from Pfeffer and Baron’s (1988) description of the three forms of attachment to organizations: temporal, physical, and administrative. Pfeffer and Baron noted that workers vary in terms of the length of time they are attached to organizations, the physical location in which they work, and who has administrative and supervisory responsibility for them. There are two aspects to their temporal attachment to organizations. First, workers might have contracts with the organization that are clearly intended to last over a finite period of time. This period of time could vary from a day to several years. Of most importance, this time boundary is explicitly part of the written or unwritten contract between the individual and the organization. Examples of this type of nonstandard worker include the temporary office assistant in the United States, the contract engineer in Nigeria, and the daily wage earner in Bangladesh, who we mentioned in the opening paragraph of this chapter. References to this category of worker can be found in studies in which they have been described as temporary workers (Parker, Griffin, Sprigg, & Wall, 2002) or workers on fixed-term contracts (Uzzi & Barsness, 1998). In contrast, standard workers have contracts with organizations that could ostensibly last indefinitely. Alternatively, workers can vary in terms of the length of time that they work per week in the organization. Thus, some workers work on a part-time basis over the longer term, whereas others work for 574

longer durations per week. Part-time workers occupy a unique place among nonstandard workers because they could have all the attributes of standard workers (i.e., are employed on indefinite contracts, work at their employers’ site, and are administered by the employer) but for the fact that they work fewer hours than the typical standard worker. National labor statistics often categorize part-time workers as a group that is different from standard workers, yet they are not included in the count of other nonstandard workers, such as independent contractors or those who work for temporary help agencies (e.g., U.S. Bureau of Labor Statistics, 2005). Nevertheless, part-time workers are seen to be different enough from nonstandard workers to warrant independent investigations of their commitment to organizations (McGinnis & Morrow, 1990), their satisfaction with their jobs (Fenton-O’Creevy, 1995), and the effect of their presence on their standard coworkers (Broschak & Davis-Blake, 2006). In this chapter, we pay limited attention to part-time workers on account of this liminal position between standard and nonstandard work. Physical attachment to the organization refers to the individual’s work location. Whereas standard workers are those who work at a specified organizational location, nonstandard workers engage in their job-related responsibilities at sites that have traditionally not been associated with work. The banker in our earlier example, who lives in Hong Kong but telecommutes to Switzerland, is an example of this type of nonstandard worker. These workers have been labeled at-home workers (Ammons & Markham, 2004), virtual workers (Hill, Ferris, & Martinson, 2003), or teleworkers (Bailey & Kurland, 2002). Finally, Pfeffer and Baron (1988) argued that workers can vary in the degree that they are administratively attached to organizations. Nonstandard workers who are employed by temporary help agencies or workers who provide services to an organization without being on the payroll can be included in this category. The contract engineer in Nigeria in our opening example typifies this kind of worker who provides specialized services to the organization without being an employee. Researchers have studied this type of nonstandard worker under the label of contract workers (George & Chattopadhyay,

Copyright American Psychological Association. Not for further distribution.

Nonstandard Workers

2005), independent contractors or freelancers (Kunda, Barley, & Evans, 2002; Stanworth & Stanworth, 1997), and agency temps (Van Breugel, Van Olffen, & Olie, 2005). Although Pfeffer and Baron’s (1988) categories provided useful handles to distinguish between forms of attachment to the organization, they are limited in a practical sense because jobs and workers often are low on more than one of these dimensions of attachment. A contract worker could have both limited administrative and temporal attachment to the organization. Our definition of nonstandard work and workers is guided by Ashford et al.’s (2007) practical recommendation for determining whether workers can be categorized as nonstandard. First, they suggested that we examine whether the worker is low on any one of the three forms of attachment to organizations: temporal, physical, or administrative. If the answer is in the affirmative, then Ashford et al. suggested that we examine whether that kind of job has traditionally been organized in this fashion. In other words, do others in the same job work with the same form of attachment to the organization? Nonstandard work thus is work that is weakly temporally, physically, or administratively connected to organizations, and this form of attachment is different from how work is typically done in that context. Nonstandard workers are those workers who engage in these nonstandard jobs. Two key attributes of this definition are that there are many different ways in which work can be nonstandard and that the labeling of work as nonstandard is socially determined such that the same job can be called standard or nonstandard, depending on the context in which it is conducted. A plethora of terms have been used to describe nonstandard work, such as contingent work (McLean Parks, Kidder, & Gallagher, 1998), temporary work (De Cuyper et al., 2008), casual employment (Campbell & Burgess, 2001), external employment and contract work (Kalleberg, Reynolds, & Marsden, 2003). Each of these terms is useful in that it highlights different aspects of this form of work that distinguishes it from more standard work. The term contingent work focuses attention on the fact that this form of work is seen as a solution for organizations to deal with the contingences of their environ-

ments in a flexible manner (Uzzi & Barsness, 1998). Temporary work and casual work highlight the temporal nature of this form of work (Foote & Folta, 2002) and the associated idea that the commitment of individuals and organizations to each other is limited (McDonald & Makin, 2000). Similarly, contract work makes explicit the notion of a defined task or role that workers perform (Lepak & Snell, 1999) and highlights the often transactional relationship between nonstandard workers and organizations (Millward & Brewerton, 1999). Finally, the use of the term externalized work (Masters & Miles, 2002) or outsourced work (Harrison & Kelley, 1993) draws attention to the boundary of the organization, making explicit that some work is conducted outside the physical and administrative boundaries of the organization. We believe that nonstandard work can take all of these aspects of work into account: Nonstandard work implies work that is contingent on the organization’s needs and that can be limited in time, space, or administration from how work is traditionally organized and conducted. Of most importance, the term nonstandard work makes explicit the idea of a standard and suggests the possibility that the same work arrangement might be seen as standard or nonstandard, depending on the context. ORGANIZATIONAL RATIONALE FOR NONSTANDARD WORK Organizations use nonstandard workers for a wide range of reasons. For instance, von Hippel et al. (1997) reported cutting cost and increasing flexibility as two common reasons why organizations use temporary employees. Similar findings were found for outsourcing. Managers in the U.S. metalworking sector in Harrison and Kelley’s (1993) study reported that their use of outsourcing was to reduce cost, to manage fluctuation in demand that exceeds current capacity, and to meet the requirement of specialized equipment or skills that the organization does not have in house. More recently, a survey conducted by InformationWeek (“In Depth: Customers Analyze Outsourcing,” 2006a) reported that 65% of their sample of business technology professionals listed cost saving as the most important factor driving outsourcing, followed by flexibility to adjust information 575

Copyright American Psychological Association. Not for further distribution.

George and Ng

Firms use nonstandard workers to gain flexibility in both the type and amount of human resources that they have (Davis-Blake & Hui, 2003; Harrison & Kelley, 1993). For example, managers in Kalleberg et al.’s (2003) study reported acquiring special skills and accommodating varying work demands as the two top reasons for deploying nonstandard work arrangements. We discuss next the two forms of flexibility that nonstandard workers afford to organizations: numerical and functional flexibility.

Empirical studies provide evidence of different aspects of the numerical flexibility strategy. A number of studies have shown that there is a significant and positive relationship between the use of nonstandard workers and variability in the demand for workers. For example, industry seasonality is positively associated with the use of nonstandard work arrangements (Kalleberg et al., 2003; b = 1.02, p < .05), variability in organizational employment needs has a positive relationship with the number of temporary workers used (Abraham, 1988; DavisBlake & Uzzi, 1993), and uncertainty concerning the demand for workers in a given job position in the future increases the likelihood that the position will be filled via external labor arrangements (Masters & Miles, 2002; b = −0.47, p < .01). Case studies by Houseman et al. (2003) found that hospitals use agency temporaries to fill vacancies pending the recruitment of permanent staff, and auto parts manufacturers use agency temporaries to buffer core workers and accommodate increases in the workload. Finally, findings pertaining to organization size also give indirect evidence of how numerical flexibility drives the use of nonstandard workers. Davis-Blake and Uzzi (1993) found support for their argument that organizations with larger size have more employees and slack to meet temporary needs and thus are less likely to use temporary workers (b = −0.001, SE = .00005, p < .05).

Numerical flexibility. Nonstandard work arrangements facilitate numerical flexibility, which refers to the ability to adjust workforce size so that it can be matched to the volume of work (Atkinson, 1984; Davis-Blake & Hui, 2003). Harrison and Kelley (1993) referred to this as a form of advantage achieved by capacity subcontracting in the context of outsourcing, in which “management turns to an outside source to temporarily supplement existing capacity” (p. 217). Numerical flexibility also allows organizations to use nonstandard employees to buffer their core employees from the vicissitudes of the labor market (Ko, 2003). Some organizations intentionally hire a peripheral workforce consisting of nonstandard workers who are laid off first in case of downturns in labor demand. This periphery protects standard workers from job loss (Houseman, Kalleberg, & Erickcek, 2003).

Functional flexibility. Davis-Blake and Hui (2003) defined functional flexibility as the extent to which firms can redefine work tasks, redeploy resources, and reconfigure relationships with workers so that the organization does not maintain on its staff employees for whom it has “insufficient and irrelevant demand to warrant the development of an internal capacity to do that work” (Harrison & Kelley, 1993, pp. 216–217). Functional flexibility is often mentioned by managers as a driver for nonstandard work arrangements. For example, 54% of the respondents in a study reported that the desire to access specialized resources was an important reason for outsourcing in the U.S. metal working and machinery sector (Harrison & Kelley, 1993). The most commonly cited managerial reason for using temporary help agencies and contract companies is that they provide special skills that the firms’

technology (IT) capacity (41%). Also, firms differ in terms of the degrees to which the use of nonstandard workers is planned in advance. For instance, Stanworth and Druker (2006) found that 5 out of 12 United Kingdom organizations that they studied used agency labor as reactive measures to fulfill shortterm external needs, whereas other organizations did so with more long-term planning. Regardless of whether the use of nonstandard workers is planned in advance, our review of the research on why firms use nonstandard workers suggests that the two most commonly cited reasons are cost reduction and flexibility enhancement. We discuss each of these reasons in the sections that follow and review the brief literature that evaluates the extent to which firms have achieved the desired outcomes.

Nonstandard Work to Enhance Organizational Flexibility

576

Copyright American Psychological Association. Not for further distribution.

Nonstandard Workers

regular employees lack (Kalleberg et al., 2003). However, Kalleberg et al. (2003) found no empirical support for this claim. Their study showed that the perceived difficulty of hiring workers with necessary skills is negatively correlated with the use of parttime, on-call, and direct-hire temporary workers as well as on-site intermediaries among U.S. establishments. Kalleberg et al. explained that the failure to find statistical support for the functional flexibility argument in their study was due to the inability of the establishment-level measure of hiring difficulty to be specific in terms of occupations and jobs and to capture relevant intraorganizational variation. They suggested that “a more targeted shortage (perceived hiring difficulty) measure might yield findings more supportive” of the functional flexibility argument (Kalleberg et al., 2003, p. 545).

Nonstandard Work and Cost Reduction Conventional wisdom suggests that because nonstandard workers receive lower pay and benefits relative to standard workers, organizations should benefit financially from the use of nonstandard work arrangements. Researchers also have argued that hiring workers from temporary help agencies lowers recruitment and screening costs (von Hippel et al., 1997) and training costs (Krueger, 1993). In cases where the temporary help agencies have established long-term relations with clients, the agencies take over some human resource functions for the clients, including supervisory and performance monitoring for the temporary worker (Peck &Theodore, 1998), and help to simplify the clients’ administrative work by becoming “an extension of the client firm’s human resource department” (Kalleberg, 2000, pp. 347–348). Kalleberg et al. (2003) found indirect evidence supporting this cost reduction perspective. They found that nonstandard work arrangements are common in nonprofit organizations in the United States and speculated that because those organizations are likely to be under greater pressure to economize than profit-oriented organizations, they attempt to save cost by hiring nonstandard workers. Kalleberg et al. (2003) also found that firms with better benefits for standard workers are more likely to use nonstandard workers because the high cost of benefits for standard workers forces the firms to lower

total employment costs by using nonstandard workers. The cost reduction reasoning for the use of nonstandard workers is likely to be provided by firms that have a cost–leadership strategy. Consistent with the contingency perspective of strategic human resource management (Arthur, 1992; Sherer & Leblebici, 2001; Youndt, Snell, Dean, & Lepak, 1996), which proposes that firms with a low-cost strategy are more likely to be matched with lowcommitment human resource management practice, Gramm and Schnell (2001) found that firms with a low-cost production strategy (vs. a market specialization strategy) were more likely to use flexible staffing arrangements, including using independent contractors, freelancers, temporary agency workers, and outsourcing (b = 1.40, p < .01). Firms that compete on the basis of low cost are also 1.34 times more likely than firms with other strategies to use temporary help agencies, which provide labor with lower cost than other types of employment intermediaries (Nesheim, Olsen, & Kalleberg, 2007).

Evaluating Flexibility and Costs in Firms That Have Nonstandard Work Arrangements Despite flexibility enhancement and cost reduction being cited as important reasons why organizations use nonstandard work arrangements, there are few empirical examinations of the relationship among this employment practice, flexibility, costs, and firm performance. One study, done by Lepak, Takeuchi, and Snell (2003), found that reliance on contract work arrangements led to better firm financial performance (b = 0.27, p < .01), providing indirect evidence of the flexibility enhancement argument. Another empirical study, done by Nollen and Axel (1996), found that when costs relating to other factors like turnover and productivity levels are taken into account, the use of nonstandard workers was not cost effective in two out of three companies they studied. More recently, a Deloitte Consulting (2005) survey of 25 large international corporations found that one fourth of the companies had brought business functions back in-house after realizing that they could do the work themselves more successfully and at 577

Copyright American Psychological Association. Not for further distribution.

George and Ng

lower cost. InformationWeek (“In depth: When outsourcing goes bad,” 2006b) surveyed 420 business technology professionals in 2006 and found that 45% of them attributed failure in outsourcing to poor service and lack of flexibility, and 39% attributed it to hidden cost. Given these preliminary pieces of evidence that nonstandard work arrangements might not be giving organizations the cost and flexibility advantages they have sought, we ask two questions. First, why has the evaluation of the effectiveness of nonstandard work arrangements for organizations not received more rigorous attention? We speculate that this gap might exist on account of the difficulty in measuring the direct costs and benefits associated with nonstandard work (Nollen & Axel, 1998). Also, the costs of nonstandard work may not be measurable in the short term. For instance, if the firm loses corporate knowledge through the constant turnover of nonstandard workers, this might not be detected until well after the nonstandard workers have left. For this question to be investigated in a meaningful manner, we need to have better ways of conceptualizing and measuring the direct and indirect costs and benefits of nonstandard work. Second, why might organizations fail in the implementation of nonstandard work arrangements? In the following section, we address this question with a specific focus on the strategic importance of work that is conducted by nonstandard workers and the role of technology in making nonstandard work arrangements feasible and beneficial to firms.

Strategic Value of Nonstandard Work and Workers Theorists have suggested that firms should not use nonstandard workers for tasks that are core to the organization and that are critical to its competitive advantage (Matusik & Hill, 1998; Shi, 2007). This is because core tasks involve knowledge that is of high value to the organization and also may be specific to that organization. It is to the organization’s advantage to make sure that this kind of knowledge does not reside in a workforce that is relatively transitory and might knowingly or inadvertently pass this knowledge on to competitors. Empirical work has examined the use of nonstandard work arrangements for core organizational 578

tasks (Davis-Blake & Uzzi, 1993; Lepak & Snell, 2002; Masters & Miles, 2002; Mayer & Nickerson, 2005), as well as the relationship between the use of valuable and proprietary technology, nonstandard work, and firm performance (Mayer & Nickerson, 2005). Lepak and Snell (2002) found that the skills of workers in contract work (M = 2.89, SD = .49) and alliance-based (M = 3.33, SD = .54) employment modes are less valuable and are less specific to firms than are those of standard workers (M = 4.02, SD = .53, F = 42.39, p < .001). On-the-job training, as an indicator of the firm-specific skills required by job positions, is found to be negatively associated with nonstandard work arrangements (DavisBlake & Uzzi, 1993; Masters & Miles, 2002). Finally, using data from a project-based high-technology service firm, Mayer and Nickerson (2005) showed that the more proprietary the technology required for a project, the less likely the project is to be externalized (model R2 = .25, p < .10). Furthermore, projects that involved more proprietary technologies were found to perform financially better when insourced than when outsourced, perhaps because of the additional costs from training independent contractors to use the firm’s technology and from safeguarding the technology against expropriation. In sum, nonstandard work arrangements are more suitable for filling jobs with low levels of firmspecific skills. As this list of examples suggests, many studies recommend a cautious stance on the use of nonstandard workers in areas that provide competitive advantage to firms. This conclusion, however, may not be generalizable to firms in dynamic environments (Matusik & Hill, 1998). In an environment characterized by extreme competition and rapid change, nonstandard workers can bring updated industry- or occupation-relevant knowledge and skills into the firm to prevent the firm’s stock of private knowledge from becoming obsolete relative to their competitors. Supporting this perspective, Nesheim et al. (2007) found that organizations that need continuous innovation of ideas use contingent workers from consulting firms in their core activities (b = 0.24, SE = .10, p < .05), with a belief that those workers bring special competencies that improve their competitive advantages.

Nonstandard Workers

Copyright American Psychological Association. Not for further distribution.

Role of Technology in Enabling Nonstandard Work Advances in production and IT have generally facilitated the adoption of nonstandard work arrangements. New production technologies help to standardize production procedures and simplify the skills needed to perform the job (Appelbaum, 1989; Coates, 1988; Hunter, McGregor, Mclnnes, & Sproull, 1993; Osterman, 1987; Smith, 1994; Tilly, 1991), thus lowering the skill requirements for nonstandard workers and reducing the time needed for them to learn their jobs. Supporting this line of argument, empirical results show that the more widely computer technology is used in organizations, the higher the use of fixed-term contractors and part-time workers (Uzzi & Barsness, 1998; b = 0.15, SE = .03, p < .001). Advancement in IT also facilitates telework and working from home by improving communication (Huws, Kork, & Robinson, 1990; Townsend & Bennett, 2003) and overcoming geographical distance so that many kinds of work can be done from almost anywhere (Carey & Hazelbaker, 1986; Davenport & Pearlson, 1998; Kalleberg, 2000; Shamir, 1992). Improved ITs also reduce costs of coordination with external partners, making firms more willing to move activities outside the firms’ boundaries by outsourcing (Brews & Tucci, 2004; Hitt & Brynjolfsson, 1997). Although technological advancement enables the use of more nonstandard work arrangements in general, the speed of technological change might impose a boundary condition on these positive outcomes. For example, Li, Lam, Sun, and Liu (2008) examined multinational enterprises in Chinese electronics and garment industries and found that organizations in the electronics industry are less likely to adopt an external employment mode than are those in the garment industry. They argued that rapid technological changes in the electronics industry make it difficult to obtain qualified human resources from outside. Sahaym, Steensma, and Schilling (2007) also found that the relationship between organizational investment in IT and the hiring of agency workers is dependent on the broader industrial context, such that IT investment facilitates use agency workers to the greatest extent when the level of technological change is low. They found there was a 52% increase

in the use of contingent workers when the level of technological change was high, in contrast with a 63% increase when the level of change was low. EMPLOYEES’ RATIONALE FOR ENGAGING IN NONSTANDARD WORK Just as organizations have multiple motivations for the use of nonstandard workers, individuals have varying reasons for being in nonstandard work arrangements. One stream of research that has attempted to capture these motivations has distinguished between whether nonstandard workers are pushed into this work arrangement involuntarily in response to constraints (e.g., financial or time constrains) or are pulled toward nonstandard work arrangements voluntarily by opportunities (e.g., balance family and work life, gain task variety and autonomy). This framework has been referred to as the push–pull framework (Cohen & Bianchi, 1999; Wiens-Tuers & Hill, 2002) or the voluntary– involuntary framework (Ellingson, Gruys, & Sackett, 1998). Empirical studies suggest that this conceptualization of nonstandard workers as voluntary or involuntary (e.g., Feldman, Doerpinghaus, & Turnley, 1995; Krausz, Brandwein, & Fox, 1995; Nardone, 1995) is inaccurate because an individual’s decision to be a nonstandard worker can be based on both push and pull factors simultaneously. For example, although contractors in Kunda et al.’s (2002) study valued the opportunity to escape from organizational politics, their entry into the contractor workforce was also triggered by being laid off and perceived future loss of jobs. Similarly, Mallon and Duberley (2000) observed that out of the 25 interviewees in their study, only three participants had been made compulsorily redundant and, conversely, only three described a wholly positive, career development decision to opt for working independently. Hence any simplistic push/pull dichotomy to explain the move to contingent work should be resisted. (p. 38) Ellingson et al. (1998) also found empirical support for the idea that voluntariness and involuntariness 579

Copyright American Psychological Association. Not for further distribution.

George and Ng

are two separate dimensions of nonstandard workforce entry. Perhaps a more parsimonious way to classify the reasons for being in a nonstandard work arrangement is provided by the U.S. Bureau of Labor Statistics (2005; Polivka, 1996), which distinguishes between economic reasons (e.g., “could only find this type of job,” “this job might lead to permanent one”) and personal reasons (e.g., “flexibility of work schedule,” “child care problems,” “other family or personal obligations,” “in school or training”). These economic and personal reasons can be both voluntary and involuntary, involving a combination of push and pull factors. We discuss these reasons in greater detail in the sections that follow.

Economic Motivators of Nonstandard Workers Two major economic reasons for engaging in nonstandard work arrangements are workers’ desire to escape from financial hardship (Wiens-Tuers & Hill, 2002) and their wish to transition to a permanent job (Morris & Vekker, 2001). Amuedo-Dorantes and Bansak (2003) found that being on welfare increases the likelihood of individuals being temporary workers. Nonstandard work, however, does not provide the poor an effective means to escape from poverty, primarily because of the lower wages and benefits received by nonstandard workers compared with standard workers. Research in the United States (Kalleberg, Reskin, & Hudson, 2000) and Great Britain (McGovern, Smeaton, & Hill, 2004) indicates that nonstandard workers are more likely to be exposed to bad job characteristics, including low pay, no access to health insurance, and lack of pension benefits, even after controlling for workers’ personal characteristics, family status, occupation, and industry. For example, Kalleberg et al. (2000) found that jobs held by female temporary help agency workers had 103% more bad job characteristics compared with standard workers. The situation is worse for workers in semi- and unskilled occupational classes and for those at a lower educational level (McGovern et al., 2004). In fact, scholars have documented that wage disparity between standard and nonstandard workers holds true in the United States (Belman & Golden, 2000; Houseman & 580

Osawa, 2003a; McGrath & Keister, 2008), Japan (Nagase, 2003), Western Germany, and Britain (Gustafsson, Kenjoh, & Wetzels, 2003), even after controlling for personal and job characteristics. Besides the desire to escape from poverty, some nonstandard workers enter this work arrangement because they are unable to find a permanent job and they hope nonstandard work might provide a transition to a permanent job. For example, Morris and Vekker (2001) reported that two thirds of temporary workers surveyed in the Current Population Survey in the United States preferred a permanent position but were not able to find permanent jobs that they considered superior to their current temporary position. Figures from Spain reveal a similar pattern: Amuedo-Dorantes (2000) reported that 85% of their sample of Spanish temporary workers said they were involuntarily temporary because of their inability to find a permanent job. Given nonstandard workers’ (especially temporary workers) desire to transition to standard jobs, it is worth examining whether nonstandard jobs are effective means with regard to this objective. The empirical evidence is mixed, as the reported transition rates vary. Lenz (1996) reported that in a 1995 survey of former temporary workers in United States, the National Association of Temporary and Staffing Services found that 72% of former temporary workers got permanent jobs and 56% learned new skills from their temporary positions, suggesting that temporary employment enhances the prospect of finding a permanent job; these findings should be viewed with caution because the baseline excluded workers who remained temporaries. Other studies found lower temporary-to-permanent transition rates: 57% using Current Population Survey data in the United States (Segal & Sullivan, 1997), 48% using Labor Force Survey data in Britain (Forde & Slater, 2006), and 12% using Labor Force Survey data in Spain (Amuedo-Dorantes, 2000). On the basis of these relatively dismal findings, AmuedoDorantes (2000) concluded that “temporary work is more likely to become a trap than a bridge to permanent employment” (p. 324). In a later study using quasi-experimental matching between temporary help agency employees and direct-hire temps in Spain, Amuedo-Dorantes, Malo, and Muñoz-Bullón

Nonstandard Workers

(2008) found that the transition rate was even worse for agency temporary workers than for their directhire counterparts.

Copyright American Psychological Association. Not for further distribution.

Personal Motivators of Nonstandard Workers Almost all types of nonstandard workers appreciate the flexibility of their work arrangement in terms of when and where they work. This flexibility allows them to balance work with other demands on their time. For instance, Dale and Bramford (1988) and Feldman, Doerpinghaus, and Turnley (1994) found that college students had nonstandard jobs during time off from classes to obtain additional sources of finance. Cohany (1998) reported that in 1997, the proportion of U.S. female on-call workers and temporary help agency workers who preferred their current work arrangement was higher than that of their male counterparts. This, she argued, could be a reflection of the fact that women generally have primary responsibility for the care of their families and consequently prefer nonstandard work to accommodate their household demands (Carr, 1996; Feldman et al., 1994; Hakim, 2002; Houseman, 1995). Casey and Alach (2004) also found that female agency temporary workers in New Zealand valued the flexibility provided by their work arrangement because they could combine participation in the workforce with family or other personal obligations. Female at-home workers are also found to mention family responsibilities as a key factor in the decision to work at home (Ammons & Markham, 2004; Jurik, 1998). Individuals also choose nonstandard work arrangements on account of the autonomy and control they experience (Gerson & Kraut, 1988; Kunda et al., 2002). The notion of a boundaryless or portfolio career (Marler, Barringer, & Milkovich, 2002) emphasizes the independence that workers can have in managing their own careers through a series of short-term work assignments (Handy, 1994; Knell, 2000). Benefits associated with a broadened horizon are also seen in Kirkpatrick and Hoque’s (2006) study of professional United Kingdom agency social workers. Respondents in their study used agency work as a means to explore different social work functions before committing to a permanent position in one function and to build up a stock of skills and experi-

ences by moving between assignments, thus enhancing upward mobility (Kirkpatrick & Hoque, 2006). An opportunity to escape from organizational politics (Kunda et al., 2002) and the boredom associated with routine clerical jobs (Casey & Alach, 2004) are other benefits associated with nonstandard work. This is particularly true for highly skilled nonstandard workers, such as professionals and managers (Ammons & Markham, 2004; Barley & Kunda, 2004; Kunda et al., 2002; Mallon & Duberley, 2000; Rogers, 2000), and may account for their high preference for this employment arrangement over the standard one. All the benefits of nonstandard work are perceived to be more accessible by highly skilled workers and professionals because they are more aware of the market value of their scarce skills and the alternative jobs available in the nonstandard work market (Albert & Bradley, 1997). Compared with other types of nonstandard workers, the Current Population Survey conducted from 1995 to 2005 consistently revealed that about 80% of independent contractors, who are usually older and have more schooling than the average worker (Cohany, 1998), preferred their current nonstandard arrangement to a standard one (vs. 30–40% of contingent workers, oncall workers, and temporary help agency workers). To what degree have individuals been successful in achieving their personal goals related to nonstandard work arrangements? Empirical evidence is mixed. For example, although telecommuting provides greater flexibility in scheduling work so as to better integrate work and family–personal life (Hornung, Rousseau, & Glaser, 2008), family members and household work impose limits on telecommuters’ experience of freedom, especially for those working at home (Baines, 2002; Tietze & Musson, 2003). Although at-home workers experienced more decision-making autonomy and more freedom from supervision (Ross & Wright, 1998), they also experienced difficulty in managing the boundaries between home and work (Ammons & Markham, 2004). Contractors’ use of time is also found to be constrained by the cyclic structure of employment, in which they need to work long unbillable hours to maintain their marketability (Evans, Kunda, & Barley, 2004). It is also unclear whether nonstandard workers are successful in escaping from organizational 581

Copyright American Psychological Association. Not for further distribution.

George and Ng

constraints and boredom. Although the greater task autonomy, variety of jobs, and higher skill utilization of self-employed jobs all contribute to the higher job satisfaction of the self-employed, this satisfaction advantage is found to be relatively small or nonexistent among managers and members of established professions because organizational workers in these occupations enjoy relative higher autonomy and skill utilization (Hundley, 2001). Contractors might also be given tasks of less variety, autonomy, and significance (Ang & Slaughter, 2001). Moreover, highly skilled contractors ironically spend a considerable amount of time building and maintaining social capital to generate job referrals (Barley & Kunda, 2004; O’Mahony & Bechky, 2006; Osnowitz, 2006), despite their hope to be free from politics. NONSTANDARD WORKERS IN ORGANIZATIONS

Focus on Nonstandard Workers’ Dispositions Are there relatively stable dispositional differences between standard and nonstandard workers? There is a small body of research that has considered the relationship between personality and nonstandard work arrangements. Virtanen et al. (2005) found evidence in two longitudinal studies of a relationship between individual characteristics at Time 1 and the likelihood that the individual was in temporary work at Time 2. Virtanen et al. argued that when there is competition for employment, individuals who are healthy and have good social skills are more likely to get permanent jobs. Thus, they would expect that chronically ill people or those who display traits of aggressiveness or anxiety are more likely to find themselves in temporary jobs. Their results provide general support for this line of reasoning. Specifically, temporary workers who reported that they had a psychiatric disorder at the start of the study had an odds ratio of 2.4 of staying in temporary employment 2 years later. In a second study, Virtanen et al. reported that children with a high level of anxiety at age 8 (1 standard deviation above the mean) were 2.8 times more likely to be temporary workers at age 42 than were children with low anxiety levels at age 8 (1 standard deviation below 582

the mean). It is interesting to note that aggressiveness at the age of 8 did not predict individuals’ work arrangements at age 42. Similarly, Bauer and Truxillo (2000) found that the likelihood that temporary workers would be made permanent after a year is predicted by individual differences such as selfmonitoring, tolerance for ambiguity, and self-efficacy (.20 of the variance was explained by these variables). The research linking dispositions and nonstandard work arrangements is relatively sparse, perhaps on account of the fact that individuals often move between nonstandard and standard arrangements.

Focus on Nonstandard Workers’ Attitudes Much of the research on nonstandard workers has focused on a comparison of the attitudes and behaviors of nonstandard and standard workers. The key argument made by most of these researchers is that nonstandard workers, on account of their transitory or remote relationship with organizations, would be less committed to the organization (Van Dyne & Ang, 1998), identify less strongly with the organization (Wiesenfeld, Raghuram, & Garud, 1999), be less satisfied with various job aspects (R. Hall, 2006; Miller & Terborg, 1979), exhibit fewer citizenship behaviors, or generally perform worse (Ang & Slaughter, 2001; Belous, 1989) than standard workers. The research provides mixed support for these arguments. For instance, some studies demonstrate that nonstandard workers had lower organizational commitment to their client firms than the standard workers in those firms (e.g., Van Dyne & Ang, 1998; b = 0.24, p < .001), and others show the opposite results (e.g., McDonald & Makin, 2000; t = −4.09, p < .001). Still, other studies have found comparable levels of commitment between standard and nonstandard workers (e.g., Pearce, 1993). Similarly, some studies have found higher job satisfaction among standard workers than among nonstandard workers (e.g., Forde & Slater, 2006; D. T. Hall & Gordon, 1973), but others have revealed the opposite pattern (e.g., De Cuyper & De Witte, 2007; Eberhardt & Shani, 1984; Galup, Saunders, Nelson, & Cerveny, 1997; Katz, 1993; McDonald & Makin, 2000; Parker et al., 2002). In addition, some studies do not find significant differences (e.g., Broschak, Davis-Blake, & Block, 2008).

Copyright American Psychological Association. Not for further distribution.

Nonstandard Workers

Researchers have attempted to reconcile these inconsistent findings by identifying a variety of conditions under which nonstandard workers’ attitudes vary. One argument is that these seemingly contradictory results may arise from the heterogeneity in characteristics of nonstandard workers. For instance, Marler et al. (2002) made a distinction between boundaryless and traditional contingent employees and found that the boundaryless cluster, characterized by higher preference for being temporary and higher levels of skills and earnings, has higher satisfaction with work and pay. In addition, workers who pursue temporary work involuntarily are less satisfied than those who are voluntarily temporary (Ellingson et al., 1998; model R2 = .34, p < .05; Feldman et al., 1995; Krausz, 2000). Using a similar argument, Broschak et al. (2008) found within the financial service firm that they studied that agency temporaries had a lower level of continuance commitment than standard workers (b = −2.27, SE = .83, p < .05), whereas retention part-time workers were no different from standard workers on their level of continuance commitment. Another potential moderator is suggested by Holtom, Lee, and Tidd (2002), who found that the higher the work status congruence, the higher the affective commitment to the organization, with work status congruence defined as the degree to which employers match employee preferences for full-time or part-time status, schedule, shift, and number of hours. De Cuyper and De Witte (2006) provided a third explanation, suggesting that we should pay attention not only to temporary workers but also to the work conditions of the standard workers with whom we are comparing them. They found that temporary workers are more committed to the organization in general than standard workers but that this is more the case when standard workers face high job insecurity (b = −0.17, p < .01). Task characteristics may also play a role in accounting for the mixed results, such that nonstandard workers handling tasks with greater autonomy and variety are likely to have higher job satisfaction (Hundley, 2001), whereas those with less task autonomy and variety are less satisfied (Ang & Slaughter, 2001). In addition to studies that compare the attitudes of standard and nonstandard workers, others have

been interested in the attitudes of nonstandard workers toward work and organizations and to this form of employment (Torka, 2004). Gallagher and McLean Parks (2001) observed that “the growth of ‘contingent’ or ‘alternative’ forms of work relationships highlights the need for researchers to examine work-related commitments outside of the traditional employer–employee framework” (p. 204). In response to this call, in their study including full-time, temporary, and self-employed workers as respondents, Felfe, Schmook, Schyns, and Six (2008) found that affective commitment to forms of employment explains variance in job satisfaction (model adjusted R2 = .37, p < .001) and organizational citizenship behaviors (model adjusted R2 = .18, p < .001), even after controlling for affective, continuance, and normative commitment to the organization. Scholars have also examined the effect of perceived organizational support on commitment of nonstandard workers (Connelly, Gallagher, & Gilley, 2007; Coyle-Shapiro & Morrow, 2006; CoyleShapiro, Morrow, & Kessler, 2006; Liden, Wayne, Kraimer, & Sparrowe, 2003; van Breugel et al., 2005) and have found a similar pattern of results as those based on standard workers, that is, individuals follow the norm of reciprocity such that higher perceived organizational support leads to higher organizational commitment (Rhoades & Eisenberger, 2002). The issue of organizational identification (Ashforth & Mael, 1989; Dutton, Dukerich, & Harquail, 1994) in the context of nonstandard work is getting more attention recently. Researchers suggest that nonstandard workers (McLean Parks et al., 1998) and teleworkers (Thatcher & Zhu, 2006; Wiesenfeld et al., 1999) are likely to identify less with their employing organizations because of the limited interactions with those organizations. George and Chattopadhyay (2005), however, found that contract workers did identify with both the employing (model adjusted R2 = .62, p < .01) and client (model adjusted R2 = .53, p < .01) organizations. Identification with the client was driven by personal interactions with supervisors (b = 0.21, p < .01) and colleagues (b = 0.28, p < .001), whereas identification with the principal employer was also predicted by abstract characteristics of that organization, such as its prestige and distinctiveness (b = 0.24, p < .01). 583

George and Ng

Johnson and Ashforth (2008) found that temporary workers’ identification with employing organizations and with their client mediated the effect of employment status on customer-oriented service behavior (model R2 = .11, p < .01).

Copyright American Psychological Association. Not for further distribution.

Focus on Nonstandard Workers’ Behaviors One set of researchers has compared organizational citizenship behavior and extra-role behavior among standard and nonstandard workers (e.g., Ang & Slaughter, 2001; Coyle-Shapiro et al., 2006; Feldman & Turnley, 2004; Holtom et al., 2002; Liden et al., 2003), and their studies have contradictory findings. For example, Pearce (1993) reported higher levels of extra-role behavior for contractors in one organization in the aerospace industry than for standard workers. However, van Dyne and Ang (1998) found that contingent workers in a bank and a hospital in Singapore engaged in fewer citizenship behaviors than their standard counterparts (b = 0.39, p < .01). Feather and Rauter (2004) argued that participants in van Dyne and Ang’s study experienced less pressure to perform these behaviors because of the severe labor shortage at the time the study data were collected. Consistent with their argument that job insecurity motivates citizenship behaviors, Feather and Rauter demonstrated that contract teachers in Australia who perceived high job insecurity reported more citizenship behaviors (b = 0.61, p < .01) than permanent teachers. A second point of difference between standard and nonstandard workers that has garnered significant research is related to worker safety. There is growing evidence indicating that the use of nonstandard workers is associated with deterioration in worker safety. Quinlan, Mayhew, and Bohle (2001) reviewed studies published in 12 countries plus the European Union from 1984 to 2000 and found that nonstandard workers, compared with standard workers, were associated with worse occupational health and safety outcomes, such as higher injury rates. This is particularly problematic in hazardous high-risk industries, such as the mining industry and the petrochemical industry (Kochan, Smith, Wells, & Rebitzer, 1994; Rebitzer, 1995; Rousseau & Libuser, 1997). Why are nonstandard workers associated with deterioration in occupational health and safety? 584

Quinlan and Bohle (2004) summarized factors that lead to higher injury rates among nonstandard workers into three board categories: economic pressures and reward systems, disorganization, and regulatory failure. Evidence indicates that nonstandard workers who experience economic pressure from job insecurity, or who work under an incentive pay system, often tend to work overload and overtime and to continue working when injured, leading to more injuries (Mayhew & Quinlan, 1997). Disorganization at work, such as lack of training, lack of experience, and inadequate communication and supervision for nonstandard workers, also increases risks associated with workplace safety (Rebitzer, 1995). Last, Quinlan and Bohle suggested that safety laws and compliance programs in most countries place little emphasis on safeguarding nonstandard workers, although regulators are now attempting to change this. Another behavior of nonstandard workers that has received some interest is the extent to which they share knowledge in organizations. Most theoretical treatments of this issue are based on the assumption that nonstandard workers have important information or expertise because they usually move between numerous organizational settings and consequently are likely to have higher knowledge levels in industry- and occupation-based best practices than are their standard counterparts (Matusik & Hill, 1998). There are, however, few studies that actually investigate information-seeking and information-giving behaviors of nonstandard workers. A notable exception is Sias, Kramer, and Jenkins’s (1997) study, which found that new temporary workers provided information to others less frequently than newly hired regular employees. They speculated that it was because the transitory nature of temporary work limits the incentive for temporary workers to build relationships with standard workers by sharing information with them. Finally, research on the performance of nonstandard workers has focused on manager-rated performance and peer assessment of nonstandard workers’ in-role behavior. The results of this line of research do not allow an unequivocal conclusion. On the one hand, Ang and Slaughter (2001) reported that contractors displayed lower in-role behavior as rated by

Copyright American Psychological Association. Not for further distribution.

Nonstandard Workers

their peers (M = 5.21 SD = .19) and lower performance as rated by their supervisors (M = 4.75, SD = .32) than their permanent colleagues (M = 5.81, SD = .11, and M = 5.74, SD = .80, respectively). Researchers account for the lower productivity of temporary workers by suggesting that temporary workers are new to the job and thus need to learn work processes (Kalleberg, 2000; Nollen & Axel, 1998). On the other hand, Broschak et al. (2008) found that both retention part-time and agency temporary employees received higher performance appraisal scores than their standard counterparts (b = 1.46, SE = .64, p < .05, and b = 1.11, SE = .18, p < .01, respectively). However, other studies found no significant differences between standard and nonstandard workers with respect to performance (De Cuyper & De Witte, 2006; Ellingson et al., 1998; Feldman & Turnley, 2004; Jarmon, Paulson, & Rebne, 1998). The divergent results may be due to managers’ expectation of nonstandard workers. Ho, Ang, and Straub (2003) found that the higher the supervisors’ expectations of contract workers, the higher the subsequent performance ratings that these managers gave the contractors. This result suggests the possibility that performance of nonstandard workers is affected by the self-fulfilling prophecy. As reviewed within this section, the research findings regarding nonstandard workers’ attitudes and behaviors in comparison with those of the standard workers appear to be inconsistent. Part of the inconsistency can be attributed to methodological issues, such as the different types of nonstandard worker studied (e.g., temporary workers or retention part-time workers) and the differences in other occupational and organizational contexts within which the nonstandard worker operates across the empirical settings. We believe that these factors might function as moderators or boundary conditions for the effects of being nonstandard, and we discuss how they need to be taken into account later in our section on future research directions.

Blending Standard and Nonstandard Workers Compared with the long-standing interest in the attitudes and behaviors of nonstandard workers, researchers have more recently turned their atten-

tion to the potential impact of using nonstandard work on the job content, attitudes, and behaviors of standard workers. The use of nonstandard work is generally found to be associated with negative consequences for standard workers in terms of their workload and mobility, as well as psychological outcomes, such as perceptions of job insecurity. Nonstandard workers can impose additional demands on standard workers’ job responsibilities without increases in their rewards. Temporary workers who are usually hired on short notice and with minimal screening place additional supervisory demands on standard workers (Geary, 1992). Similarly, Pearce (1993) and Smith (1994) reported that managers of the blended workforce assigned the most complex tasks to standard employees and held them responsible for errors made by their nonstandard peers. Nonstandard workers may also impact standard workers’ opportunities for upward mobility in the organization. For example, Barnett and Miner’s (1992) study of a Fortune 500 company suggested that the use of temporary workers had an adverse impact on the mobility of standard workers in lower level jobs by increasing the pool of applicants for these jobs (b = 0.22, SE = .05, p < .05) but had the opposite effect on mobility for those at the higher end of the organizational hierarchy because of the reduced presence of standard workers who are eligible to compete for advancement opportunities (b = −0.31, SE = .08, p < .05). The presence of nonstandard workers may also negatively affect standard employees’ relationship with organizations (Geary, 1992; Pearce, 1993; Smith, 1997). For instance, studies have found that standard workers with nonstandard colleagues report lower trust in the organization (Pearce, 1993), feeling that their psychological contract with the organization has been violated (George, 2003), and poorer relations with the management and their coworkers (Davis-Blake et al., 2003). There are a number of explanations for these findings. One is that organizations are seen to be exploitative of nonstandard workers and thus their use of this type of worker decreases the trustworthiness of the organization (Pearce, 1993). Another is that the use of nonstandard workers signals an organization’s decreasing commitment to its standard workforce, 585

Copyright American Psychological Association. Not for further distribution.

George and Ng

leading to psychological contract violation (George, 2003; model adjusted R2 = .29, p < .01). This detrimental effect of psychological contract violation of standard workers is greater for those with low job security than for those with high job security (b = −0.24, p < .01). Kraimer, Wayne, Liden, and Sparrowe (2005) found support for their argument that standard employees who experience low levels of job security are more likely to attribute the firm’s use of temporary workers to cost cutting via internal structure changes that threaten their jobs (b = 0.12, SE = .04, p < .01). The resulting threat perception lowers the performance of standard workers (Kraimer et al., 2005). Despite the general consensus among earlier research that nonstandard work has deteriorative effects on standard workers, there is a growing body of research that demonstrates that the effect is contingent upon a number of factors, including type of nonstandard work arrangement, salary and supervisory responsibility of standard workers, and the amount and nature of contact between standard and nonstandard employees. Different types of nonstandard work arrangements might have different impacts on standard employees. Using a national sample of employees in the United States, Davis-Blake et al. (2003) found that blending standard and temporary workers who are administratively closer to an organization is more disruptive to the relationship between standard workers and managers (b = −1.20, SE = .43, p < .01) than blending standard and contract workers who are more administratively separate from an organization (b = −0.68, SE = .34, p < .01). Another piece of evidence is provided by Broschak and Davis-Blake (2006), who found that higher proportions of nonstandard workers were associated with less favorable attitudes toward supervisors and peers (model adjusted R2 = .33, p < .01), increased turnover intentions (model adjusted R2 = .15, p < .01), and decreased work-related helping behaviors (model adjusted R2 = .32, p < .01). Of more importance, they found that the deteriorative effect is stronger when firms use temporary workers rather than retention part-time workers. They argued that the results were due to greater conflict over mobility opportunities and allocation of tasks in the case of temporary workers than in the case of retention part-time workers. 586

Salary and supervisory responsibility of the standard workers can also moderate the effects of nonstandard worker use on outcomes of standard workers. George (2003) found support for her argument that because standard employees with supervisory responsibilities may feel that they have greater control over the potential threat presented by externalization, they are less affected by the presence of external workers in terms of trust in the organization (b = 0.12, p < .10), commitment (b = 0.17, p < .05), and perception of psychological contract violation (b = 0.20, p < .01). Davis-Blake et al. (2003) found similar results, and they further suggested that “formally delegating managerial responsibilities to nonmanagerial employees may be a necessary precondition for successful use of temporary workers” (p. 483). Besides, empirical results demonstrated that having a higher salary and being senior in a firm’s mobility system can also buffer standard workers from the detrimental effects of having nonstandard peers (Broschak & Davis-Blake, 2006; Davis-Blake et al., 2003). Finally, the effect of using nonstandard workers can be alleviated or aggravated, depending on the amount and nature of interaction between standard and nonstandard employees. The literature on demographic diversity suggests that non-taskrelated interactions facilitate the exchange of social information between diverse groups and reduce negative feelings created by conflicts between groups (Williams & O’Reilly, 1998). Consistent with this line of reasoning, Broschak and Davis-Blake (2006) found that after controlling for nonstandard worker heterogeneity, non-task-related interactions were positively related to relations with supervisors (b = 0.55, SE = .16, p < .01) and coworkers (b = 0.42, SE = .13, p < .01). In contrast, task-related interactions make salient the problems of blending standard and nonstandard workers, thus having a negative impact on supervisor–subordinate relations (Broschak & Davis-Blake, 2006; b = −0.42, SE = .17, p < .01). In summary, the research we have reviewed has drawn from a wide variety of theoretical positions to examine the antecedents and outcomes of nonstandard work arrangements. In a very general sense, there are two positions taken on understanding non-

Copyright American Psychological Association. Not for further distribution.

Nonstandard Workers

standard work. On the one hand, much of the research on when and why organizations use nonstandard workers is based on either an economic or technological rationale (e.g., Davis-Blake & Uzzi, 1993; Harrison & Kelley, 1993). On the other hand, research that has focused on either the attitudes or behaviors of nonstandard workers or on the consequence of using nonstandard workers typically relies on more sociological or psychological theories (e.g., Chattopadhyay & George, 2001; Davis-Blake et al., 2003). These differences in approach have presented both opportunities and challenges. The wide variety of theories used to study nonstandard work has resulted in research that crosses intellectual boundaries and, perhaps through its diversity, is best able to capture nuances of this phenomenon. The diversity of approaches, however, has also produced a scattered body of work that has not successfully built a cumulative understanding of nonstandard work and workers (Ashford et al., 2007). We develop these observations in our next section, which discusses future directions for research. THE WAY FORWARD: DIRECTIONS FOR FUTURE RESEARCH AND PRACTICE

Research Directions Like Ashford et al. (2007) and Connelly and Gallagher (2004), we call for more research on nonstandard work and workers. In addition to the research directions that those authors identified, we propose that research in this area needs to move forward to four directions. First, as our opening discussion on the definition of nonstandard work highlighted, this type of work is shaped by its context. As Kunda et al. (2002) pointed out, not all nonstandard work is undesirable and not all nonstandard workers are involuntarily in that situation. We need more research that can help us map out the nuances of the experience of nonstandard work and workers and can help resolve some of the contradictory findings from previous research. For instance, it would be useful to understand the occupational and organizational norms within which nonstandard workers operate. Hill et al. (2003), in their study of a technology company, found that contrary to expectations, virtual office workers and home office work-

ers were more likely than traditional office workers to view their opportunity for career advancement optimistically. They argued that one possible explanation is that flexible work arrangements have been used in technology companies so extensively and for so long that this way of working has been normalized. Other researchers could examine whether there are such differences in norms across settings. Professional norms of behavior that guide nonstandard workers’ behaviors could also be studied (e.g., Blatt, 2008). Another contextual condition that warrants attention is the role of the labor market—nonstandard work is likely to be less stigmatized or poorly paid when there are skill shortages (Fraser & Gold, 2001; Hunter et al., 1993). Researchers might also consider whether context affects measures such that nonstandard and standard workers’ responses on questionnaires reflect not just differences between these types of workers but also differences in the meaning of the same construct to the two types of workers. A second area of research deals with the frames of reference of nonstandard workers. As the researchers who examined blended workforces have noted, it is very instructive to know with whom standard and nonstandard workers compare themselves (e.g., Chattopadhyay & George, 2001). Although nonstandard workers might expect that they are to work with standard workers, what are their expectations of their working conditions relative to those of their standard coworkers? Also, we need to know whether this is true across all types of nonstandard workers. We might find that the contract workers studied by Kunda et al. (2002) or Pearce (1993) are very different from the temporary workers studied by Davis-Blake et al. (2003). This point also has important methodological implications. Our measures on employee attitudes are likely to be more accurate if we know who the respondents use as referents when they develop attitudes about their jobs or their work situations. Third, researchers often treat nonstandard workers as a homogenous and static group. However, we now know that nonstandard workers vary on many dimensions, and these differences can affect their relationships with organizations (Pfeffer & Baron, 1988) and their coworkers (Broschak & Davis-Blake, 2006; Davis-Blake et al., 2003). We need more 587

Copyright American Psychological Association. Not for further distribution.

George and Ng

empirical work to explore differences between these types of nonstandard workers. In addition, we need longitudinal studies that explore nonstandard careers or careers that cross between standard and nonstandard boundaries. Although the boundaryless career has received some theoretical attention (Marler et al., 2002), we did not find any longitudinal studies that explore the unfolding of this type of career. Although we recognize that these longitudinal studies could be difficult to undertake because individuals who move in and out of nonstandard work might be difficult to track, we nevertheless believe that more work is warranted on this issue. Finally, as we noted previously, research on nonstandard work needs to build cumulatively, and a meta-analysis might be warranted at this point. However, as Thorsteinson (2003) noted in a metaanalytic review of attitudes of part-time and fulltime workers, many of the primary studies do not include information necessary for examining more specific relationships in meta-analyses. To resolve conflicting findings from previous studies, we need additional primary studies that take into account different types of nonstandard work arrangements (e.g., part-time work, agency temp, contract work), varying occupational and organizational norms within which the nonstandard worker operates, and varying labor markets. Together, these studies might provide enough data for comprehensive meta-analyses.

Notes on Being a Nonstandard Worker As our earlier discussion of the literature on nonstandard work arrangements suggests, nonstandard workers experience problems in a number of areas, such as managing the home–work boundary or identity and dealing with job insecurity. What strategies can nonstandard workers adopt to overcome those challenges? First, working at home creates work–home conflicts based on demands for time and space (Baines, 2002). At-home workers have developed strategies for managing such conflicts by screening incoming calls (Ammons & Markham, 2004), scheduling work for times when distractions are less likely (Jurik, 1998; Tietze & Musson, 2003), telling children and friends to not disturb them when they need quiet (Ammons & Markham, 2004), ignoring some dis588

traction and temptations (Kinsman, 1987), and scheduling specific times for work and nonwork activities (Tietze & Musson, 2003). Second, nonstandard workers need to deal with their experience of being marginalized, stigmatized, and treated as outsiders or invisible (Barker, 1998) as well as being treated as “second-class citizen[s] by both employers and permanent workers” (Castro, 1993, p. 44). This stigmatization often causes nonstandard workers’ identity to be devalued (Boyce, Ryan, Imus, & Morgeson, 2007; Smith 1998). Nonstandard workers can respond to this identity threat by actively resisting the dehumanizing practices of organizations (Jordan, 2003), renarrating and redefining their identities in a positive light (Pink, 2001; Rogers, 1995; Tieize, 2005; Zuboff & Maxmin, 2002), and engaging in interactions beyond the content of work to build a sense of community that can counteract the alienation, depersonalization, and loneliness associated with temporary work (Blatt & Camden, 2006). Occupational communities can provide nonstandard workers a valued, nonorganizational source of identity (Kunda et al., 2002). Finally, to be successful, workers in nonstandard work arrangements need to accumulate transferable skills and experience to ensure their marketability and employability. Nonstandard employment is far from cost free to workers because they often lack training and development opportunities, which may cause the lowering of their skill base and, consequently, their marketability (Baines, 2002; Hoque & Kirkpatrick, 2003). The literature has many examples of how nonstandard workers deal with this issue: At-home workers have ordered journals or industry publications and attended workshops, seminars, or conferences (Ammons & Markham, 2004); and contactors have cultivated professional communities where technical knowledge and information on jobs reside (Kunda et al., 2002) and managed their careers by acquiring referrals from their professional networks, differentiating and framing their competence, and even accepting pay that is below market rate (O’Mahony & Bechky, 2006).

Managing Nonstandard Workers Ashford et al. (2007) argued that there are four major aspects to managing nonstandard workers:

Copyright American Psychological Association. Not for further distribution.

Nonstandard Workers

managing job design by determining the fit between tasks and work arrangements; managing the exchange between workers and organizations by determining what workers want from their jobs; managing relationships by determining how workers can be incorporated into the social fabric of the organization; and managing identification by determining how nonstandard workers define themselves in the work context. The managerial task thus starts with determining when it is most appropriate to use nonstandard workers or what types of tasks are amenable to this work arrangement. A key consideration here is the nature of control that the organization deems to be appropriate. Virtual work, for instance, would be more amenable to tasks where it is more appropriate to monitor output than behaviors, as virtual workers cannot easily be monitored in their day-to-day tasks. Next, managers of nonstandard workers need to know both why the organization needs the skills of workers as well as the workers’ motivation for nonstandard work. This could help them determine which workers could be offered nonstandard work arrangements (e.g., part-time work offered to experienced workers to retain them) or which inducements would work for different types of workers (e.g., contract work might be more appropriate for professionals with specialist skills rather than for workers with low human capital). Managers also need to have strategies to include nonstandard workers in the organization such that they are compliant with laws that necessitate the separate treatment of standard and nonstandard workers while simultaneously destigmatizing nonstandard work and workers. Finally, managers need to understand the extent to which work organizations and the form of the work contract are central to the identity of nonstandard workers. The extent to which individuals want to be part of the organization might determine managerial actions aimed at them. CONCLUSION Nonstandard work arrangements are ubiquitous in modern organizations and attract both research and managerial attention. In this chapter, we attempted to shed some light on this work arrangement and those who are employed in it by examining four key

questions: How is nonstandard work defined? Why do organizations and individuals choose this type of work arrangement? What do we know about the attitudes and behaviors of nonstandard workers and their standard colleagues? and What are the research and managerial implications of our current state of knowledge on this topic? Our review suggests that although research aimed at understanding the rationale and consequences of nonstandard work arrangements is growing, the outcomes from the research are mixed. We draw three conclusions from these informative but often mixed findings: Nonstandard work is not a homogenous construct, and we need to investigate each type of nonstandard work independently; the attitudes and behaviors of nonstandard workers are contingent on contextual factors that need to be considered; and far more work needs to be done to establish that organizations and individuals are gaining the benefits that purportedly result from nonstandard work arrangements. We need to continue building a cumulative body of research to understand and benefit from this increasingly popular form of employment.

References Abraham, K. G. (1988). Flexible staffing arrangements and employers’ short-term adjustment strategies. In R. Hart (Ed.), Employment, unemployment, and labor utilization (pp. 288–313). Boston: Unwin Hyman. Albert, S., & Bradley, K. (1997). Managing knowledge: Experts, agencies, and organizations. New York: Cambridge University Press. Ammons, S. K., & Markham, W. T. (2004). Working at home: Experiences of skilled white collar workers. Sociological Spectrum, 24, 191–238. Amuedo-Dorantes, C. (2000). Work transitions into and out of involuntary temporary employment in a segmented market: Evidence from Spain. Industrial & Labor Relations Review, 53, 309–325. Amuedo-Dorantes, C., & Bansak, C. (2003). The role of contingent work in the war against poverty. Chicago Policy Review, 7(1), 77–98. Amuedo-Dorantes, C., Malo, M., & Muñoz-Bullón, F. (2008). The role of temporary help agency employment on temp-to-perm transitions. Journal of Labor Research, 29, 138–161. Ang, S., & Slaughter, S. A. (2001). Work outcomes and job design for contract versus permanent information systems professionals on software development teams. MIS Quarterly, 25, 321–350. 589

George and Ng

Appelbaum, E. (1989). The growth in the US contingent labor force. In R. Drago & R. Perlman (Eds.), Microeconomic issues in labor economics: New approaches (pp. 62–82). Hemel Hempstead, England: Harvester. Arthur, J. B. (1992). The link between business strategy and industrial relations systems in American steel minimills. Industrial & Labor Relations Review, 45, 488–506. Ashford, S. J., George, E., & Blatt, R. (2007). Old assumptions, new work. The Academy of Management Annals, 1, 65–117.

Copyright American Psychological Association. Not for further distribution.

Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20–39. Atkinson, J. (1984). Manpower strategies for flexible organizations. Personnel Management, 16, 28–31. Bailey, D. E., & Kurland, N. B. (2002). A review of telework research: Findings, new directions, and lessons for the study of modern work. Journal of Organizational Behavior, 23, 383–400. Baines, S. (2002). New technologies and old ways of working in the home of the self-employed teleworker. New Technology, Work and Employment, 17(2), 89–101.

Ragins (Eds.), Exploring positive relationships at work: Building a theoretical and research foundation (pp. 243–264). Mahwah, NJ: Erlbaum. Boyce, A. S., Ryan, A. M., Imus, A. L., & Morgeson, F. P. (2007). “Temporary worker, permanent loser?” A model of the stigmatization of temporary workers. Journal of Management, 33, 5–29. Brews, P. J., & Tucci, C. L. (2004). Exploring the structural effects of internetworking. Strategic Management Journal, 25, 429–451. Broschak, J. P., & Davis-Blake, A. (2006). Mixing standard work and nonstandard deals: The consequences of heterogeneity in employment arrangements. Academy of Management Journal, 49, 371–393. Broschak, J. P., Davis-Blake, A., & Block, E. S. (2008). Nonstandard, not substandard: The relationship among work arrangements, work attitudes, and job performance. Work and Occupations, 35, 3–43. Campbell, I., & Burgess, J. (2001). Casual employment in Australia and temporary employment in Europe: Developing a cross-national comparison. Work, Employment & Society, 15, 171–184. Carey, M., & Hazelbaker, K. L. (1986). Employment growth in the temporary help industry. Monthly Labor Review, 109(4), 37–44.

Barker, K. (1998). Toiling for piece-rates and accumulating deficits: Contingent work in higher education. In K. Barker & K. Christensen (Eds.), Contingent work: American employment relations in transition (pp. 195–220). Ithaca, NY: ILR Press.

Carr, D. (1996). Two paths to self-employment? Women’s and men’s self-employment in the United States, 1980. Work and Occupations, 23, 26–53.

Barley, S. R., & Kunda, G. (2004). Gurus, hired guns, and warm bodies: Itinerant experts in a knowledge economy. Princeton, NJ: Princeton University Press.

Casey, C., & Alach, P. (2004). “Just a temp?” Women, temporary employment and lifestyle. Work, Employment & Society, 18, 459–480.

Barnett, W. P., & Miner, A. S. (1992). Standing on the shoulders of others: Career independence in job mobility. Administrative Science Quarterly, 37, 262–281.

Castro, J. (1993, March 29). Disposable worker. Time, pp. 43–47.

Bauer, T., & Truxillo, D. (2000). Temp-to-permanent employees: A longitudinal study of stress and selection success. Journal of Occupational Health Psychology, 5, 337–346. Belman, D., & Golden, L. (2000). Nonstandard and contingent employment: Contrasts by job type, industry, and occupation. In F. Carré, M. A. Ferber, L. Golden, & S. A. Herzenberg (Eds.), The nature and challenges of changing employment arrangements (pp. 167–212). Champaign, IL: Industrial Relations Research Association. Belous, R. S. (1989). How human resource systems adjust to the downshift toward contingent workers. Monthly Labor Review, 112(3), 7–12. Blatt, R. (2008). Organizational citizenship behavior of temporary knowledge employees. Organization Studies, 29, 849–866. Blatt, R., & Camden, C. T. (2006). Positive relationships and cultivating community. In J. E. Dutton & B. R. 590

Chattopadhyay, P., & George, E. (2001). Examining the effects of work externalization through the lens of social identity theory. Journal of Applied Psychology, 86, 781–788. Coates, V. T. (1988). Office automation technology and home-based work. In K. Christensen (Ed.), The new era of home based work: Directions and policies (pp. 114–125). Boulder, CO: Westview Press. Cohany, S. R. (1998). Workers in alternative employment arrangements: A second look. Monthly Labor Review, 121(11), 3. Cohen, P. N., & Bianchi, S. M. (1999). Marriage, children, and women’s employment: What do we know? Monthly Labor Review, 122(12), 22. Connelly, C. E., & Gallagher, D. G. (2004). Emerging trends in contingent work research. Journal of Management, 30, 959–983. Connelly, C. E., Gallagher, D. G., & Gilley, K. M. (2007). Organizational and client commitment among contracted employees: A replication and extension with

Nonstandard Workers

temporary workers. Journal of Vocational Behavior, 70, 326–335. Coyle-Shapiro, J. A. M., & Morrow, P. C. (2006). Organizational and client commitment among contracted employees. Journal of Vocational Behavior, 68, 416–431. Coyle-Shapiro, J. A. M., Morrow, P. C., & Kessler, I. (2006). Serving two organizations: Exploring the employment relationship of contracted employees. Human Resource Management, 45, 561–583.

Copyright American Psychological Association. Not for further distribution.

Dale, A., & Bramford, C. (1988). Temporary workers: Cause for concern or complacency? Work, Employment & Society, 2, 191–209.

toward specific organizational characteristics and overall job satisfaction. Academy of Management Journal, 27, 893–900. Ellingson, J. E., Gruys, M. L., & Sackett, P. R. (1998). Factors related to the satisfaction and performance of temporary employees. Journal of Applied Psychology, 83, 913–921. Evans, J. A., Kunda, G., & Barley, S. R. (2004). Beach time, bridge time, and billable hours: The temporal structure of technical contracting. Administrative Science Quarterly, 49, 1–38.

Davenport, T. H., & Pearlson, K. (1998). Two cheers for the virtual office. Sloan Management Review, 39(4), 51–65.

Feather, N. T., & Rauter, K. A. (2004). Organizational citizenship behaviors in relation to job status, job insecurity, organizational commitment and identification, job satisfaction and work values. Journal of Occupational and Organizational Psychology, 77, 81–95.

Davis-Blake, A., Broschak, J. P., & George, E. (2003). Happy together? How using nonstandard workers affects exit, voice, and loyalty among standard employees. Academy of Management Journal, 46, 475–485.

Feldman, D. C., Doerpinghaus, H. I., & Turnley, W. H. (1994). Managing temporary workers: A permanent HRM challenge. Organizational Dynamics, 23(2), 49–63.

Davis-Blake, A., & Hui, P. P. (2003). Contracting talent for knowledge-based competition. In S. E. Jackson, M. A. Hitt, & S. D. Angelo (Eds.), Managing knowledge for sustained competitive advantage: Designing strategies for effective human resource management (pp. 198–206). San Francisco: Jossey-Bass. Davis-Blake, A., & Uzzi, B. (1993). Determinants of employment externalization: A study of temporary workers and independent contractors. Administrative Science Quarterly, 38, 195–223. De Cuyper, N., de Jong, J., De Witte, H., Isaksson, K., Rigotti, T., & Schalk, R. (2008). Literature review of theory and research on the psychological impact of temporary employment: Towards a conceptual model. International Journal of Management Reviews, 10, 25–51. De Cuyper, N., & De Witte, H. (2006). The impact of job insecurity and contract type on attitudes, well-being and behavioral reports: A psychological contract perspective. Journal of Occupational and Organizational Psychology, 79, 395–409. De Cuyper, N., & De Witte, H. (2007). Job insecurity in temporary versus permanent workers: Associations with attitudes, well-being, and behavior. Work & Stress, 21, 65–84. Deloitte Consulting. (2005). Calling a change in the outsourcing market. Retrieved June 4, 2008, from http://www.deloitte.com/dtt/cda/doc/content/ us_outsourcing_callingachange.pdf Dutton, J. E., Dukerich, J. M., & Harquail, C. V. (1994). Organizational images and member identification. Administrative Science Quarterly, 39, 239–263. Eberhardt, B. J., & Shani, A. B. (1984). The effects of fulltime versus part-time employment status on attitudes

Feldman, D. C., Doerpinghaus, H. I., & Turnley, W. H. (1995). Employee reactions to temporary jobs. Journal of Managerial Issues, 7, 127–141. Feldman, D. C., & Turnley, W. H. (2004). Contingent employment in academic careers: Relative deprivation among adjunct faculty. Journal of Vocational Behavior, 64, 284–307. Felfe, J., Schmook, R., Schyns, B., & Six, B. (2008). Does the form of employment make a difference? Commitment of traditional, temporary, and selfemployed workers. Journal of Vocational Behavior, 72, 81–94. Fenton-O’Creevy, M. (1995). Moderators of differences in job satisfaction between full-time and part-time female employees: A research note. Human Resource Management Journal, 5, 75–81. Foote, D. A., & Folta, T. B. (2002). Temporary workers as real options. Human Resource Management Review, 12, 579–597. Forde, C., & Slater, G. (2006). The nature and experience of agency working in Britain: What are the challenges for human resource management? Personnel Review, 35, 141–157. Fraser, J., & Gold, M. (2001). “Portfolio workers:” Autonomy and control amongst freelance translators. Work, Employment & Society, 15, 679–697. Gallagher, D. G., & McLean Parks, J. (2001). I pledge thee my troth . . . contingently: Commitment and the contingent work relationship. Human Resource Management Review, 11, 181–208. Galup, S., Saunders, C., Nelson, R. E., & Cerveny, R. (1997). The use of temporary staff and managers in a local government environment. Communication Research, 24, 698–730. 591

Copyright American Psychological Association. Not for further distribution.

George and Ng

Geary, J. F. (1992). Employment flexibility and human resource management: The case of three American electronics plants. Work, Employment & Society, 6, 251–270.

Hitt, L. M., & Brynjolfsson, E. (1997). Information technology and internal firm organization: An exploratory analysis. Journal of Management Information Systems, 14(2), 81–101.

George, E. (2003). External solutions and internal problems: The effects of employment externalization on internal workers’ attitudes. Organization Science, 14, 386–402.

Ho, V. T., Ang, S., & Straub, D. (2003). When subordinates become IT contractors: Persistent managerial expectations in IT outsourcing. Information Systems Research, 14(1), 66–86.

George, E., & Chattopadhyay, P. (2005). One foot in each camp: The dual identification of contract workers. Administrative Science Quarterly, 50(1), 68–99.

Holtom, B. C., Lee, T. W., & Tidd, S. T. (2002). The relationship between work status congruence and workrelated attitudes and behaviors. Journal of Applied Psychology, 87, 903–915.

Gerson, J., & Kraut, E. (1988). Clerical work at home or in the office: The difference it makes. In K. Christensen (Ed.), The new era of home based work: Directions and policies (pp. 49–64). Boulder, CO: Westview Press. Gramm, C. L., & Schnell, J. F. (2001). The use of flexible staffing arrangements in core production jobs. Industrial & Labor Relations Review, 54, 245–258. Greenwald, C. S., & Liss, J. (1973). Part-time workers can bring higher productivity. Harvard Business Review, 51(5), 20. Gustafsson, S., Kenjoh, E., & Wetzels, C. (2003). Employment choices and pay differences between nonstandard and stand work in Britain, Germany, the Netherlands, and Sweden. In S. N. Houseman & M. Osawa (Eds.), Nonstandard work in developed economies: Causes and consequences (pp. 215–266). Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. Hakim, C. (2002). Lifestyle preferences as determinants of women’s differentiated labor market careers. Work and Occupations, 29, 428–459. Hall, D. T., & Gordon, F. E. (1973). Career choices of married women: Effects on conflict, role behavior, and satisfaction. Journal of Applied Psychology, 58, 42–48. Hall, R. (2006). Temporary agency work and HRM in Australia: “Cooperation, specialization and satisfaction for the good of all?” Personnel Review, 35, 158–174. Handy, C. (1994). The empty raincoat: Making sense of the future. London: Hutchinson. Handy, C. (1995). Trust and the virtual organization. Harvard Business Review, 73(3), 40–50. Harrison, B., & Kelley, M. R. (1993). Outsourcing and the search for “flexibility.” Work, Employment & Society, 7, 213–235. Hill, E. J., Ferris, M., & Martinson, V. (2003). Does it matter where you work? A comparison of how three work venues (traditional office, virtual office, and home office) influence aspects of work and personal/family life. Journal of Vocational Behavior, 63, 220–241. 592

Hoque, K., & Kirkpatrick, I. (2003). Non-standard employment in the management and professional workforce: Training, consultation and gender implications. Work, Employment & Society, 17, 667–689. Hornung, S., Rousseau, D. M., & Glaser, J. (2008). Creating flexible work arrangements through idiosyncratic deals. Journal of Applied Psychology, 93, 655–664. Houseman, S. N. (1995). Part-time employment in Europe and Japan. Journal of Labor Research, 16, 249–262. Houseman, S. N., Kalleberg, A. L., & Erickcek, G. A. (2003). The role of temporary agency employment in tight labor markets. Industrial & Labor Relations Review, 57, 105–127. Houseman, S. N., & Osawa, M. (2003a). The growth of nonstandard employment in Japan and the United States: A comparison of causes and consequences. In S. N. Houseman & M. Osawa (Eds.), Nonstandard work in developed economies: Causes and consequences (pp. 175–214). Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. Houseman, S. N., & Osawa, M. (2003b). Introduction. In S. N. Houseman & M. Osawa (Eds.), Nonstandard work in developed economies: Causes and consequences (pp. 1–14). Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. Hundley, G. (2001). Why and when are the self-employed more satisfied with their work? Industrial Relations, 40, 293–316. Hunter, L., McGregor, A., Mclnnes, J., & Sproull, A. (1993). The “flexible firm:” Strategy and segmentation. British Journal of Industrial Relations, 31, 383–407. Huws, U., Kork, W. B., & Robinson, S. (1990). Telework: Towards the elusive office. Chichester, England: Wiley. In depth: Customers analyze outsourcing vendors and strategies. (2006a, June 19). InformationWeek. Retrieved June 4, 2008, from http://www.informationweek.com/story/showArticle.jhtml?articleID=189 500052 In depth: When outsourcing goes bad. (2006b, June 19). InformationWeek. Retrieved June 4, 2008, from

Nonstandard Workers

http://www.informationweek.com/story/showArticle. jhtml?articleID=189500043 Jarmon, R., Paulson, A. S., & Rebne, D. (1998). Contractor performance: How good are contingent workers at the professional level? IEEE Transactions on Engineering Management, 45, 11–19. Johnson, S. A., & Ashforth, B. E. (2008). Externalization of employment in a service environment: The role of organizational and customer identification. Journal of Organizational Behavior, 29, 287–309.

Copyright American Psychological Association. Not for further distribution.

Jordan, J. W. (2003). Sabotage or performed compliance: Rhetorics of resistance in temp worker discourse. Quarterly Journal of Speech, 89(1), 19–40. Jurik, N. C. (1998). Getting away and getting by. Work and Occupations, 25, 7–35. Kalleberg, A. L. (2000). Nonstandard employment relations: Part-time, temporary and contract work. Annual Review of Sociology, 26, 341–365. Kalleberg, A. L., Reskin, B. F., & Hudson, K. (2000). Bad jobs in America: Standard and nonstandard employment relations and job quality in the United States. American Sociological Review, 65, 256–278. Kalleberg, A. L., Reynolds, J., & Marsden, P. V. (2003). Externalizing employment: Flexible staffing arrangements in US organizations. Social Science Research, 32, 525–552. Katz, J. A. (1993). How satisfied are the self-employed? A secondary analysis approach. Entrepreneurship: Theory & Practice, 17(3), 35–51. Kinsman, F. (1987). The telecommuters. Chichester, England: Wiley. Kirkpatrick, I., & Hoque, K. (2006). A retreat from permanent employment? Accounting for the rise of professional agency work in UK public services. Work, Employment & Society, 20, 646–646. Knell, J. (2000). Most wanted: The quiet birth of the free worker. London: The Industrial Society. Ko, J. R. (2003). Contingent and internal employment systems: Substitutes or complements? Journal of Labor Research, 24, 473–490. Kochan, T. A., Smith, M., Wells, J. C., & Rebitzer, J. B. (1994). Human resource strategies and contingent workers: The case of safety and health in the petrochemical industry. Human Resource Management, 33, 55–77. Kraimer, M. L., Wayne, S. J., Liden, R. C., & Sparrowe, R. T. (2005). The role of job security in understanding the relationship between employees’ perceptions of temporary workers and employees’ performance. Journal of Applied Psychology, 90, 389–398. Krausz, M. (2000). Effects of short- and long-term preferences for temporary work upon psychological outcomes. International Journal of Manpower, 21, 635.

Krausz, M., Brandwein, T., & Fox, S. (1995). Work attitudes and emotional responses of permanent, voluntary, and involuntary temporary-help employees: An exploratory study. Applied Psychology: An International Review, 44, 217–232. Krueger, A. B. (1993). How computers have changed the wage structure: Evidence from microdata. The Quarterly Journal of Economics, 108, 33–60. Kunda, G., Barley, S. R., & Evans, J. (2002). Why do contractors contract? The experience of highly skilled technical professionals in a contingent labor market. Industrial & Labor Relations Review, 55, 234–261. Lenz, E. A. (1996). Flexible employment: Positive work strategies for the 21st century. Journal of Labor Research, 17, 555–566. Lepak, D. P., & Snell, S. A. (1999). The human resource architecture: Toward a theory of human capital allocation and development. Academy of Management Review, 24, 31–48. Lepak, D. P., & Snell, S. A. (2002). Examining the human resource architecture: The relationships among human capital, employment, and human resource configurations. Journal of Management, 28, 517–543. Lepak, D. P., Takeuchi, R., & Snell, S. A. (2003). Employment flexibility and firm performance: Examining the interaction effects of employment mode, environmental dynamism, and technological intensity. Journal of Management, 29, 681–703. Li, J., Lam, K., Sun, J. J. M., & Liu, S. X. Y. (2008). Strategic human resource management, institutionalization, and employment modes: An empirical study in China. Strategic Management Journal, 29, 337–342. Liden, R. C., Wayne, S. J., Kraimer, M. L., & Sparrowe, R. (2003). The dual commitments of contingent workers: An examination of contingents’ commitment to the agency and the organization. Journal of Organizational Behavior, 24, 609–625. MacDuffie, J. P. (2007). HRM and distributed work. The Academy of Management Annals, 1, 549–615. Mallon, M., & Duberley, J. (2000). Managers and professionals in the contingent workforce. Human Resource Management Journal, 10, 33–47. Marler, J. H., Barringer, M. W., & Milkovich, G. T. (2002). Boundaryless and traditional contingent employees: Worlds apart. Journal of Organizational Behavior, 23, 425–453. Masters, J. K., & Miles, G. (2002). Predicting the use of external labor arrangements: A test of the transaction costs perspective. Academy of Management Journal, 45, 431–442. Matusik, S. F., & Hill, C. W. L. (1998). The utilization of contingent work, knowledge creation, and competitive advantage. Academy of Management Review, 23, 680–697. 593

George and Ng

Mayer, K. J., & Nickerson, J. A. (2005). Antecedents and performance implications of contracting for knowledge workers: Evidence from information technology services. Organization Science, 16, 225–242. Mayhew, C., & Quinlan, M. (1997). Subcontracting and OHS in the residential building sector. Industrial Relations Journal, 28, 192–205. McDonald, D. J., & Makin, P. J. (2000). The psychological contract, organizational commitment and job satisfaction of temporary staff. Leadership & Organizational Development Journal, 21, 84–91.

Copyright American Psychological Association. Not for further distribution.

McGinnis, S. K., & Morrow, P. C. (1990). Job attitudes among full- and part-time employees. Journal of Vocational Behavior, 36, 82–96. McGovern, P., Smeaton, D., & Hill, S. (2004). Bad jobs in Britain: Nonstandard employment and job quality. Work and Occupations, 31, 225–249.

Contingent work: American employment relations in transition (pp. 126–143). Ithaca, NY: ILR Press. O’Mahony, S., & Bechky, B. A. (2006). Stretchwork: Managing the career progression paradox in external labor markets. Academy of Management Journal, 49, 918–941. Osnowitz, D. (2006). Occupational networking as normative control: Collegial exchange among contract professionals. Work and Occupations, 33, 12–41. Osterman, P. (1987). Choice of employment systems in internal labor markets. Industrial Relations, 26, 47–67. Parker, S. K., Griffin, M. A., Sprigg, C. A., & Wall, T. D. (2002). Effect of temporary contracts on perceived work characteristics and job strain: A longitudinal study. Personnel Psychology, 55, 689–719.

McGrath, D. M., & Keister, L. A. (2008). The effect of temporary employment on asset accumulation processes. Work and Occupations, 35, 196–222.

Pearce, J. L. (1993). Toward an organizational behavior of contract laborers: Their psychological involvement and effects on employee co-workers. Academy of Management Journal, 36, 1082–1096.

McLean Parks, J., Kidder, D. L., & Gallagher, D. G. (1998). Fitting square pegs into round holes: Mapping the domain of contingent work arrangements onto the psychological contract. Journal of Organizational Behavior, 19, 697–730.

Peck, J., & Theodore, N. (1998). The business of contingent work: Growth and restructuring in Chicago’s temporary employment industry. Work, Employment & Society, 12, 655–674.

Miller, H. E., & Terborg, J. R. (1979). Job attitudes of part-time and full-time employees. Journal of Applied Psychology, 64, 380–386. Millward, L. J., & Brewerton, P. M. (1999). Contractors and their psychological contract. British Journal of Management, 10, 253–274. Morris, M. D. S., & Vekker, A. (2001). An alternative look at temporary workers, their choices, and the growth in temporary employment. Journal of Labor Research, 22, 373–390. Nagase, N. (2003). Standard and nonstandard work arrangements, pay difference, and choice of work by Japanese mothers. In S. N. Houseman & M. Osawa (Eds.), Nonstandard work in developed economies: Causes and consequences (pp. 267–305). Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. Nardone, T. (1995). Part-time employment: Reasons, demographics, and trends. Journal of Labor Research, 16, 275–291. Nesheim, T., Olsen, K. M., & Kalleberg, A. L. (2007). Externalizing the core: Firms’ use of employment intermediaries in the information and communication technology industries. Human Resource Management, 46, 247–264. Nollen, S. D., & Axel, H. (1996). Managing contingent workers: How to reap the benefits and reduce the risks. New York: AMACOM. Nollen, S. D., & Axel, H. (1998). Benefits and costs to employers. In K. Barker & K. Christensen (Eds.), 594

Pfeffer, J., & Baron, J. N. (1988). Taking the workers back out: Recent trends in the structuring of employment. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 10, pp. 257–303). Greenwich, CT: JAI Press. Pink, D. H. (2001). Free agent nation: The future of working for yourself. New York: Warner Books. Polivka, A. E. (1996, October). Into contingent and alternative employment: By choice? Monthly Labor Review, 119, 55–74. Quinlan, M., & Bohle, P. (2004). Contingent work and occupational safety. In J. Barling & M. R. Frone (Eds.), The psychology of workplace safety (pp. 81–105). Washington, DC: American Psychological Association. Quinlan, M., Mayhew, C., & Bohle, P. (2001). The global expansion of precarious employment, work disorganization, and consequences for occupational health: A review of recent research. International Journal of Health Services, 31, 335–414. Quinn, J. B. (1999). Strategic outsourcing: Leveraging knowledge capabilities. Sloan Management Review, 40(4), 9–21. Rebitzer, J. B. (1995). Job safety and contract workers in the petrochemical industry. Industrial Relations, 34(1), 40–57. Rhoades, L., & Eisenberger, R. (2002). Perceived organizational support: A review of the literature. Journal of Applied Psychology, 87, 698–714.

Nonstandard Workers

Rogers, J. K. (1995). Just a temp: Experience and structure of alienation in temporary clerical employment. Work and Occupations, 22, 137–166. Rogers, J. K. (2000). Temps: The many faces of the changing workplace. Ithaca, NY: Cornell University Press. Ross, C. E., & Wright, M. P. (1998). Women’s work, men’s work, and the sense of control. Work and Occupations, 25, 333–355.

Copyright American Psychological Association. Not for further distribution.

Rousseau, D. M., & Libuser, C. (1997). Contingent workers in high risk environments. California Management Review, 39, 103–123. Sahaym, A., Steensma, H. K., & Schilling, M. A. (2007). The influence of information technology on the use of loosely coupled organizational forms: An industry-level analysis. Organization Science, 18, 865–880. Segal, L. M., & Sullivan, D. G. (1997). The growth of temporary services work. Journal of Economic Perspectives, 11, 117–136. Shamir, B. (1992). Home: The perfect workplace? In S. Zedeck (Ed.), Work, families, and organizations (pp. 272–311). San Francisco: Jossey-Bass. Sherer, P. D., & Leblebici, H. (2001). Bringing variety and change into strategic human resource management research. Research in Personnel and Human Resources Management, 20, 199–230. Shi, Y. (2007). Today’s solution and tomorrow’s problem: The business process outsourcing risk management puzzle. California Management Review, 49(3), 27–44. Sias, P. M., Kramer, M. W., & Jenkins, E. (1997). A comparison of the communication behaviors of temporary employees and new hires. Communication Research, 24, 731–754. Smith, V. (1994). Institutionalizing flexibility in a service firm: Multiple contingencies and hidden hierarchies. Work and Occupations, 21, 284–307. Smith, V. (1997). New forms of work organization. Annual Review of Sociology, 23, 315–339. Smith, V. (1998). The fractured world of the temporary worker: Power, participation, and fragmentation in the contemporary workplace. Social Problems, 45, 411–430. Stanworth, C., & Druker, J. (2006). Human resource solutions? Dimensions of employers’ use of temporary agency labor in the UK. Personnel Review, 35, 175–190. Stanworth, C., & Stanworth, J. (1997). Managing an externalized workforce: Freelance labor-use in the UK book publishing industry. Industrial Relations Journal, 28, 43–55.

Thatcher, S. M. B., & Zhu, X. (2006). Changing identities in a changing workplace: Identification, identity enactment, self-verification, and telecommuting. Academy of Management Review, 31, 1076–1088. Thorsteinson, T. J. (2003). Job attitudes of part-time vs. full-time workers: A meta-analytic review. Journal of Occupational and Organizational Psychology, 76, 151–177. Tietze, S. (2005). Discourse as strategic coping resource: Managing the interface between “home” and “work.” Journal of Organizational Change Management, 18, 48–62. Tietze, S., & Musson, G. (2003). The times and temporalities of home-based telework. Personnel Review, 32, 438–455. Tilly, C. (1991). Reasons for the continuing growth of part-time employment. Monthly Labor Review, 114(3), 10. Torka, N. (2004). Atypical employment relationships: Wishful thinking or HR challenge. Management Revue, 16(3), 5–25. Townsend, A. M., & Bennett, J. T. (2003). Human resources and information technology. Journal of Labor Research, 24, 361–363. U.S. Bureau of Labor Statistics. (2005). Contingent and alternative employment arrangements. Retrieved May 26, 2008, from http://www.bls.gov/news.release/ conemp.toc.htm U.S. Census Bureau. (2000). Working at home: 2000. Retrieved May 26, 2008, from http://www.census. gov/population/www/cen2000/briefs/phc-t35/ index.html Uzzi, B., & Barsness, Z. I. (1998). Contingent employment in British establishments: Organizational determinants of the use of fixed-term hires and part-time workers. Social Forces, 76, 967–1005. Van Breugel, G., Van Olffen, W., & Olie, R. (2005). Temporary liaisons: The commitment of “temps” towards their agencies. Journal of Management Studies, 42, 539–566. Van Dyne, L., & Ang, S. (1998). Organizational citizenship behavior of contingent workers in Singapore. Academy of Management Journal, 41, 692–703. von Hippel, C., Mangum, S. L., Greenberger, D. B., Heneman, R., & Skoglind, J. D. (1997). Temporary employment: Can organizations and employees both win? Academy of Management Executive, 11(1), 93–104. Virtanen, M., Kivimaki, M., Elovainio, M., Vahtera, J., Kokko, K., & Pulkkinen, L. (2005). Mental health and hostility as predictors of temporary employment: Evidence from two prospective studies. Social Science and Medicine, 61, 2084–2095. 595

George and Ng

Wiens-Tuers, B. A., & Hill, E. T. (2002). How did we get here from there? Movement into temporary employment. Journal of Economic Issues, 36, 303–311. Wiesenfeld, B. M., Raghuram, S., & Garud, R. (1999). Communication patterns as determinants of organizational identification in a virtual organization. Organization Science, 10, 777–790.

Copyright American Psychological Association. Not for further distribution.

Williams, K. A., & O’Reilly, C. A. (1998). Demography and diversity in organizations: A review of 40 years

596

of research. In R. I. Sutton & B. M. Staw (Eds.), Research in organizational behavior (pp. 77–140). Stamford, CT: JAI Press. Youndt, M. A., Snell, S. A., Dean, J. W. J., & Lepak, D. P. (1996). Human resource management, manufacturing strategy, and firm performance. Academy of Management Journal, 39, 836–866. Zuboff, S., & Maxmin, J. (2002). The support economy: Why corporations are failing individuals and the next episode of capitalism. New York: Viking.

CHAPTER 19

TEAM DEVELOPMENT AND FUNCTIONING

Copyright American Psychological Association. Not for further distribution.

Janis A. Cannon-Bowers and Clint Bowers

Teams are a fundamental part of almost all human endeavors. As such, interest in teams and team functioning in the workplace became a topic of serious investigation as early as the 1920s. Since then, scholarly work in the area of teams and teamwork has grown dramatically. Indeed, the modern pressures of a global economy have increased the need for organizations to optimize the use of teams, and introduced new challenges such as so-called virtual teams. The purpose of this chapter is to provide a synthesis of past work into team functioning and development, highlighting important historical milestones and drawing conclusions where appropriate. It is organized around several key topics: the history of team research in organizations, models and taxonomies of teams and team tasks, defining team requirements, team selection, team processes and emergent states, measuring team performance, team training, and other emerging issues in teams. We conclude with recommendations for future team researchers. TEAMS AND TEAMWORK: HISTORY, DEFINITIONS, AND MODELS Teams of people working together for a common purpose have been a centerpiece of human social organization ever since our ancient ancestors first banded together to hunt game, raise families, and defend their communities. Human history is largely a story of people working together in groups to explore, achieve and conquer (Kozlowski & Ilgen, 2006, p. 77).

As this quote aptly states, teams are a fundamental part of almost all human endeavors. In light of this, it is perhaps somewhat surprising to realize that the study and use of teams in the workplace is a relatively new development (Kozlowski & Ilgen, 2006). Despite this rather slow start, the field of teams and team performance has literally exploded in recent years, and there is now a voluminous literature on the subject. This has presented a rather daunting task in preparing this chapter because we are really able to only scratch the surface in reviewing this important area. As such, we have focused on important historical foundations as well as the more recent developments and have attempted to synthesize past work and distill important conclusions where possible. The remainder of this chapter is structured as follows: First, we provide a brief history of work team and work group research and follow with a discussion of important definitional issues. Next we present several examples of models of teamwork that have guided empirical work and also several major taxonomies of teams and team tasks. After this, we turn our attention to establishing team requirements, discussing issues including team task analysis and team competencies. This leads to a review of research into team selection, including personality and individual differences, team composition, heterogeneity and diversity, and team size. The next major section covers team processes, summarizing theoretical and empirical work regarding the processes that make teams effective (or ineffective). Following this, we present a review of team performance measurement and team-training strategies. Finally, we address some other, more

http://dx.doi.org/10.1037/12169-019 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

597

Cannon-Bowers and Bowers

emerging issues associated with teams: virtual teams, teams and technology, and multicultural teams.

Copyright American Psychological Association. Not for further distribution.

A Brief History of Team Research In their comprehensive review, Sundstrom, McIntyre, Halfhill, and Richards (2000) traced the history of team research to the seminal “Hawthorne studies” in the 1920s. According to these authors, among the results of this unprecedented work was the recognition that work groups were an important topic of study within organizations. In particular, the notion that work groups developed informal work structures and norms that had a direct impact on performance was born (Sundstrom et al., 2000). Over the next few decades, the interest in work groups grew rather sporadically, with a few notable exceptions (see Sundstrom et al., 2000, for a more detailed review). For example, work in the 1950s on sociotechnical systems (Rice, 1953; Trist & Bamforth, 1951) emphasized the use of autonomous work groups. The participative management movement beginning in the 1960s (Likert, 1961; McGregor, 1960) advocated expanded use of work groups, and in the 1970s, several well-publicized case studies highlighted the use of team-based organizations (e.g., General Motors’ Tarrytown plant, Volvo, Saab; see Sundstrom et al., 2000, for more detail). By the mid 1980s, the use of quality circles had taken hold as an offshoot of the total quality management movement (Cannon-Bowers, Oser, & Flanagan, 1992). These groups were not part of direct work structures; rather, they were groups of employees who came together to discuss quality improvement across the company. At the same time, the use of teams as a basis for work organization also grew in the form of production groups and project teams (Sundstrom et al., 2000), a trend consistent with popular organizational change movements of the time (e.g., Kanter, 1983; Peters, 1988). In the late 1980s and into the 1990s, the emphasis on teams and team performance grew exponentially, with impetus from at least two sectors: commercial aviation and the military (see Salas, Bowers, & Cannon-Bowers, 1995). With respect to aviation, the late 1980s witnessed development of several cockpit resource management (CRM), sometimes referred to more recently as crew resource management (also 598

CRM), programs (Wiener, Kanki, & Helmreich, 1993). In response to a number of well-publicized aviation incidents and accidents that were attributed at least in part to faulty teamwork, the commercial aviation industry began to institute CRM programs aimed at improving teamwork in the cockpit (Weiner et al., 1993). Among the important implications of this work was the recognition that “softer” skills such as teamwork attitudes, interpersonal skills, communication, and assertiveness could all have an important impact on team performance (Helmreich, Wilhelm, Gregorich, & Chidester, 1990). Another event in the summer of 1988—this one involving a U.S. Navy warship (the USS Vincennes)— also focused attention and resources onto team performance (see Cannon-Bowers & Salas, 1998, for more detail). The ship mistakenly shot down an Iranian airbus, believing that it was a hostile aircraft. After considerable investigation, it was determined that the incident could not be attributed to any system or hardware failure, prompting the establishment of the Tactical Decision Making Under Stress (TADMUS) project by the U.S. Navy (Cannon-Bowers & Salas, 1998). TADMUS was a multimillion dollar, multidisciplinary program of research that spanned almost 10 years. It had two overarching purposes: to improve the design of computer interfaces and to improve training strategies in high performance teams. According to Kozlowski and Ilgen (2006), TADMUS was successful in generating many effective team training approaches and is a good example of how theory and research can translate into organizational applications. By the year 2000, the study of teams and their application in organizations had come of age. Rather than question the value of team-based organizations, researchers in this decade have turned their attention to how best to select, train, and develop effective teams (Kozlowski & Ilgen, 2006). Emphasis has shifted to more specialized topics, such as the cognitive underpinnings of teamwork, virtual (i.e., physically dispersed) teams, and multicultural teams. The 2000s have also seen a large number of meta-analyses of different aspects of teamwork and team performance (described later), which is not surprising because the body of empirical work has now grown to the point where such analyses are possible. Before

Team Development and Functioning

we can begin a meaningful discussion of this work into teams and teamwork, it is important to address some important definitional issues.

Copyright American Psychological Association. Not for further distribution.

Definitions and Defining Characteristics The astute reader will note that we have been using the terms team and group interchangeably to this point, a practice that has led to some debate in the literature. According to Sundstrom et al. (2000), although some researchers distinguish between these terms, “the distinction has been neither consistent or widely recognized” (p. 44). We tend to share this view, because it is the defining characteristics and features of the construct that are important, not strictly the label. Hence, our discussion focuses on the definitions of work teams and groups that have been offered in the literature, which range from fairly simple and straightforward, to more complex. Among the more straightforward definitions, Sundstrom, DeMeuse, and Futrell (1990) defined work teams as “small groups of interdependent individuals who share responsibility for outcomes for their organizations” (p. 120). In a similar vein, Salas, Dickinson, Converse, and Tannenbaum (1992) defined a team as a distinguishable set of two or more people who interact dynamically, interdependently, and adaptively toward a common and valued goal/objective/mission, who have each been assigned specific roles or functions to perform, and who have a limited life-span membership. (p. 4) In an even more comprehensive treatment, Kozlowski and Ilgen (2006) synthesized past work to arrive at the following definition of teams: (a) Two or more individuals who (b) socially interact (face to face, or increasingly, virtually); (c) possess one or more common goals; (d) are brought together to perform organizationally relevant tasks; (e) exhibit interdependencies with respect to workflow, goals, and outcomes; (f) have different roles and responsibilities; and (g) are together embedded in an encompassing organi-

zational system, with boundaries and linkages to the broader system context and task environment. (p. 79) Parsing these definitions yields several important features that we believe represent the essence of teamwork: interdependence of action; shared responsibility; and common, meaningful goals. It also highlights the fact that team members often have specialized roles and that teams exist in a broader organizational context (which presumably effects their performance). Other authors have also added to the list of essential features the fact that teams have members who see themselves as, and are recognized by others as, a group (e.g., Stevens & Campion, 1994) and that teams display adaptive strategies that allow them to respond to change (Paris, Salas, & Cannon-Bowers, 2000). Types of teams. The rather broad, overarching definition presented earlier is useful in establishing the boundaries of theorizing and research on teams. However, to better appreciate the nature of teams and team performance, a more detailed description is needed. This need has given rise to several typologies of teams and their work in recognition of the fact that the nature of the work being completed by the team has an important impact on their functioning. It has also served to organize thinking about teams by highlighting commonalities among the kinds of teams employed by organizations. With respect to taxonomies of teams (as opposed to team tasks, which are reviewed subsequently), Sundstrom et al. (1990) offered one popular framework. This view categorized teams into: advice/ involvement, production/service, project/development, and action/negotiation. Each of these was defined in terms of its work-team differentiation (i.e., the degree of specialization, independence, or autonomy of the team with respect to other organizational units) and external integration (i.e., the extent of integration into the larger organizational system and degree of coordination and synchronization needed). Also included in the taxonomy were work cycles (which can be brief, recurring, enduring) and typical outputs generated by the team. In subsequent updates of this work, Sundstrom and colleagues (Sundstrom, 1999; Sundstrom et al., 599

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

2000) added that these various types of teams vary on four main characteristics: level of authority within the organization, time horizon (i.e., time until the team disbands), degree of specialization, and degree of interdependence within and external to the team. Their most recent work also offered a revised set of team types to include production groups, service groups, management teams, project groups, action and performing groups, and advisory groups (see Sundstrom, 1999, for more detail). In other work, Klimoski and Jones (1995) presented a taxonomy of decision-making teams that included command-and-control teams, production teams, customer service teams, professional (technical) decision-making teams, and executive teams. In contrast to Sundstrom (1999), this categorization was based more on an analysis of the various types of teams that exist in different organizational sectors. Klimoski and Jones’s purpose in making a distinction among these types of teams was that the unique requirements inherent in each would drive decisions about team member selection and staffing. (We have more to say about team selection in a later section.) Although there is no single, universally accepted taxonomy of teams, the point here is that finding meaningful bases on which to classify teams can be useful in setting boundaries for both research and application. Rather than trying to generalize to all types of teams (which is almost akin to generalizing to all kind of work), a taxonomic approach can better couch the pertinent questions about a team’s performance and help to establish generalizability of results. Hence, we recommend that researchers adopt and make explicit the way they view the type of teams with which they are dealing. Types of team task demands. Another approach to organizing team research and practice is to focus on the nature of the team’s task or work. According to McGrath (1984), analyzing any team performance situation should begin with an understanding of the nature of the group task to be performed. Only then can the organization and design of the team and its work be optimized. In one influential formulation of this sort, Steiner (1972) proposed that there were several types of team tasks: 600

■

■

■

■

additive tasks require summing of each team members’ effort or output, where each member has an equal part to play; disjunctive tasks require that only one team member performs well to achieve success, for example when the team’s output is dependent on the most knowledgeable member; conjunctive tasks require that every team member perform at a minimally acceptable level, as when the output of one member is dependent on the output of another; and discretionary tasks allow team members to combine their individual inputs in any of the ways described earlier or another strategy devised by the group.

The implications of this type of scheme for functions such as proper selection and staffing of team members, development of training programs, and design of teamwork environments (e.g., communication networks), as well as studies of team effectiveness, are vast. For example, when selecting team members for a disjunctive task it would make sense to place emphasis on ensuring that at least one team member had sufficient expertise, whereas conjunctive tasks would require a closer look at how the team is composed across members. We revisit this issue in more detail in later sections. In addition to Steiner (1972), McGrath (1984) offered an often-cited taxonomy of team tasks or processes. These included generative tasks (i.e., members must generate ideas or alternatives such as in brainstorming), executing tasks (i.e., members perform psychomotor or physical activities according to a prespecified plan or model), negotiating tasks (i.e., members must reach agreements by resolving conflict or reconciling different viewpoints), and choosing or decision-making tasks (i.e., members must construct or select a solution). Once again, the value of this type of framework stems from its ability to help generate hypotheses about how various team and environmental features affect team performance and determine the generalizability of results regarding team functioning. In a slightly different vein, several groups of scholars have theorized about the nature of interdependence in a team’s task (Saavedra, Earley, & Van Dyne, 1993;

Team Development and Functioning

Thompson, 1967; Van de Ven, Delbecq, & Koenig, 1976). According to this line of thinking, team tasks can be organized into the following four categories on the basis of the type of interdependence required:

Copyright American Psychological Association. Not for further distribution.

■

■

■

■

Pooled interdependence—independent workflow where each group member contributes separately to the group’s output without interacting directly with other members. In such situations, group members typically hold similar roles and complete the entire task. Group output is the sum of the individual group members’ contributions. Sequential interdependence—one-way workflow where the input from one group member is necessary to the functioning of another. Typically, group members have specialized roles and perform different parts of the task, and performance is a function of the correct actions being completely in the correct sequence. An example here would be a traditional assembly line. Reciprocal interdependence—two-way workflow where two team members interact such that the output of one becomes the input to the other and vice versa. In this case, the roles of members are typically specialized, but the order of individual actions can vary. Group performance requires coordination among members because there is flexibility in the sequence of steps. Commandand-control teams are examples of this type of interdependence because they must coordinate actions among individual experts in support of overall task performance. Team interdependence—simultaneous, multidirectional workflow where group members must act collaboratively to complete the task. This category was added by Van de Ven et al. (1976) to the original Thompson (1967) scheme. Groups in this category have the discretion to jointly decide how the task will be completed and have the power to modify allocation of resources to reach their goals. Self-managed work teams typify this category.

Saavedra et al. (1993) expanded these original notions about types of task interdependence by making the point that to optimize performance, other performance-shaping functions within the team needed to be congruent with the nature of

interdependence. These researchers defined complex interdependence as comprising task, goal, and feedback interdependence, and they hypothesized that the congruence among these factors would affect performance. This hypothesis was supported in a laboratory study; specifically, congruent complex interdependence was associated with higher levels of group performance (measured as quantity and quality of group output on a management task; r = .22 for quantity, r = 25. for quality; Saavedra et al., 1993). Finally, based on a good deal of prior work, Fleishman and Zaccaro (1992) developed a taxonomy of team functions. This formulation identifies classes of functions performed within a task and is based on the notion that team performance is the result of four antecedents: the external conditions imposed on the team, team member resources, task demands and characteristics, and team characteristics. It consists of seven categories of functions: ■

■

■

■

■

■

■

orientation functions—exchanging information about member resources and constraints; team tasks and goals; task priorities; resource distribution functions—distributing the task or workload across members; matching member resources to task requirements; timing functions—influencing how activities are paced overall and at the individual level; response coordination functions—synchronizing and timing of tasks and output; response sequencing; motivational functions—managing individual and team-level goals; developing team norms; establishing performance–reward contingencies; resolving conflict; balancing team versus individual orientation; systems monitoring function—monitoring of the task to enable adjustments and error correction; general and individual activity monitoring; and procedure maintenance—monitoring of general and individual activities and procedures; adjusting nonstandard activities.

Models of Teamwork and Team Effectiveness Flowing directly from the definition of work teams and taxonomies of teams and tasks is consideration of teamwork—what it is and how it is affected by 601

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

factors both within and outside the team. Related to this is the question of team effectiveness—simply put, the identification of factors that facilitate or inhibit a team’s performance. In this regard, a number of models or frameworks of teamwork have been offered; several of these are reviewed next. Perhaps the most influential of team frameworks was offered by McGrath (1964), in which team performance is viewed as involving inputs, processes, and outputs. This line of thinking led to specification of a series of so-called input–process–output (I–P–O) models. According to this framework, inputs include a variety of individual and team factors, resources, and organizational/environmental variables. Process refers to activities engaged in by the team to accomplish task demands, essentially the transformation of inputs into outputs. Outputs include resulting team performance as well as other outcomes such as team satisfaction, cohesion, and turnover. One popular version of the I–P–O model genre was proposed by Hackman (1987). This model

(which has been updated and revised over the years but still endures as a viable representation of group effectiveness) emphasizes organizational context, as well as team members’ efforts, skills, and performance strategies (see Figure 19.1). It also emphasizes that group effectiveness includes the notion that members’ needs are satisfied and that their capability to work together in the future is maintained or strengthened. This is defined as team viability (i.e., the willingness of team members to remain a part of the team) and is an important group output. In another, more recent, I–P–O model proposed by Salas and colleagues (Salas et al., 1992; Tannenbaum, Beard, & Salas, 1992), the host of variables considered as important to team effectiveness was expanded significantly. Called the team effectiveness model (TEM), it contains four classes of input factors: task characteristics (e.g., workload, stress), work characteristics (e.g., role assignment, communication structure), individual characteristics (e.g., individual proficiencies, attitudes, competencies), and team characteristics

FIGURE 19.1. Hackman’s model of team effectiveness. From Handbook of Organizational Behavior (p. 331) by J. Lorsch (Ed.), 1987, Englewood Cliffs, NJ: Prentice-Hall. Copyright 1987 by Prentice-Hall. Adapted with permission. 602

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

(e.g., cohesiveness, familiarity, role clarity). Processes include such things as coordination, communication, decision making, and backup behavior. Finally, outcomes include team performance and attitudes. A feedback loop ensures that outcomes affect subsequent input and performance. Recently, the use of I–P–O models as a guide for empirical work has been criticized because these models are not reflective of the dynamic nature of team performance and tend to be descriptive rather than causal in nature (Kozlowski & Ilgen, 2006). Further, Ilgen, Hollenbeck, Johnson, and Jundt (2005) maintained that the I–P–O perspective falls short on three grounds: (a) many of the mediational processes cited by researchers as responsible for transforming inputs into outputs are not processes but emergent cognitive or affective states, for example, collective efficacy, cohesion, and situation awareness (see Marks, Mathieu, & Zaccaro, 2001), (b) I–P–O models are limited because they imply a single cycle, linear path from inputs through outcomes, and (c) recent work indicates that there are interactions between and among inputs, processes, and emergent states, suggesting that a main-effect progression from one to the next may not hold. In response to these deficiencies, Ilgen et al. (2005) proposed an alternative conceptualization that involves input–mediator–output–input (IMOI). The main differences between this modified version and the original formulation is that the term “mediator” is used as opposed to “process” to imply that there is a broader range of mediational influences that transform inputs into outcomes (e.g., emergent states). The addition of input at the end implies that there is an explicit cyclical feedback loop. Finally, the elimination of hyphens between letters is meant to imply that the causal linkages may not be linear, but can be nonlinear or conditional (see Ilgen et al., 2005, for more detail). In other formulations, several team scholars have presented models that attempt to describe a team over time, focusing on developmental stages. These models differ from I–P–O models in that they seek to explain what happens to a team across its life cycle rather than at any single point in time. Kozlowski, Gully, Nason, and Smith (1999) provided a summary and review of several of these types of models.

According to these authors, there is a high level of agreement with respect to developmental stages and emphasis on interpersonal processes and outcomes (see Kozlowski et al., 1999, for more detail). For example, the classic “norming, storming, forming, performing” framework presented by Tuckman (1965) is consistent with several others that propose separate stages of development (e.g., Caple, 1978; Francis & Young, 1974). In a slightly different formulation, Gersick (1988) proposed a two-stage punctuated equilibrium model. This model holds that in the first stage, an immediate pattern of activity persists to the halfway point of the team’s performance followed by a significantly altered pattern of group activity that focuses attention on task completion. Combining this view with that of Tuckman’s (1965), Morgan and colleagues (Morgan, Glickman, Woodward, Blaiwes, & Salas, 1986; Morgan, Salas, & Glickman, 1993) presented the team evolution and maturation (TEAM) model. It posits nine stages of development: one performing stage that represents forces external to the team that cause the team to be formed, seven central or core stages (i.e., forming, storming, norming, performingI, reforming, performing-II, conforming), and a final stage where, having completed its task, the team disbands (i.e., deforming). Gersick’s notion of a change in activity at the halfway point is represented in this model after the first performance stage (i.e., performing-I) when the team re-forms to complete the task. Also represented in the TEAM model is the notion that two separate tracks of activity develop as the team progresses through its developmental cycle. The first of these involves activities that have to do with the specific tasks being performed (hence, the label taskwork), whereas the second is concerned with those activities necessary to ensure effective functioning of the team (labeled teamwork). This second category includes such activities as developing social interaction patterns, coordination strategies, and relationships among members that are crucial to the success of the team. Recognition of this track highlighted the fact that teams require a unique, separate set of skills that are not pertinent in individual settings and led to a more direct focus on teamwork skills than 603

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

had been apparent previously (e.g., see CannonBowers, Tannenbaum, Salas, & Volpe, 1995). Kozlowski et al. (1999) also presented a view of teams from a developmental standpoint. Focusing on more complex tasks and structured environments, these authors sought to describe the development of teams who can adapt to changing task demands. This model proposes four phases of team development that vary with respect to the focus (i.e., individual vs. team), content, processes, and outcomes (i.e., cognitive, affective, behavioral) that change over time. They conceptualize team compilation as a developmental process that involves the building of knowledge, skills, and performance capabilities over time and levels. It incorporates notions such as socialization; team orientation; skill acquisition; task mastery; role knowledge, negotiation, identification, and routinization; self-regulation; continuous improvement; and team adaptability.

Summary The taxonomies, frameworks, and models of team performance described here are but a sampling of the many representations that exist in the literature. Perhaps the best conclusion that can be drawn from these various (and sometimes competing) conceptualizations is that there is no one best way to view teams or team performance. Teams are, in fact, complex and dynamic systems that affect, and are affected by, a host of individual, task, situational, environmental, and organizational factors that exist both internal and external to the team. What is probably most lacking are causal models that relate specific sets of features to predicted team outcomes and the conditions under which they hold. Although somewhat tedious to develop, such predictive frameworks should be useful guides for both researchers and practitioners. DEFINING TEAMWORK REQUIREMENTS The issue of defining teamwork requirements is fundamental to the development of effective teams because it precludes selection, training, and other personnel decisions. According to Stevens and Campion (1994), establishing teamwork competencies, or knowledge, skills, and abilities (KSAs), is 604

essential to a variety of human resource management (HRM) functions. Exhibit 19.1 summarizes the implications of teamwork knowledge, skills, and abilities (KSAs) for HRM systems based on the original presented by Stevens and Campion. As shown in this exhibit, teamwork competencies are fundamental to a number of crucial functions within an organization. The sections that follow first focus on methods to establish teamwork competencies (i.e., team task analysis) and a related discussion regarding how task demands effect team requirements and performance. Next, we move to a review of what has been found regarding which competencies are required for effective teamwork. Other issues raised in Table 16.1 (e.g., selection, staffing, training) will be addressed in more detail in subsequent sections.

Team Task Analysis As is the case with all personnel systems, it is crucial to establish the job and task requirements that confront a team. (See also Vol. 2, chap. 1, this handbook.) Typically, job/task analysis (JTA) methods focus on individual skills; however, in recent years, several researchers have called for a more direct emphasis on the teamwork demands of the task (Bowers, Baker, & Salas, 1994). This line of thinking is consistent with the model of teamwork offered by Morgan et al. (1993) described earlier. Unfortunately, there has not been much attention paid to team task analysis methods; indeed, what has been written focuses on techniques to augment more traditional JTA methods. The following sections summarize current work in this area. Several approaches to team task analysis have been proposed (e.g., McNeese & Rentsch, 2001; Swezey, Owens, Bergondy, & Salas, 1998). Burke (2004) summarized the essential elements of these approaches and others. She proposed seven steps that should be included in a thorough team task analysis. The first step is to conduct a requirements analysis. The goal of this step is to clearly describe the specific job that is to be trained. This step is important because simply relying on existing job titles or historical descriptions may not yield an accurate description of the job, especially if the organization has recently changed to a team-based operation.

Team Development and Functioning

Exhibit 19.1 The Implications of Teamwork Knowledge, Skills, and Abilities for the Design of Human Resources Systems

■ ■ ■ ■ ■ ■

Copyright American Psychological Association. Not for further distribution.

■ ■ ■ ■

■

■

■

■

■

■ ■ ■ ■

■ ■

■ ■ ■

Selection Selection procedures for jobs in team environments should assess teamwork KSAs. Selection procedures should be tailored to the types of teams within the organization. Employment tests assessing teamwork KSAs may be valid predictors of teamwork-related job performance. Teamwork KSAs may be measurable by selection procedures such as written tests, interviews, assessment centers, and biodata. Recruiting for teams should emphasize the importance of teamwork KSA requirements. Team staffing decisions should also consider differences in employee preferences for working in groups. Training Organizations with team environments should train teamwork KSAs as part of their development programs. There are a broad variety of potentially useful approaches to the training of teamwork KSAs. Managers of teams should also be trained in teamwork KSAs and in how to develop these KSAs in employees. Different training strategies may be more or less appropriate for various teamwork KSAs. Performance Appraisal To motivate teamwork in organizations with team environments, performance appraisals should be modified to assess and reward the behavioral and performance indicators of teamwork. An organization-specific job analysis may be needed to identify the behavioral or performance indicators of teamwork across various jobs. The specific requirements (i.e., KSAs) associated with the team task should be expressed in terms of measurable aspects of performance. Based on teamwork KSAs, valid and reliable measures of teamwork should be developed so that they can be used in performance appraisals. Goals to improve team-level KSAs should be incorporated into the appraisal process. Career Development Promotion criteria may need to be modified to consider the opportunities to develop teamwork contributions. Career planning systems may need to consider the opportunities to develop teamwork KSAs that jobs offer. Teamwork KSAs may be needed for proper socialization and, in turn, be enhanced by socialization. Mastery of teamwork KSAs should be considered in advancement decisions. Compensation Compensation systems in organizations with team environments should include compensable factors reflecting teamwork KSAs. Pay for skills programs in team environments should consider team-level KSAs. Job and Task Analysis Job and task analysis procedures should explicitly include team-level KSAs. Job and task analysis procedures should include measures of teamwork KSAs. The conditions and standards associated with teamwork across organizational tasks should be built into job and task analysis.

Note. From “The Knowledge, Skill, and Ability Requirements for Teamwork: Implications for Human Resource Management,” by M. Stevens and M. Campion, 1994, Journal of Management, 20, p. 505. Copyright 1994 by Sage. Adapted with permission.

The requirements analysis phase is also the one in which the analyst chooses the knowledge-gathering approaches that will be used in the task analysis. A variety of approaches are available, such as observation, interviews, surveys, and so forth. For the most part, the particular choice of methods depends on the nature of the task and the resources available to the

analyst. For example, surveys are relatively cheap and easy to administer but may not capture the complexity of a highly complex job (see Goldstein, 1993, for more detail regarding the relative merits of data-gathering approaches). In the team area, it is probably the case that interviews and possibly observation are needed to fully understand the team’s task and procedures. 605

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

Step 2 of the team task analysis is to identify the specific tasks that compose the targeted job using methods selected in Step 1. This process is done in a manner that is nearly identical to that used in the analysis of individual-level jobs. As noted earlier, this can be done in a number of ways, including directly interviewing job incumbents (in the case of team task analysis this can be done individually or with the entire team present), using previously generated task lists as a basis for discussion, sending out surveys to elicit task statements, observing performance directly (including video taping), and conducting workshops that involve a number of job incumbents. (The interested reader should consult Goldstein, 1993, for more detail.) The result of this phase is a list of specific tasks, including how they are done. The third step described by Burke (2004) is the identification of a teamwork taxonomy. This is a listing of teamwork behaviors that are frequently required in team performance situations. As noted previously, several such taxonomies are available (e.g., Cannon-Bowers et al., 1995; Fleishman & Zaccaro, 1992; Smith-Jentsch, Zeisig, Acton, & McPherson, 1998). The analyst should choose the taxonomy that best captures the teamwork needs of the targeted job. This taxonomy is used as the foundation for identifying the teamwork requirements of the job. The fourth step in Burke’s (2004) team taskanalysis approach is to conduct a coordination analysis. This step is designed to augment traditional job analysis techniques by focusing specifically on teamwork behaviors that are apt to be overlooked because they are not the focus of traditional (i.e., individual) JTA methods. One survey-based approach to accomplish coordination analysis has been described by Bowers, Morgan, Salas, and Prince, (1993). These researchers used a task list based on aviation tasks and the Cannon-Bowers et al. (1995) taxonomy to create a survey of coordination demands in military aviation. Arthur, Edwards, Bell, Villado, and Bennett (2005) recently described an expanded version of this approach that was effective in identifying teamwork requirements for a simulated combat task. The fifth step in the team task-analysis procedure is to select the relevant tasks for training. The product 606

of the preceding steps is typically a large list of tasks associated with the targeted job. This list is usually so long that it would be impractical to attempt to specifically train each task. Consequently, there is a need to select the most important tasks to be trained. In individual task analysis, this selection is done with the assistance of various task importance indices (e.g., Levine & Dickey, 1990; Sanchez & Levine, 1989). Bowers et al. (1994) evaluated the utility of several of these in predicting the overall importance of team tasks and found that a combination of only two variables, criticality and importance, were significant predictors. The sixth step described by Burke (2004) is the translation of tasks into KSAs or competencies that will become the actual targets for selection, staffing, training, and development. The translation process, described by Goldstein and Ford (2002), is accomplished by the analyst but may also involve team performance experts. Because many job analysts are not well versed in team performance, calling in such experts is prudent. In addition, many common teamwork behaviors have been translated into KSAs by Cannon-Bowers and her colleagues (1995) and can be used to guide the translation process; this work will be covered further in a subsequent section. The final step in team task analysis is to link the KSAs back to the team tasks. The goal of this step is to validate the outcome of the team task analysis before beginning selection or training development. This step is typically accomplished through a survey in which subject matter experts evaluate each KSA in terms of importance. Those for which there is high agreement are retained for further use. Others are either eliminated or revised so that they are useful.

Team Competencies As noted, the output of a successful team task analysis is a set of required competencies, or KSAs, for effective teamwork. Because much of the work in this area has been done in support of training system design, some researchers have substituted attitudes for abilities as the “A” in KSAs because abilities are typically considered fixed attributes and, therefore, cannot be affected much by training interventions. For clarity, we will use the acronym KSA traditionally here (i.e., A = abilities) and spell the phrase out when A = attitudes.

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

Recently, Mohammed, Cannon-Bowers, and Foo (2010) synthesized a great deal of literature regarding team competencies; the portion of this work pertaining to knowledge, skills and attitudes is displayed in Table 19.1. This summary can be considered a useful source for understanding what is required in effective teams and we attempted to sharpen this understanding further by indicating the percentage of variance accounted for in efforts to validate the construct’s importance. Specifically, we computed effect sizes where data were available to do so. As is evident from this table, there have been many different labels and definitions of teamwork knowledge, skills, and attitudes. In a similar review several years ago, Cannon-Bowers et al. (1995) argued that team researchers often used different labels to refer to similar constructs or used similar-sounding labels for different constructs. They concluded that this conceptual confusion has made it difficult to generalize and consolidate findings in the team performance area. Efforts such as those represented by Mohammed et al. (2010) are helping to address this problem. Other authors have been more specific in discussing team competencies. According to CannonBowers et al. (1995), team competencies apply in different ways depending on the task demands and the situation. These researchers offered a framework for thinking about team competencies that conceptualizes them as being either specific or generic with respect to a particular team or task situation. Essentially, team competencies can be either task-specific or task-generic on the one hand and either team-specific or team-generic on the other. Task-specific competencies are those that are applicable in a specific task or type of task. That is, the specific manner in which the team knowledge, skill, or attitude is expressed varies as a function of the task at hand. Hence, team members must understand how the teamwork competency is applied in context. An example of a task requiring task-specific competencies would be in aviation, where cockpit crews must complete a highly complex set of interdependent tasks to be successful. In such cases, the specific application of constructs such as communication or task sequencing constitutes the nature of the requirement. In contrast, task-generic competencies can be used effectively across a variety of tasks.

For example, possessing effective team planning skills can be useful in different tasks and can therefore be “transported” from one task to another, so the requirement in this case is more general. Further, team competencies can also be specific or generic with respect to the team members. This distinction concerns whether the competency requires knowledge of a specific teammate or is applicable to any set of team members. In more complex tasks where team members have some discretion regarding how they perform the task, it is important for them to predict how teammates will behave. This is especially true when the task does not allow team members to discuss strategies and must rely on what they know about one another to help predict how teammates will react. For example, many anecdotal accounts of professional sports teams relate that teams seem to perform better when the team members have had a chance to practice together (perhaps explaining why “dream teams” composed of the best players in the world often do not win). Such teams require team-specific competencies according to the Cannon-Bowers et al. (1995) framework. In contrast, team-generic competencies are those that contribute to team effectiveness regardless of which team members are present. For example, interpersonal skills can be developed and applied across a variety of teams, regardless of membership. Teamgeneric competencies are important because many teams within organizations are ad hoc, meaning that team members must interact effectively with teammates they do not know well. Hence, task forces, committees, and other similar structures depend on well-developed team-generic competencies. Taking these two factors together results in a matrix shown in Table 19.2. Specifically, teams (and associated competency requirements) can be broken into four categories: team-contingent, taskcontingent, context-driven, and transportable. Each of the four categories in Table 19.2 describes a different type of competency required for teamwork; Cannon-Bowers et al. (1995) classified a series of teamwork KSAs into these categories. The significance of this scheme is that it can provide useful guidance to team selection and training design efforts based on the nature of the competencies required. For example, if a competency is team607

Cannon-Bowers and Bowers

TABLE 19.1 Team Competencies Attribute

Definition

Related and subsidiary constructs

Validation and measurement issues

Effect sizes

Copyright American Psychological Association. Not for further distribution.

Knowledge Knowledge of teamwork skills

Understanding of the necessary underpinnings and behavioral requirements of effective team performance

Understanding teamwork, familiarity with teamwork, knowledge of teamwork KSAs

Assessed through Teamwork KSA test. Validation data show that this variable predicts effective teamwork (Hirschfeld et al., 2006; McClough & Rogelberg, 2003; Stevens & Campion, 1999).

.25 (Hirschfeld et al., 2006) to .56 (Stevens & Campion, 1999)

Knowledge of team roles

Knowledge of team roles and their situational contingencies

Intrapositional knowledge, knowledge of teammates

Assessed through Team Role Test. Validation data show that this test predicts role performance (Mumford et al., 2008).

.39 (Mumford et al., 2008)

Skills Adaptability

Ability of team members to adjust their strategies in response to task demands, by reallocating team resources

Compensatory behavior, backing-up behavior, ability to effect dynamic reallocation of function, ability to effect mutual adjustment, ability to balance workload

Best assessed in a work sample or other simulation. Most likely a function of past experience (so not easily trained). Some data to suggest that adaptability improves teamwork (Salas et al., 2007).

Insufficient data to compute.

Interpersonal factors

Ability of team members to optimize the quality of team member interactions through resolution of dissent, motivational reinforcement, and cooperative behaviors

Morale building behavior; conflict resolution behavior; ability to negotiate; cooperativeness; ability to consult with others; interpersonal trust; social perceptiveness; persuasiveness; desire to help others; social skills

May be assessed through a combination of paper and pencil and behavioral measures. Some validation data suggests that interpersonal skills predict teamwork (e.g., see Morgeson et al., 2005).

.17 (Morgeson et al., 2005)

Team management and leadership factors

Ability of team members to direct and coordinate activities; assign tasks; organize workflow among members; plan, organize, and establish a positive climate

Task motivation, goal setting abilities, planning and task coordination, ability to establish roles and expectations, ability to instruct others, planning abilities, organizing abilities

Best assessed in a work sample or other simulation, although paper and pencil instruments may add value. Not easily trained because probably develops with experience. Some validation indicates that individual leadership skills are associated with teamwork effectiveness (e.g., Ahearn et al., 2004; Burke, Stagl, Klein, Goodwin, Salas, & Halpin, 2006).

.18 (Ahearn et al., 2004); r = .33 for team effectiveness and r = .20 for productivity (Burke et al., 2006)

608

Team Development and Functioning

TABLE 19.1 (Continued) Team Competencies

Copyright American Psychological Association. Not for further distribution.

Attribute

Definition

Related and subsidiary constructs

Validation and measurement issues

Effect sizes

Assertiveness

Capacity of team members to communicate effectively by sharing ideas clearly and directly in interpersonal situations

Task-related assertiveness, a component of extraversion

Can be assessed with paper and pencil tests, but behavioral measures are better. Some validation data exists (Pearsall & Ellis, 2006; SmithJentsch et al., 1996).

.30 (Pearsall & Ellis, 2006)

Mutual performance monitoring factors

Ability of team members to accurately monitor and assess the work of others; ability to give, seek, and receive taskclarifying feedback in a constructive manner and to offer advice

Acceptance of suggestions and criticism; ability to give suggestions and criticism; ability to provide intrateam feedback; ability to monitor and give feedback; cross checking ability; error correction ability; team maintenance ability

Best assessed through a combination of paper and pencil and behavioral measures. Has been linked to team performance (e.g., Marks & Panzer, 2004).

.42 (Marks & Panzer, 2004)

Communication factors

Ability to clearly and accurately articulate and exchange information among team members, using accepted terminology; ability to acknowledge receipt of information; ability to clarify message when needed

Active listening ability, ability to exchange information, ability to engage in closed-loop communication, ability to share information, ability to engage in open exchange and to consul with others

Best assessed through a combination of paper and pencil and behavioral measures. Closedloop communication has been shown to predict teamwork (e.g., see Bowers et al, 1998; Mesmer-Magnus & DeChurch, 2009).

.11 to .45 (MesmerMagnus & DeChurch, 2009)

Cross-boundary factors

External, task-related actions directed to other teams or the larger organizational context

Organizational awareness, organizational resourcefulness, ability to build relationships with other teams

Paper and pencil measure developed by Druskat & Kayes (1999).

Insufficient data to compute.

Has been assessed with paper and pencil measures. Some evidence to suggest that a collective orientation leads to better teamwork (Driskell & Salas, 1992) and that those who enjoy working in a team engage in less social loafing (Stark et al., 2007) and have better team performance (Helmreich & Foushee, 1993).

.14 (Driskell & Salas, 1992: Stark et. al., 2007)

Attitudes Preference for teamwork

Inclination and desire to be part of a team, willingness to engage with other people in pursuit of task success, appreciation for the importance of teamwork in accomplishing challenging tasks

Team or collective orientation, importance of teamwork; appreciation for teamwork; desire to work in a team; collectiveness; preference for teamwork

(continued)

609

Cannon-Bowers and Bowers

TABLE 19.1 (Continued) Team Competencies Attribute

Copyright American Psychological Association. Not for further distribution.

Self-efficacy for teamwork

Related and subsidiary constructs

Definition Degree to which individuals believe that they have the requisite knowledge, skills, and other attributes to be successful team members

Teamwork self-efficacy

Validation and measurement issues Has been measured with paper and pencil measure (e.g., McClough & Rogelberg, 2003). Some data supports the link to effective teamwork (e.g., Tasa, Seijts, & Taggar, 2007).

Effect sizes .25 (Tasa et al., 2007)

Note. KSA = knowledge, skills, and abilities. From Handbook of Employee Selection (pp. 806–808), by J. L. Farr and N. T. Tippins (Eds.), 2010, New York: Routledge. Copyright 2010 by Routledge. Adapted with permission.

TABLE 19.2 Types of Team Competencies Relation to task Relation to team

Specific

Generic

Specific

Context-driven competencies: Description: Specific to both the task and team; when a task is complex and requires team members to respond quickly to changing demands, team members must hold competencies that are tailored to both the specific team members involved and task at hand. These represent the most extreme form of teams for which intimate knowledge of both the task and team are required. Examples: Sports teams, combat teams. Implications: Best developed through practice in realistic task environments with actual team members; not good candidates for selection because they are highly dependent on the particular team members.

Team-contingent competencies: Description: Specific to the team but not to the task, these are teams whose members work together across a variety of tasks. Examples: Functional teams who stay together across projects; management teams; self-managed work teams. Implications: Best trained with actual team members across a variety of tasks; not good candidates for selection since they are highly dependent on the particular team members.

Generic

Task-contingent competencies: Description: Specific to the task but not to the team. In many organizations, team membership cannot remain intact; instead, team members must perform the same task with different teammates. Examples: Cockpit crews, surgical teams. Implications: Best trained in the realistic task environment with actual or ad hoc teammates. May be useful in selecting new members.

Transportable competencies: Description: Generic to both the task and team. In situations where task interdependence is relatively low and team members have time to negotiate mutually beneficial performance strategies, transportable competencies may be sufficient. In addition, it is often the case in organizations that teams are called on to work on a variety of tasks with a variety of teammates. Examples: Task forces, advisory groups, temporary project teams. Implications: Can be trained across a variety of tasks and team members. May be useful in selection. Can be considered a precursor to more specific team training.

610

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

specific, this implies that practice must occur with the actual teammates because performance depends, at least in part, on team members’ knowledge of one another. It also implies that team-specific competencies are not good candidates for selection criteria because they depend on the particular characteristics of the team. However, it might be possible to select for task-specific competencies because potential employees may have experience in a similar task environment (e.g., surgical nurses who have prior experience in surgical teams). In addition, taskspecific competencies are going to be best trained in a setting that resembles the actual task environment, for example, by using simulation. Cannon-Bowers

et al. (1995) provided a number of propositions along these lines. We will continue the discussion of team training in subsequent sections. In other work, Campion and his colleagues (McClough & Rogelberg, 2003; Stevens & Campion, 1994; Stevens & Campion, 1999) summarized various literatures and conceptualized teamwork KSAs as falling into several categories. These include: interpersonal KSAs, which can be broken down further into conflict resolution KSAs, collaborative problem solving KSAs, and communication KSAs; and self-management KSAs, including goal setting and performance management KSAs, and planning and task coordination KSAs (see Exhibit 19.2). These

Exhibit 19.2 Knowledge, Skills, and Abilities (KSAs) Requirements for Teamwork Interpersonal KSAs Conflict Resolution KSAs 1. The KSA to recognize and encourage desirable, but discourage undesirable, team conflict. 2. The KSA to recognize the type and source of conflict confronting the team and to implement an appropriate conflict resolution strategy. 3. The KSA to use an integrative (i.e., win–win) negotiation strategy rather than the traditional distributive (i.e., win–lose) strategy.

Collaborative Problem-Solving KSAs 1. The KSA to identify situations requiring participative group problem solving and to use the proper degree and type of participation. 2. The KSA to recognize the obstacles to collaborative group problem solving and implement appropriate corrective actions. Communication KSAs 1. The KSA to understand communication networks, and to use decentralized networks to enhance communication where possible. 2. The KSA to communicate openly and supportively, that is, to send messages which are: (a) behavior- or event-oriented, (b) congruent, (c) validating, (d) conjunctive, and (e) owned. 3. The KSA to listen nonevaluatively and to appropriately use active listening techniques. 4. The KSA to maximize consonance between nonverbal and verbal messages and to recognize and interpret the nonverbal messages of others. 5. The KSA to engage in ritual greetings and small talk, and a recognition of their importance. Self-Management KSAs Goal Setting and Performance Management KSAs 1. The KSA to help establish specific, challenging, and accepted team goals. 2. The KSA to monitor, evaluate, and provide feedback on both overall team performance and individual team member performance.

Planning and Task Coordination KSAs 1. The KSA to coordinate and synchronize activities, information, and task interdependencies between team members. 2. The KSA to help establish task and role expectations of individual team members and to ensure proper balancing of workload in the team.

Note. From “The Knowledge, Skill, and Ability Requirements for Teamwork: Implications for Human Resource Management,” by M. Stevens and M. Campion, 1994, Journal of Management, 20, p. 514. Copyright 1994 by Sage. Adapted with permission.

611

Cannon-Bowers and Bowers

Copyright American Psychological Association. Not for further distribution.

researchers developed a situational judgment test to measure the degree to which individuals possess each of the KSAs. As such, they are best thought of as knowledge competencies because even the skills and abilities are measured cognitively rather than more behaviorally. For this reason, the Campion and Stevens competencies are represented in Exhibit 19.2 as “knowledge of teamwork skills.” Subsequent validation efforts have demonstrated the efficacy of this test in predicting teamwork performance (Campion & Stevens, 1999; McClough & Rogelberg, 2003). ISSUES IN SELECTING TEAM MEMBERS As noted in Table 19.1, Stevens and Campion (1994) argued that understanding teamwork competencies is an essential step in developing selection systems for teams. In fact, the issue of team selection involves a number of important questions. These include the following: ■ ■

■

■

What makes a good team member? Which factors should be considered when selecting team members? What is the best composition of team member knowledge and skill? How does diversity play into effective teamwork?

The first two of these issues—what makes a good team member and what team-related factors should be considered in selecting team members—were addressed, at least in part, in the sections on competencies. That is, the requirements for effective teamwork discussed in those sections form the basis for specifying which competencies should be sought in the applicant pool. However, recent work suggests that in addition to individual KSAs, other issues may affect selection in teams. These include personality variables and individual differences as well as team composition. We cover these issues in the following sections.

Personality and Individual Differences Interest in determining the contribution of personality factors to work performance began a more than a decade ago with seminal work into the structure of personality (e.g., Barrick & Mount, 1991; Costa & McCrae, 1992; Hough, 1992; McCrae & Costa, 1996), 612

with some success. That work led to the establishment of the so-called Big Five, which refers to the set of personality factors that appear to predict job performance. These include: conscientiousness, extraversion, neuroticism (emotional stability), agreeableness, and openness to experience. Recently, several researchers have attempted to generalize these findings to the team level. Table 19.3, an expansion of the work of Mohammed et al. (2010), summarizes the findings related to personality variables and team performance, including effect sizes where available. Indeed, several researchers have made the argument that personality factors will affect teamwork (e.g., Barry & Stewart, 1997; Driskell & Mullen, 2006; Kichuk & Weisner, 1997; Mohammed, Mathieu, & Bartlett, 2002), a contention that has been supported in recent meta-analytic reviews (Bell, 2007; Morgeson, Reider, & Campion, 2005; Peeters, Van Tuijl, Rutte, & Reymen, 2006; Prewett, Gray, Stilson, Rossi, & Brannick, 2009). Moreover, at least one meta-analysis found that measures of personality (i.e., conscientiousness, extraversion, agreeableness, emotional stability) added incrementally to the prediction of contextual performance when used in conjunction with team knowledge (vis-à-vis Stevens & Campion, 1994; as assessed by a situational judgment test) and social skills (i.e., assessed in a structured interview), prompting the authors to conclude that “these constructs are uniquely important for performance in teamwork settings” (Morgeson et al., 2005, p. 602). Clearly, including selected personality measures into a selection system for team-based work seems justified.

Team Composition Thus far, we have presented the discussion of predicting performance using individual attributes (i.e., both KSAs and personality variables) as if it were simply an application of individual-level selection methods to the team level. However, the situation is far more complicated in a team setting because it is not only the individual attributes that lead to effective or ineffective performance but also the unique manner in which these attributes are configured across team members (Levine & Moreland, 1990). Hence, the composition of the team is considered an important contributor to team effectiveness, and past researchers have debated

Team Development and Functioning

TABLE 19.3 Personality Variables Associated With Effective Teamwork Definition

Related and subsidiary constructs

Validation and measurement issues

Conscientiousness

Extent to which a person is self-disciplined and organized

Need for achievement; ambition; responsibility; dependability

Has been assessed with paper and pencil measures. Positively related to contextual performance in team settings (Morgeson et al., 2005) and team performance in field settings (Bell, 2006; Peeters et al., 2006).

.04 to .30 (Bell, 2007), .25 (Morgeson et al., 2005)

Extraversion

Extent to which an individual is social, outgoing, and talkative

Enthusiasm, optimism, assertiveness, dominance, gregariousness

Has been assessed with paper and pencil measures. Positively related to contextual performance in team settings (Morgeson et al., 2005); relationship small to moderate.

.04 (Peeters et al., 2006) to .15 (Bell, 2007); .18 (Morgeson et al., 2005)

Agreeableness

Extent to which an individual is gentle and cooperative

Likeability, interpersonal facilitation, trust worthiness, tolerance, courteousness

Has been assessed with paper and pencil measures. Small, positive relationship to contextual performance in team settings (Morgeson et al., 2005) and team performance in field settings (Bell, 2006; Peeters et al., 2006)

.24 (Peeters et al., 2006) to .31 (Bell, 2007); .12 (Morgeson et al., 2005)

Emotional stability

Extent to which an individual is calm and poised

Neuroticism (negative relationship), adjustability, lack of nervous tendencies, lack of anxiousness, security

Has been assessed with paper and pencil measures. Positively (but only marginally) related to contextual performance in team settings (Morgeson et al., 2005). Small, positive relationship in field studies (Bell, 2006).

.04 (Peeters et al., 2006) to .21 (Bell, 2006); .15 (Morgeson et al., 2005)

Openness to experience

Degree to which an individual is willing to experience new things

Has been assessed with paper and pencil measures. Positively related to team performance in field settings (Bell, 2006); not supported in all studies.

.03 (Peeters et al., 2006) to .20 Bell (2006)

Copyright American Psychological Association. Not for further distribution.

Attribute

Effect sizes (r )

Note. From Handbook of Employee Selection (pp. 806–808), by J. L. Farr and N. T. Tippins (Eds.), 2010, New York: Routledge. Copyright 2010 by Routledge. Reprinted with permission.

the best way to represent such team-level constructs (e.g., see Mohammed et al., 2010). It should be noted that this discussion is highly dependent on the way team effectiveness is defined and measured; we address this issue in a subsequent section. The crux of the debate over how to best compose a team is related to whether the team-level construct is represented as the mean of individual scores or

more representative of the distribution of scores (e.g., standard deviation). Conceptually, these are important distinctions. On the one hand, using the mean score suggests that more of the trait (collectively) is better (or worse) for team performance. On the other hand, using a measure of distribution suggests that the combination of a trait across members is important. It is also logical to consider the minimum 613

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

or maximum scores on a trait (collectively across members). The latter suggests a compensatory model; in other words, as long as the trait is well represented in one member, it does not need to be present in others. Even more complex is the notion of a threshold model, where presence of the trait is beneficial to a point, but not in excess. This might be the case for extraversion, where having one or two extraverted members may be effective but having all team members high on extraversion may actually disrupt overall team performance. According to several authors, (Bell, 2007; Mohammed et al., 2010) the task typology offered by Steiner (1972), which was described earlier, is perhaps the most popular basis to determine which aggregation method is best. For additive tasks, using the mean (or simply summing trait levels) is appropriate because team output is the sum of each individual team members’ contributions. Using the mean or sum is also best for compensatory tasks because higher performing members can compensate for the low output of poorer performing members. When considering conjunctive tasks, the minimum score is considered appropriate because performance depends on the weakest member; whereas for disjunctive tasks, the maximum score is best because it reflects the highest performer on the team. Strict reliance on a task-based model such as Steiner’s (1972) has been criticized on conceptual and empirical grounds (Day, Arthur, Miyashiro, Edwards, & Hanson, 2004; Hollenbeck, DeRue, & Guzzo, 2004). Other approaches to this issue advocate deeper consideration of the trait itself. For example, Kozlowski and Klein (2000) suggested that researchers should consider how the variable might manifest itself at the team level as a guide to aggregation method. To complicate matters, besides mean, standard deviation, minimum and maximum, even more complex manifestations may also be necessary. Hence, configural models, which specify more complex interactions among variables, may be appropriate. For example, one disagreeable member could disrupt team performance disproportionally or, as Barry and Stewart (1997) found, a curvilinear relationship between extraversion and team performance may hold such that extremes of low or high are detrimental to team performance. 614

Several recent meta-analyses have attempted to address this issue, but with limited success. First, Bell (2007) found that the method of operationalizing the composition variables moderated the relationship between the variables and team performance. Further, Bell found little support for the Steiner typology. Instead, she concluded that “no single operationalization was best for all composition variables; rather, the best operationalization was dependent on the specific team composition variable of interest” (Bell, p. 607). In a second meta-analysis, Peeters et al. (2006) used theoretical arguments (and associated findings) to generate hypotheses for extraversion, agreeableness, conscientiousness, and emotional stability based on their prediction regarding whether the elevation of the trait (i.e., mean or sum) or variability of the trait would have a greater impact on team performance. They found, as expected, that higher elevations in agreeableness and conscientiousness (but not variability) were related to team performance; but extraversion, openness, and emotional stability (i.e., elevation or variability) were not. Finally, Prewett et al. (2009) included the type of criterion (i.e., team behavior vs. team outcomes) and type of workflow patterns (i.e., pooled, reciprocal, intensive) in their meta-analysis to test for moderating effects. Among the conclusions drawn by these authors is that team personality variables seem to be more closely related to team behaviors (i.e., processes) than to more distal outcomes. Second, they found that the pattern of workflow does moderate the personality-performance relationship in the team, specifically, that situations requiring intensive (i.e., interdependent) workflow were more closely related to personality. Their conclusions regarding the superiority of task versus trait-based methods of aggregation were equivocal. They offer some insights as to why this may have been the case, including the correlation among measures, the reliability of measures and the role of the team members (i.e., some roles may require higher or lower degrees of a trait). One variable that seems to be important regardless of how it is aggregated is cognitive ability (Bell, 2007; Devine & Phillips, 2001; Stewart, 2006). The relationship seems to hold across task types and “has the strongest, most consistent positive relationship with team performance” (Stewart, p. 45).

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

Related to this, Stewart also concluded that member expertise has a small, positive relationship with team performance. A final consideration with respect to individual differences and team performance has to do with combined (i.e., interactive) effects. In a field study of intact military teams, Halfhill, Nielsen, Sundstrom, and Weilbaecher (2005) hypothesized that agreeableness and conscientiousness would be associated with higher group performance. These authors found that mean and minimum agreeableness and mean and minimum conscientiousness were associated with better performance and that variability in agreeableness (but not conscientiousness) correlated with better performance. Perhaps more interesting, they also found that the interaction of agreeableness and conscientiousness was significantly related to performance, such that those groups highest on both traits outperformed all others. Halfhill et al. concluded that there might be a possible synergy among complementary personality traits on team performance.

Diversity and Heterogeneity Somewhat in parallel to investigating the issue of how to aggregate personality characteristics in composing teams, a great deal of attention has been paid to the impact of heterogeneity (i.e., trait variability) of team members on team performance. (See also chap. 20, this volume.) Indeed, a popular topic in modern organizational settings is diversity, which typically refers to differences in social-category variables such as ethnicity, gender and age. Traditionally, both sides of the heterogeneity and team performance argument have been made. On the one hand, heterogeneity has been hypothesized to facilitate team performance by promoting creativity and knowledge sharing across diverse perspectives. On the other hand, heterogeneity has also been assumed to be detrimental to team performance by disrupting communication and cohesion, and increasing conflict (Mohammed et al., 2010). In summarizing the literature regarding demographic diversity, Hollenbeck et al. (2004) concluded that the preponderance of findings do not support the contention that heterogeneity (or diversity) has a reliable impact on team performance. This conclusion is bolstered by results of several recent meta-analyses including Bell (2007); Bowers, Pharmer, and Salas

(2000); Horowitz and Horowitz (2007); Stewart (2006); and Webber and Donahue (2001). This is not to say that heterogeneity is not important; indeed, it may be more reflective of the fact that other variables, such as task type, time, attitudes, and the type of diversity being considered may affect its relationship to team performance. For example, Harrison, Price, Gavin, and Florey (2002) found that over time, the impact of demographic diversity was diminished. Similarly, a recent meta-analysis by Bell, Villado, Lukasik, Briggs, and Belau (2007) showed that race and gender diversity effects were lowered over time (i.e., the longer the team was together). Bowers et al. (2000) concluded that the impact of heterogeneity varied as a function of task type. This meta-analysis demonstrated that composing teams of similar individuals might improve performance in low-difficulty tasks but decrease performance on high-difficulty tasks. Other researchers have shown that the specific type of diversity being considered (i.e., demographic vs. functional or task-related) has differential effects on team task performance (Pelled, Eisenhardt, & Xin, 1999). Specifically, functional diversity (i.e., functional background, tenure) was more likely to be associated with task conflict, whereas demographic (i.e., age, race, gender) diversity was more likely to lead to emotional conflict. Moreover, task conflict was found to enhance performance whereas emotional conflict had the opposite effect. A metaanalysis conducted by Horowitz and Horowitz (2007) found similar results; specifically, task-related diversity was positively related to quality and quantity of team performance, whereas demographic diversity was unrelated to performance. Recently, Homan, van Knippenberg, Van Kleef, and De Dreu (2007) investigated how team members’ attitudes toward diversity affected team performance. As expected, these researchers found that informational diversity had a positive impact on performance only when team members held prodiversity beliefs as opposed to pro-similarity beliefs. In this case, pro-diversity and pro-similarity beliefs were manipulated through experimenter induction prior to performance. Collectively, the results of this study and others cited earlier indicate that the manner in which 615

Cannon-Bowers and Bowers

individual attributes best combine to enhance team performance is complex and depends on a number of mediating and intervening factors. In fact, the recent meta-analysis by Prewett et al. (in press) confirmed that the pattern of workflow in the team moderates the relationship between personality and team outcomes. Clearly, further research is needed to more fully elucidate these important relationships.

Copyright American Psychological Association. Not for further distribution.

Team Size Another feature that has been of some interest to team researchers interested in maximizing performance is team size. Conceptually, it seems reasonable to hypothesize that too few team members can result in excessive workload, whereas a team with too many members may become unwieldy. In either case, team performance will suffer. Obviously, the type of task the team must perform will have a direct impact on how big the team needs to be. Hence, the question becomes, what is the optimal team size given the nature of the team task demands? To date, attempts to answer this question have not been terribly successful (Mohammed et al., 2010). This is perhaps because the answer is so task dependent. For example, despite finding that project teams benefited from more members, Stewart (2006) concluded that a “clear prescription for optimum team size is thus difficult and appears to depend on the purpose and responsibilities of the team” (p. 45). Obviously, additional work is needed to better address the team size issue, particularly in light of efforts to reduce the size of the workforce in many organizations. TEAM PROCESSES According to the definitions offered earlier, teams are required to coordinate their actions to accomplish their task or mission. This necessitates a variety of personal interactions among the members of the team. These interpersonal behaviors have been referred to broadly as team processes. A variety of team process variables have been discussed in the literature and there have also been a number of taxonomies proposed to organize them (e.g., CannonBowers et al., 1995; Marks et al., 2001; Rousseau, 616

Aubé, & Savoie, 2006; Salas, Sims, & Burke, 2005). Of these, perhaps the best validity data are available for the model posited recently by Marks et al. For example, a recent manuscript by LePine, Piccolo, Jackson, Mathieu, and Saul (2008) evaluated the Marks et al. (2001) model using confirmatory factor analysis and subsequent meta-analysis. The results demonstrated that the three-factor model proposed by Marks et al. fit the data better than competing models. Subsequent meta-analytic techniques demonstrated that the three factors were positively related to team performance, with effect sizes averaging .29 for each of the three factors. Give the strength of this support, we use the Marks et al. model to organize the review of teamwork processes. This approach describes three dimensions of teamwork behaviors: transition processes, action processes, and interpersonal processes (see Figure 19.2). Each dimension, and its incumbent process behaviors, is reviewed briefly next.

Transition Behaviors One cluster of team-process behaviors is called transition behaviors. According to Marks and her colleagues (2001), this dimension includes behaviors related to “when teams focus primarily on evaluation and/or planning to guide their accomplishment of a team goal or objective” (p. 364). Two team processes within the cluster have received substantial research attention: mission analysis and planning. Mission analysis. Mission analysis is “the interpretation and evaluation of the team’s mission, including identification of its main tasks as well as the operative environmental conditions and team resources available for mission execution” (Marks et al., 2001, p. 365). This dimension includes selfappraisal behaviors as well as mission appraisal behaviors. The self-appraisal behaviors allow the teams to consider their previous performance, diagnose shortcoming, and discuss approaches to remediate them before engaging the new task. This is akin to the notion of “team self-correction” (to be discussed further later; Blickensderfer, CannonBowers, & Salas, 1997). These behaviors have been shown to be positively related to team performance. For example, Smith-Jentsch, Cannon-Bowers, Tannenbaum, and Salas (2008) reported that a

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

FIGURE 19.2. Summary of team processes in transition and action phases. From “A Temporally Based Framework and Taxonomy of Team Processes,” by M. A. Marks, J. E. Mathieu, and S. J. Zaccaro, 2001, Academy of Management Review, 26, p. 361. Copyright 2001 by the Academy of Management. Adapted with permission.

guided self-correction intervention was related to both more accurate mental models and better team performance in a study using Navy commanders (r = .57). Mission analysis also involves “forward visioning” and the development of a shared mental model of the team’s goals. Developing a shared understanding of team goals has been suggested as an important determinant of subsequent team performance (McComb & Green, 1999). However, this effect requires additional empirical study. Goal specification and strategy formulation. Another critical transitive process is goal specification and strategy formulation. The process behaviors in this dimension are associated with creating and articulating goals for the team and the strategy that will be used to accomplish them. The benefits of an effective goal specification process have been reported in many settings. For example, Senecal, Loughead, and Bloom (2008) reported that a shared

goal-setting intervention was related to improved team cohesion in a longitudinal study of a sports team. Similar positive relationships were noted in distributed project teams (Forester, Thoms, & Pinto, 2007), performance appraisal teams (Resick & Bloom, 1997), and computerized business simulations (Fandt, Richardson, & Connor, 1990). Although this effect appears consistent and of reasonable magnitude (i.e., effect size was .25 in Forester et al.), the mechanisms that underlie observed performance improvements may be mediated by such things as improved states, including cohesion and shared mental models (discussed in a later section). Marks and her colleagues (2001) also emphasize the importance of a variety of strategy formulation behaviors. These behaviors involve three types of team planning behaviors: creating plans for an initial course of action; contingency planning for likely disruptions; and reactive, on-the-fly planning 617

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

in reaction to novel stimuli. This dimension also includes planning the specific roles and activities of the team members. Several authors have discussed the importance of initial and contingency planning. For example, Patrick, James, and Ahmed (2006) reported that team planning behaviors were an important predictor of subsequent team situation awareness. McClennan, Holgate, Omodei, and Wearing (2006) reported similarly positive results between planning and effectiveness of firefighting teams. Finally, Stout, Cannon-Bowers, Salas, and Milanovich (1999) described a well-controlled laboratory study that also provides support for the notion that initial planning behaviors are related to improved team process and performance (r = .58). Less empirical study has been dedicated to the issue of backup planning. However, a recent study by DeChurch and Haas (2008) may shed some light on the importance of this type of planning. These researchers used a laboratory “scavenger hunt” task specifically designed to evaluate the types of planning described by Marks and her colleagues (2001). Their data provided evidence that the three types of planning are largely nonoverlapping and each adds predictive value at varying phases of the team’s performance (change in r2 = .10). In fact, in early phases of performance, contingency planning was the best predictor of subsequent team coordination during task performance (r = .24). Yet another aspect of strategy formulation is the creation and communication of roles within the teams. As noted earlier, several authors (e.g., Hackman, 1987; Sundstrom et al., 1990) have described the importance of team member roles in teamwork. Despite the frequency with which the construct of roles is invoked, there is relatively little empirical research to guide interventions to support team role clarity (Mumford, Morgeson, Van Iddekinge, and Campion, 2008). However, the available evidence seems to support the hypothesis that increased role clarity is associated with improved team processes and performance. For example, Volpe, Cannon-Bowers, Salas, and Spector (1996) demonstrated that an intervention designed to improve team members’ interpositional knowledge was associated with better performance in a simulated combat aviation task. Cannon-Bowers 618

and her colleagues later replicated these effects, with a reported effect size of .34 (Cannon-Bowers, Salas, Blickensderfer, & Bowers, 1998). Kraut, Fussell, Lerch, and Espinosa (2004) reported similarly positive results with teams participating in a business simulation. Finally, in validating their new test of team role knowledge, Mumford et al. (2008) investigated the relationship between knowledge of team roles and subsequent team performance (i.e., measured as rated performance on academic projects). They found that role knowledge was an effective predictor of team performance over and above well-known predictors such as member personality and cognitive ability (r = .39). Further, they demonstrated this relationship in both laboratory and work settings. It should be noted that these planning behaviors are not orthogonal. For example, Ellis Bell, Ployhart, Hollenbeck, and Ilgen (2005) demonstrated that planning about team member roles was related to better initial mission planning. It is probable that these planning behaviors also interact with several of the other process behaviors in this section. However, the nature of these interactions is not well known. It is likely that discovering the nature of these interactions will be a topic of future research as scientists pursue ways to increase team training effectiveness.

Action Behaviors The second category of process behaviors discussed by Marks et al. (2001) are action behaviors or “activities leading to goal accomplishment” (p. 366). Within the dimensions of action behaviors, there are several smaller clusters of process behaviors. These smaller clusters are discussed next. Monitoring behaviors. Marks and her colleagues (2001) described several types of monitoring behaviors. These include monitoring of progress toward goals, which refers to behaviors that support “tracking task and progress toward mission accomplishment, interpreting system information in terms of what needs to be accomplished for goal attainment, and transmitting progress to team members” (Marks et al., 2001, p. 366). They also described systemmonitoring behaviors. These behaviors facilitate the following factors:

Team Development and Functioning

Copyright American Psychological Association. Not for further distribution.

Tracking team resources and environmental conditions as they relate to mission accomplishment; and involve (1) internal systems monitoring, i.e., tracking team resources such as personnel, equipment, and other information that is generated or contained with the team, and (2) environmental monitoring, i.e., tracking environmental conditions relevant to the team. (p. 367) These types of monitoring behaviors are discussed frequently in descriptions of best practices in team performance (e.g., Gaddy & Wachtel, 1992; Weisband, 2002); however, evaluating the empirical support for them is difficult because these behaviors are rarely studied in isolation. Rather, they are typically studied in the context of “emergent states” such as team situational awareness, as opposed to team performance itself. In that regard, however, the importance of these monitoring behaviors seems well supported (e.g., Prince & Salas, 2000; Salas, Prince, Baker, & Shresta, 1995). For example, in an analysis of aviation incidents and accidents (reported in the National Transportation Safety Board accident database), a lack of monitoring behaviors appeared to be associated with an increased likelihood of accidents attributed to lapses in situational awareness (Jentsch, Barnett, Bowers, & Salas, 1999). Marks and Panzer (2004) found a positive relationship between monitoring behaviors and team coordination in a similar task. Blandford and Wong (2004) also reported that these behaviors are associated with effective behaviors in emergency medical dispatch teams. Reported effect sizes have been relatively large (e.g., .43; Marks & Panzer). Backup behaviors. Marks et al. (2001) also described a set of team backup behaviors within the action processes cluster. These are behaviors performed with the intention of “assisting team members to perform their tasks which may occur by (1) providing a teammate verbal feedback or coaching, (2) assisting a teammate behaviorally in carrying out actions, or (3) assuming and completing a task for a teammate” (Marks et al., 2001, p. 367). Backup behaviors are often considered be a critical behavior in high-performing teams. In fact, Salas,

Sims, and Burke (2005) included it in their Big Five behaviors that are critical in teamwork. Ideal backup behavior would occur when team members observe when their teammates are overloaded or experiencing some other factor that decreases their performance. The members would then redeploy their own resources to assist the struggling team member. Despite the widely held opinion that this is a critical team characteristic, there is little empirical data with which to evaluate the claim. DeChurch and Haas (2008) found that reactive adjustment planning (and, one would presume, the execution of those plans) was related to positive team performance, with relationships ranging between .15 and .44. Porter, Hollenbeck, Ilgen, Ellis, and West (2003) and Porter (2005) also demonstrated that backup behaviors seem to be related to a team’s ability to cope with high workload situations. However, other researchers offer some cautions about the use of backup behaviors. Specifically, Barnes et al. (2008) pointed out that the execution of backup behaviors uses cognitive resources that could be applied to the core task. Further, they expressed concerns that these behaviors might encourage negative social behaviors such as social loafing or dependence. They conducted a study designed to assess these potential negative consequences using a computerized simulation. Consistent with their fears, Barnes et al. found that the frequency of backup behaviors provided by team members was related to incidences of them neglecting their own taskwork. Further, the provision of backup behaviors was related to subsequent decreases in effort by the member receiving the help. As such, Barnes and his colleagues suggested that scientists exercise more caution before recommending the training or encouragement of these behaviors in all teams. Coordination behaviors. Within the dimension of action behaviors, coordination is another group of behaviors related to the manner in which the team organizes and shares its resources among members to accomplish the tasks. A key behavior within this group is team communication. Communication is a critical behavior used to exchange information among members. Moreover, the importance of communication is particularly salient in more interdependent 619

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

teams where there is an increased need to share information (Tesluk, Mathieu, Zaccaro, & Marks, 1997). Researchers have studied the nature of different kinds of team communication processes in an attempt to identify elements most closely associated with effectiveness. For example, Achille, Schulze, and Schmidt-Nielson (1995) reported that encouraging teams to use standardized communication terms actually led to wordier communications, opposite from what was intended. PatrashkovaVolzdoska, McComb, Green, and Compton (2003) investigated the link between overall communication and team performance. It is of interest to note that a curvilinear relationship was found with the two extremes of communication frequency associated with poorer performance. Too few communication episodes may not allow sufficient information to be passed. Conversely, overabundant communication may add so much workload to the team that it detracts from performance (McMillan, Entin, & Serfaty, 2004). Other authors have discussed the importance of specific speech types within the broader team communications events. For example, Svensson and Andersson (2006) reported that communication types such as metacommunication and tactics were observed more often in winning teams. Urban, Bowers, Monday, and Morgan (1995) reported that effective teams asked fewer questions and fewer “implied questions” than did ineffective teams. Others have noted that effective teams tend to make fewer explicit requests for assistance during stressful or high-workload periods (Kleinman & Serfaty, 1989; Manser, Howard, & Gaba, 2008). Another approach to investigating communication behaviors in teams has been to analyze the patterns of information flow. For example, Bowers and Jentsch (1998) used a variety of sequential analysis techniques in an attempt to discriminate between good and poor teams. They found that effective teams were more likely to use “closed-loop” communication patterns during the execution of a task, whereas poorer teams were prone to insert extraneous information while processing problems. Patrick et al. (2006) used a process-tracing approach to study supervisory teams. This analysis indicated that the generation of hypotheses and the discussion of performance 620

progress were key discriminating behaviors. Lingard and colleagues (2004) used pattern analysis to study the flow of communication events in surgical procedures. One interesting finding is that periods of conflict seemed to lead to reduced contributions by younger members of the team. Clearly, the need to understand the nature of effective communication is a key challenge to team performance researchers. It has been suggested that the lack of progress in this regard is due to the difficulty and expense associated with this kind of research. However, recent advances in automatic discourse analysis hold great promise in facilitating this type of research in the future (see Foltz & Martin, 2008, for a review of these approaches). Although not specifically mentioned by Marks et al., (2001), researchers have also described the importance of coordination behaviors that are not communication-based. Several authors have described the potential importance of nonverbal communication in teams as an important form of coordination (Fowlkes, Lane, Salas, Franz, & Oser, 1994; Harris & Sherblom, 2002; Stevens & Campion, 1994). However, there has been little empirical work to establish which nonverbal behaviors under which conditions are related to team performance. Fowlkes and her colleagues included nonverbal behaviors in their Targeted Acceptable Responses to Generated Events or Tasks (TARGETS) team measurement approach. Although overall scores on this measure have been related to team performance, there are no reported data that allow one to specifically evaluate particular nonverbal behaviors, such as gestures or facial expressions. Other scales used in team performance also combine verbal and nonverbal factors into one construct, making it impossible to discern their individual contributions (see Prince, Brannick, Prince, & Salas, 1997). Cole, Walter, and Bruch (2008) indicated that nonverbal behaviors, in this case, expressions of negative emotions, were a mediating variable in the relationship between overall negative affective tone and subsequent team performance. Clearly, this is an area in which more research is required before meaningful guidance can be provided to the practice community.

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

One coordination behavior associated with nonverbal communication is implicit coordination. Kleinman and Serfaty (1989) coined the term implicit coordination to describe a phenomenon that they observed in a study of team performance. Under low workload, effective performance occurred when teams made clear requests for assistance (i.e., explicit coordination). However, under conditions of high workload, effective teams adjusted their behavior so that they began to provide needed resources without an explicit request. This pattern, described as implicit coordination, has since been used to describe the provision of resources of all types in the absence of an explicit request. The construct of implicit coordination has been replicated in a few different settings (Serfaty, Entin, & Johnston, 1998) and is frequently used to describe processes executed by effective teams (e.g., CannonBowers & Salas, 1998; Fiore, Salas, Cuevas, & Bowers, 2003). Recently, Rico, Sánchez-Manzanares, Gil, and Gibson (2008) provided a number of propositions designed to facilitate the emergence of implicit coordination behaviors in teams. They also provided a set of research directions that should be helpful in designing further investigations of this important process.

Interpersonal Processes A final dimension of team process behaviors presented by Marks et al. (2001) is interpersonal processes. They point out that these behaviors are likely to co-occur with behaviors in the other dimensions. In fact, they may mediate the effectiveness of those behaviors. In any event, they are important in managing the affect that is likely to occur in teambased work. The three clusters of behaviors within this dimension are described later. Conflict management. The ability to manage the affective responses that are likely to emerge is a critical team-level skill. Marks and her colleagues (2001) pointed out that there are behaviors that are likely instrumental in preventing conflict whereas others might be effective in managing conflicts once they occur. The available literature supports the contention that these process behaviors can be helpful in promoting effective performance. Preventive behaviors, such as creating standard operating procedures

for managing stressful times, seem to be effective in avoiding negative teamwork behaviors (Smolek, Hoffman, & Moran, 1999). Similarly, the presence of effective conflict management behaviors has been associated with effective team performance (Porter & Lilly, 1996). As noted, it has been suggested that there may be a difference between emotional (i.e., interpersonal) conflict and task conflict on team functioning, and several studies show this effect. However, a metaanalysis by De Dreu and Weinart (2006) did not bear this out. They found the expected negative effect of interpersonal conflict on team performance (r = −.23) but no evidence that task conflict has a facilitative effect. Alper, Tjosvold, and Law (2000) suggested that one way to manage task-related conflict is to translate it into intrateam competition, which might become a motivational factor. Their data suggested that this type of interpretation is associated with more effective performance. Motivation and confidence building. Marks and her colleagues (2001) included a cluster of behaviors which involved generating and preserving a sense of collective confidence, motivation, and taskbased cohesion with regard to mission accomplishment (p. 368). These behaviors may work by creating emergent states such as collective efficacy (described elsewhere in this chapter). However, they may have their own, separate impact in some cases. For example, expressions of encouragement might be a mechanism for helping teams cope with stressors such as fatigue (Harville, Elliott, & Barnes, 2007). These types of behaviors might also be ways in which self-managed work teams reinforce positive behaviors (Pearce & Manz, 2005). However, as noted earlier, because these behaviors are often grouped within larger assessments of communication, it is difficult to identify the specific contribution of these behaviors to performance. Affect management. The final cluster of process behaviors described by Marks et al. (2001) is affect management behaviors. These are behaviors involved in “regulating member emotions during mission accomplishment, including (but not limited to) social cohesion, frustration, and excitement” (Marks et al., 2001, p. 369). This cluster includes a number 621

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

of specific teamwork behaviors, such as setting norms for behavior, calming stressed members, and increasing morale. Many of the behaviors are subsumed in the increasingly important construct of emotional intelligence. Several studies have recently reported positive relationships between team member emotional intelligence and team processes and performance (Côté & Miners, 2006; Offermann, Bailey, Vasilopoulos, Seal, & Sass, 2004). The percentage of variance accounted for in these studies was relative small, however, ranging from .10 to .18. Nonetheless, it is important to determine whether emotional intelligence can be improved through team training or some other means. Although theorists have suggested that team training should improve emotional intelligence (Druskatt & Wolff, 2001), there is little empirical data with which to confirm this hypothesis; it warrants future attention. OTHER EMERGENT STATES In addition to the processes described by Marks et al. (2001), several other issues related to team performance are worth discussing: team adaptability, shared mental models and cohesion. These are not processes (i.e., behaviors) under the Marks et al. definition, but cognitive and affective emergent states that have a direct impact on team performance. Like process behaviors, they are important to understand because they form the causal chain that links team, task, and situational demands to desired team outcomes.

Adaptability Teams are often used in complex, evolving environments. As teams attempt to do their work in these challenging settings, it becomes clear that one competency that is required is that of being able to deviate from a plan of action in response to changing events. This competency is referred to as team adaptability or team adaptation. This construct was defined by Cannon-Bowers and her colleagues (1995) as “the process by which a team is able to use information gathered from the task environment to adjust strategies through the use of compensatory behaviors and reallocation of intrateam resources” (p. 344). Kozlowski, Toney, Mullins, Weissbein, Brown, and 622

Bell (2001) added to this definition by emphasizing the importance of generalizing existing knowledge. Thus, team adaptability has important cognitive and behavioral elements. Several other definitions of team adaptability have been offered, but the concepts of reallocation of resources and generalization of knowledge seem to capture the essence of this construct (see Burke, Stagl, Salas, Pierce, & Kendall, 2006, for a review of these definitions). Theories of team adaptability. Although a number of theoretical positions are relevant to the concept of team adaptability, perhaps the most comprehensive of these was recently offered by Burke, Stagl, Salas, and colleagues (2006). This complex theory is illustrated in Figure 19.3. As illustrated in Figure 19.3, the foundation of the Burke, Stagl, Salas, et al. (2006) model is a four-stage adaptive cycle. The first stage in this cycle is situation assessment, which includes identifying the important cues in the environment and ascribing the appropriate meaning to them. Once the situation has been diagnosed, teams enter the second phase, plan formulation. It is in this phase that the team decides how best to reallocate its resources to cope with the new situation. The third phase, plan execution, refers to the behaviors the team performs to accomplish the new plan. For the most part, the activities in this phase are identical to any other team performance situation. However, there is an additional need to monitor the new plan and to react to any shortcomings as necessary. Finally, Burke and her colleagues describe team learning as the fourth stage in the adaptive cycle. This refers to the process of identifying errors and successes, discussing incorrect assumptions, and crystallizing knowledge gains for subsequent performance situations. Furthermore, as illustrated in the model, Burke, Stagl, et al. (2006) described the influence of 12 variables that they hypothesized will influence team adaptability. Each of these variables is thought to influence one or more of the stages of adaptability described previously. For example, individual characteristics, such as various knowledge, skills, and abilities, are thought to exert a particular influence on Stage 1 performance.

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

FIGURE 19.3. Model of team adaptability. From “Understanding Team Adaptation: A Conceptual Analysis and Model,” by C. Burke, K. Stagl, E. Salas, L. Pierce, and D. Kendall, 2006, Journal of Applied Psychology, 91, p. 1190. Copyright 2006 by the American Psychological Association.

Most of the 12 variables have been described in earlier sections and will not be reviewed again here. However, Burke, Stagl, et al. (2006) included variables that are not often considered in other theories of team performance. For example, they highlighted the importance of psychological safety in different stages of the adaptation process. Psychological safety refers to the degree to which team members feel safe taking risks in the team performance environment (Edmondson, 1999). Burke and her colleagues suggested that the feeling of safety is a determinant of the degree to which individual team members offer suggestions, criticize other member’s offerings, and provide their full pool of resources to the team. This construct is also theorized to play a role in execution of the agreed-on plan and the monitoring of its effectiveness. A second variable included by Burke, Stagl, et al. in their (2006) theory that warrants discussion is

self-management. They described self-management as a condition within job design where members of the team have some freedom in how they allocate their resources or arrange their tasking. Burke and her colleagues argued that the perception of selfmanagement influences the degree of responsibility members feel to create a positive outcome, the alternatives that team members consider, and the degree to which they feel empowered to further alter their actions as necessary. Preparing teams to be adaptable. Preparing teams to be adaptable presents a difficult challenge. One must take into account the large number of cognitive and behavioral competencies discussed previously even though, in many ways, some of the proposed training interventions run contrary to the “training culture” that currently exists in many modern organizations. For example, one key suggestion for training 623

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

team adaptability is that teams must be allowed to explore a variety of options, fail, learn from their failures, and so forth (Smith, Ford, & Kozlowski, 1997). However, most training situations value efficiency, leading to an emphasis on procedural training to a specified criterion. This type of training does not provide a context for developing these adaptability skills (Kozlowski et al., 2001). In addition to errors, several specific training approaches might foster team adaptability. Many current approaches emphasize the training of individuals in some of the competencies described earlier. For example, Spiro, Feltovich, Jacobson, and Coulson (1992) suggested that training individuals in cognitive flexibility might be a way to enable downstream team adaptability. Marks et al. (2001) provided data to suggest that team interaction training could also facilitate adaptive behavior in teams. Other researchers have emphasized the importance of identifying the cues that might necessitate adaptive behaviors (Entin & Serfaty, 1999; Martin-Milham & Fiore, 2004). Researchers have also discussed how to use team-level training to improve the adaptability of teams. For example, Klein and Pierce (2001) suggested that teams should be instructed in how to alter their own processes, beliefs, and systems to react to external cues. Other authors have recommended the use of simulations as a way to create situations where teams can practice their adaptability skills and receive feedback (Burke, Stagl, et al., 2006; Kozlowski & Bell, 2008).

Shared Mental Models and Shared Knowledge Another important emergent state in teams has been called shared mental models, shared knowledge, or team cognition. Regardless of the label, this construct refers to a crucial aspect of team functioning and is worthy of discussion here. In describing the interaction between humans and machines or systems, it is clear that people develop an understanding of how the machine or system works. This understanding, or mental model, as it has been called, guides how the individual interacts with the machine or system. According to Johnson-Laird (1983), mental models help people draw conclusions about how things work, deduce 624

the relationship between units, and predict outcomes. Since that time, the notion of mental models has been widely used to describe internal cognitive representations of complex systems (see Westbrook, 2006, for a review of the development of mental model theory). As noted, during the late 1980s, interest in team performance increased dramatically due in part to several well-publicized incidents that were attributed to faulty teamwork (e.g., the USS Vincennes shooting down of a commercial jet, the close call at Three Mile Island, and the Air Florida airline crash in Washington, DC). Given the need to quickly understand and improve the performance of these critical teams, researchers extended the construct of mental models to teams. Most notably, CannonBowers and her colleagues described a construct called shared mental models, which refers to knowledge that is common or shared across team members. Shared mental models (SMMs) allow team members to understand not only their work requirements but also to predict the needs and actions of their teammates (Cannon-Bowers, Salas, & Converse, 1993; Rouse, Cannon-Bowers, & Salas, 1992). Although the concept of shared mental models is widely used, definitional clarity is still lacking. For example, Carley (1997) discussed team mental models, which emphasize the similarities of knowledge types among team members. Others have invoked a metaphor of a shared cognitive structure among team members, such as a collective mind (Yoo & Kanawattanachai, 2001) or team mind (Klein & Thorsden, 1989). Yet others have described a process of group cognition that differs rather drastically (Akkerman, Admiraal, Simons, & Niessen, 2006). As pointed out by Cannon-Bowers and Salas (2001), this conceptual confusion may limit theoretical and empirical progress in this area. Moreover, it leads to different approaches for measuring shared mental models; we address this issue in the general section on team performance measurement next.

Empirical Research Into Shared Mental Models As noted, empirical research on SMMs has progressed rather slowly. The majority of research has

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

focused on attempting to establish the relationship between SMMs and team performance (e.g., Espevik, Johnsen, Eid, & Thayer, 2006; Marks, Sabella, Burke, & Zaccarro, 2002; Marks, Zaccaro, & Mathieu, 2000; Stout et al., 1999). For the most part, these studies have demonstrated a small to moderate relationship. The percentage of variance in team performance accounted for by SMMs in these studies (rs ranging from .30 to .39) is impressive when considering the other factors that likely contribute to team performance, such as individual ability, experience, and so forth. Several researchers have also pointed out that the degree of “sharedness” or similarity in mental models among members is not the only condition necessary for effective performance. For example, Rentsch and colleagues (e.g., see Rentsch & Hall, 1994) argued that the accuracy of mental models is as important as their degree of sharedness, because inaccurate mental models, even if shared among members, will lead to poor performance. Along these lines, Lim and Klein (2006) found that accuracy of team mental models explained a significant amount of variance in performance when added to similarity of models. This result was recently replicated and extended by Banks and Millward (2007), who also demonstrated that shared procedural knowledge was negatively related to team performance. Only accurate procedural knowledge was a positive predictor. Similarly, Smith-Jentsch, Mathieu, and Kraiger (2005) investigated the relationship between two types of SMMs and team performance. One type of mental model, cue-strategy associations, focused on elements of the task that had related, predictable actions. The other, positional-goal interdependencies, focused more on features of the team such as roles and responsibilities. It is interesting that neither of the individual mental model measures was a significant predictor of performance. However, the two did interact to predict this performance. These results illustrate the complex interplay between the subtypes of mental models that are likely to characterize performance in complex settings.

Team Cohesion One emergent state that has received a great deal of attention is team cohesion. This construct has been

defined as “a dynamic process that is reflected in the tendency for a group to stick together and remain united in the pursuit of its instrumental objectives and/or for the satisfaction of member affective needs” (Carron, Brawley, & Widmeyer, 1998, p. 213). Several authors have theorized that team cohesion should predict a substantial portion of the variance in team performance (e.g., Carron & Brawley, 2000; Carron, Eys, & Burke, 2007; Siebold, 2006). Indeed, the belief that team cohesion is an important condition for performance is so widely held that increasing cohesion is frequently included in team-training interventions (e.g., Bloom & Stevens, 2002; Healy, Milbourne, Aaronson, & Errichetti, 2004; Wheeler, Goldie, & Hicks, 1998). Unfortunately, the empirical data regarding team cohesion and performance is much less clear than one might hope. A meta-analysis by Mullin and Cooper (1994) reported only a weak relationship between the two constructs. However, they did note that the effect was more robust, although not strong, for smaller teams. Subsequent to this analysis, researchers have explored other variables that might mediate the cohesion-performance relationship. For example, Gully and Devine (1995) found that the relationship was stronger in more highly interdependent groups. Recently, Beal, Cohen, Burke, and McLendon (2003) reported a somewhat stronger relationship when using efficiency as the outcome measure (r = .24). It is interesting that Carron et al. (2007) reported moderate to large effect sizes when limiting the analysis to sports teams. Overall, it appears that the relationship between cohesion and performance is more complicated than originally thought, but it remains an important construct for team performance researchers. MEASURING TEAM PERFORMANCE AND FUNCTIONING At various points in the discussion thus far we have alluded to the importance of being able to measure different aspects of teams and their functioning. Indeed, measurement pervades all aspects of team performance, from assessment of the degree to which team members possess requisite KSAs, to measurement of on-going team processes, to evaluation of 625

Cannon-Bowers and Bowers

Copyright American Psychological Association. Not for further distribution.

the team’s task success. In general, the issues associated with the measurement of team performance are similar to those associated with individual performance (see chap. 20, this volume, for a detailed discussion); however, there are several unique challenges associated with measurement at the team level that are worth addressing. These are summarized in the following sections, with associated discussion of the methods that have been developed to measure these aspects of performance.

Measuring Team Outcomes Ultimately, the measurement of effectiveness of a team must be based on the task being performed. An informal survey of 42 empirical studies of team performance revealed that the majority constructed measures of team effectiveness based on the task, including measures such as quality of output (usually assessed by raters), quantity of work, number of problems correctly solved, decision-making accuracy, and number of points scored. A smaller number of studies used supervisor ratings of performance and still fewer used self-ratings of performance. Other researchers have also assessed team outcomes in terms of the teams’ desire to remain together (or what Hackman, 1987, called team viability). A unique issue in assessing team outcomes is that the performance of the overall team may not represent equally the contributions of all members. Conversely, one team member may outperform teammates, essentially pulling the entire team along. At some level, this may not matter if the only thing of interest is overall team outcomes; however, much of the time, it is informative to understand these nuances. Hence, even overarching measures should be sensitive enough to distinguish the contribution of individual members.

Measuring Team Behavior and Process As just described, a unique aspect of performance measurement at the team level is that it transcends any particular individual on the team to include one or more team members. Hence, measurement of team performance must consider the behavior of any given team member relative to the behavior of his or her teammates. For example, when a team member 626

makes an error, the appropriate behavior of a teammate is to correct that error, but this is only the case if the error is made in the first place. So, as in other dynamic environments, measurement systems in teams must account for the moment-to-moment changes that occur during performance and as a consequence of interacting with other members (Cannon-Bowers & Salas, 1997). Along these lines, many team scholars have advocated the development and use of process measures— those that capture the mechanisms the team uses to accomplish its task—as well as outcome measures (Cannon-Bowers & Salas, 1997; Salas & CannonBowers, 2000). Process measures are designed to capture the team behaviors described by Marks et al. (2001) that were delineated earlier. These include such things as backup behaviors, monitoring, coordination, conflict resolution, and communication. With respect to measurement, virtually all measures of team process require that the actions and behaviors of team members be carefully recorded during performance. Typically, this is accomplished through human observers or instructors, although automated systems to aid in measuring ongoing performance are beginning to be available (Deaton et al., 2007). With respect to human observers, a good deal of effort has been devoted to developing observational protocols that aid the assessment process. For example, Dwyer, Fowlkes, Oser, Salas, and Lane (1997) initiated a line of research into a methodology that highlights what needs to be observed and provides a checklist for observers to make it easier to capture performance accurately. Others have advocated similar approaches using scenario-based training (see Baker & Salas, 1992; Rosen et al., 2008). It should be noted that because team performance assessment relies so heavily on observation, it is also advisable to ensure that raters are sufficiently trained (Baker & Salas, 1992). Because communication is considered to be such a critical aspect of team performance, it is not surprising that a considerable level of effort has been dedicated to its measurement. Several measurement approaches have been tried. Because direct observation of communication is difficult and expensive, researchers have developed a variety of paper-and-

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

pencil measures of communication. Some of these are retrospective appraisals of group processes, such as that described by Taylor and Bowers (1972). Others have focused on the group’s satisfaction with their teams’ communication (e.g., Hecht, 1978). Although these are presumed to be measures of a unitary construct, Johnston, Reed, Lawrence, and Onken (2007) recently reported that there are at least two apparent, orthogonal factors among the items. Each of these factors was a successful predictor of performance in a business simulation. Several researchers have attempted to use direct observation of speech acts to predict team performance (Bowers & Jentsch, 1998; Fischer, McDonnell, & Orasanu, 2007; Svensson & Andersson, 2006; Urban et al., 1995). However, because researchers typically use different coding schemes, it is difficult to draw an overall conclusion about the relationship between specific speech acts and team performance. Finally, some researchers have suggested that the “pattern” of communication may be more predictive of team performance than the frequency of specific speech acts. For example, Bowers and Jentsch (1998) reported that a pattern of communication that suggested a closed loop was more predictive of performance than the frequency of specific types of speech (Bowers & Jentsch, 1998). Fischer and her colleagues (2007) reported that both the frequency of individual speech acts and speech patterns discriminated between effective and ineffective teams. It is interesting that Manser, Howard, and Gaba (2008) recently reported that specific patterns of communication were predictive at varying phases of a medical team’s performance, with high effect sizes reported for many of the specific pattern-performance relationships being observed (r = .80).

Measuring Team Cognition With respect to team cognition, we have discussed the growing literature in shared mental models and associated constructs. Unfortunately, the practice of measuring shared mental models has presented some difficulty (Mohammed, Klimoski, & Rentsch, 2000). According to Cannon-Bowers and Salas (2001), some of the problem may stem from a lack of conceptual clarity in defining exactly what is

meant by shared. For example, the term shared may actually mean identical, compatible, overlapping, complementary, or common, each of which has different implications for measurement. Specifically, if shared is interpreted as identical, then the indictor of a shared mental model would be that it is the same across members. Conversely, holding partially overlapping knowledge implies that only a portion of knowledge needs to be common across members. In practice, it may well be that any or all of these definitions of shared are necessary, depending on the demands of the task. Hence, it is essential to understand and specify the shared knowledge requirements in a particular team. Once defined, the precise nature of the requirement for shared knowledge should drive the strategy for assessing whether it is sufficient in the team. With respect to measuring team cognition, Cooke, Salas, Cannon-Bowers, and Stout (2000) maintained that there is not a single method to measure team knowledge because team knowledge is multifaceted and can therefore be measured in several different ways. These authors included observation, interviews and surveys, process tracing, and conceptual methods as potential techniques. Detailed description of these methods is beyond our scope here; see Cooke et al. (2000) for further explanation. Briefly, observations, interviews, and surveys can be used in this context to infer and/or elicit team knowledge. Process tracing refers to a family of methods that seek to construct an ongoing record of performance, for example, a thinkaloud protocol where team members provide a running commentary of what they are thinking as they perform the task. Other approaches include audio or videotaping, keystroke analysis, and the like. Finally, conceptual methods refer to those that represent the domain as a series of concepts and the relationships among them. In their comprehensive review, Mohammed et al. (2000) described four such conceptual methods that seemed to represent the most promising approaches to direct mental model measurement at that time. These included Pathfinder associative networks, multidimensional scaling, interactively elicited cause mapping, and text-based cause mapping. Briefly, Pathfinder is a technique for measuring 627

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

individual knowledge structures that uses paired comparison ratings to create a network structure in which concepts are represented as nodes and the relatedness of concepts as links between the nodes (see Schvaneveldt, 1990). At the team level, the degree of similarity between the knowledge structures can be assessed. Multidimensional scaling (MDS) uses similarity ratings among concepts to spatially represent proximity data in n-dimensional space. That is, it represents the similarity among concepts geometrically and helps to reveal the underlying organizational scheme associated with a series of concepts. As with Pathfinder, MDS can be applied at the team level by assessing the degree of similarity among the cognitive representations of different team members (see Mohammed et al., 2000, for more detail). The last two methods reviewed by Mohammed et al. (2001) both fall within the category of concept mapping. Concept mapping produces graphic representations of an individual’s cognitive structure. A particular type of concept map, called cause maps, specifically focuses on eliciting an individual’s belief about the causal relationship among concepts. Mohammed et al. distinguished two types of cause maps that have been used to assess shared cognition: interactively elicited cause mapping, where the concepts are generated interactively with team members through surveys and questionnaires; and text-based cause mapping, where concepts are inferred by assessing ongoing transcripts or records of performance. At the team level, the cause maps of individual team members can be compared to assess similarity, coverage, and so forth. In a similar vein, Hall, Stevens, and Torralba (2002) studied the conversations among team members to find discourse elements that might indicate the creation of these group models. Given the centrality of team cognition to virtually all accepted theories of team performance, more work on measuring it is clearly warranted. Progress in clarifying the construct is needed, as well as efforts to develop and test appropriate measurement tools. Ultimately, the psychometric properties of proposed methods must be demonstrated so that their validity can be established. 628

TEAM TRAINING The increased attention focused on teams in recent years has led to a number of advancements in teamtraining theory and practice. Several theoretically based interventions have been developed and tested (Salas & Cannon-Bowers, 2000). The sections that follow first set the stage by describing the settings and paradigms typically used in team-training research. Next, we review an overarching framework in which to organize thinking about the components of team training, followed by a brief review of some of the major interventions that have been developed in recent years, along with evidence of their effectiveness. Following this, we summarize a set of guidelines for team training gleaned from the literature, and then address some issues associated with the transfer of team training, in particular how it might be improved.

Prevailing Paradigms and Settings for Team-Training Research There is no doubt that much of the empirical research into team-training effectiveness has focused on two environments: military teams and aviation teams. Historically, this has been driven primarily by funding, that is, the military and the Federal Aviation Authority have been much more likely to fund studies of team-training effectiveness than other agencies or sponsors. This tendency is due, at least in part, to the fact that in high performance environments, the consequences of error are catastrophic, and improving team performance is seen as a mechanism to enhance safety. For the same reason, the past few years have seen an increase in the number of studies conducted on medical teams, although to date, funding for such work has not been as available. Because the preponderance of team-training research has been conducted in these relatively few environments, an important question becomes: How does this research translate into other types of teams in industry? The answer, we believe, rests on two related issues: the similarity of environmental and task demands between military and aviation teams and the targeted environment (to which one would like to generalize) and, in turn, how these task

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

demands translate into specific competency requirements in the teams. Obviously, the more closely the targeted environments resemble the experimental situation, the more justified one would be in generalizing. To take a simple example, one of the characteristics of many of the military and aviation teams that have been studied is time pressure. As we have noted, when teams have little time to respond or make decisions, we believe they must rely on shared knowledge (or mental models) in lieu of more deliberate planning. This phenomenon can be observed over a variety of teams, especially sports teams, when, for example, players make a no-look, or blind pass, essentially predicting the position of teammates on the court or field. Presumably, players in this situation are relying on the preexisting, shared knowledge they have of the situation and of their teammates (Cannon-Bowers et al., 1993). Hence, if the targeted environment is characterized by time pressure, we would conclude that research investigating how to build shared team knowledge would apply, regardless of the setting in which it was tested. This is not to say that we would not like to see more empirical studies of team training in a wider variety of organizational settings—we clearly would—but until such studies are conducted, we believe that careful generalization from the existing literature is possible. Another feature of team-training studies that has actually been raised as a limitation is that many are conducted in laboratory settings with artificial teams as opposed to “real world” environments. As with other laboratory-based research, the argument here is that ad hoc teams do not exhibit behaviors and processes found in teams performing in real organizations because of their limited life span and the relative simplicity of the lab environment. It is interesting to note that a meta-analysis by Mullen, Driskell, and Salas (1998) actually concluded that studying groups in controlled settings does not necessarily lead to less valid or robust findings. In fact, their meta-analysis concluded that similar, but stronger, effects were found in studies conducted with real teams compared with those conducted in the laboratory. According to Mullen et al., the implication of this finding was that more controlled lab

environments did not produce unrealistically strong or spurious effects. Instead, these authors concluded the following: The weight of available evidence strongly suggests that the increased experimental control obtained through the use of artificial groups in the laboratory does not come at the cost of realism, ecological validity, or the ability to generalize beyond the laboratory to real groups in the real world. (p. 228) The argument put forth by Mullen et al. (1998) notwithstanding, it probably follows that some research questions may be more amenable to study in the lab than others. For example, if a crucial question involves familiarity with teammates (i.e., how well they know each other), then trying to study them in a 2-hour lab session is probably not feasible. However, creating artificial groups to study decision making or communication patterns may be justified if team members are properly trained on the experimental task. For example, a number of research groups have used computer-based aviation or command and control simulations to study teams in the laboratory (Ellis et al., 2003; Leedom & Simon, 1995; Proctor, Panko, & Donovan, 2004; Shebilske, Jordon, Goettl, & Paulus, 1998). Task simulations such as these can be constructed carefully so that they represent important aspects of the environmental and task demands (sometimes called cognitive fidelity). As we argued earlier, generalizing on the basis of similarities in task demands and associated team competency requirements between the laboratory and real world is justified if done carefully and systematically. Hence, we conclude this discussion of research and experimental settings for studying team training by again calling for more empirical work that spans a broader array of environments and tasks. Until such results are available, prudent generalization of existing literature seems justified.

Framework for Understanding Team Training Figure 19.4 displays a framework for conceptualizing team training (adapted from Salas & CannonBowers, 1997) that was developed based largely on 629

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

FIGURE 19.4. Team training model. From Training and Retraining: A Handbook for Businesses, Industry, Government, and Military (p. 314), by S. Tobias & D. Fletcher (Eds.), 2000, Farmington Hills, MI: Macmillan. Copyright 2000 by Gale, a part of Cengage Learning, Inc. Reproduced by permission. www.cengage.com/permissions

work aimed at improving performance in military combat teams and military and commercial aviation teams (see Cannon-Bowers & Salas, 1998; Salas, Burke, Bowers, & Wilson, 2001). We have updated the original slightly to streamline the presentation and reflect more recent thinking and findings. Despite its origins, the framework has wide applicability in describing and organizing the components of team training. It specifies several categories of features that define a team-training program, including tools, methods, strategies, objective, and content. To begin with, Salas and Cannon-Bowers (2000) offered team task analysis, performance measurement, simulations and exercises, feedback, and principles as tools that are necessary to team-training development. We have already discussed the necessity of team task analysis and performance measurement; both are crucial for training. Beyond this, simulations and exercises are important because they provide a context in which hands-on practice can occur (see Salas & Cannon-Bowers, 2000, for more detail), whereas feedback is a primary mechanism by which learning can occur. Finally, principles and 630

guidelines for team training are useful to explain how team training should be implemented (we cover these in more detail later). The framework also delineates a variety of methods that are useful for team training. In many ways, these are similar to those applied at the individual level. But when combined with team-training content, these methods result in a series of specific team-training strategies. For example, using handson practice (i.e., a method) for increasing knowledge of team roles (i.e., a competency) could be developed into cross-training (i.e., a strategy). Obviously, a training program can combine several tools, methods, competencies, and strategies; and the model can be used to guide the development process. We now turn our attention to the specific team-training interventions that have been proposed and evaluated in recent years.

Effectiveness of Team Training At some level it makes no more sense to pose the question, “Is team training effective?” than it is to ask, “Is training effective?” because the answer to

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

both depends on a huge number of factors that are related to the task, team, particular training methods used, quality of implementation, training and organizational context, and so forth. For reasons delineated elsewhere (see Cannon-Bowers & Bowers, 2010), general questions such as these are not meaningful because the number of ways that training interventions and situations can vary make it difficult to determine the generalizability of results. Hence, we have tried to limit our review to teamtraining strategies that are tightly defined and theoretically justified and for which at least a few investigations have been conducted. These are summarized in Table 19.4 and are described briefly later. Where possible, we tried to include effect sizes in reporting findings in Table 19.4. Where meta-analyses were available we reported only those effect sizes, otherwise we attempted to report effect sizes for individual studies. These are shown where available. Cross-training. Cross-training involves training team members about the roles and responsibilities of other members. The notion is that giving team members a chance to experience the task from another’s point of view can provide important insight into how the task is accomplished from the perspective of teammates. Such understanding should help them to better anticipate the needs of teammates, and even enable them to empathize so that they are better able to interpret teammates’ behavior. As is clear from Table 19.4, there have been a few empirical investigations of cross-training, with mixed results. Although at least three empirical investigations have supported the technique (Cannon-Bowers et al., 1998; Marks, Sabella, Burke, & Zacarro, 2002; Volpe et al., 1996), one did not (McCann, Baranski, Thompson, & Pigeau, 2000). In the Volpe et al. and Cannon-Bowers et al. studies, which were both conducted in the context of simulated military tasks, team performance was defined as overall team score, time spent accessing information, and accuracy in prosecuting targets. Similarly, Marks et al. used the number of targets shot and time teams remained alive in a military simulation as team outcome measures. McCann et al. used an almost identical task and team outcome score.

A meta-analysis by Salas, Nichols, and Driskell (2007) also failed to find an effect of cross-training on team performance. However, the meta-analysis was based on few studies, so results should be interpreted with caution until more data are available. To complicate matters, the studies reported by Marks et al. (2002) tested three different levels of crosstraining that varied in depth: positional rotation, positional modeling, and positional clarification. Results were inconclusive regarding what level of cross-training is needed to achieve benefits, although some evidence suggested that full positional rotation might not be necessary. More empirical work is needed to refine and validate this approach. Team self-correction training. Guided team selfcorrection—in which teams are instructed to observe teammates’ behavior and provide and accept performance-enhancing feedback—is often mentioned as a promising means to improve shared knowledge. In fact, there is mounting evidence that this strategy is effective in improving team performance (Blickensderfer, Cannon-Bowers, & Salas, 1998; Smith-Jentsch et al., 2008). The Smith-Jentsch et al. (2008) study was conducted with active duty military personnel, and the measure of effectiveness was a rated assessment of the team’s situational awareness. Moreover, the approach is consistent with efforts to build continuous learning environments in teams because it emphasizes the role of each team member in monitoring the performance of teammates and providing constructive feedback. These mechanisms can be applied to all episodes of performance and are not limited only to training exercises. Scenario-based training. The value of scenariobased training stems from its ability to place team members in a realistic context while learning, a notion that has extensive support in the learning sciences (Bransford, Brown, & Cocking, 1999; Cannon-Bowers & Bowers, 2007; Cannon-Bowers, Bowers, & Sanchez, 2008). If properly implemented, scenario-based training can be effective in allowing trainees to experience the consequences of their actions, make adjustments based on feedback, successfully accomplish the task, and build collective efficacy. In fact, Cannon-Bowers, Bowers and Sanchez recently provided guidelines for how to 631

632 Insufficient data to compute

.45 (Salas et al., 2007, meta analysis)

.37 (Marks et al., 2000)

.61 (Salas et al., 2007, metaanalysis)

.41 to .56 (Ellis et al., 2005)

Alinier et al. (2006); Cannon-Bowers and Bowers (2007); Cannon-Bowers, Bowers, and Sanchez (2008); Stedman et al. (2006) Blickensderfer, CannonBowers and Salas (1997); Salas et al. (2007); SmithJentsch et al. (1998); Smith-Jentsch et al. (2008) Marks et al. (2000); Tanenbaum et al. (1998)

Entin and Serfaty (1999); Marks et al. (2000); Salas et al. (2007) Ellis et al. (2005)

Very little data to support or refute

Good empirical support for the approach in field and in laboratory studies

Positive, but limited, support has been found for this type of training in one field and one laboratory study. Few studies; good empirical support; meta-analytic support

Supported in at least one study.

Communication, coordination, mutual performance monitoring, backup behavior

Mutual performance monitoring, giving and receiving feedback

Team leader prebriefing and debriefing behaviors

Adaptability, communication, coordination

Communication, planning and task coordination; collaborative problem solving

Expertise is developed by building a repertoire of task instances through meaningful exposure to the task environment.

Team members are in the best position to catch and correct each other’s errors.

Enriched briefings by the team leader will help the team construct shared meaning and better understand the task and their role in it. Teams who are able to increase use of implicit coordination strategies will adapt better in high workload situations. Team members who must perform across different tasks and with different team members would benefit from training in generic skills.

Expose trainees to realistic scenarios so that they can receive feedback to improve performance.

Teach team members to engage in mutual performance monitoring and constructive team feedback.

Teach team members to provide better preand postexercise briefings.

Help team members to anticipate the needs of teammates and achieve effective task synchronization.

Training in generic teamwork knowledge and skills

Team selfcorrection training

Team leader training

Team coordination and adaptation training

Generic teamwork skills training

Scenario-based training

−.09 (Salas et al., 2007 meta analysis)

Cannon-Bowers et al. (1998); Marks et al. (2002); McCann et al. (2000); Salas et al. (2007); Volpe et al. (1996)

Mixed results; has received some empirical support but not supported in one study and in a meta analytic study

Shared task knowledge, interpositional knowledge, knowledge of roles and responsibilities

Team members are required to perform the task while taking the role of one of his or her teammates.

Cross-training

Building interpositional knowledge will allow team members to better predict the informational needs of teammates.

Effect sizes (r )

Sources

Empirical findings

Rationale

Description

Intervention

Targeted competencies

Team Training Interventions

TABLE 19.4

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

develop scenarios and implement scenario-based training to target specific teamwork KSAs. For example, these authors advocated exposing trainees to increasingly time-pressured scenarios so that they can practice implicit coordination strategies (which is more efficient than explicit verbalizing). Surprisingly, relatively few empirical studies of this technique have been reported; the few that have been conducted have been supportive. For example, Stedman et al. (2006) found that simulation-based training was superior to problem-based learning in medical students. Likewise, Alinier, Hunt, Gordon, and Harwood (2006) found that knowledge and confidence in nursing students was enhanced through scenario-based training. Team-leader training. Obviously, there have been volumes devoted to leadership and leadership training, and reviewing this work is beyond the scope of this chapter. However, a subset of this attention has focused on training team leaders in specific behaviors that support team performance. For example, Tannenbaum et al. (1998) tested an approach to team leader training that prepared leaders by teaching them to conduct effective prebriefs (i.e., before a hands-on exercise) and debriefs (i.e., after a hands-on exercise). According to these authors, prebriefs are essential opportunities for leaders to guide the team in planning, setting mutual goals, developing contingency plans, and clarifying roles and responsibilities. Debriefs also serve multiple purposes: to provide specific feedback on team performance, encourage active team member participation, emphasize teamwork as well as taskwork, and accept feedback from team members. In a study with active-duty naval officers, they found that the training was successful in improving the briefing behaviors of team leaders. Moreover, the teams who had trained leaders outperformed teams with untrained leaders in realistic task simulations. This is fairly convincing evidence that leader behavior can be improved and that it makes a difference when it is improved. A more recent analysis by Marks et al. (2000) also provided support for the notion that training leader briefing behaviors would improve team performance.

Team coordination and adaptation training. The notion that team members must exercise implicit coordination strategies (i.e., those that rely on shared knowledge rather that verbal communication) to be effective has been raised several times in this chapter. Training team members to increase their use of implicit coordination strategies is the subject of team coordination and adaptation training (Entin & Serfaty, 1994). Regarding validation of this strategy, the Salas et al. (2007) meta-analysis mentioned earlier revealed that this intervention had the best results (i.e., compared with cross-training and team selfcorrection) in improving team performance. In a somewhat different interpretation, Marks, Zaccaro, and Mathieu (2000) tested team interaction training, and found that it was effective in improving team performance in a simulated tank war-game (i.e., number of targets correctly prosecuted). Generic team-skills training. In an earlier section, we described the difference between task- and teamspecific competencies (see Table 19.2). Using this framework, Ellis et al. (2005) reasoned that action teams would benefit from training in generic team skills because they regularly transition across different teams and tasks. In a study of command and control teams, they found that generic team skills training was effective in teaching declarative team knowledge, as well as planning and task coordination, collaborative problem solving, and communication. Among other things, these researchers concluded that generic teamwork training might serve a crucial foundation for developing more taskor team-specific competencies.

Guidelines for Team Training Salas and Cannon-Bowers (2000) assembled a set of guidelines for team training (see Exhibit 19.3). On the basis of work conducted primarily under the TADMUS project (described earlier; see CannonBower & Salas, 1998, for details), these authors complied this set of guidelines as a mechanism to help practitioners better implement team-training programs. We have updated and modified the original set of guidelines for presentation here. For the most part, the guidelines shown in Exhibit 19.3 are best applied to teams who must 633

Cannon-Bowers and Bowers

Exhibit 19.3 Guidelines for Team Training ■

■

■ ■

■ ■ ■

Copyright American Psychological Association. Not for further distribution.

■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■

Design training to foster both individual and team skills; individual skills are a necessary but not sufficient condition of effective team performance Develop team members’ knowledge to a threshold level before beginning team training so that they can focus on acquiring teamwork skills Make team members aware of the interrelatedness of individual development and team performance early in training Develop exercises, cases and/or examples that link critical task cues and cue patterns to appropriate response strategies and consequences Provide team members with motivational guidance and seek to remove or ameliorate factors that hinder motivation Familiarize team members with each other’s roles and create realistic expectations for task requirements Build opportunities for success early in training to help develop collective efficacy Teach team members the importance of effective communication strategies Teach team members to seek clarification of ambiguous information Teach team members how to monitor each other’s performance to identify errors Instruct team members in constructive strategies for providing feedback to teammates Help team members to accept constructive feedback from teammates Provide supplemental instruction for team leaders so that they can better guide team members Teach team members to maintain high situational awareness in the team by providing updates Exercise teams in a variety of situations to build their adaptability Develop exercises/cases/scenarios that contain critical events around which the team’s performance can be measured Allow team members to experience the task from the perspective of teammates Instruct teams to how to reflect on their performance and generate self- and team-corrective strategies Teach team members how to plan collaboratively to build shared task expectations Provide opportunities for team members to practice synchronizing and coordinating their input Help team members to recognize and ameliorate interpersonal and task-related conflict Help team members to respect the views and inputs of teammates

From Training and Retraining: A Handbook for Businesses, Industry, Government, and Military (p. 322), by S. Tobias & D. Fletcher (Eds.), 2000, Farmington Hills, MI: Macmillan. Copyright 2000 by Gale, a part of Cengage Learning, Inc. Reproduced by permission. www.cengage.com/permissions.

perform tasks with reciprocal or team interdependencies (as discussed in the framework described earlier). In some senses, this represents the most challenging team-training situation because team members must be flexible and adaptable in their behavior patterns and must adeptly coordinate and communicate implicitly. Hence, not all of these guidelines will be applicable to every team-training situation (especially less complex ones). However, they can provide a good starting point for practitioners and can be a useful mechanism to foster transition of scientific results into the workplace.

Transfer of Team Training Apart from the training content and targeted level, team training may also differ from individual-level training in terms of its mechanisms of transfer. Whereas the general issue of training transfer is 634

addressed elsewhere in this volume, some have argued that the team situation adds unique requirements for transfer that are not present at the individual level (Cannon-Bowers, Salas, & Milham, 2003). The framework in Exhibit 19.4 summarizes some of these (see Cannon-Bowers et al., 2003, for a detailed discussion). Exhibit 19.4 organizes the discussion of teamtraining transfer into three broad categories: those activities that occur before training, those activities that occur during training, and those activities that occur after training. Before training, there are a host of team characteristics that can affect learning and transfer. These include the team’s ability and aptitude (which are consistent with what is known at the individual level), history and experience, and several team attitudes (i.e., collective orientation, collective efficacy, and attitudes toward teamwork). According to the authors, these factors should be

Team Development and Functioning

Exhibit 19.4 Factors That Affect the Transfer of Team Training

■ ■

■

■

Copyright American Psychological Association. Not for further distribution.

■ ■ ■ ■

■ ■ ■ ■ ■ ■

Before Training: Team ability/aptitude—should be sufficiently high to support learning Team history & experience—low expectations should be adjusted before training; gaps in knowledge/experience should be remediated Team attitudes: collective efficacy, collective orientation, attitudes towards teamwork—attitudes should be assessed prior to training and adjusted if necessary During Training: Advance organizers—should explain complex team and task relationships Pre-practice briefs—should explain team roles and clarify expectations Team level metacognitive skills—should be supported during training Training environment fidelity (realism) of the task and/or team—should be matched to the nature of the task demands Feedback—intrateam feedback and self-correction should be fostered After Training: Climate for teamwork—should be established and maintained Performance appraisals—should emphasis teamwork and collective achievement Reward and performance management systems—should reflect team accomplishment Team goals—should be developed at the team level Leader/team member support—should support transfer of newly learned skills Relapse prevention—should involved the entire team

From Improving Learning Transfer in Organizations (pp. 219–220), by E. Holton and T. Baldwin (Eds.), 2003, San Francisco: Jossey-Bass. Copyright 2003 by Jossey-Bass. Adapted with permission.

assessed prior to training so that remedial activities can occur if they are found to be a problem. For example, if team members hold negative attitudes toward teamwork (and hence do not value being part of a team), it may be possible to improve these attitudes before the training begins. Cannon-Bowers et al. (2003) recommended that during training, advance organizers are available to explain complex team and task relationships and prebriefs that help to clarify roles and responsibilities in the team. They also advocated matching the level of team and task fidelity to the specific task demands. For example, if the task allows for high degrees of behavioral discretion (i.e., where team members have considerable control over how they complete their tasks), then it is advisable to train with actual teammates whenever possible (so that team members can gain required knowledge of teammates preferences, strengths, weaknesses, tendencies, etc). In other cases (e.g., where the task is highly interdependent but proceduralized), task fidelity may be more important, suggesting that

team members should be trained in a simulation or work sample environment. Cannon-Bowers et al. (2003) also made recommendations for activities that occur after training as means to improve transfer. In many organizations that have adopted team-based organizations to accomplish work, associated human resources (HR) practices are still based at the individual level. This is a problem because it can set up conflicting messages for the team members. For example, if team members are taught in team training sessions that they should feel responsible to their team and that team performance is valued but then are evaluated and rewarded on an individual basis, it can create conflict. Hence, transfer of team training to the workplace will be hindered until alignment of performance management practices and systems is achieved. EMERGING ISSUES IN TEAM PERFORMANCE The future of teams in organizations is likely to continue to grow and will be marked by a set of new 635

Cannon-Bowers and Bowers

challenges. To begin with, globalization of the modern workforce, coupled with advances in technology, has led to a rise in the use of teams who are physically dispersed. Referred to as distributed or virtual teams, these work groups exist across functional, organizational, and even national boundaries. Hence, the issues that confront such teams are complex; they involve altered team processes, multicultural issues, and supporting technologies. The following sections review some of these emerging issues.

Copyright American Psychological Association. Not for further distribution.

Virtual Teams and Their Processes Driskell, Radtke, and Salas (2003) defined virtual teams as those whose members “are mediated by time, distance, or technology” (p. 297). Hertzel and colleagues (Hertzel, Geister, & Konradt, 2005) added the definitional requirement that these teams must coordinate their work through electronic communication media. It should be pointed out, however, that all virtual teams are not the same. As pointed out by Driskell et al. (2003), there are a variety of technologies that can support virtual teams, ranging from simple e-mail to real-time video and audio communications. Each of these technologies is associated with a set of cues available to team members. Although some of these technologies provide few of the cues available to face-to-face teams, others provide almost all of the cues available in traditional colocated teams. Similarly, these teams can differ in temporal distribution, roles, and other variables (see Priest, Stagl, Klein, Salas, & Burke, 2006). Consequently, it has been suggested that the processes required for effective performance might be different depending on the capabilities of the particular technology used to enable the team’s work. In the next section, we focus on the consequences of distance on the team’s processes and emerging states and cover technology issues in a subsequent section. The definition of virtual teams dictates that these teams must work under conditions that are far different from traditional teams. The fact that these teams must coordinate their actions, without the benefit of face-to face interactions, to be effective has led several researchers to suggest that these teams must somehow alter their processes to accommodate for the lack of cues. Consequently, there has been consider636

able interest by researchers in the people and processes that are associated with effective virtual team performance. In their review, Driskell and colleagues (2003) proposed four processes that they hypothesized to be particularly important for virtual teams: cohesiveness, status, counternormative behavior, and communication. First, as described earlier, cohesiveness is thought to be an important element in the performance of traditional teams. Given the reduced cues available to virtual teams, it has been suggested that this important team characteristic might be less likely to occur in virtual teams. In general, this hypothesis has been supported by the available empirical literature; specifically, the degree of disruption of cohesiveness is related to the degree to which the communication medium approximates face-to-face teams. For example Bos, Gergle, Olson, and Olson (2001) reported that teams who communicated only via text experienced greater difficulty in establishing trust than either video-conferencing or face-to-face teams. Similarly, the video-conferencing teams experienced more difficulty than did the face-to-face teams. Doherty-Sneddon et al. (1997) and Straus (1996) reported similar results. It should be pointed out, however, that the relationship between cohesiveness and virtual team performance is likely more complex than described here. For example, Gonzalez, Burke, Santuzzi, and Bradley (2003) reported that their data better fit a model where cohesiveness mediated the relationship between collective efficacy and performance than did a model examining the direct effect between cohesiveness and performance itself. A second process highlighted by Driskell and his colleagues (2003) is status. From an interpersonal standpoint, perceived status is a key factor that determines the processes that emerge in teams, such as leadership. Because of the degraded stimulus array in virtual teams, is has been hypothesized that these status effects might be altered because it is less clear who is more experienced, wealthy, older, and so forth. Several studies supported this contention (e.g., Sproull & Kiesler, 1986, 1991). However, other researchers have reported that status and leadership behaviors are not always disrupted in

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

these kinds of teams (Saunders, Robey, & Vaverek, 1994). Moreover, there are a number of methodological concerns that preclude a thorough understanding of status differences in distributed teams; hence, a dedicated program of research is needed. The third important process is counternormative behavior. It has been widely reported that computermediated communications tend to include more flamboyant, inappropriate statements than do personal communications. This aberrant behavior is attributed to the reduced inhibition one feels in the relatively anonymous situation that often characterizes computer-mediated teamwork (Sproull & Kiesler, 1991). Such phenomena have led to a concern that teamwork in virtual teams might be disrupted by rude, negative behaviors. Early studies seemed to support this hypothesis (e.g., Siegal, Dubrovsky, Kiesler, & McGuire, 1986). However, Driskell and his colleagues (2003) concluded that the more recent literature is largely equivocal. Again, there are a host of potential moderator variables that must be accounted for, and a great deal of research will be required to determine the likelihood and extent of this phenomena. Finally, Driskell et al. (2003) proposed communication as the fourth important process variable in distributed teams. The importance of communication has been described earlier in this chapter. However, the limited cue set in virtual teams is likely to change the nature of effective communication in these teams. Reviews of this literature have largely supported this notion (Culnan & Markus, 1987). In large part, these differences are related to reduced knowledge sharing and feedback (Ashford & Tsui, 1991; Thompson & Coovert, 2003). Consequently, researchers have recently extracted a set of communication guidelines for leaders to help counteract these negative effects (Rosen, Furst, & Blackburn, 2007; Walvoord, Redden, Elliott, and Coovert, 2008). In another review, Bowers, Smith, CannonBowers, and Nicholson (2008) delineated possible differences between collocated and distributed teams in terms of behavioral, attitudinal, and cognitive aspects. Focusing first on those behaviors not covered by Driskell et al. (2003), Bowers and colleagues noted that leadership behaviors are also

affected by distance among team members. Leaders of distributed teams face unique challenges when it comes to directing the teams’ work, assessing performance, developing and encouraging members (including setting a climate for teamwork), and planning and communicating effective strategies for task accomplishment. Zhang, Fjermestad, and Tremaine (2005) hypothesized that different leadership behaviors may be needed in virtual teams. With respect to team attitudes, Bowers et al. (2008) noted that trust, collective efficacy, and team orientation (in addition to cohesion, which was discussed by Driskell et al., 2003) may all be affected by distance. Beginning with trust, several authors have suggested that dispersion is likely to reduce trust in virtual teams (Ardichvili, Page, & Wentling, 2003; Staples & Webster, 2008) because mistrust can stem from misattributions due to technological anomalies or reduced cues. Likewise, collective efficacy (i.e., the team’s shared belief in its ability to accomplish the task) may not develop as it does in collocated teams. The issue here is that distributed teams may have difficulties experiencing shared success or correctly attributing such success to the team’s efforts. Finally, team members in a distributed team may not see themselves as a team in the same way that face-to-face teams do. This may reduce the team members’ ability to correctly recognize that teamwork is important and a component of team orientation (Bowers et al., 2008). Regarding cognitive differences among virtual and collocated teams, Bowers and colleagues (2008) argued that memory processes can be affected by altered information flow, task cues, and communication mechanisms when teams are distributed. For example, it is likely that workload will be higher in virtual teams because commonly shared visual and auditory cues are not available, whereas retrieval and prospective memory may be altered because of a lack of normally robust cue sets (see Fiore, Cuevas, Schooler, & Salas, 2006, for more detail). In addition to the processes and emerging states already discussed, other researchers have recently described phenomena that might be altered in virtual teams. For example, Fiore and his colleagues (2003) described a construct of team opacity. This refers to the awareness that team members have of 637

Copyright American Psychological Association. Not for further distribution.

Cannon-Bowers and Bowers

one another. These researchers argued that the nature of distributed teams reduces the clarity of this awareness so that other teamwork behaviors are required for these teams to be successful. Other researchers have described constructs such as interpersonal perceptions (Fiore, Salas, & CannonBowers, 2001) and adaptability (Fiore et al., 2006) as issues worthy of consideration. Obviously, additional research (particularly empirical investigations) is needed to better understand the workings of virtual teams so that recommendations for important HR functions can be generated. This is especially true in light of the fact that virtual teams are already being used in many organizations and that trend is growing. Recently, Rosen et al. (2007) conducted a survey of 440 training and development specialists regarding their organizations’ use and treatment of virtual teams. Among other things, this survey indicated that little is being done to train virtual team members or the leaders who must manage them. Further, Rosen et al. (2007) conducted a logistic regression analysis to determine the differences among organizations whose respondents perceived their virtual team-training programs to be effective versus those who perceived it to be ineffective. Results indicated that in organizations with (perceived) effective training, virtual teams were considered a strategic advantage in maintaining competitiveness, communication technologies for collaboration were viewed as critical, and high levels of top management support were present. Rosen and colleagues went on to offer a prototypical virtual team-training program based on the synthesis of survey findings along with anecdotal accounts and reports of current training practices. It included things such as training leaders to set expectations, measure and reward performance, coach and mentor in a virtual environment, and model effective distributed team behaviors. For virtual team members, these authors recommended a face-to-face teambuilding session prior to virtual team launch, followed by modules aimed at training in the use of technology, communication skills, and team management.

Technology and Teams There is little doubt that technology has had a major impact on teams in the modern workplace 638

(see Bowers, Salas, & Jentsch, 2006). Many of the issues associated with technology have already been addressed in discussing virtual teams. However, we thought that a separate section was warranted so that a fuller spectrum of issues could be addressed. According to Hoeft, Jentsch, and Bowers (2006), technology can affect teams in at least three ways; technology can (a) be designed to support teamwork, (b) actually affect the environment in which team members interact with one another, and (c) permeate a team to become an integral part of that team; in essence, technology can become another team member. With respect to technologies to support teamwork, Mittleman and Briggs (1999) delineated a list of technologies (based on work by McGrath & Hollingsworth, 1994; Nunamaker, Briggs, Mittleman, Vogel, & Balthazard, 1997) that support internal team communication, external team communication, process deliberation, task deliberation, and information access. These include such things as presentation support systems, audio and video conferencing, application sharing, group support systems, and, of course, e-mail. In addition, other authors have suggested ways to use technology to improve the functioning of virtual teams. In particular, technology may have value as a means to enrich the cuing environment so that shared knowledge and perceptions can be enhanced. For example, Walvoord et al. (2008) advocated use of multimodal displays (i.e., those that provide visual, auditory and tactile information) as a means to enrich the cue environment in distributed teams. Schatz, Cannon-Bowers, and Bowers (2006) discussed the benefits of data visualization techniques as a means to foster shared cognition among team members. Finally, Bowers et al. (2008) contended that virtual world technologies (in which team members represent themselves as realistic avatars and interact in real time with teammates in a virtual workspace) hold promise as a means to improve virtual teamwork. According to these authors, team members in such environments can communicate more easily (especially when auditory channels are available), better observe and assess each other, share resources, and possibly improve attitudes (i.e., trust, cohesion). Moreover, they may also improve cognitive functions (i.e., working memory, activation) by mimicking the

Copyright American Psychological Association. Not for further distribution.

Team Development and Functioning

cue sets of face-to-face teams more closely than with typical groupware. In the second category posited by Hoeft et al. (2006) (changing the team environment through teamwork), they described a variety of high tech contexts in which teams must operate, including command and control, nuclear power, and space. These environments have a fundamental impact on the way the team accomplishes its tasks. Finally, on the subject of computer-generated or synthetic team members, automation and robotics are maturing to the point where such entities may soon be a reality (Groom & Nass, 2007). As this work progresses, it will be imperative that team researchers investigate the unique consequences of this technology on teamwork (see Scielzo, Fiore, Jentsch, & Finkelstein, 2006, for more details).

Multicultural Issues in Teams As noted, distance technologies are enabling groups of employees from widely dispersed geographical locations to work collaboratively as teams. As noted, such teams confront obstacles and challenges that are not present in face-to-face teams; these challenges are even more acute when team membership crosses national and cultural boundaries. In their review of multinational, multicultural (MNMC) teams, Connaughton and Shuffler (2007) contended that such teams present both opportunities and challenges to organizations. On the one hand, MNMC teams have access to a wide variety of resources, but on the other hand, many virtual teams fail to meet their objectives because of challenges in communication and coordination. Based on a qualitative review of 25 articles on MNMC teams (a mix of empirical and theoretical), Connaughton and Shuffler (2007) concluded that the definition of culture has been inconsistent across the literature. Indeed, many authors recognize that drawing national boundaries is not sufficient in defining culture because it can extend to include ethnic, racial, gender, and other characteristics. Connaughton and Shuffler also reported a number of recurring themes (defined as those investigated in at least three studies) from their review: communication frequency and mode, conflict and its manage-

ment, and temporality (in terms of stages of MNMC team development and its cultural interpretation). Adopting a multilayered view of culture, Cseh (2003) delineated a number of issues associated with culture that can affect the performance of multicultural teams: individualism versus collectivism and communitarianism; power distance; achievement versus ascription in according status; universalism versus particularism in relationships and rules; neutral versus affective in feelings and relationships; specific versus diffuse in involvement and relationships; relation to time; relation to nature; and uncertainty avoidance and masculinity. According to Cseh, these factors can also have an impact on team learning at various stages of development. In another review—this one aimed at team leaders in multicultural teams—Burke, Hess and Salas (2006) argued that the team leader can help or hinder performance in multicultural teams. They presented a model of multicultural team leadership that capitalizes on Graen, Hui, and Gu’s (2004) notion that leaders can create a “third culture,” which is a synthesis of the diverse cultural backgrounds represented in the group that is acceptable to group members. This third culture is seen as beneficial to team performance by creating a flexible additional structure (Graen, Hui, & Taylor, 2004). Burke et al. provided a series of propositions related to team leadership in multicultural teams, including suggestions for improving performance through interpersonal skills (e.g., active listening, empathy, negotiation and conflict resolution, communication), team building, and decision making. On a more practical level, Brett, Behfar, and Kern (2006) argued that multicultural teams present a unique set of challenges to managers. Difficulties can be attributed to cultural differences that arise from difference in directness of communications, language (i.e., accents and fluency), attitudes toward hierarchy and authority, conflicting norms for decision making, and adaptation. On the basis of interviews with managers and members of multicultural teams, Brett et al. describe four strategies for dealing with multicultural challenges: adaptation (i.e., acknowledging and working around cultural issues), structural interventions (i.e., changing the structure of the team), managerial interventions (i.e., setting early 639

Cannon-Bowers and Bowers

norms or engaging management), and exit (i.e., removing a team member when all else fails). These suggestions are worthy of empirical evaluation.

may be ameliorated through effective leadership, training, and/or structural interventions. FUTURE DIRECTIONS

Copyright American Psychological Association. Not for further distribution.

CONCLUSIONS As stated at the onset of this chapter, teams have been of interest as a mechanism of organization in the workplace for decades. Over this time, literally hundreds of studies have been conducted to study teams and their performance. In this chapter, we have attempted to represent the major findings of this body of work. On the basis of our review, we are able to draw several conclusions. These include the following: 1. Several models and taxonomies of teams, team tasks, and teamwork have been offered. These devices have value in helping to organize the host of factors that affect team performance. 2. Teams require unique competencies that must be identified through team task analysis and are related to the nature of the task and team. 3. Personality factors in teams do have an influence on team performance. These influences are complex, possibly interactive, and nonlinear. Further, personality factors may interact with task characteristics. 4. Teamwork consists of many complex processes and emergent states that express themselves in a variety of ways depending on the nature of the task. 5. Team training can be thought of as a combination of tools and methods, which when combined with specific content, create discrete training strategies. Studies of team training indicate that several training strategies hold promise as a means to improve performance. 6. Virtual teams present unique challenges by altering the sets of cues available to team members that typically underlie coordinated performance. There is no doubt that virtual teams can be effective, and new technologies such as virtual worlds may be useful in this regard. 7. Multicultural teams are growing in popularity and need further empirical investigation. Clearly, differences in culture can lead to conflict and other disruptions to team coordination, but these 640

As we noted at the onset of this chapter, the scholarly study of work teams is a relatively recent phenomenon. And given the complexity of teams and team performance, it is not surprising that we concluded many of the sections of this review with a phrase such as “more research is needed to fully understand this relationship.” Indeed, if we pose the question, “Which areas of team performance do we understand so well, that no more investigation is necessary,” none come to mind. But rather than conclude with the general call for more research (especially empirical studies), we would like to offer the following recommendations for how such work should be approached. These include the following: 1. Make explicit the assumptions you are making about team performance; ideally select (as opposed to create) a theoretical model that best fits your way of thinking. 2. Specify the type of team you are testing, preferably using one of the prevailing taxonomies offered in the literature (e.g., Sundstrom et al., 1999). 3. Include extensive description of the type of task the team is performing, preferably using one of the taxonomies that currently exist. 4. Describe the processes you are studying (or assuming) using an accepted model of team process (we recommend the one developed by Marks et al., 2002). 5. Delineate (where applicable) the organizational or situational factors that impinge on the performance of the team you are studying. 6. Describe in detail how you are defining and measuring team performance and effectiveness; report effect sizes where appropriate. 7. If studying team training, describe the competencies you are addressing, preferable using a referent such as the one described by Mohammed et al. (2010). 8. If studying team training, describe the intervention (i.e., tools, methods, strategies) in enough detail that others can understand, replicate, and generalize the results.

Team Development and Functioning

Copyright American Psychological Association. Not for further distribution.

9. When examining virtual teams, describe their mechanisms of support (especially technology) in detail. Obviously, not all of these suggestions will fit every situation. The point we are making is that the field of team performance has matured to the stage where a viable foundation of theoretical work has been accomplished. At this point, adding to this foundation may not be as fruitful as using it to guide future empirical work. Moreover, as we have stated, it is challenging to get a handle on this area because it is difficult to make meaningful comparisons across studies. This is because researchers often do not include enough information or create their own terminology so that generalization back to the extant literature is impossible. Overall, we are encouraged by the body of work we have reviewed here as a basis on which to begin making improvements to the performance of teams in organizations. We are particularly heartened at the trend toward integration, meta-analytic reviews, and empirical investigations. We are confident that the future will hold advances that continue to inform both the science and practice of team performance and training.

References

Arthur, W., Edwards, B., Bell, S., Villado, A., & Bennett, W. (2005). Team task analysis: Identifying tasks and jobs that are team based. Human Factors, 47, 654–669. Ashford, S., & Tsui, A. (1991). Self-regulation for managerial effectiveness: The role of active feedback seeking. Academy of Management Journal, 34, 251–280. Baker, D. P., & Salas, E. (1992). Principles for measuring teamwork skills. Human Factors, 34, 469–475. Banks, A., & Milward, L. (2007). Differentiating knowledge in teams: The effect of shared declarative and procedural knowledge on team performance. Group Dynamics: Theory, Research, and Practice, 11, 95–106. Barnes, C., Hollenbeck, J., Wagner, D., DeRue, D., Nahrgang, J., & Schwind, K. (2008). Harmful help: The costs of backing-up behavior in teams. Journal of Applied Psychology, 93, 529–539. Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A metaanalysis. Personnel Psychology, 44, 1–26. Barry, B., & Stewart, G. (1997, January 1). Group Performance Measure. Journal of Applied Psychology, 82, 62–78. Beal, D., Cohen, R., Burke, M., & McLendon, C. (2003). Cohesion and performance in groups: A metaanalytic clarification of construct relations. Journal of Applied Psychology, 88, 989–1004. Bell, S. (2007). Deep-level composition variables as predictors of team performance: A meta-analysis. Journal of Applied Psychology, 92, 595–615.

Achille, L., Schulze, K., & Schmidt-Nielsen, A. (1995). An analysis of communication and use of military terms in navy team training. Military Psychology, 7, 95–107.

Bell, S. T., Villado, A. J., Lukasik, M. A., Briggs, A., & Belau, L. (2007). Revisiting the team demographic diversity and performance relationship: A metaanalysis. Paper presented at the meeting of the Society for Industrial/Organizational Psychology, New York.

Ahearn, K. K., Ferris, G. R., Hochwarter, W. A., Douglas, C., & Ammeter, A. P. (2004). Leader political skill and team performance. Journal of Management, 30, 309–327.

Blandford, A., & Wong, W. (2004). Situation awareness in emergency medical dispatch. International Journal of Human–Computer Studies, 61, 421–452.

Akkerman, S., Admiraal, W., Simons, R., & Niessen, T. (2006). Considering diversity: Multivoicedness in international academic collaboration. Culture & Psychology, 12, 461–485. Alinier, G., Hunt, B., Gordon, R., & Harwood, C. (2006). Effectiveness of intermediate-fidelity simulation training technology in undergraduate nursing education. Journal of Advanced Nursing, 54, 359–369.

Blickensderfer, E., Cannon-Bowers, J., & Salas, E. (1997). Theoretical bases for team self-corrections: Fostering shared mental models. In M. M. Beyerlein & D. A. Johnson (Eds.), Advances in interdisciplinary studies of work teams (Vol. 4, pp. 249–279). Greenwich, CT: JAI Press.

Alper, S., Tjosvold, D., & Law, K. (2000). Conflict management, efficacy, and performance in organizational teams. Personnel Psychology, 53, 625–642.

Blickensderfer, E., Cannon-Bowers, J., & Salas, E. (1998). Cross-training and team performance. In J. A. Cannon-Bowers & E. Salas (Eds.), Making decisions under stress: Implications for individual and team training (pp. 299–311). Washington, DC: American Psychological Association.

Ardichvili, A., Page, V., & Wentling, T. (2003). Motivation and barriers to participation in virtual knowledge-sharing communities of practice. Journal of Knowledge Management, 7, 64–77.

Bloom, G., & Stevens, D. (2002). Case study: A teambuilding mental skills training program with an intercollegiate equestrian team. Athletic Insight: The Online Journal of Sport Psychology, 4. Retrieved 641

Cannon-Bowers and Bowers

August 1, 2009, from http://www.athleticinsight. com/Vol4Iss1/EquestrianTeamBuilding.htm Bos, N., Gergle, D., Olson, J., & Olson, G. (2001). Being there versus seeing there: Trust via video. CHI ’01 extended abstracts on human factors in computing systems (291–292). New York: ACM. Bowers, C., Baker, D., & Salas, E. (1994). Measuring the importance of teamwork: The reliability and validity of job/task analysis indices for team-training design. Military Psychology, 6, 205–214.

Copyright American Psychological Association. Not for further distribution.

Bowers, C., & Jentsch, F. (1998). Analyzing communication sequences for team training needs assessment. Human Factors, 40, 672. Bowers, C., Jentsch, F., Salas, E., & Braun, C. (1998). Analyzing communication sequences for team training needs assessment. Human Factors, 40, 672–679. Bowers, C., Morgan, B., Salas, E., & Prince, C. (1993). Assessment of coordination demand for aircrew coordination training. Military Psychology, 5, 95–112. Bowers, C., Pharmer, J., & Salas, E. (2000). When member homogeneity is needed in work teams: A metaanalysis. Small Group Research, 31, 305–327. Bowers, C., Salas, E., & Jentsch, F. (2006). Creating hightech teams: Practical guidance on work performance and technology. Washington, DC: American Psychological Association. Bowers, C., Smith, P. A., Cannon-Bowers, J. A., & Nicholson, D. (2008). Using virtual worlds to assist distributed teams. In P. Zemliansky (Ed.), The handbook of research on virtual workplaces and the new nature of business practices (pp. 408–423). Hershey, PA: IGI Publishing Group. Bransford, J. D., Brown, A. L., & Cocking, R. R. (1999). How people learn: Brain, mind, experience, and school. Washington, DC: National Academy Press. Brett, J., Behfar, K., & Kern, M. (2006). Managing multicultural teams. Harvard Business Review, 84, 84–91. Burke, C. S. (2004). Team task analysis. In N. A. Stanton, A. Hedge, K. Brookhuis, E. Salas, & H. Hendrick (Eds.), Handbook of human factors and ergonomic methods (pp. 365–423). Boca Raton, FL: CRC Press. Burke, C., Hess, K., & Salas, E. (2006). Building the adaptive capacity to lead multicultural teams. In C. S. Burke, L. Pierce, & E. Salas (Eds.), Understanding adaptability: A prerequisite for effective performance within complex environments (pp. 175–211). Amsterdam: Elsevier. Burke, C., Stagl, K., Klein, C., Salas, E., Halpin, S., & Goodwin, G. (2006). What type of leadership behaviors are functional in teams? A meta-analysis. The Leadership Quarterly, 17, 288–307. 642

Burke, C., Stagl, K., Salas, E., Pierce, L., & Kendall, D. (2006). Understanding team adaptation: A conceptual analysis and model. Journal of Applied Psychology, 91, 1189–1207. Campion, M., & Stevens, M. (1991). Neglected questions in job design: How people design jobs, task–job predictability, and influence of training. Journal of Business and Psychology, 6, 169–191. Cannon-Bowers, J. A., & Bowers, C. A. (2007). Synthetic learning environments. In J. M. Spector, M. D. Merrill, J. J. G. van Merriënboer, & M. P. Driscoll (Eds.), Handbook of research on educational communications and technology (3rd ed., pp. 306–315). Mahwah, NJ: Erlbaum. Cannon-Bowers, J. A., & Bowers, C. A. (2010). Synthetic learning environments: On developing a science of simulation, games, and virtual worlds for training. In S. Kozlowski & E. Salas (Eds.), Learning, training, and development in organizations (pp. 229–262). London: Taylor & Francis. Cannon-Bowers, J. A., Bowers, C. A., & Sanchez, A. (2008). Using synthetic learning environments to train teams. In V. Sessa & M. London (Eds.), Work group learning: Understanding, improving, and assessing how groups learn in organizations (pp. 315–346). New York: Erlbaum. Cannon-Bowers, J., Oser, R., & Flanagan, D. (1992). Work teams in industry: A selected review and proposed framework. In R. W. Swezey and E. Salas, (Eds.), Teams: Their training and performance (pp. 355–377). Westport, CT: Ablex. Cannon-Bowers, J., & Salas, E. (1998). Team performance and training in complex environments: Recent findings from applied research. Current Directions in Psychological Science, 7, 83–87. Cannon-Bowers, J. A., & Salas, E. (1997). A framework for developing team performance measures in training. In M. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods and applications (pp. 45–62). Mahwah, NJ: Erlbaum. Cannon-Bowers, J., & Salas, E. (2001). Reflections on shared cognition. Journal of Organizational Behavior, 22, 195–202. Cannon-Bowers, J., Salas, E., Blickensderfer, E., & Bowers, C. (1998). The impact of cross-training and workload on team functioning: A replication and extension of initial findings. Human Factors, 40, 92. Cannon-Bowers, J., Salas, E., & Converse, S. (1993). Shared mental models in expert team decision making. In N. J. Castellan, Jr. (Ed.), Individual and group decision making: Current issues (pp. 221–246). Hillsdale, NJ: Erlbaum. Cannon-Bowers, J. A., Salas, E., & Milham, L. (2003). The transfer of team training. In E. Holton & T.

Team Development and Functioning

Baldwin (Eds.), Improving learning transfer in organizations (pp. 63–74). San Francisco: Jossey-Bass. Cannon-Bowers, J. A., Tannenbaum, S. I., Salas, E., & Volpe, C. (1995). Defining competencies and establishing team-training requirements. In R. A. Guzzo & E. Salas (Eds.), Team effectiveness and decision making in organizations (pp. 333–380). San Francisco: Jossey-Bass. Caple, R. B. (1978). The sequential stages of group development. Small Group Behavior, 9, 470–476.

Copyright American Psychological Association. Not for further distribution.

Carley, K. (1997). Extracting team mental models through textual analysis. Journal of Organizational Behavior, 18, 533–558. Carron, A., & Brawley, L. (2000). Cohesion: Conceptual and measurement issues. Small Group Research, 31(1), 89–106. Carron, A. V., Brawley, L. R., & Widmeyer, W. N. (1998). The measurement of cohesiveness in sport groups. In J. L. Duda (Ed.), Advances in sport and exercise psychology measurement (pp. 213–226). Morgantown, WV: Fitness Information Technology. Carron, A., Eys, M., & Burke, S. (2007). Team cohesion: Nature, correlates, and development. In D. Lavallee & S. Jowett (Eds.), Social psychology in sport (pp. 91–101). Champaign, IL: Human Kinetics. Cole, M., Walter, F., & Bruch, H. (2008). Affective mechanisms linking dysfunctional behavior to performance in work teams: A moderated mediation study. Journal of Applied Psychology, 93, 945–958. Connaughton, S., & Shuffler, M. (2007). Multinational and multicultural distributed teams. Small Group Research, 38, 387–412. Cooke, N. J., Salas, E., Cannon-Bowers, J. A., & Stout, R. J. (2000). Measuring team knowledge. Human Factors, 42, 151–173. Costa, P. T., & McCrae, R. R. (1992). Four ways five factors are basic. Personality and Individual Differences, 13, 653–665. Côté, S., & Miners, C. (2006). Emotional intelligence, cognitive intelligence, and job performance. Administrative Science Quarterly, 51, 1–28. Cseh, M. (2003). Facilitating learning in multicultural teams. Advances in Developing Human Resources, 5, 26–40. Culnan, M., & Markus, M. (1987). Information technologies. In F. M. Jablin, L. L. Putnam, K. H. Roberts, & L. W. Porter (Eds.), Handbook of organizational communication: An interdisciplinary perspective (pp. 420–443). Thousand Oaks, CA: Sage. Day, E. A., Arthur, W., Miyashiro, B., Edwards, B. D., & Hanson, T. (2004). Criterion-related validity of statistical operationalizations of group general cognitive ability as a function of task type: Comparing the

mean, maximum, and minimum. Journal of Applied Social Psychology, 34, 1521–1549. Deaton, J., Bell, B., Fowlkes, J., Jentsch, F., Bell, M., & Bowers, C. (2007). Enhancing team training and performance with automated performance assessment tools. International Journal of Aviation Psychology, 17, 317–331. DeChurch, L., & Haas, C. (2008). Examining team planning through an episodic lens: Effects of deliberate, contingency, and reactive planning on team effectiveness. Small Group Research, 39, 542–568. De Dreu, C. K. W., & Weingart, L. R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis. Journal of Applied Psychology, 88, 741–749. Devine, D. J., & Philips, J. L. (2001). Do smarter teams do better: A meta-analysis of cognitive ability and team performance. Small Group Research, 32, 507. Doherty-Sneddon, G., Anderson, A., O’Malley, C., Langton, S., Garrod, S., & Bruce, V. (1997). Face-toface and video-mediated communication: A comparison of dialogue structure and task performance. Journal of Experimental Psychology: Applied, 3, 105–125. Driskell, J., & Mullen, B. (2006). Status, expectations, and behavior: A meta-analytic review and test of the theory. In J. M. Levine & R. L. Moreland (Eds.), Small groups (pp. 73–81). New York: Psychology Press. Driskell, J., Radtke, P., & Salas, E. (2003). Virtual teams: Effects of technological mediation on team performance. Group Dynamics: Theory, Research, and Practice, 7, 297–323. Driskell, J., & Salas, E. (1992). Collective behavior and team performance. Human Factors, 34, 277–288. Druskat, V., & Kayes, D. (1999). The antecedents of team competence: Toward a fine-grained model of selfmanaging team effectiveness. In M. A. Neale & E. A. Mannix (Series Eds.) and R. Wageman (Vol. Ed.), Research on managing groups and teams: Groups in context (Vol. 2, pp. 201–231). Stamford, CT: JAI Press. Druskat, V., & Wolff, S. (2001). Building the emotional intelligence of groups. Harvard Business Review, 79, 81–90. Dwyer, D. J., Fowlkes, J. E., Oser, R. L., Salas, E., & Lane, N. E. (1997). In M. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods, and applications (pp. 137–154). Mahwah, NJ: Erlbaum. Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44, 350–383. Ellis, A., Bell, B., Ployhart, R., Hollenbeck, J., & Ilgen, D. (2005). An evaluation of generic teamwork skills 643

Cannon-Bowers and Bowers

training with action teams: Effects on cognitive and skill-based. Personnel Psychology, 58, 641–672. Ellis, A., Hollenbeck, J., Ilgen, D., Porter, C., West, B., & Moon, H. (2003). Team learning: Collectively connecting the dots. Journal of Applied Psychology, 88, 821–835. Entin, E., & Serfaty, D. (1999). Adaptive team coordination. Human Factors, 41, 312–325.

Copyright American Psychological Association. Not for further distribution.

Espevik, R., Johnsen, B., Eid, J., & Thayer, J. (2006). Shared mental models and operational effectiveness: Effects on performance and team processes in submarine attack teams. Military Psychology, 18, 23–36. Fandt, P., Richardson, W., & Conner, H. (1990). The impact of goal setting on team simulation experience. Simulation & Gaming, 21, 411–422. Farr, J. L., & Tippins, N. T. (Eds.). (2010). Handbook of employee selection. New York: Routledge.

Francis, D., & Young, D. (1979) Improving work groups: A practical manual for team building. San Diego, CA: Jossey-Bass. Gaddy, C., & Wachtel, J. (1992). Team skills training in nuclear power plant operations. In R. Swezey & E. Salas (Eds.), Teams: Their training and performance (pp. 379–396). Westport, CT: Ablex Publishing. Gersick, C. (1988). Time and transition in work teams: Toward a new model of group development. Academy of Management Journal, 31, 9–41. Goldstein, I. (1993). Training in organizations: Needs assessment, development, and evaluation (3rd ed.). Belmont, CA: Brooks/Cole. Goldstein, I., & Ford, J. (2002). Training in organizations: Needs assessment, development, and evaluation (4th ed.). Belmont, CA: Wadsworth.

Fiore, S., Cuevas, H., Schooler, J., & Salas, E. (2006). Cognition, teams, and team cognition: Memory actions and memory failures in distributed team environments. In C. A. Bowers, E. Salas, & F. Jentsch (Eds.), Creating high-tech teams: Practical guidance on work performance and technology (pp. 71–87). Washington, DC: American Psychological Association.

González, M., Burke, M., Santuzzi, A., & Bradley, J. (2003). The impact of group process variables on the effectiveness of distance collaboration groups. Computers in Human Behavior, 19, 629.

Fiore, S., Salas, E., & Cannon-Bowers, J. (2001). Group dynamics and shared mental model development. In M. London (Ed.), How people evaluate others in organizations (pp. 309–336). Mahwah, NJ: Erlbaum.

Graen, G., Hui, C., & Taylor, E. (2006). Experiencebased learning about LMX leadership and fairness in project teams: A dyadic directional approach. Academy of Management Learning & Education, 5, 448–460.

Fiore, S., Salas, E., Cuevas, H., & Bowers, C. (2003). Distributed coordination space: Toward a theory of distributed team process and performance. Theoretical Issues in Ergonomics Science, 4, 340. Fischer, U., McDonnell, L., & Orasanu, J. (2007). Linguistic correlates of team performance: Toward a tool for monitoring team functioning during space missions. Aviation, Space, and Environmental Medicine, 78, 86–95. Fleishman, E., & Zaccaro, S. (1992). Toward a taxonomy of team performance functions. In R. W. Swezey and E. Salas, (Eds.), Teams: Their training and performance (pp. 31–56). Westport, CT: Ablex Publishing. Foltz, P., & Martin, M. J. (2008). Automated communication analysis of teams. In E. Salas, E. Goodwin & C. S. Burke (Eds.), Team effectiveness in complex environments. Danvers, MA: CRC Press. Forester, G., Thoms, P., & Pinto, J. (2007). Importance of goal setting in virtual project teams. Psychological Reports, 100, 270–274. Fowlkes, J., Lane, N., Salas, E., Franz, T., & Oser, R. (1994). Improving the measurement of team performance: The TARGETS methodology. Military Psychology, 6, 47–61. 644

Graen, G. B., Hui, C., & Gu, Q. L. (2004). A new approach to intercultural cooperation. In G. B. Graen (Ed.), New frontiers of leadership. Greenwich, CT: Information Age.

Groom, V., & Nass, C. (2007). Can robots be teammates? Benchmarks in human-robot teams. Interaction Studies: Social Behaviour and Communication in Biological and Artificial Systems, 8, 483–500. Gully, S., & Devine, D. (1995). A meta-analysis of cohesion and performance. Small Group Research, 26, 497–517. Hackman, J. R. (1987). The design of work teams. In J. Lorsch (Ed.), Handbook of organizational behavior (pp. 315–342). Englewood Cliffs, NJ: Prentice-Hall. Halfhill, T., Nielsen, T., Sundstrom, E., & Weilbaecher, A. (2005). Group personality composition and performance in military service teams. Military Psychology, 17, 41–54. Hall, R., Stevens, R., & Torralba, T. (2002). Disrupting representational infrastructure in conversations across disciplines. Mind, Culture & Activity, 9, 179–210. Harris, T. E., & Sherblom, J. C. (2002). Small group and team communication. New York: Allyn & Bacon. Harrison, D., Price, K., Gavin, J., & Florey, A. (2002). Time, teams, and task performance: Changing effects of surface- and deep-level diversity on group

Team Development and Functioning

functioning. Academy of Management Journal, 45, 1029–1045. Harville, D., Elliott, L., & Barnes, C. (2007). Analysis of team communication and performance during sustained command-and-control operations. In R. R. Hoffman (Ed.), Expertise out of context: Proceedings of the sixth international conference on naturalistic decision making (pp. 453–472). Mahwah, NJ: Erlbaum.

Copyright American Psychological Association. Not for further distribution.

Healey, K, Milbourne, G., Aaronson, W., & Errechetti, A. (2004). The use of standardized patients in international consultation. Journal of Families, Systems and Health, 22, 368–375. Hecht, M. (1978). The conceptualization and measurement of interpersonal communication satisfaction. Human Communication Research, 4, 253–264. Helmreich, R., & Foushee, H. (1993). Why crew resource management? Empirical and theoretical bases of human factors training in aviation. In E. Wiener, B. Kanki, & R. Helmreich (Eds.), Cockpit resource management (pp. 3–45). San Diego, CA: Academic Press. Helmreich, R., Wilhelm, J., Gregorich, S., & Chidester, T. (1990). Preliminary results from the evaluation of cockpit resource management training: Performance ratings of flightcrews. Aviation, Space, and Environmental Medicine, 61, 576–579. Hertzel, G., Geister, S., & Konradt, U. (2005). Managing virtual teams: A review of current empirical research. Human Resource Management Review, 15, 69–95. Hirschfeld, R., Jordan, M., Feild, H., Giles, W., & Armenakis, A. (2006). Becoming team players: Team members’ mastery of teamwork knowledge as a predictor of team task proficiency and observed teamwork effectiveness. Journal of Applied Psychology, 91, 467–474. Hoeft, R., Jentsch, F., & Bowers, C. (2006). Creating high-tech teams: A conclusion. In C. A. Bowers, E. Salas, & F. Jentsch (Eds.), Creating high-tech teams: Practical guidance on work performance and technology (pp. 283–287). Washington, DC: American Psychological Association. Hollenbeck, J., DeRue, D., & Guzzo, R. (2004). Bridging the gap between I/O research and HR practice: Improving team composition, team training, and team task design. Human Resource Management, 43, 353–366.

Horowitz, S. K., & Horowitz, I. B. (2007). The effects of team diversity on team outcomes: A meta-analytic review of team demography. Journal of Management, 33, 987–1015. Hough, L. (1992). The “Big Five” personality variables— Construct confusion: Description versus prediction. Human Performance, 5, 139. Ilgen, D., Hollenbeck, J., Johnson, M., & Jundt, D. (2005). Teams in organizations: From input–process–output models to IMOI Models. Annual Review of Psychology, 56, 517–543. Jentsch, F., Barnett, J., Bowers, C., & Salas, E. (1999). Who is flying this plane anyway? What mishaps tell us about crew member role assignment and aircrew situation awareness. Human Factors, 41, 1–14. Johnston, M., Reed, K., Lawrence, K., & Onken, M. (2007). The link between communication and financial performance in simulated organizational teams. Journal of Managerial Issues, 19, 536–553. Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language, inference, and consciousness. Cambridge, England: Cambridge University Press. Kanter, R. M. (1983). The change masters: Innovation for productivity in the American corporation. New York: Simon & Schuster. Kichuk, S., & Wiesner, W. (1997). The Big Five personality factors and team performance: Implications for selecting successful product design teams. Journal of Engineering & Technology Management, 14, 195. Klein, G., & Pierce, L. (2001, June). Adaptive teams. Paper presented at the International Command and Control Research and Technology Symposium, Alexandria, VA. Klein, G. A., & Thorsden, M. L. (1989). A cognitive model of team decision making. Yellow Springs, OH: Klein Associates. Kleinman, D. L., & Serfaty, D. (1989). Team performance assessment in distributed decision making. In R. Gilson, J. P. Kincaid, & B. Goldiez (Eds.), Proceedings of the Interactive Networked Simulation for Training Conference (pp. 22–27). Orlando, FL: Naval Training Systems Center.

Holton, E., & Baldwin, T. (Eds.). (2003). Improving learning transfer in organizations. San Francisco: JosseyBass.

Klimoski, R., & Jones, R. G. (1995). Staffing for effective group decision making: Key issues in matching people and teams. In R. A. Guzzo & E. Salas (Eds.), Team effectiveness and decision making in organizations (pp. 291–332). San Francisco: Jossey-Bass.

Homan, A., van Knippenberg, D., Van Kleef, G., & De Dreu, C. (2007). Bridging faultlines by valuing diversity: Diversity beliefs, information elaboration, and performance in diverse work groups. Journal of Applied Psychology, 92, 1189–1199.

Kozlowski, S., & Bell, B. (2008). Team learning, development, and adaptation. In V. Sessa & M. London (Eds.), Work group learning: Understanding, improving and assessing how groups learn in organizations (pp. 15–44). New York: Taylor & Francis. 645

Cannon-Bowers and Bowers

Kozlowski, S. W. J., Gully, S. M., Nason, E. R., & Smith, E. M. (1999). Developing adaptive teams: A theory of compilation and performance across levels and time. In D. R. Ilgen & E. D. Pulakos (Eds.), The changing nature of work and performance: Implications for staffing, personnel actions, and development (pp. 240–292). San Francisco: Jossey-Bass.

Copyright American Psychological Association. Not for further distribution.

Kozlowski, S., & Ilgen, D. (2006). Enhancing the Effectiveness of Work Groups and Teams. Psychological Science in the Public Interest, 7, 77–124. Kozlowski, S., & Klein, K. (2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 3–90). San Francisco: Jossey-Bass.

Lorsch, J. (Ed.). (1987). Handbook of organizational behavior. Englewood Cliffs, NJ: Prentice-Hall. MacMillan, J., Entin, E., & Serfaty, D. (2004). Communication overhead: The hidden cost of team cognition. In E. Salas & S. M. Fiore (Eds.), Team cognition: Understanding the factors that drive process and performance (pp. 61–82). Washington, DC: American Psychological Association. Manser, T., Howard, S., & Gaba, D. (2008). Adaptive coordination in cardiac anaesthesia: A study of situational changes in coordination patterns using a new observation system. Ergonomics, 51, 1153–1178. Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based framework and taxonomy of team processes. Academy of Management Review, 26, 356–376.

Kozlowski, S., & Salas, E. (Eds.). (in press). Learning, training, and development in organizations. London: Taylor & Francis.

Marks, M., & Panzer, F. (2004). The influence of team monitoring on team processes and performance. Human Performance, 17, 25–41.

Kozlowski, S., Toney, R., Mullins, M., Weissbein, D., Brown, K., & Bell, B. (2001). Developing adaptability: A theory for the design of integrated-embedded training systems. In E. Salas (Ed.), Advances in human performance and cognitive engineering research (pp. 59–123). Greenwich, CT: JAI Press.

Marks, M., Sabella, M., Burke, C., & Zaccaro, S. (2002). The impact of cross-training on team effectiveness. Journal of Applied Psychology, 87, 3–13.

Kraut, R. E., Fussell, S. R., Lerch, F. J., & Espinosa, A. (2004). Coordination in teams: Evidence from a simulated management game. Manuscript submitted for publication. Leedom, D., & Simon, R. (1995). Improving team coordination: A case for behavior-based training. Military Psychology, 7, 109–122. LePine, J., Piccolo, R., Jackson, C., Mathieu, J., & Saul, J. (2008). A meta-analysis of teamwork processes: Tests of a multidimensional model and relationships with team effectiveness criteria. Personnel Psychology, 61, 273–307. Levine, E. L., & Dickey, M. T. (1990, August). Measuring task importance: A replication and extension. Paper presented at the 98th Annual Convention of the American Psychological Association, Boston. Levine, J., & Moreland, R. (1990). Progress in small group research. Annual Review of Psychology, 41, 585. Likert, R. (1961). New patterns of management. New York: McGraw-Hill. Lim, B., & Klein, K. (2006). Team mental models and team performance: A field study of the effects of team mental model similarity and accuracy. Journal of Organizational Behavior, 27, 403–418. Lingard, L., Espin, S., Whyte, S., Regehr, G., Baker, G., Reznick, R., et al. (2004). Communication failures in the operating room: An observational classification of recurrent types and effects. Quality & Safety in Health Care, 13, 330–334. 646

Marks, M., Zaccaro, S., & Mathieu (2000). Performance implications of leader briefings and team-interaction training for team adaptation to novel environments Journal of Applied Psychology, 85, 971–986. Martin-Milham, L., & Fiore, S. (2004). Team situation assessment training for adaptive coordination. In N. Stanton, A. Hedge, K. Brookhuis, E. Salas, & H. Hendrick (Eds.), Handbook of human factors and ergonomic methods. Boca Raton, FL: CRC Press. McCann, C., Baranski, J., Thompson, M., & Pigeau, R. (2000). On the utility of experiential cross-training for team decision making under time stress. Ergonomics, 43, 1095–1110. McClennan, J., Holgate, A., Omodei, M., & Wearing, A. (2006). Decision making effectiveness in wildfire incident management teams. Journal of Contingencies and Crisis Management, 14, 27–37. McClough, A., & Rogelberg, S. (2003). Selection in teams: An exploration of the Teamwork Knowledge, Skills, and Ability Test. International Journal of Selection & Assessment, 11, 56–66. McComb, S., & Green, S. (1999). Project goals, team performance, and shared understanding. Engineering Management Journal, 11, 7. McCrae, R. R., & Costa, P. T., Jr. (1996). Toward a new generation of personality theories: Theoretical contexts for the five-factor model. In J. S. Wiggins (Ed.), The five-factor model of personality: Theoretical perspectives (pp. 51–87). New York: Guilford Press. McGrath, J. E. (1984). Groups: Interaction and performance. Englewood Cliffs, NJ: Prentice-Hall.

Team Development and Functioning

McGrath, J., & Hollingsworth, A. (1994). Groups interacting with technology. Thousand Oaks, CA: Sage. McGregor, D. (1960). The human side of enterprise. New York: McGraw-Hill. McNeese, M., & Rentsch, J. (2001). Identifying the social and cognitive requirements of teamwork using collaborative task analysis. In M. D. McNeese, E. Salas, & M. Endsley (Eds.), New trends in cooperative activities: Understanding system dynamics in complex environments (pp. 96–113). Santa Monica, CA: Human Factors and Ergonomics Society.

Copyright American Psychological Association. Not for further distribution.

Mesmer-Magnus, J., & DeChurch, L. (2009). Information sharing and team performance: A meta-analysis. Journal of Applied Psychology, 94, 535–546. Mittleman, C., & Briggs, R. O. (1999). Communication technologies for traditional and virtual teams. In E. Sundstrom (Ed.), Supporting work team effectiveness (pp. 246–270). San Francisco: Jossey-Bass. Mohammed, S., Cannon-Bowers, J., & Foo, S. (2010). Selection for team membership: A contingency and multilevel perspective. In J. L. Farr & N. T. Tippins (Eds.), Handbook of employee selection (pp. 801–823). New York: Routledge. Mohammed, S., Klimoski, R., & Rentsch, J. (2000). The measurement of team mental models: We have no shared schema. Organizational Research Methods, 3, 123. Mohammed, S., Mathieu, J., & Bartlett, A. (2002). Technical-administrative task performance, leadership task performance, and contextual performance: Considering the influence of team- and task-related composition variables. Journal of Organizational Behavior, 23, 795. Morgan, B. B., Glickman, A. S., Woodard, E. A., Blaiwes, A. S., & Salas, E. (1986). Measurement of team behaviors in a Navy environment (Technical Report No. NTSC TR–86–014). Orlando, FL: Naval Training Systems Center. Morgan, B. B., Salas, E., & Glickman, A. S. (1993). An analysis of team evolution and maturation. Journal of General Psychology, 120, 277–291. Morgeson, F., Reider, M., & Campion, M. (2005). Selecting individuals in team settings: The importance of social skills, personality characteristics, and teamwork knowledge. Personnel Psychology, 58, 583–611. Mullen, B., & Cooper, C. (1994). The relation between group cohesiveness and performance: An integration. Psychological Bulletin, 115, 210–227. Mullen, B., Driskell, J., & Salas, E. (1998). Meta-analysis and the study of group performance. Group Dynamics, 2, 213–229. Mumford, T., Morgeson, F., Van Iddekinge, C., & Campion, M. (2008). The Team Role Test:

Development and validation of a team role knowledge situational judgment test. Journal of Applied Psychology, 93, 250–267. Nunamaker, J., Jr., Briggs, R., Mittleman, D., Vogel, D., & Balthazard, P. (1996). Lessons from a dozen years of group support systems research: A discussion of lab and field findings. Journal of Management Information Systems, 13, 163–207. Offermann, L., Bailey, J., Vasilopoulos, N., Seal, C., & Sass, M. (2004). The relative contributions of emotional competence and cognitive ability to individual and team performance. Human Performance, 17, 219–243. Paris, C., Salas, E., & Cannon-Bowers, J. (2000). Teamwork in multi-person systems: A review and analysis. Ergonomics, 43, 1052–1075. Patrashkova-Volzdoska, R., McComb, S., Green, S., & Compton, W. (2003). Examining a curvilinear relationship between communication frequency and team performance in cross-functional project teams. IEEE Transactions on Engineering Management, 50, 262–269. Patrick, J., James, N., & Ahmed, A. (2006). Human processes of control: Tracing the goals and strategies of control room teams. Ergonomics, 49, 1395–1414. Pearce, C., & Manz, C. (2005). The new silver bullets of leadership: The importance of self- and shared leadership in knowledge work. Organizational Dynamics, 34, 130–140. Pearsall, M., & Ellis, A. (2006). The effects of critical team member assertiveness on team performance and satisfaction. Journal of Management, 32, 575–594. Peeters, M., Van Tuijl, H., Rutte, C., & Reymen, I. (2006). Personality and team performance: A metaanalysis. European Journal of Personality, 20, 377–396. Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. (1999). Exploring the black box: An analysis of work group diversity, conflict, and performance. Administrative Science Quarterly, 44, 1–28. Peters, T. (1988). Thriving on chaos: Handbook for a management revolution. New York: Knopf. Porter, C. (2005). Goal orientation: Effects on backing-up behavior, performance, efficacy, and commitment in teams. Journal of Applied Psychology, 90, 811–818. Porter, C., Hollenbeck, J., Ilgen, D., Ellis, A., & West, B. (2003). Backing up behaviors in teams: The role of personality and legitimacy of need. Journal of Applied Psychology, 88, 391–403. Porter, T., & Lilly, B. (1996). The effects of conflict, trust, and task commitment on project team performance. International Journal of Conflict Management, 7, 361–376. 647

Cannon-Bowers and Bowers

Prewett, M. S., Gray, A. A., Stilson, R. B., Rossi, M. E., & Brannick, M. T. (2009). A meta-analysis of team personality composition and team performance: The impact of criterion choice, task type, and interdependence. Human Performance, 22, 273–296.

Copyright American Psychological Association. Not for further distribution.

Priest, H., Stagl, K., Klein, C., Salas, E., & Burke, C. S. (2006). Virtual teams: Creating context for distributed teamwork. In C. A. Bowers, E. Salas, & F. Jentsch (Eds.), Creating high-tech teams: Practical guidance on work performance and technology (pp. 185–212). Washington, DC: American Psychological Association. Prince, A., Brannick, M., Prince, C., & Salas, E. (1997). The measurement of team process behaviors in the cockpit: Lessons learned. In M. T. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods, and applications (pp. 289–310). Mahwah, NJ: Erlbaum. Prince, C., & Salas, E. (2000). Team situation awareness, errors, and crew resource management: Research integration for training guidance. In M. R. Endsley & D. J. Garland (Eds.), Situation awareness analysis and measurement (pp. 325–347). Mahwah, NJ: Erlbaum. Proctor, M., Panko, M., & Donovan, S. (2004). Considerations for training team situation awareness and task performance through PC-gamer simulated multiship helicopter operations. International Journal of Aviation Psychology, 14, 191–205. Rentsch, J., & Hall, R. J. (1994). Members of great teams think alike: A model of team effectiveness and schema similarity among team members. Advances in Interdisciplinary Studies of Work Teams, 1, 223–261. Resick, C., & Bloom, A. (1997). Effects of goal setting on goal commitment, team processes, and performance. Psychology: A Journal of Human Behavior, 34, 2–8. Rice, A. K. (1953). Productivity and social organization in an Indian weaving shed: An examination of the socio-technical system of an experimental automatic loomshed. Human Relations, 6, 297–329. Rico, R., Sánchez-Manzanares, M., Gil, F., & Gibson, C. (2008). Team implicit coordination processes: A team knowledge-based approach. Academy of Management Review, 33, 163–184. Rosen, B., Furst, S., & Blackburn, R. (2007). Overcoming barriers to knowledge sharing in virtual teams. Organizational Dynamics, 36, 259–273. Rosen, M. A., Salas, E., Wilson, K. A., King, H. B., Salisbury, M., Augenstein, J. S., et al. (2008). Measuring team performance in simulation-based training: Adopting best practices in healthcare. Simulation in Healthcare, 3, 33–41. Rouse, W., Cannon-Bowers, J., & Salas, E. (1992). The role of mental models in team performance in complex systems. IEEE Transactions on Systems, Man, & Cybernetics, 22, 1296–1308. 648

Rousseau, V., Aubé, C., & Savoie, A. (2006). Teamwork Behaviors: A Review and an Integration of Frameworks. Small Group Research, 37, 540–570. Saavedra, R., Earley, P., & Van Dyne, L. (1993). Complex interdependence in task-performing groups. Journal of Applied Psychology, 78, 61–72. Salas, E., Bowers, C., & Cannon-Bowers, J. (1995). Military team research: 10 years of progress. Military Psychology, 7, 55–75. Salas, E., Burke, C., Bowers, C., & Wilson, K. (2001). Team training in the skies: Does crew resource management (CRM) training work?. Human Factors, 43, 641. Salas, E., & Cannon-Bowers, J. (1997). Methods, tools, and strategies for team training. In M. Quinones & A. Ehrenstein (Eds.), Training for a rapidly changing workplace: Applications of psychological research (pp. 249–279). Washington, DC: American Psychological Association. Salas, E., & Cannon-Bowers, J. A., (2000). The anatomy of team training. In S. Tobias & D. Fletcher (Eds.), Training and retraining: A handbook for businesses, industry, government, and military (pp. 312–335). Farmington Hills, MI: Macmillan. Salas, E., Dickinson, T., Converse, S., & Tannenbaum, S. (1992). Toward an understanding of team performance and training. In R. Swezey & E. Salas (Eds.), Teams: Their training and performance (pp. 3–29). Westport, CT: Ablex Publishing. Salas, E., Nichols, D., & Driskell, J. (2007). Testing three team-training strategies in intact teams: A metaanalysis. Small Group Research, 38, 471–488. Salas, E., Prince, C., Baker, D., & Shrestha, L. (1995). Situation awareness in team performance: Implications for measurement and training. Human Factors, 37, 123–136. Salas, E., Sims, D., & Burke, C. (2005). Is there a Big Five in teamwork? Small Group Research, 36, 555–599. Sanchez, J., & Levine, E. (1989). Determining important tasks within jobs: A policy-capturing approach. Journal of Applied Psychology, 74, 336–342. Saunders, C., Robey, D., & Vaverek, K. (1994). The persistence of status differentials in computer conferencing. Human Communication Research, 20, 443–472. Schatz, S., Cannon-Bowers, J., & Bowers, C. (2006). Effects of data visualizations on group decision making. In C. A. Bowers, E. Salas, & F. Jentsch (Eds.), Creating high-tech teams: Practical guidance on work performance and technology (pp. 51–67). Washington, DC: American Psychological Association. Schvaneveldt, R. W. (1990). Proximities, networks, and schemata. In R. Schvaneveldt (Ed.), Pathfinder asso-

Team Development and Functioning

ciative networks: Studies in knowledge organization (pp. 135–148). Norwood, NJ: Ablex. Scielzo, S., Fiore, S., Jentsch, F., & Finkelstein, N. (2006). Cognition and collaboration in hybrid human-robot teams: Viewing workload and performance through the lens of multimedia cognitive load. In N. J. Cooke, H. Pringle, & H. Pedersen, H. (Eds.), Human factors of remotely piloted vehicles (pp. 329–342). Amsterdam, the Netherlands: Elsevier.

Copyright American Psychological Association. Not for further distribution.

Senecal, J., Loughead, T., & Bloom, G. (2008). A seasonlong team-building intervention: Examining the effect of team goal setting on cohesion. Journal of Sport & Exercise Psychology, 30, 186–199. Serfaty, D., Entin, E., & Johnston, J. (1998). Team coordination training. In J. Cannon-Bowers & E. Salas (Eds.), Making decisions under stress: Implications for individual and team training (pp. 221–245). Washington, DC: American Psychological Association. Shebilske, W., Jordon, J., Goettl, B., & Paulus, L. (1998). Observation versus hands-on practice of complex skills in dyadic, triadic, and tetradic training-teams. Human Factors, 40, 525–540. Siebold, G. (2006). Military group cohesion. In T. W. Britt, C. A. Castro, & A. B. Adler (Eds.), Military life: The psychology of serving in peace and combat: Vol. 1. Military performance (pp. 185–201). Westport, CT: Praeger Security International. Siegel, J., Dubrovsky, V., Kiesler, S., & McGuire, T. (1986). Group processes in computer-mediated communication. Organizational Behavior and Human Decision Processes, 37, 157–187. Smith, E., Ford, J., & Kozlowski, S. (1997). Building adaptive expertise: Implications for training design strategies. In M. A. Quinones & A. Dudda (Eds.), Training for a rapidly changing workplace: Applications of psychological research (pp. 89–118). Washington, DC: American Psychological Association. Smith-Jentsch, K., Cannon-Bowers, J., Tannenbaum, S., & Salas, E. (2008). Guided team self-correction: Impacts on team mental models, processes, and effectiveness. Small Group Research, 39, 303–327. Smith-Jentsch, K., Mathieu, J., & Kraiger, K. (2005). Investigating linear and interactive effects of shared mental models on safety and efficiency in a field setting. Journal of Applied Psychology, 90, 523–535. Smith-Jentsch, K., Salas, E., & Baker, D. (1996). Training team performance-related assertiveness. Personnel Psychology, 49, 909–936. Smith-Jentsch, K., Zeisig, R., Acton, B., & McPherson, J. (1998). Team dimensional training: A strategy for guided team self-correction. In J. A. Cannon-Bowers & E. Salas (Eds.), Making decisions under stress: Implications for individual and team training (pp. 271–297). Washington, DC: American Psychological Association.

Smolek, J., Hoffman, D., & Moran, L. (1999). Organizing teams for success. In E. Sundstrom and Associates (Ed.), Supporting work team effectiveness: Best management practices for fostering high performance (pp. 24–62). San Francisco: Jossey-Bass. Spiro, R., Feltovich, P., Jacobson, M., & Coulson, R. (1992). Cognitive flexibility, constructivism, and hypertext: Random access instruction for advanced knowledge acquisition in ill-structured domains. In T. Duffy & D. Jonassen (Eds.), Constructivism and the technology of instruction: A conversation (pp. 57–75). Hillsdale, NJ: Erlbaum. Sproull, L., & Kiesler, S. (1986). Reducing social context cues: Electronic mail in organizational communication. Management Science, 32, 1492–1512. Sproull, L., & Kiesler, S. (1991). Two-level perspective on electronic mail in organizations. Journal of Organizational Computing & Electronic Commerce, 1, 125. Staples, D., & Webster, J. (2008). Exploring the effects of trust, task interdependence and virtualness on knowledge sharing in teams. Information Systems Journal, 18, 617–640. Stark, E., Shaw, J., & Duffy, M. (2007). Preference for group work, winning orientation, and social loafing behavior in groups. Group & Organization Management, 32, 699–723. Steadman, R. H., Coates, W. C., Huang, Y. M., Matevosian, R., Larmon, B. R., McCullough, L., & Ariel, D. (2006). Simulation-based training is superior to problembased learning for the acquisition of critical assessment and management skills. Critical Care Medicine, 34, 151–157. Steiner (1972). Group process and productivity. New York: Academic Press. Stevens, M., & Campion, M. (1994). The knowledge, skill, and ability requirements for teamwork: Implications for human resource management. Journal of Management, 20, 503–530. Stevens, M., & Campion, M. (1999). Staffing work teams: Development and validation of a selection test for teamwork settings. Journal of Management, 25, 207–228. Stewart, G. L. (2006). A meta-analytic review of relationships between team design features and team performance. Journal of Management, 32, 29. Stout, R., Cannon-Bowers, J., Salas, E., & Milanovich, D. (1999). Planning, shared mental models, and coordinated performance: An empirical link is established. Human Factors, 41, 61–71. Straus, S. (1997). Technology, group process, and group outcomes: Testing the connections in computermediated and face-to-face groups. Human-Computer Interaction, 12, 227. 649

Cannon-Bowers and Bowers

Sundstrom, E. (1999). Supporting work team effectiveness. San Francisco: Jossey-Bass. Sundstrom, E., DeMeuse, K., & Futrell, D. (1990). Work teams: Applications and effectiveness. American Psychologist, 45, 120–133. Sundstrom, E., McIntyre, M., Halfhill, T., & Richards, H. (2000). Work groups: From the Hawthorne studies to work teams of the 1990s and beyond. Group Dynamics: Theory, Research, and Practice, 4, 44–67.

Copyright American Psychological Association. Not for further distribution.

Svensson, J., & Andersson, J. (2006). Speech acts, communication problems, and fighter pilot team performance. Ergonomics, 49, 1226–1237. Swezey, R., Owens, J., Bergondy, M., & Salas, E. (1998). Task and training requirements analysis methodology (TTRAM): An analytic methodology for identifying potential training uses of simulator networks in teamwork-intensive task environments. Ergonomics, 41, 1678–1697. Tannenbaum, S., Beard, R., & Salas, E. (1992). Team building and its influence on team effectiveness: An examination of conceptual and empirical developments. In K. Kelley (Ed.), Issues, theory, and research in industrial/organizational psychology (pp. 117–153). Oxford, England: North-Holland. Tannenbaum, S., Smith-Jentsch, K., & Behson, S. (1998). Training team leaders to facilitate team learning and performance. In J. A. Cannon-Bowers & E. Salas (Eds.), Making decisions under stress: Implications for individual and team training (pp. 247–270). Washington, DC: American Psychological Association. Tasa, K., Seijts, G., & Taggar, S. (2007). The development of collective efficacy in teams: A multilevel and longitudinal perspective. Journal of Applied Psychology, 92, 17–27. Taylor, J., & Bowers, D. (1972). Survey of organizations: A machine-scored standardized questionnaire instrument. Ann Arbor: University of Michigan. Tesluk, P., Mathieu, J., Zaccaro, S., & Marks, M. (1997). Task and aggregation issues in the analysis and assessment of team performance. In M. T. Brannick, E. Salas, & C. Prince (Eds.), Team performance assessment and measurement: Theory, methods, and applications (pp. 197–224). Mahwah, NJ: Erlbaum. Thompson, J. D. (1967). Organizations in action. New York: McGraw-Hill. Thompson, L., & Coovert, M. (2003). Teamwork online: The effects of computer conferencing on perceived confusion, satisfaction, and postdiscussion accuracy. Group Dynamics: Theory, Research, and Practice, 7, 135–151. Tobias, S., & Fletcher, D. (Eds.). (2000). Training and retraining: A handbook for businesses, industry, government, and military. Farmington Hills, MI: Macmillan.

650

Trist, E., & Bamforth, K. (1951). Some social and psychological consequences of the Longwall method of coal-getting. Human Relations, 4, 3–38. Tuckman, B. (1965). Developmental sequence in small groups. Psychological Bulletin, 63, 384–399. Urban, J., Bowers, C., Monday, S., & Morgan, B. (1995). Workload, team structure, and communication in team performance. Military Psychology, 7, 123–123. Van de Ven, A. H., Delbecq, A. L., & Koenig, R., Jr. (1976). Determinants of coordination modes within organizations. American Sociological Review, 41, 322–338. Van de Ven, A. H., & Delbecq, A. L., & Koenig, R., Jr. (1976). Determinants of coordination modes within organizations. American Sociological Review, 41, 322–338. Volpe, C., Cannon-Bowers, J., Salas, E., & Spector, P. (1996). The impact of cross-training on team functioning: An empirical investigation. Human Factors, 38, 87–100. Walvoord, A., Redden, E., Elliott, L., & Coovert, M. (2008). Empowering followers in virtual teams: Guiding principles from theory and practice. Computers in Human Behavior, 24, 1884–1906. Webber, S., & Donahue, L. (2001). Impact of highly and less job-related diversity on work group cohesion and performance: A meta-analysis. Journal of Management, 27, 141. Weisband, S. (2002). Maintaining awareness in distributed team collaboration: Implications for leadership and performance. In P. Hinds & S. Kiesler, Distributed work (pp. 311–333). Cambridge, MA: MIT Press. Westbrook, L. (2006). Mental models: A theoretical overview and preliminary study. Journal of Information Science, 32, 563–579. Wheeler, S., Goldie, J., & Hicks, C. (1998). Counselor training: An evaluation of the effectiveness of a residential outdoor pursuits activity weekend on the personal development of trainee counselors. Counselling Psychology Quarterly, 11, 391–405. Wiener, E., Kanki, B., & Helmreich, R. (1993). Cockpit resource management. San Diego, CA: Academic Press. Yoo, Y., & Kanawattanachai, P. (2001). Developments of transactive memory systems and collective mind in virtual teams. International Journal of Organizational Analysis, 9, 187–208. Zhang, S., Fjermestad, J., & Tremaine, M. (2005). Leadership styles in virtual team context: Limitations, solutions, and propositions. Proceedings of the 38th Hawaii International Conference on System Sciences, 1, 48c.

CHAPTER 20

WORK TEAM DIVERSITY

Copyright American Psychological Association. Not for further distribution.

Susan E. Jackson and Aparna Joshi

Most organizations in the United States recognize workplace diversity as a challenging reality. Consider the case of a Fortune 500 company that we have studied (Jackson & Joshi, 2004; Joshi, Liao, & Jackson, 2006). Nearly 3 decades ago, in response to Title VII law and the civil rights movement, this company initiated proactive management practices to increase the representation of women and minorities at entry levels as well as in management. However, despite persistent top management support for its diversity policies and external recognition for its diversity management practices, our recent analysis of company data revealed that ethnically diverse work teams experienced lower levels of cooperation and higher levels of conflict, compared to ethnically homogeneous work teams. We also found that gender- and ethnicity-based pay differentials persisted for equally performing employees in the same jobs, despite our subject company’s efforts to eliminate such inequities. As this company discovered, it is easier to create a diverse organization than it is to manage a diverse organization effectively and achieve its full potential (Jackson & Joshi, 2004; Joshi, Liao, & Jackson, 2006). The conditions that have increased diversity in organizations are many and include forces both within and external to organizations. For most U.S. employers, today’s concerns about diversity are rooted in a history that reaches back half a century or more. An achievement of America’s civil rights movement, Title VII of the 1964 Civil Rights Act

made it illegal to engage in employment practices that discriminate on the basis of race, color, religion, sex, and national origin. Subsequent federal, state, and local legislation made it illegal for employers to discriminate on the basis of other characteristics, including age, disability, and in some places, sexual orientation. In supporting such legislation, American society endorsed the principle that employers should provide equal employment opportunities to all people of similar qualifications and accomplishments. Societal mandates have helped curtail employment policies and practices that discriminated against members of various minority groups and imposed both monetary and reputational costs on organizations that failed to implement fair employment practices (Wright, Ferris, Hiller, & Kroll, 1995). The subject company’s responses to Title VII and related litigation led to the adoption of aggressive affirmative action programs and the design of new approaches to measuring employee performance. The primary goal of such organizational responses was to reduce differences in employment conditions and outcomes among employees from different demographic groups. By the late 1970s and into the 1980s, many employers began to understand that simply avoiding discriminatory employment practices was not sufficient. As workplaces became more diverse, employers saw that new management initiatives were needed to ensure that the talents of all employees were leveraged to achieve organizational goals. According to

We thank Hyuntak Roh for his excellent research assistance in preparing this chapter.

http://dx.doi.org/10.1037/12169-020 APA Handbook of Industrial and Organizational Psychology, Vol 1: Building and Developing the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

651

Copyright American Psychological Association. Not for further distribution.

Jackson and Joshi

one survey of Fortune 1000 companies, by the beginning of the 21st century, 95% of large U.S. companies had implemented diversity initiatives to address racial and gender diversity (Grensing-Pophal, 2002). Foremost among the new initiatives were training programs aimed at changing employees’ attitudes and behaviors. Other popular diversity management initiatives included mentoring for women and minorities and supporting identity-based affinity social networks. The widespread adoption of training programs intended to improve relationships among employees from diverse backgrounds quickly led to an expanded meaning for the concept of diversity. Employers realized that visible, legally protected employee attributes such as ethnicity and gender were not the only types of differences that mattered in the workplace. Today the concept of diversity is used by employers and scholars alike as shorthand to refer to the wide range of physical, cultural, psychological, and behavioral differences that can be found in most large organizations. At the company we studied for example, diversity management practices are aimed at ensuring that all employees feel they are treated fairly. In this new environment, the specific concerns of majority group members in the workforce (e.g., White men, in the United States) receive as much attention as the concerns of minority group members. In the 1990s, employers began to embrace the argument that supporting diversity made good business sense. Among the many arguments made in favor of recruiting and retaining a diverse workforce was the assertion that diversity could contribute to better decision making and innovation and thereby improve the financial bottom line (Kochan et al., 2003). To date, there is some limited empirical evidence to show that diversity promotes improved organizational performance (Richard, Barnett, Dwyer, & Chadwick, 2004). Yet, in organizations that have become flatter, less bureaucratic, and more reliant on teamwork (e.g., see Harris & Beyerlein, 2003), diversity can be disruptive. Despite the challenges, U.S. employers generally concur that effectively managing diversity is mandatory for organizations that seek to fully utilize all of the talent that is available in the workforce. Increasingly, this important management challenge is being recognized in other countries, too 652

(The Conference Board, 2006; Lester, 2006; Mangaliso & Nkomo, 2001; Nishii & Özbilgin, 2007b; Wu, Yi, & Lawler, 2000). Like employers, organizational scholars have struggled to understand how diversity shapes employees’ experiences at work. The scholarly evidence confirms what employers already know— diversity seems to be a double-edged sword. On the one hand, it can bring with it interpersonal conflict, loss of social cohesion, and greater employee turnover. On the other hand, it can spur innovation and improve decision making (for other comprehensive reviews, see Jackson, Joshi, & Erhardt, 2003; Jackson, May, & Whitney, 1995; Milliken & Martins, 1996; Webber & Donahue, 2001; van Knippenberg & Schippers, 2007; Williams & O’Reilly, 1998). A major challenge for organizational psychologists is to understand the dynamics of diversity well enough to offer practical advice about how to manage it effectively. In the past 2 decades, much of the empirical research on diversity in organizations has focused on improving our understanding of the consequences of diversity in work teams and other relatively small work units. The central research question has been: To what extent and in what ways do work teams consisting of individuals who are relatively similar to each other function differently from teams consisting of individuals who are dissimilar? Typically, studies of work team diversity document the types of diversity present in teams and then assess whether differing degrees and types of diversity are systematically related to team processes and outcomes. (In contrast, research on discrimination and fairness often compares the employment experiences of individual employees from different backgrounds.) Guiding much of the research on work team diversity is a desire to understand the conditions that enable some diverse teams to effectively pool and use their differences to achieve outstanding performance, while avoiding dysfunctional conflicts. In this chapter, we report the current state of our knowledge about the interpersonal dynamics that unfold within (usually) colocated work teams and small work units, as well as the consequences of such dynamics. Throughout this chapter, we use the term work team to refer to organizational units of at least

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

3 and seldom more than 50 employees with responsibilities that require them to work interdependently. We focus on work teams as the unit of analysis rather than organizations because the vast majority of accumulated psychological research is based on studies of work teams or similar small units within organizations. Examples of the types of work teams that have been studied include top management teams, customer service teams, production teams, sales teams, and research and development teams. (See also chap. 19, this volume; Vol. 3, chap. 9, this handbook.) To maximize the workplace applicability of conclusions drawn from this review, we focus on research conducted in employment settings. By focusing on research conducted in group dynamics, much of which was conducted in other settings (e.g., school classrooms and laboratories, college sports, juries). We acknowledge the importance of that research and refer to some of it, but we do not claim to provide a comprehensive review of it here. We also note that most of the research we review was conducted in North American organizations and published in English-language journals. Increasingly in recent years, studies of work team diversity conducted outside North America also have begun to appear in English-language journals. We include such research in our review whenever possible, while recognizing that the phenomena of diversity cannot be fully understood without taking national contexts into account (Joshi & Roh, 2007; Triandis, 1992). Our discussion proceeds as follows: We begin by describing the many types of work team diversity that have been studied in organizational settings. Next, we offer brief overviews of the major theoretical perspectives that have guided research aimed at understanding the consequences associated with various types of work team diversity. We then summarize the empirical research findings that have accumulated during the past 15 years. Consistent with our focus on research that can be applied in the workplace, our review is organized around the most frequently studied consequences of team diversity (for a recent review focused on theoretical advances in understanding diversity, see van Knippenberg & Schippers, 2007). After describing

the research findings, we conclude by providing suggestions for future research. THE NATURE OF WORK TEAM DIVERSITY The term diversity is now widely used by scholars to refer to the composition of social units (for an overview of the debates and history associated with the term, see Ashkanasy, Härtel, & Daus, 2002). Work team diversity is usually measured using compositional measures that assess the distribution of team members’ personal attributes, including their demographic and psychological characteristics.

Types of Diversity To understand how diversity influences work teams, it is useful to differentiate among various types of diversity, because different types of diversity may have different consequences. A simple taxonomy for describing the types of diversity present in work teams is shown in Exhibit 20.1. The columns in Exhibit 20.1 differentiate between relations-oriented and task-related diversity (Jackson et al., 1995; Milliken & Martins, 1996). Relations-oriented diversity refers to the distribution of attributes that are instrumental in shaping interpersonal relationships but which typically have no apparent direct implications for task performance; age, gender, and personality characteristics are examples of relations-oriented diversity. Task-oriented diversity refers to the distribution of attributes that are potentially relevant to the team’s work. Organizational tenure, formal credentials and titles, and cognitive abilities are examples of task-oriented diversity. The rows in Table 20.1 differentiate between readily detected (or surface-level) diversity and underlying or deep-level diversity. Generally, readily detected diversity refers to differences among team members on attributes such as gender, age, and nationality—attributes that are easily discerned or quickly discovered. Sociological explanations of diversity emphasize the role of memberships in easily discerned social groupings because social groups are assumed to be in competition with each other for material and social resources (e.g., Blalock, 1967). Underlying diversity refers to differences among team members on attributes that generally become 653

Jackson and Joshi

Exhibit 20.1 A Taxonomy for Describing Types of Work Team Diversity

Copyright American Psychological Association. Not for further distribution.

Diversity on relationshiporiented attributes

Diversity on task-oriented attributes

Diversity on readily detected attributes

Gender Age Ethnicity Nationality Religion

Department/unit membership Organizational tenure Formal credentials and titles Education level Memberships in professional associations

Diversity on underlying attributes

Personality Attitudes Values Racial/ethnic identity Sexual identity Other social identities

Task knowledge Organizational knowledge Experience Cognitive abilities Communication skills Mental models

Note. From International Handbook of Organizational Teamwork and Cooperative Working (p. 279) by M. A. West, D. Tjosvold, and K. Smith (Eds.), 2003, New York: Wiley. Copyright 2003 by Wiley. Adapted with permission.

known only through interaction, such as personality, attitudes, and skills. Many psychological explanations emphasize the role of individual differences in characteristics that are less easily detected and discovered only through direct interaction. Psychological explanations of diversity often assume that readily detected attributes are of little theoretical interest except to the extent that they serve as easy-to-measure correlates or indicators of underlying attributes. Contrary to some others (e.g., van Knippenberg, De Dreu, & Homan, 2004; van Knippenberg & Schippers, 2007) who favor abandoning such typologies, we believe that paying attention to the distinctions shown in Exhibit 20.1 is useful for interpreting past research and is necessary as the field continues to build a systematic body of evidence about the consequences of diversity. Those who are pessimistic about the value of such distinctions point out that there is no clear pattern of results associated with different types of diversity. Our view is that the body of evidence to date is still quite small; there are far too few studies to permit firm conclusions about whether and how the potential consequences of various types of diversity differentially influence work teams. 654

Measuring Diversity In everyday conversations, people sometimes confuse the concept of diversity, which is a characteristic of social units such as work teams, with individual attributes. For example, an African American colleague might be referred to as “a diverse employee.” Or one may use the word diverse as if it were a synonym for minority, as when a team of five African Americans is referred to as “a diverse team.” Compared with its usage in everyday conversations, the concept of diversity is used more narrowly in the scholarly literature to describe the extent to which members of social units (e.g., work teams) are dissimilar from each other on one or more attributes. High levels of team diversity exist to the extent that members of a team are different from each other (i.e., the team is heterogeneous), and low levels of diversity exist when team members are similar to each other (i.e., the team is homogeneous). Thus, in the scholarly literature, a work team of five African American sales employees would be correctly described as having a very low degree of ethnic diversity because they all have the same ethnic background. The same reasoning would apply to a team of five Caucasian sales employees. Most theoretical perspectives would expect similar team

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

dynamics to occur in the all–African American and all-Caucasian work teams, because both teams are homogeneous. Of primary interest in studies of team diversity are comparisons of fairly homogeneous versus more heterogeneous teams. A comprehensive discussion of the measures that have been used to assess work team diversity is beyond the scope of this chapter (for more detailed discussions, see Harrison & Klein, 2007; Harrison & Sin, 2005; Shaw, 2004). In general, however, most studies attempt to use relatively objective measures of diversity rather than subjective perceptions of diversity. Typically, the attributes of individual team members are assessed and used to create teamlevel statistical indicators that capture the degree of heterogeneity or dispersion present. For categorical attributes (such as gender and ethnicity), greater diversity is present when team members are equally distributed among the possible categories (e.g., a six-person team with three men and three women, or an eight-person team with two engineers, two artists, two salespeople, and two support staff). For attributes that vary along a continuum, such as age and tenure, diversity is greatest to the extent that members of a team are distributed between the highest and lowest values on the continuum.

The Complexity of Diversity Work teams found in natural settings are likely to be most accurately described as highly diverse (heterogeneous) on some attributes and less diverse (homogeneous) on other attributes. For example, in our subject company (Jackson & Joshi, 2004; Joshi, Liao, & Jackson, 2006), sales teams were typically diverse in terms of gender and ethnicity. The company’s staffing practices, however, ensured that sales teams were more homogeneous in education levels and organizational tenure. Typically, researchers either fail to measure and/or do not report statistics for all of the many types of diversity that may be present in the work teams being investigated. In fact, for the studies we located, often results were reported for just two types of diversity. Gender and ethnic diversity were reported most often. Later in this chapter we present a more detailed description of the types of work team diversity that have been studied and the

number of studies that have reported results for several types of team diversity. Those data show that there is a substantial body of evidence concerning the effects of readily detected diversity in work teams and relatively little evidence concerning the effects of underlying diversity in work teams. The study of work team diversity is further complicated by the fact that attributes of the employees who work together in a team may be correlated. For example, shortly after the company that we studied adopted new recruiting practices aimed at attracting more ethnic minorities, the association between ethnicity and job tenure increased (Jackson & Joshi, 2004; Joshi, Liao, & Jackson, 2006). Likewise, after the company adopted promotion practices designed to increase the number of women in management ranks, the association between gender and job tenure rose in the managerial ranks. Also, throughout the company, older employees tended to have fewer years of formal education regardless of which jobs they held. When correlations such as these exist, it is difficult to determine the unique effects of different types of diversity. Finally, note that most studies have focused on the main effects of work team diversity, and few studies have reported on the potential interaction effects among various types of diversity (e.g., Barrick, Stewart, Neubert, & Mount, 1998; Neuman, Wagner, & Christiansen, 1999). A study of 83 work teams in eight German organizations illustrated the value of studying such interaction effects: The results showed that teams consisting of employees with higher need for cognition (i.e., an intrinsic motivation for and enjoyment of cognitive activities) were more likely to reap the benefits of age and education diversity, compared with teams consisting of employees with lower need for cognition (Kearney, Gerbert, & Voelpel, 2009). The complexity of diversity found in naturally occurring work teams, combined with overly simplistic approaches to measuring team diversity, makes it impossible to draw conclusions about whether the different types of diversity shown in Exhibit 20.1 are associated with differing team consequences. Therefore, we encourage scholars to be vigilant about measuring, evaluating, and reporting 655

Copyright American Psychological Association. Not for further distribution.

Jackson and Joshi

the effects of as many types of diversity as is feasible when conducting their research. For employers, the best advice we can offer concerning different types of diversity is to assume that not all types of diversity are created equal—some types of diversity may have greater potential benefits than other types, and some types may require more active management to avoid potentially disruptive consequences. We describe what is known about the consequences of various types of diversity later in this chapter. Before doing so, however, we briefly summarize the theoretical underpinnings of that empirical knowledge base. THEORETICAL FOUNDATIONS OF RESEARCH ON WORK TEAM DIVERSITY Several theoretical perspectives have informed the study of work team diversity. Not all of these theoretical perspectives were proposed with the primary objective of explaining the dynamics of diverse work teams. Some theories were developed to understand organization-level phenomena, and others focus on individual-level phenomena. Nevertheless, these theoretical perspectives have proved useful for informing diversity scholars of the various consequences of work team diversity. Together, they provide many useful insights into how, why, when, and where diversity influences work team processes and outcomes.

Attraction–Selection–Attrition Model The attraction–selection–attrition (ASA) model (Schneider, 1987) is one of the most popular perspectives used to guide empirical investigations of work team diversity, perhaps because it recognizes diversity’s double-edged consequences. (See also Vol. 3, chap. 1, this handbook.) Schneider and his colleagues have focused their efforts on understanding the dynamics that tend to cause whole organizations to become less diverse and more homogeneous over time, but the ASA model is invoked to explain the effects of work team diversity, too. To describe the processes through which “the people make the place,” Schneider argued that organizations naturally evolve toward greater social homogeneity because people prefer to be 656

with others who are similar (Byrne, 1971). When looking for jobs, people are initially attracted to organizations that they believe are made up of people like themselves. When evaluating job applicants, organizational agents are more likely to form favorable impressions of applicants who seem to “fit” the organization. Once hired, perceptions of similarity continue to play a role: Employees who do not seem to fit in well are more likely to experience dissatisfaction and leave. These dynamics, unfolding year after year, gradually result in organizations that are more homogeneous than they would become through random processes (e.g., see Schneider, Smith, Taylor & Fleenor, 1998; Smith, 2008). Homogeneous organizations may function more smoothly as a consequence of the similarities shared by members, but there also are potential disadvantages, such as lack of creativity and an inability to adapt to changing circumstances. The ASA model emphasizes the role of employees’ personalities, values, and interests as forces that shape organizational life (Schneider, Goldstein, & Smith, 1995). However, some scholars have argued that ASA processes also explain the gradual demographic homogenization that occurs in organizations (Boone, van Olffen, van Witteloostuijn, & De Brabander, 2004; Jackson et al., 1991; Jackson & Chung, 2008). For a Fortune 500 company such as the one we studied, the ASA perspective suggests that proactively increasing diversity within the firm may be necessary in order to contradict the natural drift toward homogeneity. At the same time, however, efforts to increase diversity may be at odds with the propensity of employees to be more attracted to and feel more comfortable working with people who are similar to them. Thus, work teams and the company as a whole may find that increasing diversity results in higher employee turnover and the various costs associated with replacing those who leave. This cost is worth paying if the organization reaps other benefits from diversity.

Organizational Demography Perspective Diversity researchers also embrace the more sociological logic of Pfeffer’s (1983) organizational demography model. In common with the ASA

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

model, employees’ preferences for similarity provide the rationale for the relationship of the social composition of organizations with organizational phenomena. But in contrast to Schneider’s focus on the behavior of individuals, Pfeffer emphasized organization-level constructs such as cohesiveness, communication patterns, and employee flows. In place of psychological attributes such as personality and values, the organizational demography perspective highlights the importance of membership in social groups defined by attributes such as age, tenure, gender, and ethnicity. Sociological studies and marketing research have both shown that differences in people’s attitudes and values are reliably associated with differences in their standing on demographic characteristics such as these. The demography perspective recognizes that people use the social cues of demographic attributes to inform their behavior toward others. Understanding the effects of tenure cohorts in organizations has been a primary focus of studies of organizational demography (cf. Pfeffer, 1992). Research shows that tenure-based cohorts produce two types of effects in organizations: competition between cohort groups and solidarity within cohort groups. Relationships between and within tenure cohorts, in turn, have implications for outcomes such as cohesiveness and turnover. Wagner, Pfeffer, and O’Reilly (1984) argued that executives who joined a company at the same time were likely to develop similar values and patterns of communication. Consistent with this logic, they found that tenure diversity within top management teams was associated with higher turnover rates for the team. In a study of academic departments, McCain, O’Reilly, and Pfeffer (1983) assessed the extent to which department members could be clearly arrayed into distinct cohort sizes with clear gaps between the cohorts (e.g., a department with a group of long-tenured full professors and a group of short-tenured assistant professors versus a department with an array of professors across all levels of seniority). Departments with more clearly identifiable cohorts experienced more intergenerational conflict as well as higher rates of faculty resignations at senior and junior levels (McCain et al., 1983).

The organizational demography perspective has drawn attention primarily to tenure and age distributions in organizations, but the main principles of this theory are easily extended to understanding the consequences of cohorts based on other demographic characteristics, including gender, ethnicity, and educational background. The organizationallevel dynamics that were of primary interest to Pfeffer (1983) also have been observed in smaller social units, including work teams (e.g., Jackson et al., 1991; Jackson & Joshi, 2004; Joshi et al., 2006; Joshi, 2006).

Social Identity Perspective The social identity perspective encompasses social categorization theory and social identity theory (Reynolds, Turner, & Haslam, 2003). Fundamental to this perspective is the observation that individuals classify themselves and others based on overt demographic attributes, including ethnicity and gender (Ashforth & Mael, 1989; Tajfel & Turner, 1979). Demographically similar individuals classify themselves as members of the in-group; those who are demographically dissimilar are classified as the out-group. Such categorizations are fundamental to the way people understand and organize their social worlds (Hewstone, Rubin, & Willis, 2002). In-group/out-group dynamics play out even when group membership has been determined randomly on the basis of meaningless cues (Brewer & Brown, 1998). Whereas the ASA model and the organizational demography perspective assume that “objective” similarity is of primary importance, the social identity perspective recognizes that similarity is socially constructed and specific to situations. People bring many attributes to each situation, but only those that become salient shape behavior. Although many factors can influence which attributes and types of diversity become salient in a situation, the composition of the particular people present is the primary factor. For example, gender is more likely to be salient and the basis for in-group/ out-group dynamics in mixed-gender (vs. singlegender) teams. By emphasizing that it is the particular mix of people in a situation that determines which types of diversity matter most, the social 657

Copyright American Psychological Association. Not for further distribution.

Jackson and Joshi

identity perspective combines an understanding of individual-level processes with an appreciation for the role of social contexts (Chattopadhyay, Tluchowska, & George, 2004). The social identity perspective serves as a reminder that diversity dynamics must be understood within the company’s particular context. On the one hand, the basic dynamics of in-group/ out-group relations is likely to be found in any organization, and within any department or work team. On the other hand, it would be wrong to assume that the same types of diversity are equally salient and influential in all teams or departments— even those in the same organization.

Information Processing Perspective In contrast to the theoretical perspectives just described, the information processing perspective emphasizes the role of less visible attributes, such as knowledge and skills. The information processing perspective assumes that employees bring differing approaches and expertise to the decision-making activities in which most work teams engage; it focuses much more on task-oriented team activities, rather than affect-based relationships. The information processing perspective assumes that task-oriented diversity can improve team decision making in a variety of ways: More diverse teams may search more broadly for information, consider more alternative solutions, and engage in more vigorous debate before reading a decision (e.g., see Jackson, 1992). Indeed, the mere presence of a minority opinion triggers greater exchange of unshared information within teams (Nemeth, 1986, 1997). The information processing perspective has informed numerous studies designed to assess the corporate performance consequences of diversity within top management teams (Bantel & Jackson, 1989; Eisenhardt & Schoonhoven, 1990). It is central to research on teams at the upper echelons of organizations (Finkelstein & Hambrick, 1996), and it also has stimulated research on decision making and performance in lower-level work teams (e.g., Jehn & Mannix, 2001). When organizations extol the potential value of work team diversity, they often draw on the logic of the information processing perspective. 658

Social Capital Theory Social capital theory (sometimes referred to as social networks theory) also brings attention to social interactions that create value for individuals and groups (Lin, Ensel, & Vaughn, 1981). A work team’s social capital consists of the actual and potential resources embedded within the network of team members’ social and job-related relationships. Social capital theory suggests that work team diversity can be both detrimental and beneficial. The potentially detrimental effects of diversity arise because diversity is likely to inhibit the development and use of social capital among team members, which is referred to as internal social capital. Reflecting the similarity–attraction effect, social networks tend to be homophilous; that is, social interactions occur more frequently among network members who share similar attributes (Marsden, 1990; McPherson, Smith-Lovin, & Cook, 2001). Homophilous exchanges are more stable than interactions among dissimilar individuals and are characterized by greater mutual trust (Brass, 1985; Marsden, 1990). In networks characterized by diversity, dense trusting relationships are less likely to develop. Offsetting the disadvantage of diversity for developing a team’s internal social capital are advantages that diversity brings to a team’s network of relationships with people outside the team, which is referred to as the team’s external social capital. Adopting an external perspective on team work highlights the important role of the relationships between team members and people outside the team boundary who contribute to the team’s effectiveness (Ancona & Caldwell, 1992). For many work teams, effectiveness requires understanding how the team’s work is situated in the organization. Coordinating with other teams, managing relationships with external stakeholders, and obtaining access to information and other valuable resources all may be needed to achieve the team’s task (e.g., see Katz, Lazer, Arrow, & Contractor, 2004; Oh, Chung, & Labianca, 2004; Tsai & Ghoshal, 1998). For achieving task-related goals, a team’s external social capital may be as valuable as its internal social capital. As the foremost proponent of social capital theory, Burt (1992) argued that a person who bridges a

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

“structural hole” between two other unconnected actors is in an advantageous position because he or she can control the exchange of resources between the two actors. Diverse teams may be more likely to include such members. Assuming that the members of a team tend to have external connections to similar others, the external social capital of diverse teams should be greater than the external social capital of homogeneous teams. According to this perspective, diversity is a net benefit for work teams when it creates sufficiently valuable external social capital to offset the reduced internal social capital associated with diversity (Reagans & Zuckerman, 2001; Reagans, Zuckerman, & McEvily, 2004). Social capital theory suggests that managing diversity effectively requires understanding how internal and external social capital can influence the performance of various work teams. For teams working on tasks that do not benefit from the team’s external network of contacts, such as many production tasks, any internal conflicts associated with diversity may result in reduced team performance. However, for teams working on tasks that require coordination and interdependence with others outside the team, such as research and development (R&D) teams, the external social capital associated with diversity may be sufficient to offset any negative consequences of reduced internal social capital. The value of external social capital for members of corporate boards was demonstrated in a study by Westphal and Milton (2000). They found that minority board members who had social ties with other board members based on common membership in other boards were more successful in exerting influence on their boards. These findings draw attention to the embeddedness of diversity-related outcomes in the larger socio-structural context. The logic of social capital theory is consistent with the investments that major corporations have made in supporting diversity initiatives such as formal mentoring programs and caucus or affinity groups. Such programs encourage employees to build their networks of social and task-related connections across organizational levels and various other internal boundaries (Killian, Hukai, & McCarty, 2005; Noe, Greenberger, & Wang, 2002; Ragins & Kram, 2007).

Faultlines Perspective The faultlines perspective is a relatively new approach to understanding the dynamics of work team diversity. According to this perspective, a complete understanding of diversity’s consequences requires an understanding of the configuration of team members’ attributes. Rather than attend to the degree or type of diversity present in work teams, the faultlines perspective asserts that differences among team members are most likely to have significant consequences when they elicit the formation of distinct subgroups (Lau & Murnighan, 1998). That is, the focus of faultlines theory is on the structure of diversity (cf., Jackson et al., 1995). A fault line is said to exist in a team when two or more relatively homogeneous and distinct subgroups form on the basis of multiple shared attributes. For example, in a longitudinal study of internationally diverse student project teams, faultlines arose on the basis of differences in nationalities and education majors. In project teams where nationality and major covaried, the presence of strong faultlines disrupted information sharing and interfered with effective team performance (Jiang, Jackson, Shaw, & Chung, 2008). In another study of student teams, Polzer and colleagues showed that faultlines also influence the functioning of geographically distributed teams; they found that fault line effects were stronger when colocated subgroups were culturally homogeneous (Polzer, Crisp, Jarvenpaa, & Kim, 2006). The distribution of team members’ attributes partly determines the likelihood that faultlines will result in the formation of distinct subgroups. However, the likelihood that clear subgroups will be salient to team members and influence team dynamics also depends on other situational conditions, including the nature of the task. Imagine a six-person team of HR professionals with three older men and three younger women. This pattern of attributes sets the stage for a gender-based fault line to become salient. If the team is assigned to work on the development of a new family leave policy, the task would likely increase the likelihood that the fault line would shape how the team functions. However, if the same team were assigned to develop a new 659

Jackson and Joshi

recruiting plan, the faultline might be more likely to remain dormant.

Copyright American Psychological Association. Not for further distribution.

Concluding Comments No single theoretical perspective addresses the full array of affective and behavioral consequences associated with work team diversity. Scholars and managers alike can gain different insights from each perspective. Some theoretical perspectives emphasize the fact that people tend to be attracted to similar others; some focus attention on the role of task-related resources, such as skills and knowledge; some point out the importance of looking at team members’ external connections, and some emphasize the fact that work teams can be influenced by the formation of competing subgroups. Despite such differences, these perspectives all assume that the types and distribution of personal attributes among members partly determine how work teams function and, ultimately, how well they perform. Future work that addresses the differing conceptual underpinnings of past research could advance our understanding of how diversity affects the daily lives of employees and organizations (Jackson & Chung, 2008). Integrating the various models and perspectives will require attending to interpersonal processes that play out at each of several levels of analysis, including the individual, dyads, teams, and larger organizations (e.g., Boone et al., 2004; Jackson et. al, 1991). Some initial steps in this direction have already been taken (e.g., Joshi, Liao, & Jackson, 2006; Hutlin & Szulkin, 1999), and we are optimistic that continued theoretical integration will occur as this area of research continues to develop. REVIEW OF EMPIRICAL RESEARCH ON THE CONSEQUENCES OF WORK TEAM DIVERSITY We turn next to a review of the empirical research on work team diversity, focusing on studies that investigated diversity’s potential consequences. This review includes studies that assessed the relationships between several types of diversity and its short-term consequences for work team processes as well as longer-term consequences for work team performance. 660

To locate relevant studies, we searched several electronic sources (PsycINFO, ABI/INFORM and EBSCO Academic, and SocIndex) and manually searched the primary journal outlets for research on work team diversity (e.g., Journal of Applied Psychology, Personnel Psychology, Academy of Management Journal, Administrative Science Quarterly) using a variety of relevant key terms and phrases (e.g., workplace diversity, team diversity, group demography, organizational demography, work team composition). We also contacted prominent diversity researchers and invited them to share their working papers and forthcoming journal articles. These searches yielded 88 articles that reported 487 diversity–consequence relationships. The majority of studies were conducted within the United States. Increasingly, scholars in other countries also are becoming interested in the issue of workplace diversity, but there is not yet sufficient evidence to make any cross-cultural comparisons of work in this field. Table 20.1 provides a summary of the types of research published on work team diversity during the past 15 years. Listed in the left column of Table 20.1 are the types of diversity that have been studied most. The next three columns show the number (percentage) of studies that investigated each of three broad categories of consequences: affective and attitudinal responses (e.g., cohesion, conflict, commitment, satisfaction), behavioral processes (e.g., communication, use of information, learning behavior, turnover), and performance (e.g., achievement of team goals and other indicators of effectiveness). Table 20.1 shows that many types of work team diversity have been investigated in organizational settings, with the greatest accumulation of evidence pertaining to readily detected types of diversity and their relationship to performance. As this table reveals, the combination of several types of diversity and several potential consequences of interest means that a very large number of studies will be needed before we can draw firm conclusions about all of the possible empirical relationships. Here we focus on those relationships for which the most data are available, while cautioning readers that firm conclusions should be drawn only after additional research has been conducted.

Work Team Diversity

TABLE 20.1 Overview of Research Conducted by Types of Diversity Attributes and Types of Consequences Types of consequencesa

Copyright American Psychological Association. Not for further distribution.

Type of diversityb Race/ethnicity (16%) Gender (15%) Functional background (11%) Age (10%) Cognitive/mental model (10%) Tenure (9%) Cultural values (7%) Education level (5%) Composite measure (4%) Fault lines (4%) Nationality (3%) Personality (3%) Others (3%)c Total N = 487d

Affect or attitude

Behavior

Performance

18 (24%) 26 (35%) 7 (12%) 18 (36%) 11 (22%) 14 (33%) 17 (53%) 13 (52%) 1 (5%) 12 (63%) 2 (13%) 2 (14%) 3 (27%) 212 (44%)

23 (30%) 18 (24%) 20 (35%) 7 (14%) 19 (39%) 8 (19%) 11 (34%) 2 (8%) 8 (40%) 5 (26%) 6 (38%) 3 (21%) 1 (9%) 131 (27%)

35 (46%) 31 (41%) 31 (54%) 25 (50%) 19 (39%) 20 (48%) 4 (13%) 10 (40%) 11 (55%) 2 (11%) 8 (50%) 9 (64%) 7 (64%)

a

Proportion (%) of effects out of 487 for each type of outcome listed is shown in parentheses. Proportion (%) of effects reported out of 487 for the type of diversity listed is shown in parentheses. cMarital status based diversity, network density, geographic diversity, experience diversity. dA total of 88 studies were selected for review, which included 487 reported effects that were coded for this analysis. b

Communication Patterns Communication processes are fundamental to accomplishing things in organizations. (See also Vol. 3, chap. 7, this handbook.) Thus, the question of whether and how the composition of work teams influences communication patterns is of long-standing interest to those who study organizations. March and Simon (1958) noted that the frequency of communication between two people creates a shared language that enhances the efficiency of their communications. For work teams, efficient communication is considered a critical antecedent of performance (Katz, 1982), and several researchers have formulated models of how work team diversity might influence team communications (Elsass & Graves, 1997; Larkey, 1996). The ASA model, social identity theory, and social capital theory all suggest that work team diversity is likely to impede frequent and effective communication among team members, while team homogeneity should facilitate effective communication. The results of several studies support these predictions for various types of diversity.

Hoffman (1985) examined the effects of ethnic minority representation within the supervisory ranks on interpersonal, organizational, and interorganizational communication frequency. Consistent with the expected effects of dissimilarity, Hoffman’s 1985 study of 96 state agencies found that higher percentages of minority representation were associated with lower frequency of informal interpersonal communication. Furthermore, in the more diverse agencies, the reduced frequency of informal communications was offset by higher frequency of formal organizational communication. In their seminal study of engineering project teams in the research division of a medium-sized firm based in the United States, Zenger and Lawrence (1989) hypothesized that team members who were similar in age and tenure would share common attitudes and values and thus would communicate more frequently with each other than with dissimilar team members. The results revealed that age (but not tenure) similarity was significantly related to the frequency of technical communications among team members. Studies of communication networks also show that communications tend to occur more frequently 661

Copyright American Psychological Association. Not for further distribution.

Jackson and Joshi

between similar members. Ibarra’s 1992 study of male and female managers in an advertising firm found that men tended to form same-gender network connections that served both social and instrumental goals. Women’s network connections were characterized by greater gender diversity, however, perhaps because fewer women were employed in the organization and connections with other women were less easily established (Ibarra, 1992). A study of the friendship networks of MBA students found that students formed friendships with others from similar ethnic backgrounds (Mehra et al., 1998). In a study of teams working on national service projects, Klein, Lim, Saltz, and Mayer (2004) found that team members who were more similar to others in the team were also more central in the team’s advice and friendship networks. As discussed earlier, the social capital perspective recognizes that the organizational embeddedness of work teams has important implications for the effects of team diversity on external communications (see Joshi, 2006). In their study of R&D teams, Zenger and Lawrence (1989) found that engineers’ communications with people outside the team often occurred with people who entered the organization at about the same time. Consistent with the organizational demography perspective, the relationships formed among members of a tenure cohort apparently served to bind together cohort members even if their job assignments dispersed them into different areas of the organization. In a study of R&D teams, Reagans and Zuckerman (2001) found that demographic diversity negatively influenced internal team communication, but diversity was positively associated with greater external communication. The study also showed that the external communication networks of work teams contributed positively to team performance. Ancona and Caldwell (1992) found that functional diversity (but not tenure diversity) predicted the frequency of teams’ external communication. Corroborating this finding, Keller (2001) reported that functionally diverse teams had more external communication and that external communication, in turn, was associated with better team performance on dimensions such as technical quality, timeliness, and staying within budget. 662

The accumulated evidence clearly indicates that work team diversity influences communication behaviors in ways that may be detrimental to internal team dynamics yet beneficial to a team’s external relationships and team performance. Thus, for a particular team, the longer term performance consequences of team diversity may partly depend on the relative need for effective internal and external communication. For teams working on highly interdependent tasks that require speedy and efficient communication, activities that build internal cohesiveness and training that addresses the potentially negative effects of diversity on communication may be particularly helpful. On the other hand, when performance is more dependent on the team’s ability to manage external relationships, it may be useful to ensure that team members understand the importance of such external relations and the ways in which their diversity can contribute to team performance. Because information sharing is of increasing importance to organizations that compete on the basis of knowledge (see Jackson, Hitt, & DeNisi, 2003; Jackson & Hong, 2008), research that helps to unravel the specific ways in which diversity influences communication is sorely needed.

Conflict Two types of conflict that arise in teams are task conflict and emotional conflict. (See also Vol. 3, chap. 13, this handbook.) Task conflict involves disagreements about the work itself, while emotional conflict involves frustration and anger related to dealing with team members. According to the information processing perspective, conflict is a mediating process that provides an explanation for observed relationships between various types of diversity and both longer term positive outcomes, such as improved team performance, and longer term negative outcomes, such as increased turnover (Jehn, Northcraft, & Neale, 1999; Pelled, Eisenhardt, & Xin, 1999). Task conflict is likely to arise when team members bring differing knowledge, expertise and, experience to bear on task completion. Presumably, task conflict can be beneficial to team performance, for it is through their attempts to resolve such conflict that team members are likely to find creative and effective solutions. Consistent with this view, Bowers,

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

Pharmer, and Salas (2000) found that diversity and performance were more strongly related for complex tasks and less strongly related for simple tasks. Emotional conflict is generally something people prefer to avoid, if possible—it is seldom constructive and can be distressing for some individuals. Because emotional conflict is unrelated to a team’s task, it presumably provides no performance benefits. Jehn and her colleagues argued that relationship-oriented (ethnic and gender) diversity should be more likely to elicit emotional conflict, whereas task-based (functional background and education) diversity should be more likely to elicit task conflict (Jehn, 1995; Jehn et al., 1999). Support for these arguments has been mixed. In a study of employees in a household goods moving firm, functional diversity was associated with greater task conflict and unrelated to emotional conflict, as expected. Also, gender and age diversity were positively associated with emotional conflict but not task conflict (Jehn et al., 1999). Somewhat different results were found in a study of 45 work teams in the electronics industry. As predicted, both ethnic and tenure diversity were positively associated with emotional conflict; however, contrary to predictions, gender diversity was unrelated to emotional conflict. Also as predicted, functional diversity was positively related to task conflict but, contrary to their predictions, tenure diversity was unrelated to task conflict (Pelled et al., 1999). Other studies also have failed to provide consistent support for the predicted relationships between the different types of diversity and conflict (e.g., Pelled, 1996; see De Dreu & Weingart, 2003, for an overview). Several situational conditions have been shown to reduce the likelihood that relationship-oriented diversity will result in emotional conflict. A strong team orientation is one condition that appears to neutralize the negative effects of gender diversity on relationship conflict (Mohammed & Angell, 2004). Team longevity also plays a role. In newly formed teams, readily detected diversity such as gender and ethnicity may elicit conflict; for long-lived teams, diversity in underlying attributes, such as values and personality, are more predictive of how much conflict teams experience (Harrison et al., 1998; see also Pelled et al., 1999).

Task characteristics can also moderate the effects of diversity. Pelled et al. (1999) considered task routineness as a moderator of the relationship between diversity and both emotional and task conflict. Their findings showed that the effect of functional diversity on task conflict was weaker for teams working on more routine tasks; also, the effects of ethnicity and tenure diversity on emotional conflict were weaker for more routine tasks. Although findings about the effects of task routineness (vs. complexity) are somewhat mixed, there is an emerging consensus that the effects of diversity on team conflict depend on the nature of the team’s task. In general, diversity is more likely to engender emotional conflict in teams working on tasks characterized by greater uncertainty and autonomy; by comparison, the effects of diversity on task conflict are not yet clear (De Dreu & Weingart, 2003; Molleman, 2005). In other efforts to understand the mixed findings concerning the relationship between team diversity and conflict, scholars have recently turned to the faultlines perspective. In a study of 71 international joint venture teams, Li and Hambrick (2005) found that strong demographic faultlines in joint venture teams led to higher levels of emotional and task conflict, and such conflict was associated with poor joint venture performance. To date, few field studies provide evidence about the relationship between demographic faultlines and team conflict. Laboratory studies indicate that faultlines theory holds some promise, however. Molleman (2005) examined the effects of demographic-, ability-, and personalitybased faultlines on team conflict. In a sample of 99 undergraduate student teams, the strength and depth of demographically defined faultlines (but not faultlines defined by personality or ability) were directly associated with internal team conflict. When task characteristics were taken into account, the study revealed that personality and ability faultlines also were associated with internal team conflict, but only under conditions of high autonomy. Despite many years of research on the relationship between diversity and various types of team conflict, scholars’ understanding of the relationship between team diversity and team conflict remains somewhat limited. Many competing theories have 663

Jackson and Joshi

been offered, and these are matched by an array of contradictory empirical findings (for a complete review, see Jehn, Greer, & Rupert, 2008). This state of affairs may be disheartening for scholars seeking to explain the causes of team conflict. However, for managers, the research to date suggests that they need not be overly concerned about the possibility that increased diversity in teams will be associated with disruptive team conflict.

Copyright American Psychological Association. Not for further distribution.

Social Cohesion, Commitment, and Turnover The ASA, organizational demography, social identity, and social capital perspectives all predict that more diverse teams are likely to experience less positive affect. Decades of research on similarity and attraction indicate that people tend to dislike dissimilar others, all else being equal. By extension, it has been argued that diversity is likely to have negative consequences for affective reactions such as social cohesion and commitment (e.g., see Pfeffer, 1983), and as a result of weak social relationships, turnover rates are likely to be higher for diverse teams (Schneider, 1987). (See also Vol. 2, chap. 11, and Vol. 3, chaps. 1 and 4, this handbook.) The empirical results generally support the conclusion that people working in social settings characterized by various types of diversity experience weaker feelings of commitment and related affective reactions, while employee turnover rates are higher in such settings (e.g., Mueller, Finley, Iverson, & Price, 1999; McCain, O’Reilly, & Pfeffer, 1983; O’Reilly, Caldwell, & Barnett, 1989; Tsui, Egan, & O’Reilly, 1992; Wagner, Pfeffer, & O’Reilly, 1984). For example, in a study of 93 top management teams in the banking industry, Jackson and colleagues (1991) found that teams with greater diversity in terms of education, tenure, and industry experience had higher turnover rates within the executive ranks. Only a few studies have found no relationships (e.g., Iaquinto & Fredrickson, 1997) or quite weak relationships (Godthelp & Glunk, 2003) between team diversity and outcomes such as commitment and turnover. Much of the available data on which this conclusion rests come from studies of top-level management teams, and the effects of some types of diversity— especially gender and ethnic diversity—have been 664

difficult to establish for such teams. As more women and ethnic minorities move into the executive ranks, additional research will be needed to determine how these types of diversity influence the affective responses of top management teams. In addition, the relationship between team diversity and these outcomes is sometimes more complex than predicted by most theoretical perspectives (e.g., see Schippers, Den Hartog, Koopman, & Wienk, 2003). For increasingly diverse organizations, a conclusion to be drawn is that proactive measures may be needed to avoid the potential problem of increased rates of turnover. We caution managers not to make assumptions about who is most likely to leave in response to increasing diversity, however. If one accepts the logic of the ASA model, one might expect those who are in the minority to be the most likely to leave because they would be least likely to feel that they fit in well (Sacco & Schmitt, 2005; Zatnick, Elvira, & Cohen, 2003). At least a few studies have found that this is not the case. Jackson et al. (1991) found that diverse teams had more turnover, but the characteristics of individual team members did not predict the likelihood of their leaving. Tsui et al. (1992) found that majority group members (males and Caucasians) were more likely to have negative reactions to team diversity, in comparison with minority group members. Findings such as these suggest that employers who aim diversity management initiatives (such as mentoring programs and affinity networks) primarily at minority group members may be inadvertently aiming their solutions at the wrong targets. Diversity initiatives that result in an overall improved diversity climate within organizations may be as effective in reducing the negative reactions of majority group members (e.g., Caucasian men) as they are for improving retention of minorities and women (McKay et al., 2007).

Creativity and Innovation The competitive landscape of the early 21st century challenges firms to continually change and adapt to external trends and events, such as increasing globalization, new technologies, and shifting political and economic conditions. To succeed in this environment requires the capacity for creativity and innovation, as these are the processes through which firms create

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

new products, improve their services, and reduce their costs. (See also chap. 9, this volume.) Thus, understanding the conditions that facilitate creativity and innovation in work teams has been a high priority among scholars. The changing demographics of the relevant workforce make the question of how diversity influences these processes especially important. Within the United States, R&D teams have become increasingly diverse in recent years as the proportion of female, non-White, and foreign scientists in the workforce has risen. According to the National Science Foundation (2004), more than one of four recipients of American doctoral degrees in science and engineering are neither U.S. citizens nor permanent residents. Moreover, women and non-Whites, respectively, now account for about 33% of all doctorates awarded in science and engineering, double the percentages of 25 years ago. If these trends continue, the science and engineering workforce in the U.S. will soon include more people of color and women than White men. At the same time, various organizational changes are increasing other types of diversity in R&D teams. Compared with the past, today’s R&D teams are more likely to include some people with specialized knowledge in areas other than the basic sciences and engineering—such as sales, marketing, and finance—and even team members who are employed by a firm’s strategic partners (Huston & Sakkab, 2006). Furthermore, studies have found that the effectiveness of R&D teams is influenced by social and task-related communication (Dailey, 1978; Keller, Julian, & Kedia, 1996; Pirola-Merlo, Hartel, Mann, & Hirst, 2002; Van den Bulte & Moenaert, 1998). According to the information processing perspective, diversity in attributes associated with taskrelated knowledge and expertise can provide teams with valuable information and alternative approaches to problem solving. According to social capital theory, relations-oriented diversity also may promote team creativity and innovation, by providing more external connections through which the team can obtain needed knowledge and resources (Oh et al., 2004). Among managers, too, work team diversity is widely believed to be conducive to creative problem solving and innovation, and this belief is supported by studies conducted in laboratory settings (McGrath, 1984; Shaw, 1981).

Consistent with the information processing perspective, studies of top management teams have found that diversity on task-related attributes such as educational specialization and functional background predict innovative organizational practices and adaptation to change (Bantel & Jackson, 1989; Wiersema & Bantel, 1992). The results of studies involving lower level teams have been mixed, however: Sometimes diversity has been associated with improved information use and problem solving (Frink et al., 2003; Hirschfeld, Jordan, Feild, Giles, & Armenakis, 2005), while at other times diversity has been associated with disruptive processes that have negative consequences for creativity and problem solving (Pelled et al., 1999; Randel, 2002). In an attempt to resolve such inconsistent results, Reagans et al. (2004) drew on social capital theory to unpack the multitude of ways through which diversity can promote creativity and innovation. By analyzing the effects of team diversity on both internal and external social capital, Reagans et al. showed that the overall impact of task-oriented diversity in R&D teams is explained by a combination of the negative impact of diversity on internal team dynamics and the positive impact of diversity on the team’s external social network. In another effort to explain the mixed findings relating team diversity to innovation, Pearsall and colleagues (2008) applied the faultlines perspective. In an experimental setting, they found that gender diversity was disruptive when gender faultlines were activated by a gender-conscious task; gender diversity had positive consequences for creativity when teams performed a gender-neutral task (Pearsall, Ellis, & Evans, 2008). A longitudinal study of student project teams by Jiang and his colleagues also employed faultlines theory to explain how team diversity could interfere with performance on a task that required creativity and innovation (Jiang, Jackson, Shaw, & Chung, 2008). Jiang et al. (2008) argued that diversity would have negative consequences when it resulted in faultlines that disrupted information sharing and informal socializing within the team. Their results showed that task-based faultlines were detrimental to information sharing and team performance, while relations-oriented faultlines 665

Copyright American Psychological Association. Not for further distribution.

Jackson and Joshi

were detrimental to informal socializing but not team performance. For organizations that face volatile environmental forces that demand effective responses, understanding creativity and innovation is essential to longterm survival. The increasingly diverse workforce of today’s organizations may prove to be an unanticipated opportunity for meeting this challenge. However, as recent research on team faultlines shows, the mere presence of diversity does not guarantee the blossoming of creativity and innovation. When the presence of diversity inadvertently results in the presence of strong faultlines, it may interfere with rather than support effective problem solving. One implication is that organizations may find it useful to conduct faultline audits when staffing R&D teams in order to reduce the likelihood that disruptive faultlines are present in those teams. In organizations that employ many such teams, attending to the composition of teams when making staffing decisions may prove to be more efficient than relying on diversity training or other interventions designed to overcome the potential negative consequences of faultlines (e.g., see Kulik & Roberson, 2008b).

Team Performance In our discussion of innovation and creativity, we described results from studies of work teams charged with identifying, creating, and developing something new—new products, new services, and/or new means of production and administration. In this section, we consider studies that assessed team performance for other types of tasks. As noted by others (Jackson et al., 2003; van Knippenberg & Schippers, 2007), the results of such studies have been mixed. In a study of manufacturing teams that assessed team performance using measures of both productivity (output) and customer service ratings, ethnic diversity was found to be negatively related to performance (Kirkman, Tesluk, & Rosen, 2004), but this result has not been found in other work settings (e.g., Kochan et al., 2003). Studies of gender diversity have found its effects on performance are sometimes positive (Rentsch & Klimoski, 2001), sometimes negative (Jehn & Bezrukova, 2003), and sometimes not significant (Watson, Johnson, & Merritt, 1998). Likewise, 666

the results from studies of age diversity are mixed. Some studies have reported positive relationships between age diversity and team performance (Kilduff, Angelmar, & Mehra, 2000), whereas others have reported nonsignificant relationships (Simons, Pelled, & Smith, 1999; Bunderson & Sutcliffe, 2002) or negative relationships (Ely, 2004). Scholars have adopted several approaches in trying to understand these mixed findings. One common assertion is that the problem of mixed results is caused by measuring readily detected types of diversity instead of underlying types of diversity (e.g., see Lawrence, 1997). This criticism rests on the assumption that psychological differences among team members are the proximal causes of any effects attributed to demographic diversity. However, the effects of psychological diversity also are unclear. For example, in one study of 82 work teams in a large retail organization, team performance was positively related to diversity on two of the Big Five personality dimensions (Extraversion and Emotional Stability) but unrelated on the other three (Neuman, Wagner, & Christiansen, 1999). Another explanation for the mixed findings concerning team diversity and team performance is that the effect of any one type of diversity depends on the other types of diversity present in the team. In a study of 365 sales teams (Jackson & Joshi, 2004), team performance was defined as the average of team members’ actual sales compared with their assigned sales goals. In the organization studied, sales goals were carefully calculated to take into account the specific products that were being sold (e.g., type of equipment or service), characteristics of sales territories (e.g., geographic scope, density), and characteristics of clients’ accounts (e.g., size). The company’s method of measuring performance permitted it to directly compare the performance across individuals and across teams, even though they were selling different products in different locations. Jackson and Joshi (2004) found that the effects on performance of several types of diversity—gender, ethnic, or tenure diversity—depended on the other types of diversity present in the team. Their most striking finding was that team performance was lowest for teams with a combination of relatively high tenure diversity and high gender diversity and high ethnic diversity.

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

Likewise, in a study of the effects of diversity in a household goods moving firm on supervisor-rated and archival performance measures, Jehn et al. (1999) showed that the effects on performance of task-oriented (education, function, and position in the firm) diversity depended on the degree of relations-oriented (gender and age) diversity present in teams. Task-oriented diversity was associated with better performance only when social category diversity was low. Differences in the types of tasks performed by the teams are another possible explanation for the inconsistent results found across studies (Jackson, 1992). In a study of 4,538 tax employees working in 222 work units, the relationship between age diversity and performance depended on the degree of complex decision making required in the work units. Age diversity improved performance only for those units working on relatively more complex tasks (Wegge, Roth, Kanfer, Neubach, & Schmitt, 2008). In their continuing efforts to understand the conditions under which various types of team diversity influence team performance, scholars have also examined the role of other team level moderators. We discuss those research efforts later in this chapter.

Concluding Comments During the past decade, research on work team diversity has begun to mature. Dozens of studies have examined the relationship between various types of work team diversity and a variety of outcomes. However, because of the large number of possible effects to consider (due to the many types of diversity and outcomes of interest), additional research is still needed. With dozens of studies to consider, can we draw any tentative conclusions about the consequences of work team diversity? We believe the answer is yes. At least for the more frequently studied dimensions of demographic diversity, the findings are beginning to converge. Relations-oriented diversity. On the basis of the available field evidence, one tentative conclusion we draw is that relations-oriented diversity in work teams is often (but not always) of little consequence—at least for the work team outcomes that have been

examined to date. Overall, more than 50% of the reported effect sizes for ethnicity, gender, and age diversity reviewed by Joshi and Roh (2007) were nonsignificant. A meta-analytic review of 22 reported effects estimated that the average effect size for the relationship between relations-oriented diversity and social integration is −.02 (Horwitz & Horwitz, 2007). A meta-analytic review of 69 reported effects estimated that the relationship between relations-oriented diversity and team performance is −.03 (Joshi & Roh, 2009). Although relations-oriented diversity is sometimes benign, it appears to be disruptive under some specific conditions. Focusing on the relationship between team diversity and team performance outcomes, Joshi and Roh (2009) investigated a variety of potential moderating conditions in an effort to uncover explanations for the variation in reported effect sizes. Their analyses indicated that the consequences of relations-oriented team diversity depend in part on the contextual factor of occupational demography. In occupations that were maledominated, the average effect size for the relationship between gender diversity and team performance was negative (−.09). But in occupations that were gender balanced, the average effect size for the relationship between gender diversity and team performance was positive (.11). A similar pattern was found for race/ethnic diversity. For occupations in which the majority was White, the average effect size for the relationship between racial/ethnic diversity and team performance was negative (−.07). But in occupations that were racially/ethnically balanced, the average effect size for the relationship between race/ethnic diversity and team performance was positive (.11). Likewise, age diversity was more strongly associated with lower performance in occupational settings in which the majority of workers were relatively young (−.08), compared with settings in which employees ages were more balanced (−.05). Results such as these indicate that the consequences of team diversity depend partly on the degree of diversity within the larger contextual setting. For managers, such findings suggest that relations-oriented team diversity may be associated with some negative outcomes if the work teams are islands of diversity struggling to succeed in a sea of homogeneity. But 667

Jackson and Joshi

Copyright American Psychological Association. Not for further distribution.

when the sea (i.e., the organization’s larger workforce) also is diverse, diverse teams are less likely to suffer from disruptive social relationships. Task-oriented diversity. Regarding task-oriented diversity, the preponderance of evidence indicates that diversity on characteristics such as functional background, education, and job or organizational tenure is often likely to be beneficial for team performance. That is, many organizations appear to reap the often-claimed benefits of task-related diversity. The performance benefits of task-oriented diversity can apparently be enjoyed despite any negative effects that such diversity might have on social relationships among team members. A meta-analytic review of 15 effect sizes by Horwitz and Horwitz (2007) revealed a negative average relationship between task-related diversity and social integration of −.04. On the other hand, a meta-analytic review of 48 effect sizes by Joshi and Roh (2009) revealed a positive average relationship between task-related diversity and team performance. More specifically, they found that functional diversity had the strongest positive relationship with team performance (.13), followed by tenure diversity (.03) and educational diversity (−.02). What do these findings mean for employers? Is the glass half empty or half full? We believe the message is that employers should make no assumptions about how diversity influences work teams in a particular organization. Instead, employers are advised to examine the effects of specific types of diversity for the specific outcomes of most interest to the organization. The evidence clearly shows that diversity can have positive and/or negative consequences. Whether diversity is beneficial or problematic depends on the type of diversity, the nature of the team task, the specific team consequences of interest, and other local organizational conditions. By conducting organization-specific analyses, employers can learn which types of diversity are most beneficial in a particular organization and can identify any types of diversity that they may need to manage more effectively. EMERGING AND FUTURE RESEARCH DIRECTIONS The growing body of evidence suggests that the consequences of work team diversity are difficult to pre668

dict. Although much of the research summarized here was motivated by a desire to link work team diversity to attitudes and behaviors that influence performance, the results reveal a complex picture of reality. In this section we make an effort to deal with this complexity by proposing various approaches that researchers can take to unravel and thereby better understand the effects of diversity in organizations. Here we consider promising new approaches that incorporate more complex conceptualizations of diversity, draw on new theoretical perspectives, and comment on multilevel and context-sensitive perspectives for understanding the effects of work team diversity.

Understanding Types of Diversity Scholars who approach the study of diversity using the lens of psychology often assume that using readily detected attributes to assess diversity (e.g., age, tenure, educational background, ethnicity) is just a methodological shortcut for assessing variations in individual attitudes, values, beliefs, and/or workrelated knowledge. For some readily detected attributes, empirical evidence shows they are indeed related to psychological differences in attitudes, values, and so on. For example, age has been shown to be negatively correlated with risk-taking propensity (Vroom & Pahl, 1971) and approaches to problem solving (Datan, Rodeheaver, & Hughes, 1987). Also, societal conditions (e.g., economic depressions vs. booms and periods of war vs. peace) that are associated with different age cohorts appear to influence the attitudes and values of people who grew up during those eras (see Elder, 1974, 1975; Thernstrom, 1973). Regarding tenure, the evidence shows that executives tend to become more committed to the status quo the longer they stay in the same organization (Finkelstein & Hambrick, 1990; Hambrick, Geletkanycz, & Fredrickson, 1990). It is not surprising that educational curriculum choices are associated with personality, attitudes, and cognitive styles (Holland, 1976). Due partly to criticisms of studies that assessed only demographic attributes (e.g., Lawrence, 1997), research on work team diversity increasingly includes psychological measures (e.g. Barrick, Stewart, Neubert, & Mount, 1998; Harrison, Price,

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

Gavin, & Florey, 2002; Klein, Lim, Saltz, & Mayer, 2004). Such studies shed light on the effects of psychological diversity, but they too yield an incomplete picture of diversity dynamics, because demographic differences are important in and of themselves. Perceptions of in-group and out-group status often reflect nominal categorical differences. Simply knowing that another person is similar—for example, knowing that the person belongs to one’s own demographic group—is sufficient to trigger in-group categorization and cooperation (Oakes, Haslam, & Turner, 1994). Including measures of both underlying (psychological) diversity and readily detected diversity presents an opportunity for gaining new insights about the effects of work team diversity; emphasizing one type of diversity over others may lead to inaccurate conclusions about which types of diversity are most consequential (Jackson & Chung, 2008). The potential value of the combined approach was demonstrated in a study by Harrison et al. (2002), which found that readily detected diversity influenced team functioning when teams had little experience together, but over time psychological diversity was more influential. Changes in perceived similarity are one explanation for the moderating effect of how long team members had worked together (e.g., see Zellner-Bruhn, Maloney, Bhappu, & Salvador, 2008). New theoretical models as well as new empirical studies that incorporate both demographic and psychological attributes are needed in order to fully understand how the complex array of differences among employees influences the workplace.

Understanding the Structure of Diversity For a variety of reasons, many studies of diversity have focused on one or two aspects of diversity (e.g., ethnicity or gender or conscientiousness), and few studies have controlled for the many possible intercorrelations among attributes. Regardless of whether one’s focus is on demographic diversity, psychological diversity, or their combination, the structure of diversity in any situation reflects the complex attribute profiles of individuals (Jackson et al., 1995). Studies of multidimensional diversity illustrate the value of assessing multiattribute profiles, instead of focusing on one particular type of diversity. For

example, Jehn and colleagues (1999) found that informational (education and function) diversity was negatively related to team efficiency when relations-oriented (gender and age) diversity was high, but not when it was low. Pelled et al. (1999) also found that the consequences of diversity for team conflict were best understood by taking into account interactive effects among multiple types of diversity. In a study of sales team performance, Jackson and Joshi (2004) found that the effects on team performance of any one type of diversity—gender, ethnic, or tenure diversity—depended on the other types of diversity present in the team. Calling for a multidimensional approach to assessing the diversity of social units is no longer new (see, e.g., Cannella, Park, & Lee, 2008; Lau & Murnighan, 1998; Jackson & Joshi, 2001; Ofori-Dankwa & Julian, 2002), yet progress in this direction has been slow. The need for large samples and disagreements about how to test for multidimensional effects are among the barriers to progress. Lack of relevant theoretical guidance has played a role, too. Fortunately, the introduction of group faultlines theory is stimulating new ideas about how to measure multidimensional diversity. In particular, we note that Shaw and his colleagues (see Shaw, 2004; Chung, Jackson, & Shaw, 2005; Jiang et al., 2008) have developed a measure of faultline strength that takes into account both internal alignment within subgroups and differences between subgroups. In comparison with other faultline measures that have been used in field research (e.g., Li & Hambrick, 2005; Thatcher, Jehn, and Zanutto, 2003), Shaw’s measure is designed to fully capture the two aspects of faultlines identified by Lau and Murnighan (1998)—degree of homogeneity among members of subgroups and degree of dissimilarity between subgroups. The question of how to measure faultlines is not yet resolved, but the accumulating evidence suggests that strong faultlines create conflict and interfere with team performance. In contrast, weak faultlines may stimulate group learning.

Understanding the Roles of Social Status and Dominance In most U.S. organizations, men and Whites enjoy higher status than women and people of color 669

Copyright American Psychological Association. Not for further distribution.

Jackson and Joshi

(Baron & Newman, 1990). Status differences have many consequences, including the way people respond to work team diversity. In organizations, those with higher status are more sensitive to the degree to which they are in the majority (Tsui et al., 1992). People with high-status social identities, even when they are in the numerical minority, tend to maintain identification with their demographic in-group, which bolsters their self-esteem and insulates them from the negative effects of their minority position (Hewstone et al., 2002; Tajfel & Turner, 1985). People with low-status social identities tend to accept their “inferior” positions and are less likely to display discriminatory behavior against higher status out-group members even when their own proportions increase (Sachdev & Bourhis, 1985, 1987, 1991). Two perspectives that shed additional light on the role of status are social dominance and system justification. These perspectives recognize that all societies are hierarchically organized. Some social groups enjoy dominant positions and others are relegated to subordinate positions. Furthermore, individuals belonging to the dominant and subordinate groups often endorse legitimating myths that provide justification for social hierarchies (Sidanius & Pratto, 1999). Conscious or unconscious internalization of these group-based hierarchies results in behaviors that reinforce the unequal distribution of power and resources (Sidanius, 1993). Proponents of social dominance theory argue that social dominance orientation (SDO), which is “a generalized tendency to support existing hierarchical relationships between groups” (Haley & Sidanius, 2006, p. 659), motivates political attitudes. A significant body of evidence shows that SDO is associated with racism and attitudes toward race-conscious policies such as affirmative action. Introducing this individual difference into studies of workplace diversity may improve our understanding of the conditions under which diversity has workplace consequences. Whereas the social dominance perspective focuses on the motivations of those in dominant positions, systems justification theory considers the internalization and perpetuation of social hierarchies by members of subordinate groups (Jost & Banaji, 1994). Jost and colleagues found that lowstatus group members were much more likely to 670

exhibit out-group favoritism; favoring of out-group members occurred much less among high-status group members. Out-group favoritism, in turn, was associated with the lower levels of self-esteem found among low-status members (Jost, Pelham, & Carvallo, 2002). Such findings suggest that the social dominance and systems justification perspectives may prove illuminating when applied in future research aimed at understanding work team diversity.

Understanding Contextual Conditions In their continuing attempts to understand the complex pattern of findings regarding how diversity influences work teams, some diversity researchers have begun to examine context as a potential moderator of diversity effects (see Joshi & Roh, 2007, for a detailed review). In testing their theoretical arguments meta-analytically, Joshi and Roh (2009) found that weak direct relationships may be obscuring the specific conditions under which diversity can have beneficial or detrimental effects on performance outcomes. Their findings revealed that after accounting for moderating contextual variables at multiple levels, diversity effects doubled or tripled in size. In this line of research, team, organizational, and even extraorganizational contexts are viewed as influences that partially determine whether diversity is likely to be associated with positive, negative, or no consequences. Work tasks. The work itself—that is, the types of tasks that teams perform—is perhaps the most obvious and frequently studied contextual factor believed to moderate diversity dynamics. The generally accepted assumption is that the potential benefits of team diversity are greater for tasks that require creativity and innovation. For routine tasks, or when speed is the goal, diversity may interfere with performance (e.g., see Jackson, 1992; Priem, 1990; West & Schwenk, 1996; Williams & O’Reilly, 1998). The results of several early laboratory experiments supported this proposition, but clear evidence is not yet available from field studies. In fact, an alternative line of reasoning suggests that diversity is more likely to be detrimental for complex tasks because diversity makes the necessary coordination among team members more difficult (e.g., see Boone et al., 2004; Carpenter, 2002).

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

In the meta-analysis discussed above, Joshi and Roh (2009) found that relations-oriented diversity (gender, age, and ethnicity) had the most negative effects in moderately interdependent teams. Moderately interdependent tasks may impose constraints that interfere with the elaborative potential of diversity, but they may not be so demanding that team members are required to overcome categorizationbased processes and collaborate with each other. Although the effect sizes were smaller, the metaanalysis also revealed a positive relationship between relations-oriented diversity and team performance for tasks requiring low levels of interdependence. Tasks that require low interdependence can be effectively performed with little communication and information exchange; the reduced interpersonal contact that is required may mean that team members are less frustrated by their dissimilarities and can more easily recognize the contributions that other team members make in achieving team goals (Chung & Jackson, 2009; Pelled et al., 1999). Given these results, it appears that managing team diversity effectively may require organizations to take into account the degree of interdependence required to complete team tasks. Teams performing more complex and interdependent tasks may benefit more from team training or team coaching that facilitates group decision making and conflict resolution. A tailored team-focused approach to designing and delivering diversity initiatives may enhance the relevance and effectiveness of these practices within teams. Virtual teams. During the past decade, many organizations have turned to geographically dispersed teams that collaborate electronically to cut costs, increase efficiency, and improve customer responsiveness (Apgar, 1998). Given the growing prevalence of virtual and geographically dispersed teams, the role of this contextual factor is of considerable interest. On the basis of the premise that status and social influence are more likely to pervade face-to-face interactions than computer-mediated settings (which are considered to be more depersonalized), scholars have contrasted face-to-face groups with computer-mediated groups to examine whether the nature of participation varies in these two settings.

In general, such studies have found that computermediated communication reduces social inhibitions and equalizes team member participation (Sproull & Keisler, 1986). Nevertheless, other research shows that the medium of communication does not alter basic social influence processes. As in face-to-face teams, status dynamics and social influence effects occur in computer-mediated groups (see Martins, Gilson, & Maynard, 2004). Weisband, Schneider, and Connolly (1995) concluded that “even in the relatively impoverished social context of anonymous computer interaction, when high-status members were aware that a low-status member was in their group, the former made assumptions about the (latter’s) identity” (p. 1146). Findings such as these suggest that diversity dynamics in virtual teams may be similar to those in face-to-face teams, but much more research is needed before firm conclusions can be drawn about the role of this contextual factor. Temporal context. Several studies indicate that the effects of work team diversity are moderated by temporal factors. We have already mentioned the findings of Harrison et al. (2002) concerning the shifting importance of readily detected and underlying (deep-level) attributes. Likewise, Pelled et al. (1999) found that the negative effects of readily detected (ethnicity, functional background, and organizational tenure) diversity were weaker in longer-tenured teams. A study of top management teams also found that the effects of demographic diversity were weaker for teams that had spent more time working together (Carpenter, 2002). More recently, in a study of turnover among adult restaurant employees, demographic misfit was found to be more predictive of turnover during the initial weeks of employment than later (Sacco & Schmitt, 2005). Other studies have failed to replicate these effects, however (e.g., Schippers et al., 2003; Watson et al., 1998), suggesting that the role of time as a moderator of diversity’s consequences in work teams requires further research (see also Mohammed, Hamilton, & Lim, 2009). Additional longitudinal research is clearly needed to trace the changing dynamics of diversity in work teams, but the available findings have some potential implications for employers. The evidence 671

Copyright American Psychological Association. Not for further distribution.

Jackson and Joshi

suggests that readily detected diversity poses particular risks for work teams at early stages of their development. Work teams may eventually work through any problems they initially encounter on their own without outside assistance, and the process of doing so may be both natural and useful to teams. For companies that rely heavily on shortlived or temporary work teams, however, it may prove useful to speed up this process. Diversity awareness training aimed at educating employees about what to expect when working in diverse teams may prove useful for this purpose (Homan, van Knippenberg, Van Kleef, & De Dreu, 2007). Organizational demography. Two objective characteristics of an organization’s demographic composition can influence interactions in diverse teams—the overall heterogeneity of the organization (representation of demographic groups in the organization) and the level of structural integration (the representation of minority groups at higher levels in the organization; see Cox, 1991; Joshi, 2006). In structurally segregated organizations, the balance of power and status is skewed to the advantage of the dominant demographic group; such status and power differentials can undermine intergroup harmony and cooperation in work teams (Cox, 1993). The degree to which an organization is structurally segregated can be a powerful influence on social category–based identification and conflict among demographic groups (Cox, 1993; Ely, 1994; Wharton, 1992). Ely’s (1994, 1995) seminal work on the social construction of gender identity in organizations highlights the powerful influence that organizational demography can have. These studies and others (e.g., Ridgeway, 1997) show that in organizations where subordinate groups are underrepresented, negative stereotypes and biases toward these groups are more widely held. Joshi, Liao, and Jackson (2006) reported similar findings. They found that ethnicity-based and gender-based pay differentials were influenced by the combined effects of the demographic composition of one’s work team and the demographic composition of managers in one’s business unit. Commensurate with Ely’s research, Joshi and colleagues found smaller pay gaps and 672

improved performance for women in sales units with larger proportions of female managers. Research that examines how organizational demography shapes the dynamics that unfold within diverse teams is likely to flourish in the future, as scholars conquer hierarchical linear modeling and other statistical techniques for testing multilevel effects. Recent theoretical advances will likely stimulate new work, also. For example, Joshi (2006) developed a conceptual framework that outlined the effects of organizational demography on the relationship between team diversity and both internal and external team relationships. The model draws on embedded intergroup relations theory and social identity theory to understand the combined effects of team diversity and organizational demography. By recognizing the fact that work teams are embedded in both organizations and larger societal systems, Joshi’s framework draws attention to the interconnectedness of diversity dynamics across multiple levels of analysis. As scholars begin to examine diversity dynamics at multiple levels of analysis simultaneously, they are likely to also develop new multidisciplinary theoretical models. Organizational culture. Organizational culture refers to the shared values, beliefs, expectations, and norms prevalent in an organization and manifests itself in organizational rituals, practices, and behavior. Two aspects of organizational culture that can shape employee behaviors and employment outcomes in diverse work settings are culture strength and culture content. Organizational culture strength refers to the extent to which organizational norms and values are explicit and enforced. Organizations with strong cultures are characterized by uniform expectations and responses. In organizations with weak cultures, there is less agreement about appropriate behaviors and expectations. To the extent that diversity management practices are supported by a strong organizational culture, they are more likely to be implemented successfully in organizations (Cox, 1993). Organizational culture content refers to the specific norms, values, and beliefs that are prevalent (Cox, 1993). Chatman and Spataro (2005) argued that work team processes are shaped by the extent

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

to which an organization’s culture emphasizes independence versus interdependence. In organizations that emphasize independence, team diversity may be negatively associated with cooperation. In more collectivistic settings, however, team members are more likely to pick up cues that emphasize cooperation in the face of demographic dissimilarity. Among groups of MBA students performing a business simulation, Chatman, Polzer, Barsade, and Neale (1998) found that the positive effects of demographic diversity are more likely to emerge in settings that emphasize a collectivistic orientation rather than in settings that emphasize individualism and distinctiveness. Ely and Thomas (2001) argued that diversity is more likely to lead to positive outcomes when the organizational culture emphasizes integration and learning. Empirical studies that examined the effects of dissimilarity (relational demography) in organizations with differing cultures seem to support the general argument that organizations with cultures that reflect a belief that diversity is a valuable resource are more likely to realize the potential benefits of team diversity (Dass & Parker, 1999; Ely, 2004; Ely & Thomas, 2001; Gilbert & Ivancevich, 2000). On the other hand, organizational cultures that endorse a so-called color-blind approach may reinforce majority dominance and result in disengagement by minority employees (Plaut, Thomas, and Goren, 2009). What is not yet well understood is how appropriate cultures become established in organizations. Generally, responsibility for setting an organization’s cultural tone is assumed to rest with top management (see Wasserman, Gallegos, & Ferdman, 2008). Unfortunately, almost no empirical research is available to assure top executives that they will succeed in leading their diverse organizations if they follow the various prescriptions offered. Diversity climate. The concept of organizational climate is closely related to the concept of organizational culture. Diversity climate refers to employees’ perceptions of the degree to which all members of the organization are integrated into the social life of the organization and the use of fair human resource management practices (Mor Barak, Cherin, & Berkman, 1998). Diversity climate perceptions have

been shown to predict behavioral outcomes such as attendance (Avery, McKay, Wilson, & Tonidandel, 2007) and turnover (McKay et al., 2007). Limited research on the topic of diversity climate shows that perceptions of diversity climate often differ among employees of different demographic backgrounds, with ethnic minorities and women perceiving less positive diversity climates in their employment settings (Kossek & Zonia, 1993; Mor Barak et al., 1998). Research also shows that clusters of employees with similar perceptions of diversity climate share other attitudes, such as organizational commitment or job satisfaction (Nishii & Raver, 2003). To date, research on diversity climate has focused on the consequences associated with the climate perceptions of individual employees; it says little about the effects of diversity climate on work teams. Nevertheless, as for organizational culture, the empirical evidence suggests that employers may increase the beneficial consequences of work team diversity by creating positive diversity climate perceptions through the use of fair human resource management practices. Despite their best efforts, however, employers may not be in complete control of their employees’ diversity climate perceptions (Pugh, Dietz, Brief, & Wiley, 2008). The racial composition of the community where the organization is located may also shape the formation of diversity climate perceptions. In a study of 142 retail banking facilities in the United States, Pugh et al. (2008) found that the racial composition of the communities in which organizations were located moderated the relationship between the racial composition of organizations and employees’ perceptions of diversity climate. In communities with few racial minorities, the degree of diversity in work organizations was correlated with positive perceptions of diversity climate. As representation of racial minorities in the external community increased, the ethnic composition of organizations tended to lose its value as a signal of organizations’ diversity climates. Results such as these indicate that the effect of diversity in organizations is influenced by the broad social and psychological conditions of the workplace (see also Homan, van Knippenberger, Van Kleef, and De Dreu, 2007). Thus, designing interventions to improve 673

Jackson and Joshi

organizational cultures and climates is likely to be a challenge that may require organizations to tailor their interventions to reflect the conditions in the communities in which they are embedded.

Copyright American Psychological Association. Not for further distribution.

Understanding Leadership in Diverse Teams Surprisingly, little is known about the question of whether team diversity creates special challenges for team leaders. According to the social identity perspective, mismatches (dissimilarity) between leaders and team members set the stage for in-group/out-group dynamics. Kirkman and colleagues argued that mismatches between the demographic attributes of leaders and their team members would be associated with lower levels of trust among team members and stronger member biases toward their leaders. Consistent with this logic, a study of 111 work teams in four organizations found a positive relationship between degree of racial fit (leader’s racial similarity to team members) and team effectiveness as judged by the leader. The results also revealed that racial fit was associated with increased team empowerment (Kirkman et al., 2004). The question of whether leader–team similarity influences team performance is of some interest, but perhaps it will be more useful to examine the leader behaviors or styles that are most effective for diverse teams. Attempting to match leaders to team members is fraught with ethical and legal difficulties, whereas selecting and training leaders based on their leadership competencies is both feasible and desirable. A company that we studied (see Jackson & Joshi, 2004; Joshi, Liao, & Jackson, 2006), for example, would never use racial fit to assign leaders to sales teams, but it was very willing to train all sales team leaders to improve their leadership competencies and/or alter their leadership styles. A study of 136 primary care teams (Somech, 2006) suggested the potential value of additional research on the role of leadership style in diverse work teams. Somech (2006) found that functionally diverse teams with participative leaders engaged in more team reflection, which in turn was associated with team innovation; directive leadership promoted team reflection in teams with less functionally diversity. 674

Leadership behaviors may also influence diverse teams through motivational channels. Motivational leaders build linkages between team members’ selfconcepts and the team’s work, thereby enhancing identification with the team (Ellemers et al., 2004; Turner & Haslam, 2001). By emphasizing the team’s mission, shared values, and ideology, motivational leaders strive to link the interests of individuals with the interests of the team (Kark & Shamir, 2002). Research on so-called transformational leaders converges with the social identity perspective to suggest that a transformational leadership style may be especially effective for diverse work teams. A transformational leadership style includes acting as a role model and providing motivational inspiration and intellectual stimulation. Through their behaviors, transformational leaders align team goals with members’ goals and values and build the team’s sense of optimism and efficacy (Avolio & Bass, 2004; Bass & Riggio, 2006). In a longitudinal study of 62 R&D teams of a German pharmaceutical company, Kearney and Gebert (2009) found that the positive relationship between team diversity (on nationality and education) and team performance was stronger for teams with transformational leaders, compared to teams whose leaders were not perceived to be transformational. Apparently, transformational leaders were more effective in facilitating the exchange and use of task-related information, which contributed to the teams’ performance. Although it seems likely that effective leadership styles do not differ dramatically for teams that are relatively homogeneous or diverse, it is possible that effective leadership has greater importance for more diverse work teams. As the theories summarized earlier in this chapter suggest, diverse work teams hold the potential for being either very effective or very troubled. Effective leadership may be especially helpful in ensuring that such teams achieve their potential for excellent performance and avoid the interpersonal problems that may result in increased turnover rates (cf. Gibson & Vermeulen, 2003).

Understanding International Diversity A 2005 survey of large global firms found that all of the respondents rated global diversity as an important or very important issue (Dunavant & Heiss,

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

2005). Yet, as Table 20.1 reveals, very few field studies of internationally diverse work teams have been published in the English-language journals we surveyed. One explanation for the dearth of research on diversity in global work teams is that little specific theoretical attention has been given to these issues. The few studies of international teams that we located drew upon the same theoretical literatures as did studies focused on other types of diversity (e.g., see Earley & Mosakowski, 2000; Li & Hambrick, 2005; Jiang et al., 2008; Polzer et al., 2006). Another possible explanation for the dearth of research is the difficulty of conducting the studies. This is suggested by the fact that most studies of internationally diverse teams have been conducted using student samples rather than natural work teams (see Li & Hambrick, 2005, a study of joint venture teams, for an exception). We also found few studies designed specifically to examine whether or how national cultures influence the functioning or performance of teams that are culturally homogeneous but otherwise diverse. Yet cultural values are associated with a wide variety of work-related attitudes and behaviors, and it is reasonable to expect them to affect employees’ responses to team diversity (Jackson & Joshi, 2001; Triandis, 1992, 1994; Triandis, Kurowski, & Gelfand, 1994). An example of how cultures contextualize team diversity dynamics is provided by Van der Vegt, van de Vliert, and Huang (2005). They argued that since demographic diversity is associated with status differences, the effects of demographic diversity on firm-level innovative climates would differ in high versus low power-distance cultures. Supporting these propositions, the results indicated that tenure and functional (but not age or tenure) diversity were positively associated with innovative climates in low power-distance cultures, but they were negatively associated with innovative climates in high powerdistance cultures. The findings suggest that value and attitudes about equality (low power distance) and inequality (high power distance) influence the consequences of work team diversity. In an effort to address the lack of research on global diversity, Nishii and Özbilgin (2007b) developed an integrative framework for conceptualizing

global diversity that is likely to stimulate new work on this important topic. Briefly, their framework identifies characteristics of organizational policy makers and organizational culture that are likely to influence organizational approaches to global diversity management and subsequent diversity-related organizational outcomes. Nishii and Özbilgin also served as guest editors of a special journal issue devoted to this topic (Nishii & Ozbilgin, 2007a), which was an important step toward developing an international community of interested scholars. Given the growing internationalization of diversity research and the importance of this topic in large organizations, we expect scholarship will soon make progress toward understanding work team diversity in the global context.

Network Diversity Employees who work in medium to large organizations must usually rely on others for the resources and support they need to function effectively (Hackman, 1999). As they cross formal boundaries, they share and obtain tacit knowledge as well as tangible resources (Anand, Glick, & Manz, 2002; Tsai, 2002; Tsai & Ghoshal, 1998). Through these and other activities, employees become embedded in organizational networks. Just as network ties within a team tend to reflect the similarity bias, networks that form among team leaders and department managers are likely to be homogeneous rather than diverse (McPherson et al., 2001). Demographic similarity among team leaders or members of management may explain workflow networks and involvement in decision making (Bunderson, 2003). If demographic similarity of team leaders facilitates interteam cooperation, the teams working under similar leaders may achieve higher performance (e.g., see Joshi et al., 2006). Psychological similarity is important, too (e.g., Klein et al., 2004). Going forward, it seems likely that studies of workplace diversity will increasingly emphasize the informal social structures that hold organizations together. When such networks are leveraged by teams or departments to gain access to knowledge and resources or to improve coordination with other units, they are likely to enhance both team 675

Jackson and Joshi

Copyright American Psychological Association. Not for further distribution.

and organizational effectiveness. The challenge for organizations, then, is to facilitate the development of cohesiveness within work teams while also encouraging employees to build relationships beyond team boundaries. Research that sheds light on the combined effects of demographic and psychological network composition may provide insights into how to better manage the wideranging organizational networks that serve to connect the many individuals and work teams found within organizations.

Multilevel Perspectives As a whole, the empirical research on workplace diversity includes work conducted at many levels of analysis, including individuals, dyads, small groups and teams, networks, business units, and organizations. This chapter focused on findings from studies of work teams. Within the literature on work team diversity, as is true more generally in research conducted on diversity, it is apparent that scholars often assume that theoretical constructs are portable across different levels of analysis, despite cogent warnings about the danger of such logic (Klein, Dansereau, & Hall, 1994; Rousseau, 1985, 2000). Paradoxically, although they often borrow theoretical arguments developed for understanding phenomena at other levels of analysis, studies of work team diversity have usually ignored multilevel complexities when developing their hypotheses and analyzing their results. A few recent studies provide examples of the value of applying a multilevel perspective when studying work team diversity. In a study of sales teams, Jackson and Joshi (2004) modeled diversity effects at three levels of analysis: individual sales employees, sales teams, and sales districts (business units). Looking at performance as the outcome, their results revealed significant interactions between individual- and team-level predictors as well as additional effects at the district level. In an investigation of pay equity at this firm, Joshi, Liao, and Jackson (2006) found that pay equity was unrelated to team-level diversity but was significantly related to district-level diversity. In both studies, differences in effects at the team and district levels were unexpected, because the available 676

theoretical perspectives provided no rationale to suggest different effects at different levels of analysis. Looking ahead, increased use of multilevel analytic techniques may prove useful as diversity researchers strive to understand the important role of the contextual conditions within which diverse work teams are embedded. Likewise, research that addresses international diversity will almost certainly need to grapple explicitly with multilevel issues. When conducting research in organizational settings (vs. classrooms and laboratories), issues of context and the multilevel structures of organizations are ever present. Thus, often it is relatively easy in field research to incorporate such features of organizations. By doing so in future research on work team diversity, scholars may simultaneously improve their ability to speak to the concerns of managers and enrich their understanding of diversity dynamics.

Evaluating Diversity Initiatives On the basis of a comprehensive study of diversity management practices, Kalev et al. (2006) sought to determine whether the use of such practices improves organizational outcomes such as diversity among top executives or firm performance. On the basis of data from 708 private sector establishments, the authors concluded that diversity practices aimed at reducing managerial bias (e.g., diversity training) were the least effective in increasing the proportion of white women and black men and women. Practices aimed at reducing social isolation (e.g., mentoring) were modestly effective. Practices that aimed at increasing accountability in meeting diversity goals were the most effective for increasing firm diversity. They concluded, “We know a lot about the disease of workplace inequality but not much about the cure.” Although managers undoubtedly evaluate—at least subjectively—the effectiveness of the diversity practices they adopt, very few rigorous evaluations of interventions intended to improve the functioning of diverse teams or business units have been published. There has been substantially more research effort devoted to assessing the effectiveness of practices aimed at reducing unfair employment bias and increasing the representation of various minority groups (for a review, see Kulik & Roberson, 2008b).

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

The possibility that diversity initiatives may be of little value is troublesome, for they have become part of the corporate landscape in recent years. Diversity training, for example, is now used in most American organizations (Esen, 2005) and is migrating to other continents (e.g., see Süß & Kleiner, 2008). Diversity training has been introduced by employers for a variety of reasons, which include communicating organizational values, educating employees about other diversity management initiatives that are being introduced by the organization (e.g., changes in performance evaluation criteria), improving employee competencies that are believed to be necessary to establish a positive diversity climate, and addressing specific problems—like differential performance, pay, promotion, or turnover among employees from various demographic subgroups. Diversity education. A few studies have shown that diversity training can achieve basic educational goals. Informing employees of the advantages of a diverse workforce (Adler, 1986) and increasing employee acceptance of other diversity initiatives are two common educational goals for employers (Ellis & Sonnenfeld, 1994; Hanover & Cellar, 1998). Diversity education programs may set the stage for subsequent behavioral change, but they do not explicitly attempt to produce behavioral change. Diversity awareness training. Two training approaches aimed at achieving behavioral change are awareness training and skills training. In awareness training programs, the goal is to change attitudes. The assumption of such training is that awareness is a necessary first step toward reducing the negative behavioral consequences of biases and stereotypes. Often the approach involves activities designed to reveal employees’ own (and perhaps others’) biases and stereotypes and also to help employees understand the possible consequences of these. Based on a review of 20 studies conducted in organizational settings, Kulik and Roberson (2008a) concluded that diversity awareness training results in sustained improvements in overall attitudes toward diversity. However, attitudes toward specific demographic groups (e.g., defined by ethnicity, gender, age) appear to be more resistant to change and may even be at risk of a backlash effect (e.g., see Alderfer, 1992).

Diversity skills training. With diversity skills training, the focus is directly on changing behaviors that are needed to work effectively with dissimilar others. Among the skills identified as relevant for working in diverse teams are communication, conflict management, and behaving in ways that reflect sensitivity to cultural differences, as well as other skills that are generally useful for teamwork. On the basis of their review of 15 studies of diversity skill training effectiveness, Kulik and Roberson (2008b) concluded that there is little evidence to show that diversity skills training produces observable behavioral changes. Studies that relied on trainees’ self-reports generally found that employees reported self-improvement after receiving diversity skills training. However, when more objective measures of skills were used to assess training effectiveness, the results were mixed (Byington, Fischer, Walker, & Freedman, 1997; Roberson, Kulik, & Pepper, 2009; Sanchez & Medkik, 2004; Wade & Bernstein, 1991; Williams, 2005). Unfortunately, none of these studies was designed to assess whether diversity skills training improved teamwork processes (e.g., reduced conflict) or outcomes (e.g., improved performance). There is a clear need for additional research aimed at evaluating diversity initiatives that might improve the effectiveness of diverse work teams. First, rigorous needs analysis research is needed to improve our understanding of the specific attitudes and behaviors (including those of leaders) that are relevant to the effectiveness of diverse work teams. Second, evaluation studies are needed to determine the types of interventions that can be used to elicit and support those attitudes and behaviors. Also needed is research that focuses on teams as the focal unit for training to complement the usual practice of focusing on training for individuals (cf. Jackson, Chuang, Harden, & Jiang, 2006). Training designed and conducted with intact teams may prove more useful than training designed for and delivered to individual employees. Team-based training may be especially useful when it is designed to address the specific types of diversity present in the team, the specific types of tasks for which the team is responsible, and the specific form of leadership under which the team operates. 677

Jackson and Joshi

Copyright American Psychological Association. Not for further distribution.

Conclusion Despite the equivocal findings described in this chapter, we end this review on an optimistic note. Recent theoretical and methodological advancements in the field have opened up many new opportunities for diversity researchers, many of which are being explored already. As we write this chapter, we know that colleagues are forging new theoretical perspectives, crossing new frontiers in the development of diversity metrics, and employing multilevel methodologies to shed new light on the complexity of diversity dynamics at work. We know, too, that employers continue to invest in new diversity initiatives as they strive to create conditions that enable an increasingly diverse population of employees to effectively leverage their differences and achieve superior team results. Perhaps, a decade from now, a review such as this one will be able to provide answers to many of the questions we have raised.

References Adler, N. J. (1986). International dimensions of organizational behavior. Boston: Kent. Alderfer, C. P. (1992). Changing race relations embedded in organizations: Report on a long-term project with the XYZ Corporation. In S. E. Jackson & Associates (Eds.), Diversity in the workplace: Human resource initiatives (pp. 138–166). New York: Guilford Press. Anand, V., Glick, W. H., & Manz, C. C. (2002). Thriving on the knowledge of outsiders: Tapping organizational social capital. Academy of Management Executive, 16, 87–101. Ancona, D. G., & Caldwell, D. F. (1992). Bridging the boundary: External activity and performance in organizational teams. Administrative Science Quarterly, 37, 634–665. Apgar, M. (1998). The alternative workplace: Changing where and how people work. Harvard Business Review, 76, 121–137. Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20–39. Ashkanasy, N. M., Härtel, C. E. J., & Daus, C. S. (2002). Diversity and emotion: The new frontiers in organizational behavior research. Journal of Management, 28, 307–338. Avery, D. R., McKay, P. F., Wilson, D. C., & Tonidandel, S. (2007). Unequal attendance: The relationships between race, organizational diversity cues, and absenteeism. Personnel Psychology, 60, 875–902. 678

Avolio, B. J., & Bass, B. M. (2004). MLQ—Multifactor leadership questionnaire. Menlo Park, CA: Mind Garden. Bass, B. M., & Riggio, R. E. (2006). Transformational leadership. Mahwah, NJ: Erlbaum. Bantel, K. A., & Jackson, S. E. (1989). Top management and innovations in banking: Does the composition of the top team make a difference? Strategic Management Journal, 10, 107–124. Baron, J. N., & Newman, A. E. (1990). For what it’s worth: Organizations, occupations and the value of work done by women and non-Whites. American Sociological Review, 55, 155–175. Barrick M. R., Stewart G. L., Neubert, M., & Mount, M. K. (1998). Relating member ability and personality to work team processes and team effectiveness. Journal of Applied Psychology, 83, 377–391. Blalock, H. M. (1967). Toward a theory of minority group relations. New York: Wiley. Boone, C. A. J. J., van Olffen, W., van Witteloostuijn, A., & de Brabander, B. (2004). The genesis of top management team diversity: Selective turnover among top management teams in Dutch newspaper publishing, 1970–94. Academy of Management Journal, 47, 633–656. Bowers, C. A., Pharmer, J. A., & Salas, E. (2000). When member homogeneity is needed in work teams: A meta-analysis. Small Group Research, 31, 305–327. Bowers, C. A., Pharmer, J. A., & Salas, E. (2000). When member homogeneity is needed in work teams: A meta-analysis. Small Group Research, 31, 305–327. Brass, D. J. (1985). Men’s and women’s networks: A study of interaction patterns and influence in an organization. Academy of Management Journal, 28, 327–343. Brewer, M. B., & Brown, R. J. (1998). Intergroup relations. In D. T. Gilbert & S. T. Fiske (Eds.), Handbook of social psychology (pp. 553–594). Boston: McGraw-Hill. Bunderson, J. S. (2003). Team member functional background and involvement in management teams: Direct effects and the moderating role of power centralization. Academy of Management Journal, 46, 458–474. Bunderson, J. S., & Sutcliffe, K. M. (2002). Comparing alternative conceptualizations of functional diversity in management teams: Process and performance effects. Academy of Management Journal, 45, 875–893. Burt, R. S. (1992). Structural holes: The social structure of competition. Cambridge, MA: Harvard University Press. Byington, K., Fischer, J., Walker, L., & Freedman, E. (1997). Evaluating the effectiveness of multicultural

Work Team Diversity

counseling ethics and assessment training. Journal of Applied Rehabilitation Counseling, 28, 15–19.

Copyright American Psychological Association. Not for further distribution.

Byrne, D. (1971). The attraction paradigm. New York: Academic Press.

Datan, N., Rodeheaver, D., & Hughes, F. (1987). Adult development and aging. Annual Review of Psychology, 38, 153–180. De Dreu, C. K. W., & Weingart, L. R. (2003). Task and relationship conflict, team performance, and team member satisfaction: A meta-analysis. Journal of Applied Psychology, 88, 741–749.

Cannella, A. A., Park, J.-H., & Lee, H.-O. (2008). Top management team functional background diversity and firm performance: Examining the roles of team member colocation and environmental uncertainty. Academy of Management Journal, 51, 768–784.

Dunavant, B. M., & Heiss, B. (2005). Global diversity 2005. Washington, DC: Diversity Best Practices.

Carpenter, M. (2002). The implications of strategy and social context for the relationship between top management team heterogeneity and firm performance. Strategic Management Journal, 23, 275–284.

Earley, P. C., & Mosakowski, E. M. (2000). Creating hybrid team cultures: An empirical test of international team functioning. Academy of Management Journal, 43, 26–49.

Chatman, J. A., Polzer, J. T., Barsade, S. G., & Neale, M. A. (1998). Being different yet feeling similar: The influence of demographic composition and organizational culture on work processes and outcomes. Administrative Science Quarterly, 43, 749–780.

Eisenhardt, K. M., & Schoonhoven, C. B. (1990). Organizational growth: Linking founding team, strategy, environment, and growth among U.S. semiconductor ventures, 1978–1988. Administrative Science Quarterly, 35, 504–529.

Chatman, J. A., & Spataro, S. E. (2005). Using selfcategorization theory to understand relational demography-based variations in people’s responsiveness to organizational culture. Academy of Management Journal, 48, 321–331.

Elder, G. H., Jr. (1974). Children of the great depression. Chicago: University of Illinois Press.

Chattopadhyay, P., Tluchowska, M., & George, E. (2004). Identifying the ingroup: A closer look at the influence of demographic dissimilarity on employee social identity. Academy of Management Review, 29, 180–202.

Ellemers, N., De Gilder, D., & Haslam, S. A. (2004). Motivating individuals and groups at work: A social identity perspective on leadership and group performance. Academy of Management Review, 29, 459–478.

Chung, Y., & Jackson, S. E. (2009). Balancing the internal and external social capital of diverse R&D teams: The role of task routineness. Unpublished manuscript. University of Idaho, Moscow.

Ellis, C., & Sonnenfeld, J. A. (1994). Diverse approaches to managing diversity. Human Resource Management, 33(1), 79–109.

Chung, Y., Jackson, S. E., & Shaw, J. B. (2005, August). Multi-level effects of demographic faultlines on team performance and rewards. Paper presented at the annual meeting of the Academy of Management, Honolulu, Hawaii. The Conference Board. (2006). The business case for diversity: Good practices in the workplace (Executive Action Report No. A-0182-06-EA). New York: Author. Cox, T. (1991). The multicultural organization. Academy of Management Executive, 5, 34–47.

Elder, G. H., Jr. (1975). Age differentiation and the life course. Annual Review of Sociology, 1, 165–190.

Elsass, P. M., & Graves, L. M. (1997). Demographic diversity in decision-making groups: The experience of women and people of color. Academy of Management Review, 22, 946–973. Ely, R. J. (1994). The effects of organizational demographics and social identity on relationships among professional women. Administrative Science Quarterly, 39, 203–238. Ely, R. J. (1995). The power in demography: Women’s social constructions of gender identity at work. Academy of Management Journal, 38, 589–634.

Cox, T. (1993). Cultural diversity in organizations: Theory, research and practice. San Francisco: Berrett-Koehler.

Ely, R. J. (2004). A field of group diversity, participation in diversity education programs, and performance. Journal of Organizational Behavior, 25, 755–780.

Dailey, R. C. (1978). The role of team and task characteristics in R&D team collaborative problem solving and productivity. Management Science, 24, 1579–1588.

Ely, R. J., & Thomas, D. A. (2001). Cultural diversity at work: The effects of diversity perspectives on work group processes and outcomes. Administrative Science Quarterly, 46, 229–273.

Dass, P., & Parker, B. (1999). Strategies for managing human resource diversity: From resistance to learning. Academy of Management Executive, 13, 68–80.

Esen, E. (2005). Workplace diversity practices survey report. Alexandria, VA: Society for Human Resource Management. 679

Jackson and Joshi

Finkelstein, S., & Hambrick, D. C. (1990). Top management team tenure and organizational outcomes: The moderating role of managerial discretion. Administrative Science Quarterly, 35, 484–503. Finkelstein, S., & Hambrick, D. C. (1996). Strategic leadership: Top executives and their effects on organizations. St. Paul, MN: West.

Copyright American Psychological Association. Not for further distribution.

Frink, D. D., Robinson, R. K., Reithel, B., Arthur, M. M., Ammeter, A. P., & Ferris, G. R. (2003). Gender demography and organization performance: A twostudy investigation with convergence. Group and Organization Management, 28, 127–147. Gibson, C. B., & Vermeulen, F. (2003). A healthy divide: Subgroups as a stimulus for team learning behavior. Administrative Science Quarterly, 48, 202–239. Gilbert, J. A., & Ivancevich, J. M. (2000). Valuing diversity: A tale of two organizations. Academy of Management Executive, 14, 93–105. Godthelp, M., & Glunk, U. (2003). Turnover at the top: Demographic diversity as a determinant of executive turnover in the Netherlands. European Journal of Management, 21, 614–625. Grensing-Pophal, L. (2002). Reaching for diversity. HR Magazine, 47, 53–56. Hackman, J. R. (1999). Thinking differently about context. In R. Wageman (Ed.), Groups in context (pp. 233–247). Stamford, CT: JAI Press. Haley, H., & Sidanius, J. (2006). The positive and negative framing of affirmative action: A group dominance perspective. Personality and Social Psychology Bulletin, 32, 656–668. Hambrick, D. C., Geletkanycz, M. A., & Fredrickson, J. W. (1990, August). Top executive commitment to the status quo: Some tests of its determinants. Paper presented at the Annual Meeting of the Academy of Management, San Francisco. Hanover, J. M., & Cellar, D. F. (1998). Environmental factors and the effectiveness of workforce diversity training. Human Resource Development Quarterly, 9, 105–124.

Harrison, D. A., Price, K. H., Gavin, J. H., & Florey, A. T. (2002). Time, teams, and task performance: Changing effects of surface- and deep-level diversity on group functioning. Academy of Management Journal, 45, 1029–1045. Harrison, D. A., & Sin, H-S. (2005). What is diversity and how should it be measured? In A. M. Konrad, P. Prasad, & J. K. Pringle (Eds.), Handbook of workplace diversity (pp. 191–216). Newbury Park, CA: Sage. Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup-bias. Annual Review of Psychology, 53, 575–604. Hirschfeld, R. R., Jordan, M. H., Feild, H. S., Giles, W. F., & Armenakis, A. A. (2005). Teams’ female representation and perceived potency as inputs to team outcomes in a predominantly male field setting. Personnel Psychology, 58, 893–924. Hoffman, E. (1985). The effect of race-ratio composition on the frequency of organizational communication. Social Psychology Quarterly, 48, 17–26. Holland, J. L. (1976). Vocational preferences. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 521–570). Chicago: Rand McNally. Homan, A. C., van Knippenberg, D., Van Kleef, G. A., & De Dreu, C. K. W. (2007). Bridging faultlines by valuing diversity: Diversity beliefs, information elaboration, and performance in diverse work groups. Journal of Applied Psychology, 92, 1189–1199. Horwitz, S. K., & Horwitz, I. B. (2007). The effects of team diversity on team outcomes: A meta-analytic review of team demography. Journal of Management, 33, 967–1015. Hutlin, M., & Szulkin, R. (1999). Wages and unequal access to organizational power: An empirical test of gender discrimination. Administrative Science Quarterly, 44, 453–472. Huston, L., & Sakkab, N. (2006). Connect and develop: Inside Procter & Gamble’s new model for innovation. Harvard Business Review, 84, 58–66.

Harris, C. L., & Beyerlein, M. M. (2003). Team-based organization. In M. A. West, D. Tjosvold, & K. G. Smith (Eds.), Handbook of organizational teamwork and cooperative working (pp. 187–209). Chichester, England: Wiley.

Iaquinto, A. L., & Fredrickson, J. W. (1997). Top management team agreement about the strategic decision process: A test of some of its determinants and consequences. Strategic Management Journal, 18, 63–75.

Harrison, D. A., & Klein, K. J. (2007). What’s the difference? Diversity constructs as separation, variety, or disparity in organizations. Academy of Management Review, 32, 1199–1228.

Ibarra, H. (1992). Homophily and differential returns: Sex differences in network structure and access in an advertising firm. Administrative Science Quarterly, 37, 422–447.

Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond relational demography: Time and the effects of surface- and deep-level diversity on work group cohesion. Academy of Management Journal, 41, 98–107.

Jackson, S. E. (1992). Consequences of group composition for the interpersonal dynamics of strategic issue processing. In P. Shrivastava, A. Huff, & J. Dutton (Eds.), Advances in strategic management (pp. 345–382). Greenwich, CT: JAI Press.

680

Work Team Diversity

Jackson, S. E., Brett, J. F., Sessa, V. I., Cooper, D. M., Julin, J. A., & Peyronnin, K. (1991). Some differences make a difference: Individual dissimilarity and group heterogeneity as correlates of recruitment, promotions, and turnover. Journal of Applied Psychology, 76, 675–689.

Copyright American Psychological Association. Not for further distribution.

Jackson, S. E., Chuang, C.-H., Harden, E. E., & Jiang, Y. (2006). Toward developing human resource management systems for knowledge-intensive teamwork. In J. J. Martocchio (Ed.), Research in personnel and human resources management (Vol. 25, pp. 1–26). New York: JAI Press.

Jehn, K. A., Northcraft, G. B., & Neale, M. A. (1999). Why differences make a difference: A field study of diversity, conflict, and performance in workgroups. Administrative Science Quarterly, 44, 741–63. Jiang, Y., Jackson, S. E., Shaw, J. B., & Chung, Y. (2008). A longitudinal study of the effects of faultlines on team processes and performance (Working paper). New Brunswick, NJ: Rutgers University. Joshi, A. (2006). The influence of organizational demography on the external networking behavior of teams. Academy of Management Review, 31, 583–595.

Jackson, S. E., & Chung, Y. (2008). The people make the place complicated. In D. B. Smith (Ed.), The people make the place: Dynamic linkages between individuals and organizations (pp. 37–52). Mahwah, NJ: Erlbaum.

Joshi, A., & Jackson, S. E. (2003). Managing workforce diversity to enhance cooperation in organizations. M. A. West, D. Tjosvold, and K. Smith (Eds.), International handbook of organizational teamwork and cooperative working (pp. 277–296). New York: Wiley.

Jackson, S. E., Hitt, M. A., & DeNisi, A. S. (Eds.) (2003). Managing knowledge for sustained competitive advantage: Designing strategies for effective human resource management. San Francisco: Jossey-Bass.

Joshi, A., Liao, H., & Jackson, S. E. (2006). Cross-level effects of workplace diversity on sales performance and pay. Academy of Management Journal, 49, 459–481.

Jackson, S. E., & Hong, Y. (2008). Knowledge work. In C. L. Cooper & J. Barling (Eds.), The Sage handbook of organizational behavior: Vol. 1. Micro approaches (pp. 371–392). London: Sage. Jackson, S. E., & Joshi, A. (2001). Research on domestic and international diversity in organizations: A merger that works? In N. Anderson, D. S. Ones, H. K. Sinangil, & C. Viswesvaran (Eds.), Handbook of work, industrial and organizational psychology (pp. 206–231). Thousand Oaks, CA: Sage. Jackson, S. E., & Joshi, A. (2004). Diversity in social context: A multiattribute, multilevel analysis of team diversity and sales performance. Journal of Organizational Behavior, 25, 675–702. Jackson, S. E., Joshi, A., & Erhardt, N. L. (2003). Recent research on team and organizational diversity: SWOT analysis and implications. Journal of Management, 29, 801–830. Jehn, K. A. (1995). A multimethod examination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40, 256–282. Jehn, K. A., Greer, L. L., & Rupert, J. (2008). Diversity, conflict, and their consequences. In A. P. Brief (Ed.), Diversity at work (pp. 127–174). New York: Cambridge University Press. Jehn, K., & Bezrukova, K. (2004). A field study of group diversity, group context, and performance. Journal of Organizational Behavior, 25, 1–27. Jehn, K. A., & Mannix, E. A. (2001). The dynamic structure of conflict: A longitudinal study of intragroup conflict and group performance. Academy of Management Journal, 44, 238–251.

Joshi, A., & Roh, H. (2007). Context matters: A multilevel framework for work team diversity research. In J. J. Martocchio (Ed.), Research in personnel and human resources management (Vol. 26, pp. 1–48). Greenwich, CT: JAI Press. Joshi, A., & Roh, H. (2009). The role of context in work team diversity research: A meta-analytic review. Academy of Management Journal, 52, 599–628. Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in system-justification and the production of false consciousness. British Journal of Social Psychology, 33, 1–27. Jost, J. T., Pelham, B. W., & Carvallo, M. (2002). Nonconscious forms of system justification: cognitive, affective, and behavioral preferences for higher status groups. Journal of Experimental Social Psychology, 38, 586–602. Kalev, A., Dobbin, F., & Kelly, E. (2006). Best practices or best guesses? Assessing the efficacy of corporate affirmative action and diversity policies. American Sociological Review, 71, 589–617. Kark, R., & Shamir, B. (2002). The dual effect of transformational leadership: Priming relational and collective selves and further effects on followers. In B. J. Avolio & F. J. Yammarino (Eds.), Transformational and charismatic leadership: The road ahead (pp. 67–91). Oxford, England: Elsevier Science. Katz, R. (1982). The effects of group longevity on project communication and performance, Administrative Science Quarterly, 27, 81–104. Katz, N., Lazer, D., Arrow, H., & Contractor, N. (2004). Network theory and small groups. Small Group Research, 35, 307–332. 681

Jackson and Joshi

Kearney, E., & Gebert, D. (2009). Managing diversity and enhancing team outcomes: The promise of transformational leadership. Journal of Applied Psychology, 94, 77–89. Kearney, E., Gebert, D., & Voelpel, S. (2009). When and how diversity benefits teams: The importance of team members’ need for cognition. Academy of Management Journal, 62, 581–598.

Copyright American Psychological Association. Not for further distribution.

Keller, R. T. (2001). Cross-functional project groups in research and new product development: Diversity, communications, job stress, and outcomes. Academy of Management Journal, 44, 547–555.

Larkey, L. K. (1996). Toward a theory of communicative interactions in culturally diverse work-groups. Academy of Management Review, 23, 325–340. Lau, D. C., & Murnighan, J. K. (1998). Demographic diversity and faultlines: The compositional dynamics of organizational groups. Academy of Management Review, 23, 325–340. Lawrence, B. S. (1997). The black box of organizational demography. Organization Science, 8, 1–22. Lester, S. (2006). Diversity and profitability: Making the connection (Executive Action Report No. A-0175-06EA). New York: The Conference Board, Inc.

Keller, R. T., Julian, S. D., & Kedia, B. L. (1996). A multinational study of work climate, job satisfaction, and the productivity of R&D teams. IEEE Transactions on Engineering Management, 42, 48–55.

Li, J. T., & Hambrick, D. C. (2005). Factional groups: A new vantage on demographic faultlines, conflict and disintegration in work teams. Academy of Management Journal, 48, 794–813.

Kilduff, M., Angelmar, R., & Mehra, A. (2000). Top management-team diversity and firm performance: Examining the role of cognitions. Organization Science, 11, 21–34.

Lin, N., Ensel, W. M., & Vaughn, J. C. (1981). Social resources and strength of ties. American Sociological Review, 46, 393–405.

Kirkman, B. L., Tesluk, P. E., & Rosen, B. (2004). The impact of demographic heterogeneity and team leader-team member demographic fit on team empowerment and effectiveness. Group and Organization Management, 29, 334–368. Klein, K. J., Dansereau, F., & Hall, R. I. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19, 195–229. Klein, K. J., Lim, B., Saltz, J. L., & Mayer, D. M. (2004). How do they get there? An examination of the antecedents of centrality in team networks. Academy of Management Journal, 47, 952–963. Kochan, T., Bezrukova, K., Ely, R., Jackson, S., Joshi, A., Jehn, K., et al. (2003). The effects of diversity on business performance: Report of the Diversity Research Network. Human Resource Management, 42, 3–21. Kossek, E. E., & Zonia, S. C. (1993). Assessing diversity climate: A field study of reactions to employer efforts to promote diversity. Journal of Organizational Behavior, 14, 61–81. Kulik, C. T., & Roberson, L. (2008a). Common goals and golden opportunities: Evaluations of diversity education in academic and organizational settings. Academy of Management Learning & Education, 7, 323. Kulik, C. T., & Roberson, L. (2008b). Diversity initiative effectiveness: What organizations can (and cannot) expect from diversity recruitment, diversity training, and formal mentoring programs. In A. P. Brief (Ed.), Diversity at work (pp. 265–317). Cambridge, England: Cambridge University Press. 682

Mangaliso, M. P., & Nkomo, S. N. (2001). HerdbuoysMcCann Erickson’s Chairman Peter Vundla on Black entrepreneurs and economic empowerment. Academy of Management Executive, 15, 16–22. March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley. Marsden, P. V. (1990). Network data and measurement. Annual Review of Sociology, 16, 435–463. Martins, L. L., Gilson, L. L., & Maynard, M. T. (2004). Virtual teams: What do we know and where do we go from here? Journal of Management, 30, 805–835. McCain, B. E., O’Reilly, C. A., & Pfeffer, J. (1983). The effects of departmental demography on turnover: The case of a university. Academy of Management Journal, 26, 626–641. McCarty-Killian, C., Hukai, D., & McCarty, C. E. (2005). Building diversity in the pipeline to corporate leadership. Journal of Management Development, 24, 155–168. McGrath, J. E. (1984). Groups, interaction, and performance. Englewood Cliffs, NJ: Prentice-Hall. McKay, P. F., Avery, D. R., Tonidandel, S., Morris, M. A, Hernandez, M., & Hebl, M. (2007). Racial differences in employee retention: Are diversity climate perceptions the key? Personnel Psychology, 60, 35–62. McPherson, J. M., Smith-Lovin, L., & Cook, J. (2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415–444. Mehra, A., Kilduff, M., & Brass, D. J. (1998). At the margins: A distinctiveness approach to social identity and social networks of underrepresented groups. Academy of Management Journal, 41, 441–452.

Copyright American Psychological Association. Not for further distribution.

Work Team Diversity

Milliken, F. J., & Martins, L. L. (1996). Searching for common threads: Understanding the multiple effects of diversity in organizational groups. Academy of Management Review, 21, 402–434.

Noe, R., Greenberger, D. & Wang, S. (2002). Mentoring: What we know and where we might go. Research in Personnel and Human Resources Management, 21, 129–173.

Mohammed, S., & Angell, L. C. (2004). Surface- and deep-level diversity in workgroups: Examining the moderating effects of team orientation and team process on relationship conflict. Journal of Organizational Behavior, 25, 1015–1039.

Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereotyping and social reality. Oxford, England: Blackwell.

Mohammed, S., Hamilton, K., & Lim, A. (2009). The incorporation of time in team research: Past, current, and future. In E. Salas, G. F. Goodwin, & C. S. Burke (Eds.), Team effectiveness in complex organizations: Cross-disciplinary perspectives and approaches (pp. 323–349). Mahwah, NJ: Erlbaum. Molleman, E. (2005). Diversity in demographic characteristics, abilities, and personality traits: Do faultlines affect team functioning? Group Decisions and Negotiations, 14, 173–195. Mor Barak, M. E., Cherin, D. A., & Berkman, S. (1998). Organizational and personal dimensions of diversity climate: Ethnic and gender differences in employee perceptions. Journal of Applied Behavioral Sciences, 31, 82–104. Mueller, C., Finley, A., Iverson, R., & Price, J. (1999). The effects of group racial composition on job satisfaction, organizational commitment, and career commitment: The case of teachers. Work and Occupation, 26, 187–219. National Science Foundation. (2004). SRS publications and data. Available at http://www.nsf.gov/statistics/ shopub.cfm?TopID=2&SubID=5 Nemeth, C. J. (1986). Differential contributions of majority and minority influence. Psychological Review, 91, 23–32. Nemeth, C. J. (1997). Managing innovation: When less is more. California Management Review, 40, 59–74. Neuman, G. A., Wagner, S. H., & Christiansen, N. D. (1999). The relationship between work-team personality composition and the job performance of teams. Group and Organization Management, 24, 28–39. Nishii, L. H., & Ozbilgin, M. F. (Eds.). (2007a). Global diversity management [Special issue] International Journal of Human Resource Management, 18(12). Nishii, L. H., & Özbilgin, M. F. (2007b). Global diversity management: Towards a conceptual framework. International Journal of Human Resource Management, 11, 1883–1894. Nishii, L. H., & Raver, J. L. (2003, April). Collective climates for diversity: Evidence from a field study. Paper presented at the 18th Annual Conference of the Society for Industrial and Organizational Psychology, Orlando, FL.

Ofori-Dankwa, J., & Julian, S. (2002). Towards diversity and similarity curves: Implications for theory, research and practice. Human Relations, 52, 199–224. Oh, H., Chung, M., & Labianca, G. (2004). Group social capital and group effectiveness: The role of informal socializing ties. Academy of Management Journal, 47, 860–875. O’Reilly, C. A., III, Caldwell, D. F., & Barnett, W. P. (1989). Work group demography, social integration, and turnover. Administrative Science Quarterly, 34, 21–37. Pearsall, M. J., Ellis, A. P. J., & Evans, J. M. (2008). Unlocking the effects of gender faultlines on team creativity: Is activation the key? Journal of Applied Psychology, 93, 225–234. Pelled, L. H. (1996). Relational demography and perceptions of group conflict and performance: A field investigation. International Journal of Conflict Management, 7, 230–246. Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. (1999). Exploring the black box: An analysis of work group diversity, conflict and performance. Administrative Science Quarterly, 44, 1–28. Pfeffer, J. (1983). Organizational demography. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (Vol. 5, pp. 299–357). Greenwich, CT: JAI Press. Pfeffer, J. (1992). Managing with power: Politics and influence in organizations. Boston, MA: Harvard Business Press. Pirola-Merlo, A., Hartel, C., Mann, L., & Hirst, G. (2002). How leaders influence the impact of affective events on team climate and performance in R&D teams. Leadership Quarterly, 13, 561–581. Polzer, J., Crisp, C. B., Jarvenpaa, S., & Kim, J. (2006). Extending the faultline model to geographically dispersed teams: How colocated groups can impair group functioning. Academy of Management Journal, 49, 579–592. Plaut, V. C., Thomas, K. M., & Goren, M. J. (2009). Is multiculturalism or color blindness better for minorities? Psychological Science, 20, 444–445. Priem, R. L. (1990). Top management team group factors, consensus, and firm performance. Strategic Management Journal, 11, 469–478. 683

Jackson and Joshi

Pugh, S., Dietz, J., Brief, A., & Wiley, J. W. (2008). Looking inside and out: The impact of employee and community demographic composition on organizational diversity climate. Journal of Applied Psychology, 93, 1422–1428. Ragins, B. R., & Kram, K. E. (2007): The handbook of mentoring at work: Theory, research and practice. Thousand Oaks, CA: Sage.

Copyright American Psychological Association. Not for further distribution.

Randel, A. E. (2002). Identity salience: A moderator of the relationship between group gender composition and work group conflict. Journal of Organizational Behavior, 23, 749–766. Reagans, R., & Zuckerman, E. W. (2001). Networks, diversity, and productivity: The social capital of corporate R&D teams. Organization Science, 12, 502–517. Reagans, R., Zuckerman, E. W., & McEvily, B. (2004). How to make the team: Social networks vs. demography as criteria for designing effective teams. Administrative Science Quarterly, 49, 101–133. Rentsch, J. R., & Klimoski, R. J. (2001). Why do “great minds” think alike? Antecedents of team member schema agreement. Journal of Organizational Behavior, 22, 107–122. Reynolds, K. J., Turner, J. C., & Haslam, S. A. (2003). Social identity and self-categorization theories’ contribution to understanding identification, salience and diversity in teams and organizations. In J. T. Polzer (Ed.), Research on managing groups and teams (Vol. 5, pp. 279–304). Oxford, England: Elsevier Science. Richard, O. C., Barnett, T., Dwyer, S., & Chadwick, K. (2004). Cultural diversity in management, firm performance, and the moderating role of entrepreneurial orientation. Academy of Management Journal, 47, 255–266. Ridgeway, C. (1997). Interaction and the conservation of gender inequality: Considering employment. American Sociological Review, 62, 218–235. Roberson, L., Kulik, C. T., Pepper, M. B. (2009). Individual and environmental factors influencing the use of transfer strategies after diversity training. Group & Organization Management, 34, 67–89. Roberson, L., Kulik, C. T., Pepper, M. B. (2009), Individual and environmental factors influencing the use of transfer strategies after diversity training. Group & Organization Management, 34, 67–89. Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level and cross-level perspectives. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (Vol. 7, pp. 1–37). Greenwich, CT: JAI Press. Rousseau, D. M. (2000). Multilevel competencies and missing linkages. In K. J. Klein & S. W. J. Kozlowski 684

(Eds.), Multilevel theory, research, and methods in organizations (pp. 572–582). San Francisco: Jossey-Bass. Sacco, J., & Schmitt, N. (2005). A dynamic multilevel model of demographic diversity and misfit effects. Journal of Applied Psychology, 90, 203–231. Sachdev, I., & Bourhis, R. Y. (1985). Social categorization and power differentials in group relations. European Journal of Social Psychology, 15, 415–434. Sachdev, I., & Bourhis, R. Y. (1987). Status differentials and intergroup behaviour. European Journal of Social Psychology, 17, 277–293. Sachdev, I., & Bourhis, R. Y. (1991). Power and status differentials in minority and majority group relations. European Journal of Social Psychology, 21, 1–24. Sanchez, J. L., & Medkik, N. (2004). The ethics of diversity awareness training on differential treatment. Group and Organization Management, 29, 517–536. Schippers, M. C., Den Hartog, D. N., Koopman, P. L., & Wienk, J. A. (2003). Reflexivity and diversity in teams: The moderating effects of outcome interdependence and group longevity. Journal of Organizational Behavior, 24, 779–802. Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437–453. Schneider, B., Goldstein, H. W., & Smith, D. B. (1995). The ASA framework: An update. Personnel Psychology, 48, 747–773. Schneider, B., Smith, D. B., Taylor, S., & Fleenor, J. (1998). Personality and organizations: A test of the homogeneity of personality hypothesis. Journal of Applied Psychology, 83, 462–470. Shaw, J. B. (2004). The development and analysis of a measure of group faultlines. Organizational Research Methods, 7, 66–100. Shaw, M. E. (1981). Group dynamics: The psychology of small group behavior (3rd ed.). New York: McGrawHill. Sidanius, J. (1993). The psychology of group conflict and the dynamics of oppression: A social dominance perspective. In S. Iyengar & W. McGuire (Eds.), Explorations in political psychology (pp. 183–219). Durham, NC: Duke University Press. Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press. Simons, T., Pelled, L. H., & Smith, K. A. (1999). Making use of difference: Diversity, debate, and decision comprehensiveness in top management teams. Academy of Management Journal, 42, 662–673.

Work Team Diversity

Sproull, L., & Keisler, S. (1986). Reducing social context cues: Electronic mail in organizational communication. Management Science, 32, 1492–1512. Süß, S., & Kleiner, M. (2008). Diversity management in Germany: Dissemination and design of the concept. International Journal of Human Resource Management, 18, 1934–1953.

Copyright American Psychological Association. Not for further distribution.

Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), Social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks-Cole. Tajfel, H., & Turner, J. C. (1985). The social identity theory of intergroup behavior. In S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (2nd ed., pp. 7–24). Chicago: Nelson-Hall.

Van der Vegt, G. S., van de Vliert, E., & Huang, X. (2005). Location-level links between diversity and innovative climate depend on national power distance. Academy of Management Journal, 48, 1171–1182. van Knippenberg, D., De Dreu, C. K. W., & Homan, A. C. (2004). Work group diversity and group performance: An integrative model and research agenda. Journal of Applied Psychology, 89, 1008–1022. van Knippenberg, D., & Schippers, M. C. (2007). Work group diversity. Annual Review of Psychology, 58, 515–541. Vroom, V., & Pahl, B. (1971). Relationship between age and risk-taking among managers. Journal of Applied Psychology, 55, 399–405.

Thatcher, S. M. B., Jehn, K. A., & Zanutto, E. (2003). Cracks in diversity research: The effects of diversity faultlines on conflict and performance. Group Decision and Negotiation, 12, 217–241.

Wade, P., & Bernstein, B. L. (1991). Cultural sensitivity training and counselor’s race: Effects on Black female clients’ perceptions and attitudes. Journal of Counseling Psychology, 38, 9–15.

Thernstrom, S. (1973). The other Bostonians: Poverty and progress in the American metropolis. 1880–1970. Cambridge, MA: Harvard University Press.

Wagner, W. G., Pfeffer, J., & O’Reilly, C. A. (1984). Organizational demography and turnover in topmanagement groups. Administrative Science Quarterly, 29, 74–92.

Triandis, H. C. (1992). The importance of context in studies of diversity. In S. E. Jackson (Ed.), Diversity in the workplace: Human resources initiatives (pp. 225–233). New York: Guilford Press. Triandis, H. C. (1994). Culture and social behavior. New York: McGraw-Hill. Triandis, H. C., Kurowski, L. L., & Gelfand, M. J. (1994). Workplace diversity. In H. C. Triandis, M. D. Dunnette, & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2, pp. 769–827). Palo Alto, CA: Consulting Psychologists Press. Tsai, W. (2002). Social structure of “coopetition” within a multiunit organization: Coordination, competition, and intra-organizational knowledge sharing. Organization Science, 13, 179–190. Tsai, W., & Ghoshal, S. (1998). Social capital and value creation: The role of intrafirm networks. Academy of Management Journal, 41, 464–476. Tsui, A. S., Egan, T. D., & O’Reilly, C. A., III. (1992). Being different: Relational demography and organizational attachment. Administrative Science Quarterly, 37, 549–579. Turner, J. C., & Haslam, A. (2001). Social identity, organizations and leadership. In M. Turner (Ed.), Groups at work: Theory and research (pp. 25–65). Mahwah, NJ: Erlbaum. Van den Bulte, C., & Moenaert, R. K. (1998). The effect of R&D team co-location on communication patterns among R&D, marketing, and manufacturing. Management Science, 44, 1–18.

Wasserman, I. C., Gallegos, P. V., & Ferdman, B. M. (2008). Dancing with resistance: Leadership challenges in fostering a culture of inclusion. In K. M. Thomas (Ed.), Diversity resistance in organizations: Manifestations and solutions (pp. 175–200). New York: Taylor & Francis. Watson, W. E., Johnson, L., & Merritt, D. (1998). Team orientation, self-orientation, and diversity in task groups: Their connection to team performance over time. Group and Organization Management, 23, 161–189. Webber, S. S., & Donahue, L. M. (2001). Impact of highly and less job-related diversity on work group cohesion and performance. Journal of Management, 27, 141–162. Wegge, J., Roth, C., Kanfer, R., Neubach, B., & Schmitt, K.-H. (2008). Age and gender diversity as determinants of performance and health in a public organization: The role of task complexity and group size. Journal of Applied Psychology, 93, 1301–1303. Weisband, S. P., Schneider, S. K., & Connolly, T. (1995). Computer-mediated communication and social information: Status salience and status differences. Academy of Management Journal, 38, 1124–1151. West, C. T., Jr., & Schwenk, C. R. (1996). Top management team strategic consensus, demographic homogeneity and firm performance: A report of resounding nonfindings. Strategic Management Journal, 17, 571–576. West, M. A., Tjosvold, D., & Smith, K. (Eds.) (2003). International handbook of organizational teamwork 685

Jackson and Joshi

and cooperative working (pp. 277–296) New York: Wiley. Westphal, J. D., & Milton, L. (2000). How experience and network ties affect the influence of demographic minorities on corporate boards. Administrative Science Quarterly, 45, 366–398. Wharton, A. (1992). The social construction of gender and race in organizations: A social identity and group mobilization perspective. In P. Tolbert & S. Bacharach (Eds.), Research in the sociology of organizations (Vol. 10, pp. 55–84). Greenwich, CT: JAI Press.

Copyright American Psychological Association. Not for further distribution.

Wiersema, M. F., & Bantel, K. A. (1992). Top management team demography and corporate strategic change. Academy of Management Journal, 35, 91–121. Williams, C. C. (2005). Training for cultural competence: Individual and group processes. Journal of Ethnic and Cultural Diversity in Social Work, 14, 11–143. Williams, K., & O’Reilly, C. A. (1998). Demography and diversity in organizations: A review of 40 years of research. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 20, pp. 77–140). Greenwich, CT: JAI Press.

686

Wright, P., Ferris, S. P., Hiller, J. S., & Kroll, M. (1995). Competitiveness through management of diversity: Effects on stock price valuation. Academy of Management Journal, 38, 272–287. Wu, J., Yi, X., & Lawler, J. J. (2000, August). Genderbased employment discrimination and multinational firms: A study of job announcements in Taiwan and Thailand. Paper presented at the annual meeting of the Academy of Management, Toronto, Ontario, Canada. Zatnick, C. D., Elvira, M. M., & Cohen, L. E. (2003). When is more better? The effects of racial composition on voluntary turnover. Organization Science, 14, 483–496. Zellner-Bruhn, M. E., Maloney, M. M., Bhappu, A. D., & Salvador, R. (2008). When and how do differences matter? An exploratory of perceived similarity in teams. Organizational Behavior and Human Decision Processes, 107, 41–59. Zenger, T. R., & Lawrence, B. S. (1989). Organizational demography: The differential effects of age and tenure distributions on technical communication. Academy of Management Journal, 32, 353–376.

Copyright American Psychological Association. Not for further distribution.

APA Handbook of

Industrial and Organizational Psychology

APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

Copyright American Psychological Association. Not for further distribution.

American Psychological Association • Handbooks in Psychology

Copyright American Psychological Association. Not for further distribution.

APA Handbook of

Industrial and Organizational Psychology volume 2 Selecting and Developing Members for the Organization

Sheldon Zedeck Editor-in-Chief

American Psychological Association • Washington, DC

Copyright © 2011 by the American Psychological Association. All rights reserved. Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, including, but not limited to, the process of scanning and digitization, or stored in a database or retrieval system, without the prior written permission of the publisher.

Copyright American Psychological Association. Not for further distribution.

Published by American Psychological Association 750 First Street, NE Washington, DC 20002-4242 www.apa.org To order APA Order Department P.O. Box 92984 Washington, DC 20090-2984 Tel: (800) 374-2721; Direct: (202) 336-5510 Fax: (202) 336-5502; TDD/TTY: (202) 336-6123 Online: www.apa.org/books/ E-mail: [email protected] In the U.K., Europe, Africa, and the Middle East, copies may be ordered from American Psychological Association 3 Henrietta Street Covent Garden, London WC2E 8LU England AMERICAN PSYCHOLOGICAL ASSOCIATION STAFF Gary R. VandenBos, PhD, Publisher Julia Frank-McNeil, Senior Director, APA Books Theodore J. Baroody, Director, Reference, APA Books Shenyun Wu, Project Coordinator, APA Books Typeset in Berkeley by Circle Graphics, Inc., Columbia, MD Printer: Edwards Brothers, Ann Arbor, MI Cover Designer: Naylor Design, Washington, DC Library of Congress Cataloging-in-Publication Data APA handbook of industrial and organizational psychology / Sheldon Zedeck, editor-in-chief. — 1st ed. p. cm. — APA Handbooks in psychology) Includes bibliographical references and index. ISBN-13: 978-1-4338-0727-5 ISBN-10: 1-4338-0727-0 ISBN-13: 978-1-4338-0732-9 (vol. 1) ISBN-10: 1-4338-0732-7 (vol. 1) [etc.] 1. Psychology, Industrial. 2. Organizational behavior. I. Zedeck, Sheldon. II. American Psychological Association. III. Title: Handbook of industrial and organizational psychology. HF5548.8.A684 2011 158.7—dc22 2009048439 British Library Cataloguing-in-Publication Data A CIP record is available from the British Library.

Printed in the United States of America First Edition

Copyright American Psychological Association. Not for further distribution.

Contents

Volume 2: Selecting and Developing Members for the Organization Editorial Board . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Part I. Foundations of Selection and Development . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Chapter 1. Work Analysis: From Technique to Theory . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Frederick P. Morgeson and Erich C. Dierdorff Chapter 2. Recruitment: A Review of Research and Emerging Directions . . . . . . . . . . . 43 Brian R. Dineen and Scott M. Soltis Chapter 3. Career Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Yehuda Baruch and Nikos Bozionelos Part II. Specific Selection Strategies and Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Chapter 4. Individual Differences: Their Measurement and Validity . . . . . . . . . . . . . . 117 Oleksandr S. Chernyshenko, Stephen Stark, and Fritz Drasgow Chapter 5. Personality and Its Assessment in Organizations: Theoretical and Empirical Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Frederick L. Oswald and Leaetta M. Hough Chapter 6. Interviews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Allen I. Huffcutt and Satoris S. Culbertson Chapter 7. Assessment Centers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 Winfred Arthur Jr. and Eric Anthony Day Chapter 8. Situational Judgment Tests: A Critical Review and Agenda for the Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 Robert E. Ployhart and William I. MacKenzie Jr. Part III. Evaluating Individuals and Performance. . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Chapter 9. The Appraisal and Management of Performance at Work . . . . . . . . . . . . . 255 Angelo S. DeNisi and Shirley Sonesh Chapter 10. Expanding the Criterion Domain to Include Organizational Citizenship Behavior (OCB): Implications for Employee Selection . . . . . 281 Dennis W. Organ, Philip M. Podsakoff, and Nathan P. Podsakoff Chapter 11. Organizational Exit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325 Peter W. Hom v

Contents

Part IV. Evaluating Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

Copyright American Psychological Association. Not for further distribution.

Chapter 12. Applicant Reactions to Organizations and Selection Systems . . . . . . . . . . 379 Donald M. Truxillo and Talya N. Bauer Chapter 13. Validation Support for Selection Procedures . . . . . . . . . . . . . . . . . . . . . . 399 Neal Schmitt and Ruchi Sinha Chapter 14. Utility of Selection Systems: Supply-Chain Analysis Applied to Staffing Decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421 Wayne F. Cascio and John W. Boudreau Chapter 15. The Unique Origins of Advancements in Selection and Personnel Psychology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 James L. Outtz Part V. Developing Members . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467 Chapter 16. Training and Employee Development for Improved Performance . . . . . . 469 Kenneth G. Brown and Traci Sitzmann Chapter 17. Mentoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505 Lillian T. Eby Chapter 18. Executive Coaching: A Critical Review and Recommendations for Advancing the Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527 David B. Peterson Chapter 19. Proactive Work Behavior: Forward-Thinking and Change-Oriented Action in Organizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567 Uta K. Bindl and Sharon K. Parker

vi

Copyright American Psychological Association. Not for further distribution.

Editorial Board

EDITOR-IN-CHIEF Sheldon Zedeck, PhD, Vice Provost for Academic Affairs and Faculty Welfare and Professor of Psychology, University of California, Berkeley ASSOCIATE EDITORS Herman Aguinis, PhD, Dean’s Research Professor, and Professor of Organizational Behavior and Human Resources, Kelley School of Business, Indiana University, Bloomington Wayne F. Cascio, PhD, Robert H. Reynolds Chair in Global Leadership, University of Colorado, Denver Michele J. Gelfand, PhD, Professor of Organizational Psychology, University of Maryland, College Park Kwok Leung, PhD, Chair Professor, Department of Management, City University of Hong Kong, Kowloon, Hong Kong Sharon K. Parker, PhD, Director of the Institute of Work Psychology and Professor of Organizational Psychology, University of Sheffield, Sheffield, England Jing Zhou, PhD, Houston Endowment Professor of Organizational Behavior, Jesse H. Jones Graduate School of Management, Rice University, Houston, TX

vii

Copyright American Psychological Association. Not for further distribution.

Copyright American Psychological Association. Not for further distribution.

PART I

FOUNDATIONS OF SELECTION AND DEVELOPMENT

Copyright American Psychological Association. Not for further distribution.

CHAPTER 1

WORK ANALYSIS: FROM TECHNIQUE TO THEORY

Copyright American Psychological Association. Not for further distribution.

Frederick P. Morgeson and Erich C. Dierdorff

Work analysis is ubiquitous in organizational settings. As is often noted, work analysis serves as the foundation for virtually every human resource (HR) activity, including job description, classification, and evaluation; selection system development; job and team design; performance management programs; training program development; compensation program development; career management systems; workforce planning; and legal compliance (Brannick, Levine, & Morgeson, 2007). In short, work analysis is an essential HR tool. Given this plethora of uses, it is likely that work analysis data are the most widely collected type of HR data in both large and small organizations. Traditionally, the analysis of work has been viewed as a process of collecting information about jobs (McCormick, 1979). As a consequence, research has tended to focus on a variety of technical and procedural issues, such as what, how, when, and from whom to collect data. More recently, however, scholars have begun exploring a range of theoretically driven issues associated with the collection of work-related information (Dierdorff & Rubin, 2007; Morgeson & Campion, 1997, 2000; Sanchez & Levine, 2000). One outcome of this expanded focus has been the suggestion that the term job analysis be replaced with the broader term work analysis (Sanchez, 1994; Sanchez & Levine, 1999, 2001). Given the recent focus on the broader world of work, coupled with our desire to move from a focus on job analysis techniques to a focus on work analysis

theory, we use the term work analysis throughout this chapter. This encompasses traditional job analysis topics as well as more recent innovations in work analysis. We seek to achieve two primary goals in this chapter. First, we offer some historical background and review of past work analysis research. This provides a sense of what research has been conducted in this area. However, we want to move beyond a simple summarization of past research. Thus, our second goal is to draw from the considerable body of work analysis research to discuss recent innovations and map out a strategy for moving work analysis research forward. Quite frankly, we want to shake things up a bit and try to stimulate some new thinking in the work analysis domain. We feel not only that work analysis is foundational to any understanding of individual and organizational performance, but also that there are still many important and interesting research questions to be answered. Thus, our goal in this chapter is to be a little provocative and approach work analysis in a slightly different way than it has been approached in the past, all in the hopes of moving this area of research forward. To do this, we first provide an extended definition of work analysis. Our goal is to define work analysis in such a way as to not only incorporate past conceptualizations but also create a more flexible and inclusive definition that helps us advance future research. Second, we briefly review the history of

We thank Mike Brannick, Wally Borman, Ed Levine, Paul Sackett, Juan Sanchez, and Olga Smit-Voskuijl for their comments on an earlier version of this chapter. We really tried to incorporate your great ideas.

http://dx.doi.org/10.1037/12170-001 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

3

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

work analysis. Such a review enables us to understand the evolution of work analysis by identifying where we have been and what is still left to be done. Third, we discuss the range of practical choices that need to be made when analyzing work. A number of considerations go into making these choices, and we discuss the pros and cons of these choices. Fourth, we then discuss the Occupational Information Network (O*NET), which is the biggest innovation in work analysis in recent years. Fifth, we discuss a recent stream of research that has sought to explore how different factors can influence the quality of work analysis information. Sixth, we then highlight a range of potential future research directions for work analysis. Finally, we conclude with a discussion of how we can take a more theoretical view of work analysis as research proceeds into the 21st century. (See also Vol. 1, chap. 13, this handbook.) DEFINING WORK ANALYSIS Work analysis can be defined as the systematic investigation of (a) work role requirements and (b) the broader context within which work roles are enacted. Because this definition differs somewhat from past definitions, further explanation is warranted. We use the term work role requirement as a short-hand way of describing both work and worker requirements. Work requirements would include such things as the tasks performed and the general responsibilities (or work activities) of those performing the work. Worker requirements would include the different types of knowledge, skill, ability, and other characteristics that are needed to perform the work (see also Dierdorff & Morgeson, 2007). Such a distinction between work and worker requirements is consistent with the “two worlds of human behavioral taxonomies” identified by Dunnette (1976) and the “activity” and “attribute” distinction more recently articulated by Sackett and Laczo (2003). In addition, we have deliberately chosen to focus on roles rather than the traditional focus on jobs for five reasons. First, as an expected pattern or set of behaviors interrelated with the behaviors of others (Biddle, 1979; Katz & Kahn, 1978; Stewart, Fulmer, & Barrick, 2005), a role subsumes the traditional 4

work requirements of both tasks and responsibilities and thus helps integrate across work requirements. This offers a more flexible language with which to describe and discuss work. Second, a focus on roles enables the explicit acknowledgment of connections to and among other role holders, as well as the embeddedness of roles in the broader work context. Although often touched on in traditional definitions, this has tended to be neglected in practice. Third, one of the traditional criticisms of work analysis is that it tends to view jobs as static entities (Guion, 1993). By focusing on roles, we move away from a more static conceptualization of jobs to a more flexible roles orientation. Thus, work analysis could consider not only prescribed or established task elements, but also discretionary or emergent task elements (Ilgen & Hollenbeck, 1991; Morgeson & Humphrey, 2008). Fourth, focusing on jobs tends to place an emphasis on work activities, leading some to conceptualize work analysis in a narrow fashion (Harvey, 1991). However, it is clear that work analysis includes the study of both work activities and worker attributes (Sackett & Laczo, 2003; Sanchez & Levine, 2001). Considering roles and role enactment leads more naturally to a consideration of worker attributes. Fifth, focusing on jobs places an emphasis on individual job incumbents. Although this is often justified given the uses of work analysis data, it tends to ignore the fact that jobs are situated in a larger team and organizational context. One problem with focusing primarily on individual jobs is that there is an insufficient link to an organization’s business goals and strategies (Schippmann et al., 2000), prompting many to pursue a quasi-work analytic approach like competency modeling. The role concept, in contrast, is implicitly multilevel. For example, a role can be described in terms of individual role holder work activities, the combination of roles that exist within a team that produces interdependent collective action, and the structure of organizations as a system of roles (Katz & Kahn, 1978). Thus, in conducting a work analysis, a focus on roles could alert the analyst to consider how individual roles connect to the broader system of roles within the organization and the implications of these connections for the specific role under consideration.

Work Analysis

Collecting work-related information has long been important to large-scale human endeavors. For example, Mitchell, Bennett, and Strickland (1999) pointed out that the first effort to document information about work could be seen over 3,000 years ago in the Imperial Court of China (circa 1115 B.C.). During the more recent times of the past century, Münsterberg (1913) pioneered systematic methods for estimating job requirements for personnel selection purposes and job design. The first history of work analysis was compiled by Uhrbrock (1922), in which he emphasized using job analysis for setting performance standards and introduced the need to identify personal attributes associated with successful job performance (Wilson, 2006). Perhaps in a bit of historical irony, Frederick Taylor actually used the term work analysis in the early 1900s (Cunningham, 2000), despite our modern day depiction of scientific management as having an exclusive emphasis on reductionism to the most molecular of behavioral elements! Even with these rich historical linkages, what we have come to currently recognize as the field of work analysis has its firmest roots in research conducted after the 1940s. Because there have been several excellent reviews of this period (e.g., Mitchell, 1988; Mitchell & Driskill, 1996; Primoff & Fine,

1988), we do not discuss these historical developments in detail. Instead, we focus on trends in work analysis research over the last 50 years. Although published work analysis research certainly predates 1960, we felt that a nearly half-century snapshot would be sufficient for depicting any important trends. We searched PsycINFO for work analysis research published since 1960 using keywords such as job analysis, work analysis, job specification, and so forth. We restricted our search to only research published in peer-reviewed journals, thus excluding dissertations, technical reports, and books. Finally, an article’s primary focus had to be work analysis to be included. Thus, articles that simply presented the results of work analysis (e.g., job description of a nursing occupation) and tangentially related articles not specifically focused on work analysis (e.g., job redesign, synthetic validity) were excluded. Figure 1.1 displays the frequency of work analysis publications across the 48-year time period. In total, the search produced 193 work analysis journal articles that have been published in peer-reviewed journals. When examined by each decade, close to one third (30%) of the articles were published during the 1980s alone. Approximately 7% of the publications were during the 1960s, and 17% were during the 1970s. The publication percentages for articles in the 1990s and 2000–2008 were nearly equivalent

13 12 11 10 9

Publications

8 7 6 5 4 3 2 1 0

19 60 19 62 19 64 19 66 19 68 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08

Copyright American Psychological Association. Not for further distribution.

HISTORY OF WORK ANALYSIS

Year

FIGURE 1.1. Frequency of work analysis publications from 1960–2008. 5

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

(about 23%). Thus, with respect to pure volume, it appears that almost as many work analysis articles have been published before and after 1990 (54% compared with 46%). The pattern of publications after this date also appears to display greater variability, as represented by the larger peaks and valleys in the figure. Additionally, around this time, it was widely discussed among work analysis scholars that work analysis research was not garnering much respect (i.e., being published) in industrial–organizational (I/O) and management journals (Cunningham, 1989). However, a more nuanced examination of previous work analysis articles reveals trends that may shed light on this historical concern. The data in Figure 1.1 clearly show that work analysis research is alive and well and is being published in peer-reviewed journals, albeit with slightly more variability in volume in recent years. Note, however, that this conclusion is in an absolute sense (i.e., exclusively focusing on work analysis research). Cascio and Aguinis (2008) recently found in their content analysis of Journal of Applied Psychology and Personnel Psychology that research within the work analysis domain has waned relative to other research domains within I/O psychology. With this in mind, we examined what journals have published work analysis articles and how the publishing outlets may have changed over time. To accomplish this, we categorized the collected work analysis articles into two broad groupings: (a) those published in one of the “top seven” journals (as identified by Podsakoff, MacKenzie, Bachrach, & Podsakoff, 2005) and (b) those published in any other journal. The results of this analysis are displayed in Figure 1.2.1 These findings show a striking trend toward proportionally fewer work analysis articles being published in the top seven journals across the 48-year time period. For example, from 1960 to 1979, approximately 77% of all work analysis articles were published in one of the top seven journals. Although this proportion decreased to 58% during the 1980s, the overall 1

number of work analysis articles in top seven journals still increased from the previous decade. The most noticeable decrease began in the 1990s, where only about 28% of work analysis articles were published in top seven journals, and this downward trend continues today (e.g., 27% since 2000). Collectively, these results suggest that work analysis research is increasingly absent from the most influential journals. Such a decrease is unfortunate, in part because of the influence the top journals have on shaping the field. For example, one might wonder whether the substantial volume of research concerning the Position Analysis Questionnaire (PAQ; McCormick, Jeanneret, & Mecham, 1972) would have been conducted (and subsequently published) if the original research had not appeared in a monograph within Journal of Applied Psychology, one of the top applied psychology journals. Or, as another example, whether there would have been such widespread acceptance and ensuing use of the critical-incident technique had it not been published in Psychological Bulletin (Flanagan, 1954), a top psychology journal. To get a better sense of publication trends over time, we qualitatively reviewed the work analysis articles to see if we could further discern any patterns in the type of research being published over the last 50 years. This examination produced 10 broad categories shown in Table 1.1. This table also provides the percentages of articles falling into each category. These data show that, with the exception of research examining rater training and rating scales, work analysis research has a relatively even distribution across the topical groupings (ranging from 8% to 15%). However, percentage differences for some categories were apparent with respect to the nearly 5 decades that the research spans. For example, all of the research focusing on specific work analysis instruments was published prior to 1990, as well as the majority of research (75%) regarding various analytic techniques (e.g., factor analysis). The majority of

Podsakoff et al. (2005) divided management-related journals, which includes top I/O psychology journals, into quartiles on the basis of the journal’s impact (as assessed by citations per article). The top quartile consisted of the Academy of Management Journal, Academy of Management Review, Administrative Science Quarterly, Journal of Applied Psychology, Organizational Behavior and Human Decision Processes, Personnel Psychology, and Strategic Management Journal. These “top seven” journals accounted for almost 61% of all citations between 1981 and 1999. Moreover, the top seven journals “averaged almost six times more citations per paper (23.93 vs. 4.54) from 1981 to 1999 than the seven bottom journals” (Podsakoff et al., 2005, p. 481). Although some of these journals do not necessarily publish work analysis articles, many of them do. These journals are, however, highly influential and thus represent a good way to examine the prominence of work analysis research in the field of psychology.

6

Work Analysis

60

Other 50

Publications

Top Tier 40

30

20

Copyright American Psychological Association. Not for further distribution.

10

0 1960-69

1970-79

1980-89

1990-99

2000-08

FIGURE 1.2. Work analysis publications in the top seven journals (top tier) and all other journals.

research on job classification and clustering occurred during the 1970s and 1980s (70%). In comparison, articles published on three of the topics (development of instruments, procedures, or taxonomies; uses for work analysis information or results; and general or topical reviews) were rather evenly spread throughout the 5 decades. Finally, research in the area of reliability and validity and in the area of factors influencing ratings was primarily conducted since 1990 (74% and 57%, respectively). This qualitative investigation yields two key insights. First, there are notable omissions in past

TABLE 1.1 Work Analysis Topics Studied Category Development of instruments, procedures, or taxonomies Reliability and validity Instrument-specific research Uses for job analysis information–results General or topical review Job classification and clustering Rater training Factors influencing ratings Rating scales Other analytic techniques

Percent 13.47 13.99 10.36 13.47 12.95 8.81 2.59 14.51 1.04 8.29

work analysis research. For example, Sackett and Laczo (2003) previously described several important changes in work analysis practice that had taken place by the time of their review of the field. These changes included personality-oriented work analysis, competency modeling, cognitive task analysis, strategic job analysis, and issues of accuracy in work analysis. However, published research on most of these changes remains largely absent. That is, the empirical work analysis literature offers little evidence regarding a host of questions surrounding the ramifications of these changes (e.g., issues of utility, reliability, validity, legality, acceptance). Furthermore, Sackett and Laczo (2003) noted this same empirical paucity over 5 years ago. For example, from our literature search, since 2003, only a single published article examined the use of strategic job analysis (e.g., Siddique, 2004), and this was merely an indirect examination. The same situation was found for personality-oriented work analysis as well (e.g., Cucina, Vasilopoulos, & Sehgal, 2005). Two exceptions to this scarcity trend are in the areas of competency modeling and issues of accuracy. Since 2003, at least four articles have included examinations related to competency modeling (e.g., Goffin & Woycheshin, 2006; Lievens & Sanchez, 2007; Lievens, Sanchez, & De Corte, 2004; Morgeson, Delaney-Klinger, Mayfield, Ferrara, & 7

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

Campion, 2004) and five articles have focused on factors related to accuracy (e.g., Dierdorff & Rubin, 2007; Dierdorff & Wilson, 2003; K. Prien, Prien, & Wooten, 2003; Van Iddekinge, Putka, Raymark, & Eidson, 2005; Wang, 2003). Thus, it appears that work analysis research needs to begin to focus research attention on some of the techniques and changes that have occurred in work analysis practice. A second implication of our analysis is that the topical focus of work analysis research has not changed all that much over the past 50 years. This is especially true for work analysis research concerned with more technical questions, such as developing new procedures or taxonomies, using work analysis data for different purposes (e.g., content-related validity for test creation), and so forth. Thus, it appears that considerable work analysis research continues to focus on technical issues rather than theoretical issues. Perhaps this can explain the relative decrease in work analysis research in the top journals. As the field of I/O psychology has matured, empirical research is expected to make stronger theoretical contributions. To the extent that work analysis research is unable to contribute theoretically, it will likely be shut out from the top journals in the field. However, there does appear to be some hope, as there has been a recent increase in the amount of work analysis research going beyond these traditional areas. One common thread among this research is the focus on a theory-driven understanding of the various nonjob factors that influence work analysis judgments. For example, this research has been conceptually driven using cognitive (schema) theory (e.g., Lievens & Sanchez, 2007), role theory (e.g., Dierdorff & Morgeson, 2007; Dierdorff & Rubin, 2007), and impression management (selfpresentation) theory (e.g., Morgeson et al., 2004). Important to work analysis research, this recent trend may indicate that a reinvigoration of the topics examined in work analysis, as well as a grounding of such research in relevant psychological theory, is both fruitful and necessary. We return to these points in greater detail within ensuing sections of this chapter.

8

WORK ANALYSIS CHOICES Conducting a work analysis involves making numerous choices. These choices reflect the different ways a work analysis can be conducted in practice. We first discuss the range of choices that can be made, including the choice of descriptor, the methods to use, the rating scales to use (if using the questionnaire method), and the sources of work analysis data. Because these choices are driven by the purpose of the work analysis, we then discuss the intersection of the purposes and work analysis choices.

Descriptor Type Broadly speaking, descriptors are simply the various features of work examined during a work analysis (Brannick et al., 2007). There are three major types of descriptors that can be used in work analysis. The first concerns the requirements of the work itself and involves the activities performed by workers (Sackett & Laczo, 2003). The two most commonly discussed work requirements are the specific tasks performed and more general work responsibilities. Tasks are collections of specific work elements and include actions, the object of the action, and the purpose or results of the action (Fine & Getkate, 1995) as individuals fulfill their work roles. Of importance, tasks are typically specific to a particular work role. For example, the tasks for an industrial machinery mechanic would include such things as disassembling machinery and equipment to remove parts and make repairs and repairing and replacing broken or malfunctioning components of machinery and equipment. Responsibilities are collections of related tasks that represent a set of generic behaviors applicable across a wide variety of work roles (Cunningham, 1996). As such, responsibilities are broad activity or behavior statements that are aggregates of several highly related behaviors used in accomplishing major work goals (Jeanneret, Borman, Kubisiak, & Hanson, 1999). Continuing our example, responsibilities for an industrial machinery mechanic would include repairing and maintaining equipment and inspecting equipment, structures, or material. The second major type of descriptor concerns worker requirements and involves a consideration of

Copyright American Psychological Association. Not for further distribution.

Work Analysis

the worker characteristics needed to successfully perform the work (Sackett & Laczo, 2003). Four commonly discussed worker requirements include knowledge, skill, ability, and other characteristics. Knowledge can be defined as collections of discrete but related facts and information about a given domain, such as biology, mathematics, or medicine (Costanza, Fleishman, & Marshall-Mies, 1999). A further distinction is often made between declarative (knowledge of what) and procedural (knowledge of how) knowledge (Campbell, McCloy, Oppler, & Sager, 1993). Skills reflect the level of proficiency or competency to perform a task or learned activity (Peterson et al., 2001). Skills can be divided into basic and crossfunctional categories. Basic skills are thought to facilitate learning or knowledge acquisition and include such things as writing or critical thinking skills. For their part, cross-functional skills are developed capabilities that foster performance across job contexts and include such things as problem solving and negotiation skills. Skills are commonly thought to improve with training and experience on a particular task. Abilities are relatively enduring basic capacities for performing a range of different activities (Fleishman, Costanza, & Marshall-Mies, 1999). This would include cognitive (e.g., verbal, quantitative), psychomotor (e.g., reaction time, manual dexterity), physical (e.g., strength, endurance), and sensory– perceptual (e.g., visual, auditory) abilities. Relative to knowledge and skill, abilities are thought to be more stable over time. Other characteristics is a catch-all category designed to encompass all other potentially relevant factors that might be important for successful performance. Other characteristics that are commonly discussed include personality and motivational traits (e.g., conscientiousness, leadership, initiative), specific forms of work and educational experience, and licensure and certification that may be required in certain fields (e.g., registered nurses, certified public accountant). The third major type of descriptor concerns the work context within which work is performed (and roles are enacted). Work context can be broadly

defined as “situational opportunities and constraints that affect the occurrence and meaning of organizational behavior as well as functional relationships between variables” (Johns, 2006, p. 386) and consists of task, social, and physical aspects (Hattrup & Jackson, 1996; Johns, 2006; Strong, Jeanneret, McPhail, Blakley, & D’Egidio, 1999). The task context reflects the structural and informational conditions under which work roles are enacted and includes such things as the amount of autonomy and task clarity, the consequence of error inherent in the work, level of accountability, and the resources available to perform the task. The social context reflects the nature of role relationships and interpersonal contingencies that exist among workers and includes such things as social density, different forms of communication, the extent and type of interdependence with others, and the degree of interpersonal conflict present in the work environment. The physical context reflects elements of the material space or built environment within which work roles are enacted and includes general environmental conditions (e.g., noise, lighting, temperature, air quality), presence of hazardous work conditions (e.g., radiation, high places, disease exposure), and overall physiological job demands (e.g., sitting, standing, walking, climbing). Although the nature of the context is not often explicitly taken into account when conducting work analysis, recent research has shown it can have a pronounced effect on work role requirements (Dierdorff & Morgeson, 2007).

Method Once a decision is made on the type of descriptor(s), the next choice involves the method to use to collect data on those descriptors. There are many different methods to use (see Ash, 1988; Brannick et al., 2007, for comprehensive lists), but some of the most common include observation, individual interviews, group meetings, and questionnaires. Note that there is very little research that compares the relative effectiveness of these different work analysis methods (see Ash & Levine, 1980, for a framework for evaluating work analysis methods). A general rule of thumb, however, would be to use multiple methods that could permit subsequent triangulation of collected

9

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

information, as well as the opportunity to capture different perspectives of the target work roles under examination. Of course, time requirements, cost effectiveness, and, most importantly, the intended purpose of the information (discussed later in this section) must be considered when choosing work analysis methods. Observation can take several forms, but the most often used method involves direct observation, whereby someone not directly involved in the task performance (e.g., a supervisor, job analyst) observes workers as they complete their tasks. Generally, an observer would record (via notes, checklists, or questionnaires) the what, why, and how of various aspects of the work. Other forms of observation include having supervisors record or recall particularly effective or ineffective worker behaviors (i.e., critical incidents) or video recording worker task performance for later analysis. Although time consuming, an advantage of observation is that it is not subject to problems of selective recall or other reporting biases on the part of workers (but there is potential bias in terms of what is recalled). However, for some jobs it may not be possible to observe key aspects of the job, particularly for work that has a large mental or knowledge component (i.e., most work processes occurs in the head of the worker). Individual interviews involve conducting interviews with respondents one at a time. Typically, interviews are conducted with multiple different types of respondents (e.g., workers, supervisors) who are asked similar types of questions about the work. Interviews enable the acquisition of detailed information, in part because the interviewer can prompt the interviewee for additional details and check or otherwise question the validity of the information being transmitted. A major challenge of interviews is that some individuals might not be able to describe what they do or what the work requires in sufficient detail. This is particularly likely to occur if an individual has been working in the role for an extended period of time and has routinized the performance of major tasks. Another potential limitation is interviewer bias in terms of faulty recording or recall of the content of the interview itself. 10

Group meetings (also called “subject matter expert” [SME] meetings) involve getting a number of workers, supervisors, or technical experts together to discuss various aspects of the work. Typically, one would conduct separate meetings for workers, supervisors, and technical experts, in part because one would likely focus on different aspects of the work with the different groups. Such meetings are usually facilitated by a job analyst and are a more efficient way to collect information than the individual interview. Common activities in group meetings include brainstorming or generating lists of activities or attributes or evaluating data that have been previously gathered. An advantage of group meetings is the possibility of consensus, which is often needed for implementation of a work analysis product. However, group meetings can be subject to numerous dysfunctional group processes, including a lack of participation by some group members and conformity to a dominant group member. Such social processes are discussed in greater detail later. Questionnaires are structured surveys (either paper and pencil or computer based) used to collect information on any of the work role requirements discussed previously. There has been a tremendous amount of research on the questionnaire method (somewhat in contrast to the other methods). The bulk of this research has focused on the presentation and evaluation of particular work analysis methods or questionnaires (although recognize that the use of custom, organization-specific work analysis questionnaires is widespread). Some examples of the questionnaire approach include the task inventory approach (e.g., Gael, 1983), the PAQ (McCormick et al., 1972), and the O*NET (Peterson et al., 2001). The evaluation of each of these questionnaire methodologies is beyond the scope of this chapter, but one advantage of this method is its ability to systematically gather a large amount of work-related information that can be quantitatively summarized. These strengths, however, should be balanced against some potential weaknesses. Questionnaire respondents can be overwhelmed by the task (some questionnaires can be several hundred questions long and involve numerous rating scales) and subsequently provide responses that are unreliable and inaccurate.

Work Analysis

Copyright American Psychological Association. Not for further distribution.

Type of Rating Scale As noted, the questionnaire method has received a great deal of attention. One of the key decisions to be made when using the questionnaire methodology is which kind of rating scale to use. In this section, we discuss some of the rating scales that have been used in the past. Although many of these scales have been used to collect task-related information, they can also be used to collect other work and worker requirement data. Table 1.2 provides examples of some of the most commonly used scales. It is often helpful to obtain estimates about how often particular tasks are performed. To do so, researchers have used different types of frequency scales. At least two different options are possible when measuring frequency. In the first, a frequency estimate is made using highly specific time-based estimates (e.g., from “about once per year” to “about

once each hour or more often”). In the second, a less specific estimate is provided (e.g., from “never” to “very often”). We are not aware of any research that has directly compared these two different types of frequency scales, but we have used both in our research with good results. We have found that sometimes respondents have difficulty making the highly specific frequency estimates. In some ways, it is almost too precise, given the way workers often view their job. We have more to say about the complexity of the judgments often made in work analysis a little bit later. One frequency scale that seems to have fallen somewhat out of favor is the relative time-spent scale. This could be due, in part, to the criticisms leveled against this kind of scaling by Harvey (1991), who suggested that such a “within-job relative” rating scale (e.g., the time spent on a particular task

TABLE 1.2 Commonly Used Job Analysis Rating Scales Type of rating scale Frequency “I perform this task . . .” (Gael, 1983)

“I perform this task . . .” (Drauden, 1988) Importance “How important is this task to the performance of your present job?” Criticality–consequence of error “Indicate the degree to which an incorrect performance would result in negative consequences.” (Brannick, Levine, & Morgeson, 2007)

Task difficulty “Indicate the difficulty in doing a task correctly relative to all other tasks within a single job.” (Brannick et al., 2007)

Required on entry “Review each task statement and ask yourself the following question: ‘When is a new employee expected to be able to possess this knowledge or skill?’ ”

Anchors 1 = about once per year, 2 = about once every six months or less, 3 = about once each month, 4 = about once each week, 5 = about every other day, 6 = about every day or more often (not each hour), and 7 = about once each hour or more often 0 = never, 1 = rarely, 2 = occasionally, 3 = sometimes, 4 = often, and 5 = very often 1 = not important, 2 = somewhat important, 3 = important, 4 = very important, and 5 = extremely important 1 = consequences of error are not at all important, 2 = consequences of error are of little importance, 3 = consequences are of some importance, 4 = consequences are moderately important, 5 = consequences are important, 6 = consequences are very important, and 7 = consequences are extremely important 1 = one of the easiest of all tasks, 2 = considerably easier than most tasks, 3 = easier than most tasks performed, 4 = approximately 1⁄2 tasks are more difficult, 1⁄2 less, and 5 = harder than most tasks performed 1 = not expected to possess immediately, but after formal training is provided, 2 = not expected to possess immediately, but can be quickly learned on the job, and 3 = should be able to possess immediately

11

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

compared with the time spent on other tasks) makes cross-job comparisons problematic. It is interesting to note that research has found that relative time spent and both absolute and relative frequency scales provide largely the same information (Friedman, 1990, 1991; Manson, Levine, & Brannick, 2000). Although this does not speak to making cross-job comparisons, it does suggest that within a job, any of these frequency scales are likely equivalent. In addition to frequency, it is often useful to obtain estimates about the importance of particular tasks to the overall work role. At least two different strategies have been used in past research. In the first, judgments of criticality or consequences of error (the extent to which the incorrect performance of a task would result in negative consequences) and difficulty (how hard it is to perform a task correctly) are combined into an overall index of importance (Sanchez & Levine, 1989). In the second, the importance of a task is directly estimated by simply asking how important the task is to performance on the job. Research has shown that direct estimates are as reliable as composites of difficulty and criticality (Sanchez & Fraser, 1992). Although the preceding rating scales have typically been used in the context of task questionnaires (and more generally in activity-based work analysis), attribute-oriented work analysis efforts have also used the questionnaire method. Of the rating scales described above, only minor modifications would be needed to adapt them for use with attribute descriptors. For example, instead of referencing the frequency of task performance, the rating scale could reference the frequency with which knowledge or skills are needed on the job. A similar adjustment can be made for importance. In fact, importance rating scales have been used in both the PAQ and O*NET. There are, however, some rating scales that take on particular relevance in attribute questionnaires. For example, a key question when conducting work analysis for the purposes of developing a selection system is the extent to which a particular attribute is needed at the point of entry (hiring) or whether it can be learned (trained) on the job. This can provide input into which attributes to focus on during selection assessments (but also note the same question could be asked about when a worker is expected to 12

be able to perform tasks) and which to include in formal training programs. The final rating scale we discuss is not included in Table 1.2 but is particularly salient to attributeoriented questionnaires. This is the level of the attribute that is required by the job. Originally developed for use in Fleishman’s Ability Requirements Scales (Fleishman, 1992), its use has been extended to multiple domains in O*NET. The basic idea is that any work role has a particular amount or level of ability or skill needed for effective performance. In practice, level rating scales range from low to high but typically use behavioral anchors that are illustrative of different levels of the attribute. For example, the ability of “reaction time” (defined as the ability to quickly respond [with the hand, finger, or foot] to a signal [sound, light, picture] when it appears) could have anchors for low, moderate, and high levels of ability as follows: “start to slow down the car when a traffic light turns yellow,” “throw a switch when a red warning light goes off,” and “hit the brake when a pedestrian steps in front of the car,” respectively. Despite the distinctions that are made among these different rating scales, there is evidence that many of these distinctions are often lost on the workers who complete work analysis questionnaires. For example, although level and importance rating scales are quite different conceptually, in the initial pilot test of O*NET (Peterson, Mumford, Borman, Jeanneret, & Fleishman, 1999), level and importance scales were often highly correlated (in the low .90s). The rating scales of importance and criticality have also shown high overlap (rs > .80), whereas correlations between difficulty to learn ratings and importance and criticality ratings have ranged from moderate to high (rs from .37 to .77; Manson et al., 2000; Sanchez & Fraser, 1992; Sanchez & Levine, 1989). Finally, a meta-analysis by Dierdorff and Wilson (2003) showed variability in interrater reliability estimates of ratings using importance (r = .71), frequency (r = .69), difficulty (r = .63), and timespent (r = .67) scales, but the 80% confidence intervals for these estimates were overlapping, indicating a lack of significant differences. In total, this evidence suggests that despite their conceptual independence, respondents who complete work analysis surveys

Work Analysis

are not always able to make the same sort of fine distinctions that are prompted by different rating scales. This suggests that if multiple response scales are to be used, then they should be chosen so as to minimize redundancy and ensure alignment with the intended purposes of the work analysis.

Copyright American Psychological Association. Not for further distribution.

Source Once a method is determined, the next choice involves deciding the source of the work analysis information. Common sources include written documentation, role incumbents, technical experts, supervisors, clients, and job analysts. A wide variety of written documentation can be used to support a work analysis effort. This would include such things as existing job descriptions, previous work analyses, published information about the work role (e.g., from publicly available databases, such as O*NET), training manuals or other documents used to prepare workers for the role, and checklists or operating guides for any of the equipment, tools, or other work aids. Collecting this kind of documentation is typically the first step in the work analysis, as one seeks to compile all the known information about the work role. One benefit of this source of information is that its collection can be very cost efficient. However, the work analysis practitioner must be aware that existing documentation could be outdated or may lack sufficient depth or breadth to be useful for the intended purpose of the work analysis. Role incumbents are another useful source of work analysis information. Incumbents are a useful source of information because of their familiarity with the role and specific knowledge about what is done on a day to day basis. However, some incumbents may not be able to effectively articulate exactly what they do, either because of a lack of verbal ability or a lack of motivation to provide accurate and reliable information. Technical experts are individuals who do not perform the role but have some sort of specialized expertise with the work that is performed. Examples might include engineers who design a manufacturing process, chemists who study the effects of drug interactions, lawyers who write and approve contracts, or professors who are experts in the discipline that underlies the work being studied. Such experts are likely to provide an important

perspective on the technical aspects of the work, particularly in terms of ideal system functioning. Supervisors (either the immediate supervisor or a higher level manager) can also provide a useful perspective on the work role requirements. Supervisors may have a higher level of verbal ability than incumbents and thus might be able to provide work role information that incumbents are unable to articulate effectively. In addition, supervisors are probably less motivated to distort or otherwise bias the information they provide. Finally, given their hierarchical position, they are likely to have a broader perspective with respect to differences among the work roles and the attributes needed for successful role performance. Despite these positive features, however, one major problem with supervisors as a source is that they may have less detailed and nuanced information about the work role because they do not actually perform the work (and may not even know how to perform the work). Work analysts are another source of information. These can be either HR professionals inside the organization who have expertise and training in work analysis methods or outside consultants or experts. In a typical work analysis, work analysts serve an integrative role by designing and implementing the variety of methods discussed earlier. Some advantages of work analysts are that they tend to produce highly reliable ratings, have no (or little) motivation to bias the results, and are able to integrate the large amounts of information that typically result from a thorough work analysis. However, unless they accumulate enough information about the work, work analysts may lack adequate information to make good decisions. Finally, because experienced work analysts often have prior exposure to similar work roles, they may be subject to preexisting stereotypes about the work. Unless they are careful, their decisions may be influenced by their stereotypes instead of the actual work itself. Although one could choose to use only one (or a subset) of these sources when conducting a work analysis, in practice, a comprehensive work analysis would entail using all sources to varying degrees. In addition to capturing different perspectives of the work role under examination, using multiple sources may have the added benefit of producing higher 13

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

quality work analysis information, as some research has shown differences across different data sources (e.g., Dierdorff & Wilson, 2003). A process commonly used in practice is to begin by reviewing existing written documentation. This documentation then informs subsequent data collection from role incumbents and technical experts. Supervisors then check and augment the data collected from incumbents and experts. Analysts then compile all the information (and likely were intimately involved in collecting the data) and draw relevant conclusions. Such an approach is often highly effective because it provides a more accurate description of the work and worker requirements. In addition, by gathering input from the relevant stakeholders, it can enhance acceptance of any HR system that is built from the findings of the work analysis.

Purpose of Work Analysis As noted at the beginning of this section (and implicitly throughout), the choices made when designing and conducting a work analysis depend on the reason or purpose of the work analysis. There are numerous reasons why one might conduct a work analysis, including selection system development, job and team design, performance management system design, training system development, compensation system development, and career management systems. Because a comprehensive review of these purposes is beyond the scope of this chapter (see Ash, 1988; Brannick et al., 2007; McCormick, 1979, for complete lists), we have chosen to focus on what goes into making such choices and providing some selective examples. Perhaps the most important consideration when making work analysis choices is how the information will be used. For example, conducting a work analysis to determine what kinds of selection tools to use would place a priority on identifying the attributes (e.g., knowledge, skill, ability, other characteristics) needed to effectively perform the work and the extent to which certain attributes (e.g., skills) are needed immediately on the job and others can be learned once on the job. Conducting a work analysis for developing a new training program, however, would place a premium on the activities performed, in part because the activities form the core of the training 14

program content. If the intention is to carry out a work analysis to produce information for job descriptions–specifications, then emphasis would be on a full breadth of descriptors (activities, attributes, and context), with attention paid to ascertaining the importance of these descriptors to role performance. As these examples illustrate, the ultimate use of the work analysis information plays a major part in any decisions that are made. Beyond the use of the information, several other ancillary considerations deserve mention, including quality, cost, acceptability, and legal defensibility. Although one would always like to obtain as high a quality of information as possible, quality considerations often must be balanced against cost considerations. All else being equal, the highest quality work analysis information will be the most costly. Organizations often have to make pragmatic decisions about when a work analysis is good enough. We return to issues of quality in more detail in a subsequent section. Acceptability is another important consideration, particularly if the work analysis has major implications for current workers. For example, if a work analysis is being conducted to redesign jobs or determine pay levels, then choices should be made to include the interested parties wherever possible. Interested parties can include incumbents whose jobs are being redesigned or whose pay is being affected and labor unions who represent job incumbents. A final consideration would be legal defensibility. If one were conducting a work analysis in an environment where the resulting HR system might be subject to legal challenge, fully documenting a detailed and thorough (i.e., high quality) work analysis would be advised. For example, if a work analysis is being conducted to revise a performance management system where there have been allegations of gender discrimination, then a complete and thorough work analysis would need to be carefully documented. O*NET Arguably the most significant innovation in work analysis of the past several decades has been the development of O*NET by the U.S. Department of Labor. Although other occupational classification

Copyright American Psychological Association. Not for further distribution.

Work Analysis

systems exist both in North America (e.g., Canada’s National Occupation Classification) and in Europe (e.g., International Standard of Occupational Classification and EurOccupations), O*NET encompasses the broadest scope of work information ranging from labor market data and wages to important knowledge, skills, and required tasks. As such, O*NET is a comprehensive system of occupational information designed to replace the Dictionary of Occupational Titles (DOT), which was first published in 1939. There were numerous reasons why the DOT was in need of replacement (Dunnette, 1999). Most salient among these reasons were (a) the lack of information to allow cross-job comparisons, which permit classification and determination of similarities and differences across a variety of work roles; (b) the primary focus on task information to the exclusion of other important work role requirements, such as knowledge, skills, abilities, and traits; (c) the limited description of the conditions under which work is performed (e.g., the DOT mainly described aspects of the physical context); and (d) the numerous difficulties of maintaining the currency of the information in a rapidly changing world of work. A special panel was commissioned by the federal government (Advisory Panel for the Dictionary of Occupational Titles, or APDOT) to review these issues surrounding the DOT and to offer recommendations for improvement and alternative approaches. As a result, APDOT released a

final report (APDOT, 1993) that outlined a roadmap toward creating what would later become O*NET. For more details regarding how this process unfolded, readers are encouraged to consult Dunnette (1999) and Dye and Silver (1999). At the heart of O*NET is its content model, which theoretically organizes the wide variety of information that can be used to describe the world of work. The content model is shown in Figure 1.3 and comprises six major areas: worker characteristics, worker requirements, experience requirements, occupation requirements, workforce characteristics, and occupation-specific information (Mumford & Peterson, 1999; Peterson et al., 2001). Of importance, this structure enables a focus on areas that describe important attributes and characteristics of both workers and the work itself. Table 1.3 displays the types of descriptors that fall within each domain of the content model. Also shown in the table are the conceptual categories of these descriptors and the sources from which data are collected. More specific information may be found in Peterson et al. (2001) or at O*NET OnLine (see http://online.onetcenter.org). With regard to the field of work analysis, several features of the content model are especially noteworthy. First, the model represents a comprehensive way to conceptualize virtually all of the types of workrelated data that are of interest to both individuals and organizations. For example, the model subsumes

Worker-oriented Worker Characteristics

Worker Requirements

Experience Requirements

Abilities Interests Work Values Work Styles

Knowledge Basic Skills Cross-functional Skills Education

Experience & Training Entry Requirements Work Licensing

Occupation Requirements

Workforce Characteristics

Occupation-specific Data

Generalized Activities Detailed Activities Organizational Context Work Context

Labor Market Data Compensation Occupational Outlook

Tasks Tools & Technology

Work-oriented

FIGURE 1.3. Content model for the Occupational Information Network. 15

Morgeson and Dierdorff

TABLE 1.3 Content Model Descriptor Types and Categories Domain and descriptor type Occupation requirements Generalized work activities

Data source

Information input, mental processes, work output, and interacting with others 2,165 activities (e.g., administer medications or treatments, analyze psychological testing data, prepare records of customer charges) Interpersonal relationships, physical work conditions, and structural job characteristics Organizational structure, human resources systems and practices, goals, roles at work, culture, and role of supervisors

Role incumbents

Content skills and process skills Social skills, complex problem-solving skills, technical skills, systems skills, and resource management skills Business and management, manufacturing and production, engineering and technology, mathematics and science, health services, education and training, arts and humanities, law and public safety, communications, and transportation Required level of education, instructional program required, and educational level in specific subjects

Role incumbents Role incumbents, analysts

Cognitive abilities, psychomotor abilities, physical abilities, and sensory abilities Achievement orientation, social influence, interpersonal orientation, adjustment, conscientiousness, independence, and practical intelligence Realistic, investigative, artistic, social, enterprising, and conventional Achievement, working conditions, recognition, relationships, support, and independence

Analysts

25–30 tasks per occupation 25,000+ equipment, tools, machines, software, and other information technology

Role incumbents Analysts

Workforce characteristics Labor market information Occupational outlook

Wages, employment statistics, and so forth Employment projections (e.g., growth, shrinkage)

Bureau of Labor Statistics Bureau of Labor Statistics

Experience requirements Experience and training

Related work experience and on-the-job training

Office of Apprenticeship and role incumbents Not currently collected Not currently collected

Detailed work activities

Work context Organizational context Copyright American Psychological Association. Not for further distribution.

Descriptor categories or details

Worker requirements Basic skills Cross-functional skills Knowledge

Education

Worker characteristics Abilities Work styles

Occupational interests Occupational values Occupation-specific information Tasks Tools and technology

Basic skill entry requirements Cross-functional skill entry requirements Licensing

16

Analysts

Role incumbents Not currently collected

Role incumbents, analysts

Bureau of Labor Statistics and Department of Education (third category not collected)

Role incumbents

Analysts Analysts

Office of Apprenticeship

Copyright American Psychological Association. Not for further distribution.

Work Analysis

labor market data, wages, and occupational forecasts, as well as the attribute and activity requirements necessary for occupational performance. Second, the model posits a taxonomic structure for most of its domains. For instance, the worker characteristic domain of abilities is grouped into four broad categories, with more specific abilities contained within each grouping. This taxonomic approach is beneficial primarily because it directly incorporates multiple levels of data specificity. This is valuable because it allows one to choose between various levels of specificity in a particular domain, depending on the intended use of the information. Third, the model establishes a common language with which to describe the world of work. Using standardized descriptors is essential for cross-occupation comparisons that seek to identify similarities or differences between occupations. The benefits of a common language are numerous, in part because it can serve as a unifying force that eliminates the potential confusion that is created when a host of similar descriptors are used to capture work role requirements. Finally, the model also allows for occupationspecific information, such as detailed task information, wage information, and so forth. Such occupation-specific data were central to the original DOT. Of importance, the incorporation of this type of information ensures that, in addition to more molar cross-occupational comparisons, more molecular within-occupation descriptions are possible. Further, occupation-specific data are necessary for a number of HR purposes, such as developing training programs or generating position descriptions (Sager, Mumford, Baughman, & Childs, 1999).

Using O*NET in Practice In addition to characteristics of the content model described above, the information contained within the O*NET database holds particular value for the HR practitioner. First, the information is nationally representative of the U.S. workforce and is “fresh” in the sense that it has been collected in the past 7 years, with nearly three quarters of the occupations updated since 2005. Second, the data available in the O*NET system are more descriptive than information typi-

cally found in the products of many work analyses in practice (e.g., job descriptions and specifications). As an example, consider the ambiguity and widely variant levels of specificity of the descriptors commonly found in online job postings. O*NET provides descriptors that are clearly defined and theoretically based. As we discussed earlier, a single right way to conduct work analysis does not exist, but, rather, the chosen approach must be congruent with the ultimate uses of the collected information. Likewise, it would be a mistake to suggest that there is only one way to effectively use O*NET in practice. With that said, we believe O*NET can make a substantial contribution to improving the effectiveness of work analysis in practice. This is probably best accomplished by utilizing O*NET as a starting point for work analysis efforts. The O*NET database would then serve as a foundation upon which to undertake one’s own work analysis, regardless of the ultimate purpose. Following this logic, a practitioner would first consult O*NET to locate the relevant occupation(s) matching the focal role(s) of his or her work analysis, as well as the desired descriptors most relevant to the intended purpose (e.g., tasks and/or skills for designing training programs, skills and/or traits for choosing selection instruments). Then, the practitioner would use these data to inform their own in-house data collection efforts, whether these efforts are as simple as SME or incumbent verification of existing O*NET information (through ratings, rankings) or as complex as customized initiatives that seek to generate more company-specific information to augment O*NET data (e.g., knowledge germane to particular software systems, responsibilities or activities described in the language of a particular business function or department, etc.). In this sense, O*NET can provide generalizable data to help ground and facilitate local work analysis projects. Considering that work analysis results are frequently the key components to establishing content-related validity evidence, coupling local work analysis results with information from the nationally representative O*NET database may bolster the defensibility of decisions based on such evidence.

17

Morgeson and Dierdorff

Copyright American Psychological Association. Not for further distribution.

Recent O*NET Developments Since its initial pilot testing and development (see Peterson et al., 1999, 2001, for greater details), O*NET has undergone a number of important revisions, updates, and additions. First, the occupational coding scheme for O*NET has been aligned with the Bureau of Labor Statistics’s Standard Occupational Classification system (available from http://www.bls. gov/SOC) to ensure compliance with the requirements of all federal statistical agencies reporting occupational data (Levine, Nottingham, Paige, & Lewis, 2000). These coding changes have adjusted the total number of occupations in the O*NET system to 949, of which 812 were included in data collection efforts as of 2006. This represents a significant departure from the roughly 12,000 titles in the DOT and the 1,120 titles in the early versions of the O*NET database. Second, the original O*NET database was populated with analyst ratings. Several domains in the O*NET database have subsequently been updated

on a semiannual basis with ratings collected from role incumbents, as noted in the previous section. Publication of these data derived from incumbents began in 2003 and continues today. Thus far, the vast majority (96%) of the 812 data-level occupations have been updated with incumbent data. Third, new data pertaining to the variety of tools and technology needed for occupational performance have been recently added to the O*NET database (Brendle, Rivkin, & Lewis, 2008). Currently, tools and technology information have been generated for 327 occupations, with 427 occupations (53%) slated for completion by 2008. Over 25,000 tools and technology objects have been collected thus far, making this portion of the O*NET database the largest in terms of sheer volume. The number of objects per occupation range from 12 to 300. In general, “tools” refer to machine, equipment, and tools, whereas “technology” refers to software. Table 1.4 shows examples of tools and technology for several occupa-

TABLE 1.4 Examples of O*NET Tools and Technology Tools and technology objects Surveying technicians Electrotapes, measuring chains, tellurometers Echosounders, fathometers Total stations, Tribrach level bubble adjusting blocks, Tribrach optical plummet adjusting cylinders MicroSurvey FieldGenius, Survey Starnet Software ESRI ArcView, Geomechanical design analysis GDA software Anesthesiologists Intra-arterial catheters, Swan Ganz artery catheters Precordial stethoscopes, pretracheal stethoscopes AetherPalm InfusiCalc, Skyscape 5-Minute Clinical Consult EDImis Anesthesia Manager, Healthpac Computer Systems H2000 Anesthesia Billing Software Marketing managers ClickTracks software, online advertising reporting software Atlas OnePoint GO TOAST, Microsoft Project Accountants Best MIP Fund Accounting, Intuit QuickBooks, Sage CPAPractice Manager ACCUCert software, Intrax ProcedureNet, tax compliance property tax management software AuditWare software, Cartesis Magnitude iAnalysis, fixed-assets depreciation software

UNSPSC classification Distance meters Sonars Theodolities Database reporting software Map creation software Arterial line catheters Electronic stethoscopes or accessories Medical software Accounting software

Analytical or scientific software Project management software Accounting software Compliance software Financial analysis software

Note. Classifications are from the lowest level of the United Nations Standard Products and Services Code (UNSPSC) taxonomy (available at http://www.unspsc.org). O*NET = Occupational Information Network. 18

Copyright American Psychological Association. Not for further distribution.

Work Analysis

tions. Because of the substantial number of objects, as well as the hierarchical structure that underlies O*NET data, a critical need is to organize the tools and technology information. Currently, the tools and technology data are classified according to an existing and established taxonomy entitled the United Nations Standard Products and Services Code (available from www.unspsc.org). Of importance, the use of this taxonomy allows for cross-occupational comparisons and further promotes the common language approach inherent in the O*NET content model. A final recent O*NET development is interesting to note. To reiterate, one of the key recommendations in the aforementioned APDOT report was the need to maintain currency in occupational information. To accomplish this, efforts have been made to identify what are termed “new and emerging” occupations (Dierdorff, Cantwell, & Nottingham, 2008). Such occupations (a) involve significantly different work than that performed by incumbents of other preexisting occupations and (b) are not adequately reflected in the current O*NET system. Efforts to identify new and emerging occupations are focused on specific industries or sectors that have been deemed as “high growth” by the Department of Labor’s Employment and Training Administration. High-growth industries are those sectors projected to add substantial numbers of new jobs or affect the growth of other industries or that have existing or emerging businesses that are being transformed by technology and innovation, requiring new skill sets (Dierdorff et al., 2008). Table 1.5 shows examples of these high-growth industries as well as examples of new and emerging occupations that have been identified for inclusion in the O*NET database. As of 2008, 102 new and emerging occupations have been generated. For work analysis in general, these efforts focused on identifying and describing new and emerging occupations highlight the value of attending to more molar forces at the labor market and economic levels that shape the way work is performed but are rarely addressed in work analysis practice.

O*NET: Some Remaining Questions The O*NET system represents the most significant theoretical development in work analysis in recent history and reflects the cumulative expertise of over

50 years of work analysis research (Campion, Morgeson, & Mayfield, 1999). The developments described in the previous section also suggest that efforts to improve, update, and extend O*NET appear promising. Nonetheless, certain areas are in need of further attention. We highlight a few of these in the next several paragraphs. Although the core of the O*NET system grew out of an extensive pilot study that sought to offer reliability, validity, and other evaluative evidence for the domains covered by the content model (Peterson et al., 1999), there has been little published empirical research in the more than 10 years since this developmental research was undertaken. For example, much of the basic research conducted in the pilot study has yet to be replicated (and extended) on the current database. This research includes examinations of reliability, discriminability, and underlying factor structures of the present O*NET database, which now has several domains based on role incumbent ratings as well as analyst ratings. Such research is essential to the broader field of work analysis, considering that O*NET represents our state-of-the-art practices. The little research that has been conducted has focused on applications of O*NET data or uses of O*NET data in other non-work-analysis investigations. One example of application-oriented research is a study by Jeanneret and Strong (2003) that examined the utility of using select generalized work activities from O*NET for estimating job component validity. These authors showed positive evidence that O*NET descriptors were significantly predictive of general cognitive ability (via General Aptitude Test Battery and Wonderlic test scores). LaPolice, Carter, and Johnson (2008) described another study using a job component validity approach, which is a validation technique where relationships between quantitative work analysis data (e.g., levels of skills required by the job) and test scores of role incumbents are assessed across various jobs. This research supported the usefulness of O*NET knowledge, skill, ability, and generalized work activity data in predicting adult literacy test scores. A third example of O*NET application research is a study conducted by Converse, Oswald, Gillespie, Field, and Bizot (2004), in which they 19

Morgeson and Dierdorff

TABLE 1.5

Copyright American Psychological Association. Not for further distribution.

O*NET New and Emerging Occupations High-growth industry

New and emerging occupation

Definition

Advanced manufacturing

Mechatronics engineers

Apply knowledge of mechanical, electrical, and computer engineering theory and methods to the design of automation, intelligent systems, smart devices, or industrial systems control.

Automotive

Fuel cell engineers

Design, evaluate, modify, and construct fuel cell components and systems for transportation, stationary, or portable applications.

Biotechnology

Geneticists

Research and study the inheritance of traits at the molecular, organism, or population level. May evaluate or treat patients with genetic disorders.

Construction

Nondestructive testing specialists

Test the safety of structures, vehicles, or vessels using radiograph (X-ray), ultrasound, fiber optic, or related equipment.

Energy

Energy auditors

Conduct energy audits of buildings, building systems, and process systems. May also conduct investment grade audits of buildings or systems.

Financial services

Risk management specialists

Analyze company balance sheets and apply mathematical models to calculate risk associated with trading or credit transactions.

Geospatial technology

Geodetic surveyors

Measure large areas of the Earth’s surface using satellite observations, global positioning systems, light detection and ranging, or related sources.

Health care

Cytotechnologists

Stain, mount, and study cells to detect evidence of cancer, hormonal abnormalities, and other pathological conditions following established standards and practices.

Homeland security

Intelligence analysts

Gather, analyze, and evaluate information from a variety of sources, such as law enforcement databases, surveillance, intelligence networks, and geographic information systems. Use data to anticipate and prevent organized crime activities, such as terrorism.

Hospitality

Spa managers

Plan, direct, or coordinate activities of a spa facility. Coordinate programs, schedule and direct staff, and oversee financial activities.

Nanotechnology

Nanosystems engineers

Design, develop, and supervise the production of materials, devices, and systems of unique molecular or macromolecular composition, applying principles of nanoscale physics and electrical, chemical, and biological engineering.

Retail trade

Loss prevention managers

Plan and direct policies, procedures, or systems to prevent the loss of assets. Determine risk exposure or potential liability and develop risk control measures.

Transportation

Logistics engineers

Design and analyze operational solutions for projects such as transportation optimization, network modeling, process and methods analysis, cost containment, capacity enhancement, routing and shipment optimization, and information management.

Note. High-growth industries identified by the U.S. Department of Labor’s Employment and Training Administration. O*NET = Occupational Information Network.

20

Copyright American Psychological Association. Not for further distribution.

Work Analysis

outlined and evaluated a process for career guidance that matched individuals to occupations using O*NET abilities. Finally, Reiter-Palmon, Brown, Sandall, Buboltz, and Nimps (2006) described research that used O*NET data in the development and implementation of a Web-based work analysis process. In terms of research that centrally focuses on O*NET itself, rather than its applications, even fewer studies have been conducted. One example is research conducted by Hadden, Kravets, and Muntaner (2004), who used exploratory factor analysis to examine a version of the O*NET database populated with analyst ratings. The authors found evidence that this database possessed a factor structure that was comparable with the DOT. Because this study was conducted with the older, analyst-version of the O*NET database, no conclusions can be made regarding the factor structure of the current incumbent-populated database. Another study by Eggerth, Bowles, Tunick, and Andrew (2005) examined the convergent validity of O*NET occupational interests (also analyst derived) as compared with the DOT Holland codes and the Strong Interest Inventory. These authors found varying levels of agreement in the scores produced across the three instruments, with the highest agreement levels between O*NET and the DOT and Strong Interest Inventory scores. Finally, Dierdorff and Morgeson (2009) provided the only study to date that directly investigates incumbent ratings from the O*NET database. These authors used variance component analysis and meta-analysis to examine sources of variance and interrater reliability of ratings on O*NET tasks, generalized work activities, knowledge, skills, and work styles (traits). Variance component analysis was used to partition rating variance into two sources: (a) variance due to the item (i.e., “true” differences) and (b) variance due to the rater (i.e., idiosyncratic differences). Using data collected from job incumbents across 309 occupations (N = 41,137), Dierdorff and Morgeson found that larger proportions of variance (more than twice the amount) were generally attributable to items rather than to raters. The one exception to this general trend was for rating of work styles, where the opposite finding was evident

(i.e., twice as much variance was due to the rater). Meta-analysis showed similar results, with lower interrater reliability for work style ratings, suggesting that incumbents are likely to show lower consensus when rating the traits that are important to performing their roles. Taken collectively, the results of this study offer generally favorable results for incumbent ratings of O*NET tasks, generalized work activities, knowledge, and skills. Broadly speaking, the O*NET-related research evidence accumulated thus far appears to support the quality and viability of the data. However, we believe there are at least three key areas that still require directed treatment. First, the need for additional evaluative research cannot be overstated. Brannick et al. (2007) raised an interesting point of comparison with regard to the predecessor of O*NET when they stated, “despite its limitations, the DOT benefited from many years of research conducted on it” (p. 122). Research investigating topics such as the factor structure underlying O*NET data, relationships between analyst and incumbent ratings, and the uniqueness or redundancy in types of ratings (importance vs. level ratings) are broad examples of such empirical needs. Second, more work is necessary to further explicate the efficacy of applying O*NET data to the wide variety of HR systems (e.g., selection, compensation). With over 3.5 million page views per month of the O*NET OnLine website (Brendle et al., 2008), it would appear that O*NET information is being widely used. In addition, O*NET data are often used by governmental agencies to form various workforce strategy initiatives, such as focused training investments (Dierdorff & Cox, 2008). However, to our knowledge, there exists no direct evidence of how extensive and for what purposes O*NET is used by organizations. It also is important to note that application-oriented research should focus less on documenting a particular process or describing case studies and instead turn attention to more useful criteria, such as the validity, utility, acceptance, and effectiveness of the systems using O*NET information. Third, it is unclear whether some content model areas for which information is not currently available will be the beneficiaries of

21

Morgeson and Dierdorff

future data collection efforts. For example, the organizational context descriptors (e.g., high performance work practices, culture) developed for O*NET would be particularly valuable not only to work analysis research, but also to many other areas of I/O psychology and management (Campion et al., 1999).

Copyright American Psychological Association. Not for further distribution.

THE QUALITY OF WORK ANALYSIS INFORMATION Because of its centrality to so many HR systems, considerable research has been focused on ensuring that work analysis data are of high quality. This has been reflected, in part, by research that has focused on the interrater reliability of work analysis data. Moreover, as one might expect, when properly conducted, work analysis data are highly reliable (see Dierdorff & Wilson, 2003, for a meta-analytic summary). However, reliability is only one component of data quality. A bigger issue is the validity and accuracy of work analysis data. In this section, we first discuss how accuracy has been conceptualized in work analysis. Next, we discuss the range of potential influences on work analysis data. Finally, we close with a discussion of the kinds of inferences that are made in work analysis and the resulting inferential leap that is often made when conducting work analysis.

Accuracy in Work Analysis The issue of the accuracy of work analysis data is a difficult one. In many respects, work is a social construction (as our focus on role enactment emphasizes). As such, it is not clear what is meant by work analysis accuracy. Part of the problem is that most work analysis research has relied on the principles of classical test theory (Campion et al., 1999; Harvey, 1991). Classical test theory would suggest that there is a “true score” for a given work role, that true scores are stable across time, and that measurement variation is error that can be eliminated by aggregating across sources and time (Nunnally & Bernstein, 1994). This has led researchers to aggregate data across sources (e.g., incumbents) to determine the true score for a given role. In this view, work analysis data quality is indexed by estimating interrater reliability. 22

However, there is considerable reason to believe that the assumptions of classical test theory are inappropriate, in part because there are potentially numerous influences on the quality of work analysis data (Morgeson & Campion, 1997). This has led some to advocate and use a generalizability theory perspective (Sanchez & Levine, 2000; Van Iddekinge et al., 2005). An advantage of generalizability theory is that it enables one to simultaneously estimate multiple sources of measurement error. Despite its advantages (i.e., it allows one to model multiple sources of variance in work analysis data), generalizability theory is also predicated on the notion of a stable true score. Other work analysis researchers have attempted to assess accuracy more directly by taking steps to identify those who might not be answering correctly. Most of these methods involve the inclusion of specific items or indices to detect such individuals (e.g., carelessness index, Green & Stutzman, 1986; infrequency index, Green & Veres, 1990; veracity items, McCormick, 1960; false reporting index, Pine, 1995). Such indices generally include two types of items: (a) veracity items considered to be requisite and thus performed by all incumbents in a given work role and (b) distractor or “bogus” items considered to be unrelated to the job and never performed by incumbents. Another approach has been to repeat particular items in a rate–rerate approach so as to assess intrarater consistency (Wilson, Harvey, & Macy, 1990). In general, such approaches to assessing accuracy are best suited for collecting work analysis information using the questionnaire method. Although offering the benefit of direct estimation of accuracy (i.e., they represent an unambiguous index of accuracy), these approaches do have some associated costs, such as increasing the overall length of the survey and reducing the face validity of the survey (e.g., respondents may wonder why bogus items are being presented or why items are being “unnecessarily” repeated). Another perspective on the issue of accuracy in work analysis data was forwarded by Morgeson and Campion (1997). They suggested that instead of focusing on any particular single true score estimate, one could simply index the accuracy of work analysis

Copyright American Psychological Association. Not for further distribution.

Work Analysis

data in multiple ways, in part because different sources of inaccuracy have different effects on work analysis data. The implications of this are that only by taking a multidimensional view of accuracy could one begin to understand the quality of the data. They identified six aspects of work analysis data quality. First was interrater reliability, which is the most commonly used measure of data quality in the work analysis domain (Dierdorff & Wilson, 2003). Interrater reliability reflects consistency across raters and indexes rater covariation (Shrout & Fleiss, 1979). Second was interrater agreement, which reflects the absolute level of agreement across raters and thus indexes the degree to which different raters make similar ratings (Kozlowski & Hattrup, 1992). Third was discriminability between jobs, which reflects between-job variance and the ability to distinguish between different jobs. Fourth was dimensionality of factor structures, which reflects the extent to which factor structures are complex or multidimensional. Fifth was mean ratings, which reflects inappropriately elevated or depressed ratings. Sixth was completeness, which reflects the extent to which the work analysis data are complete or comprehensive. Thus, one way to evaluate the accuracy of work analysis data is to focus on a broader set of criteria.

Sources of Variance in Work Analysis Data Although considerable energy has been devoted to developing work analysis methods that generate reliable and valid data, the bulk of this research rests on the implicit assumption that any error is essentially random in nature. Proceeding from this assumption, most work analysis research has sought to eliminate such error through traditional means, such as using sophisticated sampling strategies and standardizing work analysis materials. However, there is reason to believe that work analysis data are subject to systematic (and predictable) sources of variance. If this is the case, then the traditional ways of controlling error will be ineffective and resulting work analysis data will be inaccurate. We now turn to a brief review of factors that may impact work analysis data. Prior to this discussion, however, it is important to acknowledge that although some of the issues we highlight have been supported in past

work analysis research, other issues are more speculative, based on suggestive evidence, and thus require additional research. Rater influences. Researchers have long acknowledged that certain rater characteristics may influence work analysis outcomes. For example, Madden (1962, 1963) explored the role of job familiarity and E. P. Prien and Saleh (1963) explored the role of job tenure. As Harvey (1991, p. 115) noted, “one cannot simply assume that job analysis ratings will be unaffected by characteristics of the rater.” Supporting this conclusion, recent research has demonstrated that a considerable amount of variance in work analysis outcomes is indeed due to rater characteristics. For example, Van Iddekinge et al. (2005) found that 21.6% and 29.1% of the error variance in single-rater reliabilities of knowledge, skill, ability, and other characteristics importance ratings and needed-at-entry ratings (respectively) were attributable to rater idiosyncrasies. As such, it is important to explore how attributes of the raters (or source) can impact work analysis information. First, general cognitive ability may impact work analysis information in a number of ways. Within the same job, individuals of higher cognitive ability might be able to provide more accurate and complete work analysis information because of their superior job knowledge (Hunter, 1986) of the focal role than those of lower cognitive ability. Cornelius and Lyness (1980) offered additional reasons why cognitive ability might influence the quality of work analysis judgments. In work analysis, respondents are often asked to make inferences or abstract judgments about aspects of the work, or they may be asked to integrate a large amount of information. Because of the cognitive demands of these judgments, those high in cognitive ability have an advantage because of their additional mental resources. These integrative judgments can be viewed as controlled processes (W. Schneider & Shiffrin, 1977), and cognitive ability is highly predictive of success in such processes (Ackerman & Humphreys, 1990). Greater cognitive ability may also result in more accurate work information, because many questionnaires require a high reading level (Ash & Edgell, 1975; Harvey, Friedman, Hakel, & Cornelius, 1988), and cognitive ability is related to education level. Research has supported the 23

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

relationship between educational level and reliability or other differences in work analysis data (Cornelius & Lyness, 1980; Fried & Ferris, 1986; Green & Veres, 1990; Landy & Vasey, 1991). Two caveats should be considered regarding cognitive ability. First, incumbents with noticeably higher cognitive ability may create extraneous information that could lead analysts or supervisors to rate the job requirements higher for these individuals, even though the underlying work is the same. Second, incumbents with higher cognitive ability may have qualitatively different experiences in the work setting because they are assigned (or take on) additional or different (e.g., higher level or more complex) tasks. This could influence the tasks and knowledge, skills, abilities, and other characteristics they generate, as well as ratings of importance and time spent. In support of this, Morgeson, Delaney-Klinger, and Hemingway (2005) recently found that cognitive ability was positively related to the number of tasks performed. These differences may be more pronounced on jobs where there is increased autonomy or opportunity for discretionary behavior. Second, different personality characteristics may influence work analysis responding in a variety of ways. For example, individuals high in conscientiousness may be more careful and diligent in their responding, resulting in more reliable and accurate responses. Or, individuals high in extraversion may incorporate more socially oriented work elements into their focal role, thereby changing the nature of the work they perform, compared with less extraverted coworkers who are in the same role. Although there have been attempts to systematically measure the personality requirements of work (e.g., Raymark, Schmit, & Guion, 1997), there have been few attempts to explore how different personality characteristics are related to work analysis data. Future research should address this gap. Another important attribute is work experience. More experienced incumbents may provide more accurate information because they may have greater information and insight into the job. The research evidence is mixed, however, with some studies showing differences (Borman, Dorsey, & Ackerman, 1992; Landy & Vasey, 1991; Sanchez & Fraser, 1992) and others not (Mullins & Kimbrough, 1988; Schmitt & 24

Cohen, 1989; Silverman, Wexley, & Johnson, 1984). Furthermore, some of the differences in work analysis information may be due to differences in the jobs performed by more experienced incumbents. For example, Borman et al. (1992) found significant differences in 9 of 12 time-spent scores between more and less experienced stockbrokers. It appears that as stockbrokers advance in their careers, they are involved in distinctly different activities, with a relationship-building phase early and a relationshipmaintenance phase later. Landy and Vasey (1991) found similar differences for more and less experienced police officers. Finally, Sanchez and Fraser (1992) found that when rating task importance, individuals differentially weight time spent and difficulty of learning as a function of their job experience. However, Mullins and Kimbrough (1988) found no such experience differences for police officers in the generation of critical incidents, although groups were divided into very narrow bands of experience (e.g., each group constituted an increment of only 1 year of experience). Another view of work experience and work analysis has been provided by Richman and Quiñones (1996). They found that less experience with an experimental task was related to more accurate estimates of the frequency with which individual task elements had been performed and correct identification of tasks performed. They suggested that individuals have more difficulty recalling the frequency of specific events if similar events occurred frequently. Given these mixed findings, understanding the role of work experience in work analysis judgments is an important area of future research. In investigating this issue, however, it would be important to adopt a multidimensional view of work experience. Tesluk and Jacobs (1998) developed a model of work experience that specifies measurement modes of work experience (i.e., amount, time, density, timing, type) and levels of specification (i.e., task, job, work group, organization, career–occupation). Any research on experience should seek to measure multiple aspects of experience, as some (e.g., task, job) may be more logically connected to work analysis than others (e.g., organization, occupation). Future research should also explore different rating scales, as differences in experience may also depend on the rating

Copyright American Psychological Association. Not for further distribution.

Work Analysis

scale used. For example, all incumbents may identify the same tasks as critical, regardless of experience, but the amount of time they spend on different tasks may vary with experience. Also, differences with experience may be more pronounced if the jobs have some autonomy or opportunity for discretion in terms of which tasks to perform or the relative emphasis tasks are given. A final rater attribute that might be important is the performance level of workers. As with the other rater influences, empirical results have been mixed. For instance, Borman et al. (1992) found significant relationships between time-spent ratings and performance of stockbrokers. Mullins and Kimbrough (1988) also found significant differences between low- and high-performing patrolpersons in their importance ratings. In contrast, Wexley and Silverman (1978) found no performance differences in importance and time-spent ratings in a sample of retail store managers. Conley and Sackett (1987) also found no differences in terms of either task generation or ratings of knowledge, skill, and ability between high- and low-performing juvenile officers. Finally, Aamodt, Kimbrough, Keller, and Crawford (1982) found no performance-related differences in the type of critical incident categories generated by residence hall workers. As with the other attributes, differences in work analysis responses may be due to genuine differences in the jobs performed by higher performing employees. Better employees may be assigned additional or different tasks because they are more able to handle the extra work or as a reward for their high performance. In addition, low performers could be more likely to leave the organization (on a voluntary basis or by being terminated), which would introduce issues of range restriction that might affect relationships between experience and work analysis ratings. Social and cognitive influences. Although rater attributes have been previously identified as a potential influence on work analysis information, it is only more recently that other potential influences have been identified. In fact, Morgeson and Campion (1997) identified 16 distinct potential social and cognitive sources of inaccuracy. The social sources “are created by normative pressures from the social environment and reflect the fact that individuals act

and reside in a social context,” whereas the cognitive sources “reflect problems that primarily result from the person as an information processor with distinct limitations” (p. 628). Given the in-depth discussion of these processes in past research (Morgeson & Campion, 1997), we only provide an overview and selected examples. The reader is referred to the original article for a more extended discussion. Social sources are divided into social influence and self-presentation processes. Social influence processes include three distinct processes that occur when judgments are made in group settings. The first is conformity pressures, which reflects the fact that a group can exert quite a bit of normative influence to reach consensus. For example, in an SME group meeting, there are often strong pressures from a majority of group members to reach a certain conclusion (e.g., a particular aspect of the work is essential). Even if another group member disagrees, it is likely that they will go along because of the pressure for conformity that will exist. The second is extremity shifts (also called “group polarization”), which refers to the tendency for group member opinions to become more extreme following group discussion. The third is motivation loss, which reflects the tendency for individuals to exert less effort when in a group as compared with an individual setting. This can have the unfortunate result of not obtaining all the input of group members, resulting in deficient work analysis information. Self-presentation processes included three processes that reflect an individual’s attempt to present him- or herself in a particular light. The first is impression management, which reflects attempts to present oneself in such a way as to “create and maintain desired perceptions of themselves” (Gardner & Martinko, 1988, p. 321). Incumbents are likely to “inflate” the value of their job, particularly when the outcome of the work analysis might potentially benefit them (e.g., such as when a compensation system is being redesigned). The second is social desirability, which reflects “a need for social approval and acceptance and the belief that this can be attained by means of culturally acceptable and appropriate behaviors” (Marlowe & Crowne, 1961, p. 109). For example, Smith and Hakel (1979) found that incumbents and supervisors displayed considerable response 25

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

inflation on socially desirable work analysis items compared with analyst ratings. The third is demand effects, which reflects the tendency of individuals to play the “good subject” role and respond in such a way as to validate external expectations. One might imagine a situation where a work analyst conveys to role holders that a certain set of skills are particularly important, and the role holders subsequently validate this expectation by rating them as highly important. Cognitive sources are divided into limitations in information-processing systems and biases in information-processing systems. Limitations in information-processing systems include three different processes. The first is information overload, which reflects the fact that human information processing has limits when attempting to process complex or large quantities of information. For example, when confronting numerous, detailed activity and attribute statements in a work analysis questionnaire, respondents may simply be unable to effectively process all the information. The second is heuristics, which reflects the fact that individuals often rely on simplifying heuristics (such as representativeness and availability) when making judgments (Tversky & Kahneman, 1974). Because these heuristics imperfectly mirror reality, they tend to result in inaccurate judgments. The third is categorization, which reflects the fact that individuals tend to organize their experiences into distinct categories. Once categorized, subsequent inferences about the experience are made with respect to the category and not the specific experiences. Thus, if a role holder has concluded that “my work is highly complex,” then he or she is likely to make subsequent inferences consistent with this conclusion. Biases in information-processing systems include seven processes. The first is carelessness, which reflects response distortion due to inattention. For example, work analysis respondents often do not read questionnaire items closely (e.g., they do not realize that an item is reverse coded) or carefully (e.g., they indicate they perform tasks that they could not possibly perform) enough. The second is extraneous information, which can create inaccuracy when information not relevant to the work analysis is somehow included or considered. For example, in a work analysis conducted for the purpose of 26

determining pay levels, knowledge of current pay levels can influence the resulting work analysis information. The third is inadequate information, which refers to situations where raters have incomplete job information. This can occur if inexperienced (or naive) raters are used or if analysts have not conducted a systematic analysis of the work. The fourth is order and contrast effects, which involves the influence of contextual ratings effects, such as order (primacy and recency) and contrast effects. Primacy effects refer to the influence of initial information (e.g., the first interviews conducted by an analyst), whereas recency effects refer to the influence of recent information (e.g., how recently performed tasks might be overly salient). Contrast effects reflect distortions caused by differences between stimuli. For example, if a work analyst had been rating a number of low-level roles, he or she might give inappropriately high ratings to an averagelevel job because of the implicit contrast between jobs. The fifth is halo, which occurs when ratings are assigned on the basis of global impressions rather than a systematic consideration of differences among separate categories. One way that halo might affect work analysis is that if the task domain or work behavior is not sampled adequately enough, then there is likely to be more of a reliance on global impressions. The sixth is leniency and severity, which reflects a general response tendency to give consistently high (leniency) or low (severity) ratings. Leniency is more likely in work analysis, in part because of a general reluctance to be overly critical when making work analysis judgments. The seventh is method effects, which reflects the fact that when data are collected through a single method, there can be spurious covariation among responses. This is likely to be an issue in work analyses when the questionnaire method is used and all the data are collected from a single source at a single point in time. Contextual influences. Another category of factors that may influence work analysis information stems from the context within which work roles are performed. As discussed earlier, aspects of work context are one of the descriptor types that work analysis seeks to understand. Thus, features of work context can be a type of information directly collected during work analysis, such as when

Copyright American Psychological Association. Not for further distribution.

Work Analysis

elements of the task, social, and physical environments in which roles are enacted are assessed. In addition, it is important to recognize that contextual forces are also likely to shape how work roles are perceived and ultimately enacted. Katz and Kahn (1978, p. 195) explained this relationship by stating that role enactment “does not occur in isolation; it is itself shaped by additional or contextual factors.” In this sense, work context not only shapes how a work role is enacted, but also may serve as a systematic source of variance in work analysis data. Contextual influences on work analysis information can be examined using an omnibus approach or a discrete approach. Discrete descriptions of work context focus on more specific classes of variables, such as those described earlier (i.e., delineating task, social, and physical elements). In contrast, an omnibus approach entails a broader consideration of contextual influences and “refers to an entity that comprises many features or particulars” (Johns, 2006, p. 391). That is, an omnibus approach accounts for contextual effects using more molar boundary conditions. For instance, one useful entity for studying omnibus context is that of occupation. In relation to how work context may influence variance in work analysis data, a discrete approach might focus on the effects of social context (e.g., role interdependence) on work analysis ratings, whereas an omnibus approach might focus on the organizational effects on ratings (e.g., ratings of similar roles in different companies). Work analysis research has used both approaches to studying contextual effects. With regard to omnibus context effects, Van Iddekinge et al. (2005) examined whether the error variance in knowledge, skill, ability, and other characteristics ratings were impacted by the organization in which raters worked (these effects were not significant). Another study by Taylor, Li, Shi, and Borman (2008) showed that mean ratings and rank ordering of items from several O*NET domains were quite similar across four different countries. With regard to discrete context effects, Lindell, Claus, Brandt, and Landis (1998) found discrete features of organizations (e.g., size, formalized structure, technology) were correlated with time-spent ratings on tasks (average r = .32) but not importance ratings. Finally, Dierdorff, Rubin, and Morgeson (2009) examined both omnibus

and discrete context effects on managerial work role requirements. These authors found evidence of omnibus context effects, as the type of managerial occupation (e.g., financial manager, HR manager) accounted for 4% to 39% of the total variability (p < .01) in importance ratings on 18 work role requirements spanning responsibility, skills, knowledge, and trait domains. Further, discrete elements from the task, social, and physical contexts (e.g., autonomy, interdependence, hazardous work conditions) accounted for additional variance in these ratings (roughly 18% of between-occupation variance across dimensions of discrete context). From the accuracy of work analysis data to the quality of work analysis inferences. As the preceding discussion highlights, there are numerous potential influences on work analysis data. In addition to questions about the prevalence of such influences, another question centers around the extent to which any observed variability of work analysis data reflects meaningful differences in role enactment as opposed to error or inaccuracy. Because individuals often enact similar roles in slightly different ways (Biddle, 1979; Graen, 1976; Katz & Kahn, 1978), not all observed differences necessarily reflect inaccuracy. The possibility that some variance in work analysis data may be due to legitimate differences in role enactment introduces another key challenge in understanding work analysis accuracy, leading some to suggest that because work is a social construction, there is no gold standard of accuracy in work analysis (Sanchez & Levine, 2000). A potential resolution of the dilemma, however, is to shift the focus from the accuracy of work analysis data (which has been the traditional conceptualization) to a focus on the quality of work analysis inferences (Morgeson & Campion, 2000). This is a potentially useful shift for two reasons. First, it is difficult to establish the stability or objectivity of work analysis data. As such, we can only begin to approximate (via some of the criteria discussed above) the accuracy of the data. Second, work analysis data are often completely based on human judgment (Goldstein, Zedeck, & Schneider, 1993). Put another way, “The making of job ratings can be conceptualized as an inferential decision” (Sanchez & Levine, 1994, p. 48), where the process 27

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

of inductive inference involves drawing general conclusions from specific examples (Hempel, 1965). Thus, one could begin to describe the different kinds of inferences that are made in work analysis and then develop some means for estimating the quality of the inferences made. Instead of evaluating the quality of work analysis data, one would evaluate the quality of the inferences one is making on the basis of the work analysis data. The first step in such an endeavor would be to describe the different types of inferences made in work analysis. Morgeson and Campion (2000) developed an integrative framework that identifies three key inferences that specifically occur in work analysis (see Figure 1.4). First, the work descriptive inference involves the extent to which a description of work activities (i.e., tasks and responsibilities) faithfully represents the physical and mental activities underlying role performance. Second, the work specification inference involves the extent to which a specification of worker attributes (i.e., knowledge, skill, ability, and other characteristics) reflects the psychological constructs underlying role-related aptitudes. Third, the operational inference involves the extent to which the identified worker attributes are needed to perform identified work activities.2 The quality of these inferences could then be evaluated by “deriving theorybased expectations about how scores should behave under various conditions and assessing the extent to which these expectations receive support” (see Aguinis, Mazurkiewicz, & Heggestad, 2009, p. 433). One implication of this model is that some inferences require a greater inferential leap than other inferences, where the inferential leap in work analysis can be defined as the complexity of the evaluative judgments made about various work role requirements. This complexity is reflected in leaping from observations of work activities to inferences about role requirements. All types of work analysis judgments require some sort of inferential leap, in part because even the most observable aspects of a role (e.g., the performance of very specific tasks) usually 2

Psychological Constructs Underlying Role-Related Aptitudes

Work Activities Underlying Role Performance

1

Work Requirements • Tasks, Responsibilities

2

3

Worker Requirements • Knowledge, Skill • Ability, Other Characteristics

FIGURE 1.4. Key inferences that occur in work analysis. From “Accuracy in Job Analysis: Toward an InferenceBased Model,” by F. P. Morgeson and M. A. Campion, 2000, Journal of Organizational Behavior, 21, p. 823. Copyright 2000 by John Wiley & Sons, Ltd. Adapted with permission.

require one to move from observable behavior to judgments about such behavior (e.g., frequency of performance, importance to the role). Such a view has been recently recognized and supported in the work analysis literature (e.g., Lievens & Sanchez, 2007; Lievens et al., 2004; Morgeson et al., 2004; Voskuijl & van Sliedregt, 2002). Recent research by Dierdorff and colleagues (Dierdorff & Morgeson, 2007, 2009; Dierdorff & Rubin, 2007; Dierdorff & Wilson, 2003) has provided some indirect evidence of the inferential leap required by work analysis ratings. This research has shown rating differences attributable to the work descriptor being judged, such as the variation in levels of reliability, carelessness, consensus, and discriminability of work analysis ratings. Broadly speaking, this research suggests that ratings of less specific and directly observable descriptors (e.g., traits) require a larger inferential leap than more molecular and visible descriptors (e.g., tasks). This research also suggests that the inferential leap may systematically vary because of the source (analysts vs. role incumbent) as well as work context (e.g., amount of discretion in one’s role).

The operational inference is similar to what Gatewood and Feild (2001) called the “work–worker attribute leap.” In addition, Gatewood and Feild described three other types of inferential leaps pertinent to HR activities in general (see also Sanchez & Levine, 2000): (a) the worker attribute– organizational intervention leap, (b) the work–performance measure leap, and (c) the organizational intervention–performance measure leap. Because these latter three types of inferential leaps do not directly deal with the collection of work analysis data but instead refer to the development HR systems (e.g., selection systems), we do not discuss them further.

28

Copyright American Psychological Association. Not for further distribution.

Work Analysis

Understanding the nature of the inferential leap in work analysis is important for at least two reasons. First, as we discussed, work analysis judgments have been typically treated as free from systematic error. However, we now understand that work analysis judgments are subject to various systematic sources of error and inaccuracy (Morgeson et al., 2004; Van Iddekinge et al., 2005). Focusing on work analysis inferences helps us better estimate the quality of our work analyses as well as helps us make appropriate inferences about the data that are collected. Second, the popularity of competency modeling approaches to work analysis (Lucia & Lepsinger, 1999; Schippmann, 1999) has resulted in an increased emphasis on abstract, holistic work descriptors (Schippmann et al., 2000). Focusing on the nature of the inferences made when collecting this type of information helps us better understand the potential limitations that attend the use of such descriptors. For the work analysis practitioner, a better understanding of the types of inferences required and the consequences of these inferences (e.g., changes in levels of consensus and carelessness) can allow for better work analysis design decisions. For example, evidence suggests that incumbents are likely to show lower consensus when rating traits than when rating duties or skills (Dierdorff & Morgeson, 2009). Thus, using analysts to rate trait descriptors, or using multiple types of respondents (incumbents, supervisors, trainers), would be beneficial when capturing judgments about these abstract descriptors. LOOKING AHEAD: FUTURE AVENUES OF RESEARCH In this section, we offer a number of potentially fruitful areas for future work analysis researchers to pursue. These suggestions are by no means exhaustive but are intended to address some more traditional areas of work analysis research and to stimulate new thinking in areas not conventionally falling under the purview of work analysis. To accomplish this, we propose topics we believe are potentially fruitful avenues for future research, most of which would be viewed as germane to the field of 3

work analysis. We then discuss several other areas that hold the potential to meaningfully extend work analysis research into other theoretical domains. To the extent possible, we present illustrative research questions throughout the ensuing discussion.

Variance in Work Analysis Ratings Accumulating empirical evidence shows that considerable variance in work analysis ratings is attributable to idiosyncratic sources as compared with the dimension upon which a work role is being judged (e.g., skills).3 From a practical standpoint, idiosyncratic variance is generally viewed as undesirable because these rating differences are not due to consensus differences in the target work role and, if large enough, make aggregation of work analysis ratings problematic. Fortunately, recent research has shown that rater training (frame-of-reference training) can be an effective way to decrease idiosyncratic variance in attribute descriptor ratings provided by analysts (Lievens & Sanchez, 2007) and incumbents (Aguinis et al., 2009). These results are promising and suggest that additional research is warranted. Other forms of rater training shown to be effective in the performance appraisal literature, such as rater error training or performance dimension training (see Woehr & Huffcutt, 1994), should also be investigated. In addition, rater training could be applied to other common work analysis inferences, such as judgments about the linkages between tasks and knowledge, skill, ability, and other characteristics for purposes of identifying selection instruments. Of importance, future research should include not only traditional work analysis criteria (e.g., reliability, accuracy), but also criteria relevant to training interventions (e.g., affective outcomes, cost effectiveness). A second way to approach the idiosyncratic variance found in work analysis ratings is to search for variables that can account for this variation. Future work analysis research could use Morgeson and Campion’s (1997) framework discussed earlier to guide such investigations. For example, these authors offer over a dozen specific research propositions, many of which have yet to be subjected to empirical testing. In fact, to date, only two studies have applied

Although, as we discussed earlier, some of the variance is likely to be systematic and explainable.

29

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

this conceptual framework to examine potential sources of inaccuracy in work analysis ratings, and both have shown meaningful results (Dierdorff & Rubin, 2007; Morgeson et al., 2004). Beyond these two studies, however, there is substantial research that remains to be conducted using this framework. A third way to explore variance in work analysis ratings would be to draw from the broader I/O literature to understand some of the factors that might be related to differences in role enactment. This includes work attitudes (e.g., job satisfaction, commitment, fairness perceptions; Conte, Dean, Ringenbach, Moran, & Landy, 2005), the relationship between workers and their immediate supervisors (Hofmann, Morgeson, & Gerras, 2003), the amount of autonomy present in the work (Morgeson, Delaney-Klinger, & Hemingway, 2005), work experience (Borman et al., 1992), and ability (Morgeson, Delaney-Klinger, & Hemingway, 2005). A fuller understanding of how these (and other) factors relate to work analysis ratings would help us better understand the meaning of rating differences and whether differences reflect inaccuracy or meaningful differences in role enactment.

Exploring Relationships Among Different Work Role Requirements Another avenue for future research is an examination of the relationships among the different work role requirements. Such cross-domain research would include exploring the linkages between attributes and activities. Indeed, fundamental to the majority of theory in I/O psychology (and most fields of psychology) has been the notion that person attributes are antecedent to work behaviors. In this sense, cross-domain research offers valuable information regarding such relationships. Further, the pursuit of a unified theory of performance, which necessarily includes crossdomain specifications, has long been a part of the work analysis tradition (e.g., Fleishman, 1975). An example of this pursuit can be seen in the well-accepted “data–people–things” descriptive framework adopted in the DOT. One interesting question is whether this framework still holds in the contemporary world of work. Some empirical 30

evidence indicates that additional factors (e.g., organizational structure) might be required to adequately describe variance in work role requirements (Hadden et al., 2004). Research that explicates precisely how crossdomain relationships vary across different types of work roles (e.g., occupations) would also be beneficial. Such research not only would increase our theoretical understanding of how attributes link to activities, but also holds promise to improve the practices built from work analysis information. For example, cross-domain specifications are central to job component validity approaches, which seek to analyze the relationship between work analysis data and validity data across various work roles (Jeanneret, 1992). In addition, a better understanding of how cross-domain relationships vary across work roles provides valuable information pertaining to validity generalization (Schmidt & Hunter, 1977), which has long recognized the effects of different occupations on validity estimates (see Ghiselli, 1966). Here, instead of treating occupation as simply a nondifferentiated moderator of validity variability to be subsequently controlled as an artifact, one could extend work analysis information for use in meta-analyses to more meaningfully examine what particular features of occupations are exerting influence (e.g., work context, occupational complexity). Theory and research falling under the rubric of interactional psychology is a final area that crossdomain work analysis research could inform. In the interactional psychology literature, the importance of considering both the individual and the situation as joint determinants of work behavior has long been encouraged (Block & Block, 1981; Bowers, 1973; Magnussen & Endler, 1977; Terborg, 1981). Because the primary goal of work analysis is to systematically discover work role requirements and the context in which these requirements are enacted, work analysis research is particularly relevant for interactional psychology. One area to which work analysis research could contribute is a more thorough understanding of how concepts such as situation strength (Mischel, 1977), trait relevance (Tett & Burnett, 2003), and context effects (Johns, 2006) theoretically and empirically

Copyright American Psychological Association. Not for further distribution.

Work Analysis

operate to shape work behavior. For example, the notion of situation strength has been criticized for being overly broad and lacking emphasis on important qualitative aspects of context that make a given attribute relevant to role enactment. At the same time, however, both situation strength and trait relevance are necessary for understanding attribute and activity relationships (Tett & Burnett, 2003; Tett & Guterman, 2000). One way future research could empirically examine how these theoretical concepts function is to use work analysis to examine pertinent cross-domain linkages. Here, features of the work context (e.g., facets of task, social, or physical context) would represent varying levels of situation strength, whereas trait relevance could be systematically varied by testing the relationships between relevant–irrelevant attributes and work behavior. For example, research could investigate how the strength of social context shapes the relationships between socially relevant attributes (e.g., extraversion, conflict negotiation skills) or irrelevant attributes (e.g., conscientiousness, critical thinking skills) and role behaviors of an interpersonal nature (e.g., helping coworkers, teaching others). Presumably, the efficacy of simultaneously using both the situation strength and trait relevance concepts would be evident if a strong social context exerts greater influence on the relationships between socially relevant attributes (e.g., social orientation) and interpersonal role behaviors (e.g., helping others) when compared with the associations between socially irrelevant attributes and these behaviors.

Models of Role Performance The above discussion alludes to what we believe may be the most fruitful area into which work analysis theory and research could be extended, namely, how work analytic data can be used to better understand role performance. Although this is not necessarily a new connection when one considers that general theories of performance have been directly based on empirical results from work analyses (e.g., Campbell et al., 1993), nonetheless, there exists today very little cross-fertilization of theory and research between the work analysis and job performance domains. Such connections would be

valuable for a number of reasons. First, work analysis is sometimes characterized as simply an atheoretical, descriptive process that is necessitated primarily because of legal codifications (e.g., Albemarle Paper Co. v. Moody, 1975) and professional standards (e.g., Principles for the Validation and Use of Personnel Selection Procedures; Society for Industrial and Organizational Psychology, 2003). Of course, we believe there are many other reasons why work analysis is crucial to organizations, not the least of which is to improve the decisions made in various HR practices. However, this does suggest that one way to increase value perceptions of work analysis research is to demonstrate how such data relate to individual effectiveness. As authors of work analysis research who have, on numerous occasions, dealt with the “so what” question about our field of study, further justification of why work analysis matters has a certain appeal. However, there is another reason for extending work analysis theory and research into the performance domain that is perhaps more fundamental than addressing criticisms of work analysis relevance. When research is focused on the definition, measurement, or prediction of performance, it is essentially concerned with the manner with which work role requirements are fulfilled. This notion is consistent, for example, with the contrast between work analysis as identifying requisite role behaviors and performance appraisal as identifying which of these behaviors are to be subject to evaluation (i.e., deemed valuable by the organization or its agents). The key idea here is that work analytic data, which are purposefully derived to discover the requirements of work role enactment, can therefore be meaningfully brought to bear within any related research that examines the nature of job performance or attempts to account for performance differences across individuals. Furthermore, there are several ways to link work analysis data to performance data at the multiple levels that are typically of interest to organizational researchers (i.e., individual, team, and organizational). Cognitive task analysis approaches could also be potentially useful, as they focus on discovering differences between experts and novices, which is another way to conceptualize antecedents to superior performance. 31

Morgeson and Dierdorff

Copyright American Psychological Association. Not for further distribution.

Work Analysis and Organizational Performance At the organizational or firm level, future research could examine how work analysis contributes to so-called high-performance work practices (HPWP). Such practices have included different recruitment strategies, systematic personnel selection, strategic training, performance management systems, a variety of different compensation systems, use of teams, and HR planning (Huselid, 1995; Pfeffer, 1998), all of which are purported to increase individuals’ knowledge, skills, and motivation for the benefit of the organization (Becker & Huselid, 1998). Several studies have provided supportive empirical evidence that HPWP are associated with organizational performance, including reduced turnover and increased sales, profits, and market value (Delaney & Huselid, 1996; Huselid, 1995; Huselid, Jackson, & Schuler, 1997). The role of work analysis as an element of HPWP, however, has been uneven. In some of the seminal research in the area, work analysis was an explicit dimension of “employee skills and organizational structures” (Huselid, 1995, p. 646). In particular, organizations were assessed in terms of “the proportion of the workforce whose job has been subjected to a formal job analysis” (Huselid, 1995, p. 646). In addition, Delery and Doty (1996, p. 834) focused on the nature of job descriptions in use at the organization (e.g., “the duties of this job are clearly defined,” “this job has an up-to-date job description,” “the job description for this job contains all of the duties performed by individual employees”), which is one fundamental aspect of work analysis. Finally, in more recent research, Toh, Morgeson, and Campion (2008) explored how individual HR practices could be described in terms of coherent bundles of HR practices. In terms of work analysis, these HR practices included the number of selection systems in place that were based on formal work analyses and the number of training programs used that incorporated a careful, systematic training needs analysis. In other HPWP research, however, the role of work analysis has been neglected. From our standpoint, this is an unfortunate omission for two reasons. First, as past research has shown, work analysis is an important component of HPWP. Second, as 32

this chapter has shown, work analysis underlies all of these HPWP, as it provides the vital data required to effectively create and maintain such systems. Thus, an important opportunity is missed in some HPWP research, namely, the chance to move beyond merely capturing whether these practices are used and instead ascertaining how these practices are built (i.e., on what information they are based). An additional point regarding work analysis and HPWP research is that this line of future inquiry would be congruent with previous calls for empirical investigation of strategic work analysis (see Sackett & Laczo, 2003). The ultimate goal of strategic work analysis is to forecast work role requirements of new roles that are expected to exist in the future or current roles that are expected to substantially change (Cronshaw, 1998; B. Schneider & Konz, 1989). This more predictive purpose of strategic work analysis holds particular salience to activities surrounding HR planning, which also involves forecasting various human capital needs. Further, recent research indicates that of the variety of specific practices designated as HPWP, HR planning has the largest effects on organizational performance (Combs, Liu, Hall, & Ketchen, 2006). Thus, future work analysis research that examines topics exclusive to strategic work analysis, as well as how this approach relates to effective HR planning, would be quite valuable.

Extending Work Analysis to the Team Level Virtually all of the past work analysis research has focused on the individual level of analysis. Thus, another potential opportunity exists in extending work analysis research to the team level. In this respect, there are at least three areas of future research. First, work analysis research could pursue the development of a taxonomy of the work role requirements necessary for enacted roles within teams. Although several scholars have identified various requirements needed for team performance (e.g., Campion, Medsker, & Higgs, 1993), these have been primarily concerned with designing teams and determining what characteristics separate effective from ineffective teams. In addition, some attention has been devoted to conducting team task analyses, with a particular emphasis on the importance of

Copyright American Psychological Association. Not for further distribution.

Work Analysis

interdependence within a team (Arthur, Edwards, Bell, Villado, & Bennett, 2005). Thus, the opportunity exists for more systematic and comprehensive work analysis efforts seeking to identify those work role requirements germane to working in teams. Second, work analysis research could examine the validity and generalizability of existing models of team role requirements. For example, Stevens and Campion (1994) outlined 14 different worker requirements (knowledge, skills, abilities) pertinent to teamwork organized into five groupings: (a) conflict resolution, (b) collaborative problem solving, (c) communication, (d) goal setting and performance management, and (e) planning and task coordination. Although this conceptual model was based on an extensive literature review, empirical research examining or using this model remains scarce. However, there is some evidence suggesting these worker requirements do contribute to performance in team settings (Morgeson, Reider, & Campion, 2005; Stevens & Campion, 1994). From a work analysis perspective, it would be interesting to empirically investigate how well this conceptual model actually functions across various work roles that should differ with respect to teamwork characteristics. In other words, one could test whether the model can meaningfully and systematically discriminate between work roles that are embedded in team contexts versus those that are not. For example, one could examine how relationships among the worker requirements specified by the model vary in relation to enacting work roles in more team-oriented contexts (i.e., those with high interdependence, shared goals, etc.). Presumably, the worker requirements specified in the models should be more salient to enacting work roles in more team-oriented contexts. A third topic for future work analysis research to address is how consensus among individual role holders regarding important work role requirements might impact group or unit effectiveness. Role theorists use the term consensus to denote sharedness or agreement among the expectations held by various role holders (Biddle, 1986). At more molar levels, consensus has been thought to lead to more effective integration of social systems (Biddle, 1979) because roles serve the important function of coordinating and integrating the behavior of individuals (Katz &

Kahn, 1978). This has led some work analysis researchers to wonder whether greater consensus could result in overall increases in cross-role-holder effectiveness (Dierdorff & Morgeson, 2007). At the same time, some costs might be associated with too much consensus among individual role holders, such as less innovation and creativity. Future research is needed to examine the potential consequences of consensus, or disagreement, for the effective functioning of units or groups.

The Implications of Role Expectations Conceptualizing work analysis judgments made by role incumbents as representing important expectations regarding how they enact their work roles allows the field of work analysis to expand considerably into other theoretical areas. Role expectations are simply beliefs about what a given role entails (Ilgen & Hollenbeck, 1991) and are important antecedents to role enactment. With regard to work analysis, the content of role expectations is reflected in judgments of various work role requirements (Dierdorff & Morgeson, 2007). Role expectations are important to a number of individual-level outcomes. For example, clarity with regard to one’s work role has substantial positive ramifications for job performance, satisfaction, and organizational commitment (Abramis, 1994; Tubre & Collins, 2000). In addition, the breadth with which individuals define their work roles has been shown to impact job performance (Morgeson, Delaney-Klinger, & Hemingway, 2005). The above findings suggest that examining role expectations in particular is a fruitful avenue for future work analysis research. One area of research in which role expectations are especially relevant is the recent work focused on the effects of role definitions on the performance of organizational citizenship behavior (OCB). Despite its early definition of being extra-role work behavior (Organ, 1988; Organ, Podsakoff, & MacKenzie, 2006), studies have shown that individuals frequently view OCB as falling within the requirements of their work roles (Haworth & Levy, 2001; Hui, Lam, & Law, 2000; Morrison, 1994). As a result, OCB researchers have begun to investigate how OCB role definitions impact whether individuals will engage in OCB (Kamdar, McAllister, & Turban, 2006; McAllister, Kamdar, Morrison, & Turban, 2007). 33

Copyright American Psychological Association. Not for further distribution.

Morgeson and Dierdorff

Work analysis research could meaningfully contribute to this area of inquiry for at least two reasons. First, from its early descriptions, OCB has always been conceptualized as generic work behavior, applicable to a wide variety of work roles (Borman & Penner, 2001). This suggests that such behavior could easily fall within the work role requirements commonly captured in work analysis. For work analysis research, the implication is that OCB research can be meaningfully informed by results from the study of work role requirements. It is interesting to note that the premise that OCB is indeed widely applicable across work roles has yet to be empirically substantiated, which is a question that can be directly addressed by future work analysis research. Second, OCB role definitions in the extant research are typically operationalized by administering to role incumbents the same measurement scale (or with very slight variation) used to capture subsequent performance of OCB. Such operationalizations of role perceptions seek to ascertain whether employees view OCB as part of their work roles and what consequences this role definition may have on ensuing OCB performance. However, these operationalizations do increase the risk of common method bias, as the same scales are used as antecedents and criteria. An alternative way for work analysis research to study whether role perceptions influence the performance of OCB would be to examine how role expectations affect the enactment of OCB. Here, role expectations could better depict how individuals construe the entirety of their work roles (vs. only if OCB is role related) and could expand to capture role perceptions of both activity and attribute requirements. For example, role expectations regarding activities and attributes that are interpersonal in nature could be examined to determine if they predict whether role incumbents engage in OCB as part of their role enactment. This approach to focusing on broader role expectations is also consistent with recent suggestions that an individual’s orientation toward his or her work role is an encompassing concept that can include various facets, such as passive, strategic, and collective orientations (Parker, 2007). The latter role orientation has particular salience to OCB performance 34

because it pertains to how individuals construe their roles with regard to working with others toward goal attainment. Important to note is that one’s orientation and expectations toward one’s work role are known to be shaped by features of the work context, such as autonomy and interdependence (Dierdorff & Morgeson, 2007; Parker, 2007). As mentioned earlier, work context descriptors clearly fall within the scope of work analysis. Thus, future work analysis research could test whether features of the work context moderate the potential relationships between role orientations, role expectations, and OCB.

The Role of Context In addition to the research topics discussed above within different levels of analysis, there are a number of cross-level questions that work analysis research could address. Many of these possible contributions stem from the fact that work analysis research provides a systematic way of describing contextual variables (Dierdorff, 2008). Indeed, several authors have pointed to the difficulties surrounding exactly how to delineate the major elements that comprise context as a major reason for the lack of context-oriented research in I/O psychology and organizational behavior (Hattrup & Jackson, 1996; Johns, 2006). One example of using work analytic data to examine contextual effects can be seen in a recent study by Dierdorff and Ellington (2008). These authors integrated work context information from O*NET into examinations of how the nature of occupational roles shapes whether individuals experience work–family conflict. Other potential areas that could similarly benefit from work analysis research include person– environment fit (e.g., demands–abilities approach; Kristof, 1996) and work design (e.g., moderators of the design characteristics–satisfaction relationship; Morgeson & Humphrey, 2008). In addition to work context, future work analysis research could examine other broader contextual factors, such as the impact of national culture. Findings from recent research have been equivocal on the influence of national culture. For example, one study found very small effects, suggesting that work analysis data are transportable across cultures (Taylor et al., 2008), whereas another study found national culture to be related to the frequency with

Copyright American Psychological Association. Not for further distribution.

Work Analysis

which certain kinds of work behaviors were performed (Shin, Morgeson, & Campion, 2007). These mixed results suggest more research is warranted. In addition to national culture, aspects of organizational culture (e.g., values) would be an example of other broader contextual effects that could be investigated. Finally, it may be interesting to examine how larger changes that are occurring in the broader world of work might shape the outcomes of work analysis. The impact of globalization, prevalence of interorganizational relationships, use of outsourcing, and reliance on information technology for communication are all examples of shifts in the world of work. Such forces are perhaps unlikely to change the primary goal of work analysis—systematically uncovering activities and attributes and the work context in which roles are performed—but rather these forces could very well impact the salience of the various outcomes of work analysis. For example, the products of work analysis may become even more important for HR planning, where an emphasis is placed on assessing current and forecasting future human capital needs. In addition, practitioners of work analysis may need to be aware of broader forces not normally considered related to work analysis concerns (e.g., changes in industry practices, meta-technology [geospatial, biotechnology, etc.], labor economics) to contextualize their findings for use in organizational strategy and HR management decisions.

about work analysis in the course of their ongoing efforts. Perhaps the most pressing issue for future work analysis research, however, is to adopt a more theoretically grounded approach, such that research not only makes a contribution to the practice of work analysis, but also advances theory. Of course, there is considerable historical precedent in work analysis research for such a theoretical grounding. Functional job analysis was based in the data–people–things framework, and the PAQ drew from the stimulus–organism–response paradigm that underlies behaviorism. However, as science has progressed, reliance on such overarching frameworks has decreased in favor of more middlerange theories (Merton, 1949). We feel that middlerange theories hold the most promise for advancing work analysis research. Thus, instead of attempting to develop a theory of work analysis, we advocate that work analysis researchers begin to draw from the numerous theoretical frameworks discussed earlier (e.g., role theory, social and cognitive psychological theory) in their future research efforts. Such an approach is certain to reenergize the field of work analysis.

References Aamodt, M. G., Kimbrough, W. W., Keller, R. J., & Crawford, K. J. (1982). Relationship between sex, race, and job performance level and the generation of critical incidents. Educational and Psychological Research, 2, 227–234.

CONCLUSION: TOWARD A MORE THEORETICAL VIEW OF WORK ANALYSIS

Abramis, D. J. (1994). Work role ambiguity, job satisfaction, and job performance: Meta-analyses and review. Psychological Reports, 75, 1411–1433.

As we hope this chapter has made clear, work analysis not only has a long history in I/O psychology, but also has a promising future. Our goal was to not only review past research, but also point to numerous opportunities for future research, particularly in areas often not traditionally considered to be the purview of work analysis research. We recognize that many researchers conduct work analysis but may not consider themselves work analysis scholars. In a practical sense, then, future work analysis research would be well-served if those involved in substantive research that relies on work analysis data (e.g., selection researchers) were attuned to the opportunities to incorporate research questions

Ackerman, P. L., & Humphreys, L. G. (1990). Individual differences theory in industrial and organizational psychology. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 1, 2nd ed., pp. 223–282). Palo Alto, CA: Consulting Psychologists Press. Advisory Panel for the Dictionary of Occupational Titles. (1993). The new DOT: A database of occupational titles for the twenty-first century (final report). Washington, DC: U.S. Employment Service, U.S. Department of Labor Employment and Training Administration. Aguinis, H., Mazurkiewicz, M. D., & Heggestad, E. D. (2009). Using Web-based frame-of-reference training to decrease biases in personality-based job analysis: An experimental field study. Personnel Psychology, 62, 405–438. 35

Morgeson and Dierdorff

Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975). Arthur, W., Jr., Edwards, B. D., Bell, S. T., Villado, A. J., & Bennett, W. (2005). Team task analysis: Identifying tasks and jobs that are team based. Human Factors, 47, 654–669. Ash, R. A. (1988). Job analysis in the world of work. In S. Gael (Ed.), The job analysis handbook for business, industry and government (Vol. 1, pp. 3–13). New York: Wiley. Ash, R. A., & Edgell, S. L. (1975). A note on the readability of the position analysis questionnaire. Journal of Applied Psychology, 60, 765–766.

Copyright American Psychological Association. Not for further distribution.

Ash, R. A., & Levine, E. L. (1980). A framework for evaluating job analysis methods. Personnel, 57, 53–59. Becker, B. E., & Huselid, M. A. (1998). High performance work systems and firm performance: A synthesis of research and managerial implications. Research in Personnel and Human Resources Journal, 16, 53–101. Biddle, B. J. (1979). Role theory: Expectations, identities, and behavior. New York: Academic Press. Biddle, B. J. (1986). Recent developments in role theory. Annual Review of Sociology, 12, 67–92. Block, J., & Block, J. H. (1981). Studying situational dimensions: A grand perspective and some limited empiricism. In D. Magnusson (Ed.), Toward a psychology of situations: An interactional perspective (pp. 85–103). Hillsdale, NJ: Erlbaum. Borman, W. C., Dorsey, D., & Ackerman, L. (1992). Time-spent responses as time allocation strategies: Relations with sales performance in a stockbroker sample. Personnel Psychology, 45, 763–777. Borman, W. C., & Penner, L. A. (2001). Citizenship performance: Its nature, antecedents, and motives. In B. W. Roberts & R. Hogan (Eds.), Personality psychology in the workplace (pp. 45–61). Washington, DC: American Psychological Association. Bowers, K. S. (1973). Situationalism in psychology: An analysis and critique. Psychological Review, 80, 307–336. Brannick, M. T., Levine, E. L., & Morgeson, F. P. (2007). Job analysis: Methods, research, and applications for human resource management (2nd ed.). Thousand Oaks, CA: Sage. Brendle, M., Rivkin, D. W., & Lewis, P. M. (2008, April). Developing O*NET tools and technology: Information for a changing workplace. In E. Dierdorff & S. Cox (Chairs), Meeting the challenges of occupational analysis in a rapidly changing workplace: Implications for the O*NET system. Symposium conducted at the 23rd annual meeting of the Society for Industrial and Organizational Psychology, San Francisco. Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. (1993). A theory of performance. In N. Schmitt & W. Borman (Eds.), Personnel selection in organizations (pp. 35–70). San Francisco: Jossey-Bass. 36

Campion, M. A., Medsker, G. J., & Higgs, A. C. (1993). Relations between work group characteristics and effectiveness: Implications for designing effective work groups. Personnel Psychology, 46, 823–850. Campion, M. A., Morgeson, F. P., & Mayfield, M. (1999). O*NET’s theoretical contributions to job analysis research. In N. G. Peterson, W. C. Borman, M. D. Mumford, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 297–304). Washington, DC: American Psychological Association. Cascio, W. F., & Aguinis, H. (2008). Research in industrial and organizational psychology from 1963 to 2007: Changes, choices, and trends. Journal of Applied Psychology, 93, 1062–1081. Combs, J. G., Liu, Y., Hall, A. T., & Ketchen, D. J. (2006). How much do high performance work practices matter? A meta-analysis of their effects on organizational performance. Personnel Psychology, 59, 501–528. Conley, P. R., & Sackett, P. R. (1987). Effects of using high- versus low-performing job incumbents as sources of job-analysis information. Journal of Applied Psychology, 72, 434–437. Conte, J. M., Dean, M. A., Ringenbach, K. L., Moran, S. K., & Landy, F. L. (2005). The relationship between work attitudes and job analysis ratings: Do task rating scale type and task discretion matter? Human Performance, 18, 1–21. Converse, P. D., Oswald, F. L., Gillespie, M. A., Field, K. A., & Bizot, E. B. (2004). Matching individuals to occupations using aptitudes and the O*NET: Issues and an application in career guidance. Personnel Psychology, 57, 451–487. Cornelius, E. T., & Lyness, K. S. (1980). A comparison of holistic and decomposed judgment strategies in job analysis by job incumbents. Journal of Applied Psychology, 65, 155–163. Costanza, D. P., Fleishman, E. A., & Marshall-Mies, J. C. (1999). Knowledge. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 71–90). Washington, DC: American Psychological Association. Cronshaw, S. (1998). Job analysis: Changing nature of work. Canadian Psychology, 39, 5–13. Cucina, J. M., Vasilopoulos, N. L., & Sehgal, K. (2005). Personality-based job analysis and the self-serving bias. Journal of Business and Psychology, 20, 275–290. Cunningham, J. W. (1989, August). Discussion. In R. J. Harvey (Chair), Applied measurement issues in job analysis. Symposium presented at the annual meeting of the American Psychological Association, New Orleans, LA.

Work Analysis

Cunningham, J. W. (1996). Generic descriptors: A likely direction in occupational analysis. Military Psychology, 8, 247–262. Cunningham, J. W. (2000). Introduction to a new journal. Ergometrika, 1, 1–23. Delaney, J. T., & Huselid, M. A. (1996). The impact of human resource management practices on performance in for-profit and nonprofit organizations. Academy of Management Journal, 39, 949–969.

Copyright American Psychological Association. Not for further distribution.

Delery, J. E., & Doty, D. H. (1996). Modes of theorizing in strategic human resource management: Tests of universalistic, contingency, and configurational performance predictions. Academy of Management Journal, 39, 802–835. Dierdorff, E. C. (2008, April). Illuminating the “murky ground”: Linking context theory to empirical research. Symposium presented at the 23rd annual meeting of the Society for Industrial and Organizational Psychology, San Francisco. Dierdorff, E. C., Cantwell, A. R., & Nottingham, J. (2008, April). Capturing and defining new and emerging occupations in high-growth sectors. Symposium presented at the 23rd annual meeting of the Society for Industrial and Organizational Psychology, San Francisco. Dierdorff, E. C., & Cox, S. P. (2008, April). Meeting the challenges of occupational analysis in a rapidly changing workplace: Implications for the O*NET system. Symposium presented at the 23rd annual meeting of the Society for Industrial and Organizational Psychology, San Francisco. Dierdorff, E. C., & Ellington, J. K. (2008). It’s the nature of work: Examining behavior-based sources of work–family conflict across occupations. Journal of Applied Psychology, 93, 883–892. Dierdorff, E. C., & Morgeson, F. P. (2007). Consensus in work role requirements: The influence of discrete occupational context on role expectations. Journal of Applied Psychology, 92, 1228–1241. Dierdorff, E. C., & Morgeson, F. P. (2009). Effects of descriptor specificity and observability on incumbent work analysis ratings. Personnel Psychology, 62, 601–628. Dierdorff, E. C., & Rubin, R. S. (2007). Carelessness and discriminability in work role requirement judgments: Influences of role ambiguity and cognitive complexity. Personnel Psychology, 60, 597–625. Dierdorff, E. C., Rubin, R. S., & Morgeson, F. P. (2009). The milieu of managerial work: integrative framework linking context to role requirements. Journal of Applied Psychology, 94, 972–988.

analysis handbook for business, industry, and government (Vol. 2, pp. 1051–1071). New York: John Wiley & Sons. Dunnette, M. D. (1976). Aptitudes, abilities, and skills. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 473–520). Chicago: Rand McNally. Dunnette, M. D. (1999). Introduction. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 3–7). Washington, DC: American Psychological Association. Dye, D., & Silver, M. (1999). The origins of O*NET. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 9–19). Washington, DC: American Psychological Association. Eggerth, D. E., Bowles, S. M., Tunick, R. H., & Andrew, M. E. (2005). Convergent validity of O*NET Holland code classifications. Journal of Career Assessment, 13, 150–168. Fine, S. A., & Getkate, M. (1995). Benchmark tasks for job analysis: A guide for functional job analysis (FJA) scales. Mahwah, NJ: Erlbaum. Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51, 327–358. Fleishman, E. A. (1975). Toward a taxonomy of human performance. American Psychologist, 30, 1127–1149. Fleishman, E. A. (1992). Fleishman-Job Analysis Survey (F-JAS). Potomac, MD: Management Research Institute. Fleishman, E. A., Costanza, D. P., & Marshall-Mies, J. (1999). Abilities. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 175–196). Washington, DC: American Psychological Association. Fried, Y., & Ferris, G. R. (1986). The dimensionality of job characteristics: Some neglected issues. Journal of Applied Psychology, 71, 419–426. Friedman, L. (1990). Degree of redundancy between time, importance, and frequency task ratings. Journal of Applied Psychology, 75, 748–752. Freidman, L. (1991). Correction to Friedman (1990). Journal of Applied Psychology, 76, 366. Gael, S. (1983). Job analysis: A guide to assessing work activities. San Francisco: Jossey-Bass.

Dierdorff, E. C., & Wilson, M. A. (2003). A meta-analysis of job analysis reliability. Journal of Applied Psychology, 88, 635–646.

Gardner, W. L., & Martinko, M. J. (1988). Impression management in organizations. Journal of Management, 14, 321–338.

Drauden, G. M. (1988). Task inventory analysis in industry and the public sector. In S. Gael (Ed.), The job

Gatewood, R. D., & Feild, H. S. (2001). Human resource selection. Orlando, FL: Harcourt Brace. 37

Morgeson and Dierdorff

Ghiselli, E. (1966). The validity of occupational aptitude tests. New York: Wiley. Goffin, R. D., & Woycheshin, D. E. (2006). An empirical method of determining employee competencies/ KSAOs from task-based job analysis. Military Psychology, 18, 121–130.

Copyright American Psychological Association. Not for further distribution.

Goldstein, I. L., Zedeck, S., & Schneider, B. (1993). An exploration of the job analysis–content validity process. In N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations (pp. 3–34). San Francisco: Jossey-Bass.

Hui, C., Lam, S. S. K., & Law, K. K. S. (2000). Instrumental values of organizational citizenship behavior for promotion: A field quasi-experiment. Journal of Applied Psychology, 85, 822–828. Hunter, J. E. (1986). Cognitive ability, cognitive aptitudes, job knowledge, and job performance. Journal of Vocational Behavior, 29, 340–362. Huselid, M. A. (1995). The impact of human resource management practices on turnover, productivity, and corporate financial performance. Academy of Management Journal, 38, 635–672.

Graen, G. B. (1976). Role making processes within complex organizations. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1201–1245). Chicago: Rand McNally.

Huselid, M. A., Jackson, S. E., & Schuler, R. S. (1997). Technical and strategic human resource management effectiveness as determinants of firm performance. Academy of Management Journal, 40, 171–188.

Green, S. B., & Stutzman, T. (1986). An evaluation of methods to select respondents to structured jobanalysis questionnaires. Personnel Psychology, 39, 543–564.

Ilgen, D. R., & Hollenbeck, J. R. (1991). The structure of work: Job design and roles. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2, 2nd ed., pp. 165–207). Palo Alto, CA: Consulting Psychologists Press.

Green, S. B., & Veres, J. G. (1990). Evaluation of an index to detect inaccurate respondents to a task analysis inventory. Journal of Business and Psychology, 5, 47–61. Guion, R. (1993). The need for change: Six persistent themes. In N. Schmitt & W. Borman (Eds.), Personnel selection in organizations (pp. 481–497). San Francisco: Jossey-Bass.

Jeanneret, P. R. (1992). Applications of job component/ synthetic validity to construct validity. Human Performance, 5, 81–96.

Hadden, W. C., Kravets, N., & Muntaner, C. (2004). Descriptive dimensions of US occupations with data from the O*NET. Social Science Research, 33, 64–78.

Jeanneret, P. R., Borman, W. C., Kubisiak, U. C., & Hanson, M. A. (1999). Generalized work activities. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 49–69). Washington, DC: American Psychological Association.

Harvey, R. J. (1991). Job analysis. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2, 2nd ed., pp. 71–163). Palo Alto, CA: Consulting Psychologists Press.

Jeanneret, P. R., & Strong, M. H. (2003). Linking O*NET job analysis information to job requirement predictors: An O*NET application. Personnel Psychology, 56, 423–429.

Harvey, R. J., Friedman, L., Hakel, M. D., & Cornelius, E. T. (1988). Dimensionality of the job element inventory, a simplified worker-oriented job analysis questionnaire. Journal of Applied Psychology, 73, 639–646.

Johns, G. (2006). The essential impact of context on organizational behavior. Academy of Management Review, 31, 386–408.

Hattrup, K., & Jackson, S. E. (1996). Learning about individual differences by taking situations seriously. In K. R. Murphy (Ed.), Individual differences and behavior in organizations (pp. 507–547). San Francisco: Jossey-Bass. Haworth, C. L., & Levy, P. E. (2001). The importance of instrumentality beliefs in the prediction of organizational citizenship behaviors. Journal of Vocational Behavior, 59, 64–75. Hempel, C. G. (1965). Aspects of scientific explanation and other essays in the philosophy of science. New York: The Free Press. Hofmann, D. A., Morgeson, F. P., & Gerras, S. (2003). Climate as a moderator of the relationship between LMX and content specific citizenship: Safety climate as an exemplar. Journal of Applied Psychology, 88, 170–178. 38

Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.). New York: Wiley. Kamdar, D., McAllister, D. M., & Turban, D. B. (2006). “All in a day’s work”: How follower individual differences and justice perceptions predict OCB role definitions and behavior. Journal of Applied Psychology, 91, 841–855. Kozlowski, S. W. J., & Hattrup, K. (1992). A disagreement about within-group agreement: Disentangling issues of consistency versus consensus. Journal of Applied Psychology, 77, 161–167. Kristof, A. L. (1996). Person–organization fit: An integrative review of its conceptualizations, measurement, and implications. Personnel Psychology, 49, 1–49. Landy, F. J., & Vasey, J. (1991). Job analysis: The composition of SME samples. Personnel Psychology, 44, 27–50.

Work Analysis

Copyright American Psychological Association. Not for further distribution.

LaPolice, C. C., Carter, G. W., & Johnson, J. J. (2008). Linking O*NET descriptors to occupational literacy requirements using job component validation. Personnel Psychology, 61, 405–441.

Merton, R. K. (1949). Social theory and social structure: Toward the codification of theory and research. Glencoe, IL: Free Press.

Levine, J., Nottingham, J., Paige, B., & Lewis, P. (2000). Transitioning O*NET to the Standard Occupational Classification. Raleigh, NC: National Center for O*NET Development.

Mischel, W. (1977). The interaction of person and situation. In D. Magnusson & N. S. Endler (Eds.), Personality at the crossroads: Current issues in interactional psychology (pp. 333–352). Hillsdale, NJ: Erlbaum.

Lievens, F., & Sanchez, J. I. (2007). Can training improve the quality of inferences made by raters in competency modeling? A quasi-experiment. Journal of Applied Psychology, 92, 812–819.

Mitchell, J. L. (1988). History of job analysis in military organizations. In S. Gael (Ed.), The job analysis handbook for business, industry, and government (pp. 30–36). New York: Wiley.

Lievens, F., Sanchez, J. I., & De Corte, W. (2004). Easing the inferential leap in competency modeling: The effects of task-related information and subject matter expertise. Personnel Psychology, 57, 881–904.

Mitchell, J. L., Bennett, W., & Strickland, W. J. (1999, May). Current and future trends in job analysis systems and technologies: Studying the world of work in AD 2000 and beyond. Symposium sponsored by the Institute for Job and Occupational Analysis, the U.S. Air Force Research Laboratory, and the Human Resources Research Organization, San Antonio, TX.

Lindell, M. K., Clause, C. S., Brandt, C. J., & Landis, R. S. (1998). Relationship between organizational context and job analysis task ratings. Journal of Applied Psychology, 83, 769–776. Lucia, A., & Lepsinger, R. (1999). The art and science of competency models: Pinpointing critical success factors in organizations. San Francisco: Jossey-Bass. Madden, J. M. (1962). The effect of varying the degree of rater familiarity in job evaluation. Personnel Administrator, 25, 42–46. Madden, J. M. (1963). A further note on the familiarity effect in job evaluation. Personnel Administrator, 26, 52–54. Magnussen, D., & Endler, N. S. (1977). Personality at the crossroads: Current issues in interactional psychology. Hillsdale, NJ: Erlbaum. Manson, T. M., Levine, E. L., & Brannick, M. T. (2000). The construct validity of task inventory ratings: A multitrait–multimethod analysis. Human Performance, 13, 1–22. Marlowe, D., & Crowne, D. P. (1961). Social desirability and response to perceived situational demands. Journal of Consulting Psychology, 25, 109–115. McAllister, D. M., Kamdar, D., Morrison, E. W., & Turban, D. B. (2007). Disentangling role perceptions: How perceived role breadth, discretion, instrumentality and efficacy relate to helping and taking charge. Journal of Applied Psychology, 92, 1200–1211. McCormick, E. J. (1960). Effect of amount of job information required on reliability of incumbents’ check-list reports (WADD-TN-60-142). Lackland Air Force Base, TX: Personnel Research Laboratory, Wright Air Development Division. McCormick, E. J. (1979). Job analysis: Methods and applications. New York: ANACOM. McCormick, E. J., Jeanneret, P. R., & Mecham, R. C. (1972). A study of job characteristics and job dimensions as based on the Position Analysis Questionnaire (PAQ) [Monograph]. Journal of Applied Psychology, 56, 347–368.

Mitchell, J. L., & Driskill, W. E. (1996). Military job analysis: A historical perspective. Military Psychology, 8, 119–142. Morgeson, F. P., & Campion, M. A. (1997). Social and cognitive sources of potential inaccuracy in job analysis. Journal of Applied Psychology, 82, 627–655. Morgeson, F. P., & Campion, M. A. (2000). Accuracy in job analysis: Toward an inference-based model. Journal of Organizational Behavior, 21, 819–827. Morgeson, F. P., Delaney-Klinger, K. A., & Hemingway, M. A. (2005). The importance of job autonomy, cognitive ability, and job-related skill for predicting role breadth and job performance. Journal of Applied Psychology, 90, 399–406. Morgeson, F. P., Delaney-Klinger, K. A., Mayfield, M. S., Ferrara, P., & Campion, M. A. (2004). Self-presentation processes in job analysis: A field experiment investigating inflation in abilities, tasks, and competencies. Journal of Applied Psychology, 89, 674–686. Morgeson, F. P., & Humphrey, S. E. (2008). Job and team design: Toward a more integrative conceptualization of work design. In J. Martocchio (Ed.), Research in personnel and human resource management (Vol. 27, pp. 39–92). Bingley, United Kingdom: Emerald Group Publishing. Morgeson, F. P., Reider, M. H., & Campion, M. A. (2005). Selecting individuals in team settings: The importance of social skills, personality characteristics, and teamwork knowledge. Personnel Psychology, 58, 583–611. Morrison, E. W. (1994). Role definitions and organizational citizenship behavior: The importance of the employee’s perspective. Academy of Management Journal, 37, 1543–1567. Mullins, W. C., & Kimbrough, W. W. (1988). Group composition as a determinant of job analysis outcomes. Journal of Applied Psychology, 73, 657–664. 39

Morgeson and Dierdorff

Mumford, M. D., & Peterson, N. G. (1999). The O*NET content model: Structural considerations in designing jobs. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 21–30). Washington, DC: American Psychological Association. Münsterberg, H. (1913). Psychology and industrial efficiency. Boston: Houghton Mifflin. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill.

Copyright American Psychological Association. Not for further distribution.

Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: Lexington Books. Organ, D. W., Podsakoff, P. M., & MacKenzie, S. B. (2006). Organizational citizenship behavior: Its nature, antecedents and consequences. Beverly Hills, CA: Sage. Parker, S. K. (2007). “That is my job”: How employees’ role orientation affects their job performance. Human Relations, 60, 403–434. Peterson, N. G., Mumford, M. D., Borman, W. C., Jeanneret, P. R., & Fleishman, E. A. (1999). An occupational information system for the 21st century: The development of O*NET. Washington, DC: American Psychological Association. Peterson, N. G., Mumford, M. D., Borman, W. C., Jeanneret, P. R., Fleishman, E. A., Campion, M. A., et al. (2001). Understanding work using the occupational information network (O*NET): Implications for practice and research. Personnel Psychology, 54, 451–492. Pfeffer, J. (1998). Seven practices of successful organizations. California Management Review, 40, 96–124. Pine, D. E. (1995). Assessing the validity of job ratings: An empirical study of false reporting in task inventories. Public Personnel Management, 24, 451–459. Podsakoff, P. M., MacKenzie, S. M., Bachrach, D. G., & Podsakoff, N. P. (2005). The influence of management journals in the 1980’s and 1990’s. Strategic Management Journal, 26, 473–488. Prien, E. P., & Saleh, S. D. (1963). A study of bias in job analysis and evaluation. Journal of Industrial Psychology, 1, 113–117. Prien, K., Prien, E., & Wooten, W. (2003). Interrater reliability in job analysis: Differences in strategy and perspective. Public Personnel Management, 32, 125–141. Primoff, E. S., & Fine, S. A. (1988). A history of job analysis. In S. Gael (Ed.), The job analysis handbook for business, industry, and government (pp. 14–29). New York: Wiley. Raymark, P. H., Schmit, M. J., & Guion, R. M. (1997). Identifying potentially useful personality constructs for employee selection. Personnel Psychology, 50, 723–736. 40

Reiter-Palmon, R., Brown, M., Sandall, D. L., Buboltz, C., & Nimps, T. (2006). Development of an O*NET web-based job analysis and its implementation in the U.S. Navy: Lessons learned. Human Resource Management Review, 16, 294–309. Richman, W. L., & Quiñones, M. A. (1996). Task frequency rating accuracy: The effect of task engagement and experience. Journal of Applied Psychology, 81, 512–524. Sackett, P. R., & Laczo, R. M. (2003). Job and work analysis. In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.), Comprehensive handbook of psychology: Industrial and organizational psychology (Vol. 12, pp. 21–37). New York: John Wiley & Sons. Sager, C. E., Mumford, M. D., Baughman, W. A., & Childs, R. A. (1999). Occupation-specific descriptors: Approaches, procedures, and findings. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 227–236). Washington, DC: American Psychological Association. Sanchez, J. I. (1994). From documentation to innovation: Reshaping job analysis to meet emerging business needs. Human Resource Management Review, 4, 51–74. Sanchez, J. I., & Fraser, S. L. (1992). On the choice of scales for task analysis. Journal of Applied Psychology, 77, 545–553. Sanchez, J. I., & Levine, E. L. (1989). Determining important tasks within jobs: A policy-capturing approach. Journal of Applied Psychology, 74, 336–342. Sanchez, J. I., & Levine, E. L. (1994). The impact of rater’s cognition on judgment accuracy: An extension into the job analysis domain. Journal of Business and Psychology, 9, 47–57. Sanchez, J. I., & Levine, E. L. (1999). Is job analysis dead, misunderstood, or both? New forms of work analysis and design. In A. Kraut & A. Korman (Eds.), Evolving practices in human resource management (pp. 43–68). San Francisco: Jossey-Bass. Sanchez, J. I., & Levine, E. L. (2000). Accuracy or consequential validity: Which is the better standard for job analysis data? Journal of Organizational Behavior, 21, 809–818. Sanchez, J. I., & Levine, E. L. (2001). The analysis of work in the 20th & 21st centuries. In N. Anderson, D. S. Ones, H. K. Sinangil, & C. Viswesvaran (Eds.), Handbook of industrial, work and organizational psychology (Vol. 1, pp. 71–89). Thousand Oaks, CA: Sage. Schippmann, J. S. (1999). Strategic job modeling: Working at the core of integrated human resources. Mahwah, NJ: Erlbaum. Schippmann, J. S., Ash, R. A., Battista, M., Carr, L., Eyde, L. D., Hesketh, B., et al. (2000). The practice of competency modeling. Personnel Psychology, 53, 703–740.

Work Analysis

Schmidt, F. L., & Hunter, J. E. (1977). Development of a general solution to the problem of validity generalization. Journal of Applied Psychology, 62, 529–540.

Terborg, J. R. (1981). Interactional psychology and research on human behavior in organizations. Academy of Management Review, 6, 569–576.

Schmitt, N., & Cohen, S. A. (1989). Internal analyses of task ratings by job incumbents. Journal of Applied Psychology, 74, 96–104.

Tesluk, P. E., & Jacobs, R. R. (1998). Toward an integrated model of work experience. Personnel Psychology, 51, 321–355.

Schneider, B., & Konz, A. M. (1989). Strategic job analysis. Human Resource Management, 28, 51–63.

Tett, R. P., & Burnett, D. D. (2003). A personality traitbased interactionist model of job performance. Journal of Applied Psychology, 88, 500–517.

Copyright American Psychological Association. Not for further distribution.

Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: I. Detection, search and attention. Psychological Review, 84, 1–66. Shin, S. J., Morgeson, F. P., & Campion, M. A. (2007). What you do depends on where you are: Understanding how domestic and expatriate work requirements depend upon the cultural context. Journal of International Business Studies, 38, 64–83. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420–428. Siddique, C. M. (2004). Job analysis: A strategic human resource management practice. International Journal of Human Resource Management, 15, 219–244. Silverman, S. B., Wexley, K. N., & Johnson, J. C. (1984). The effects of age and job experience on employee responses to a structured job analysis questionnaire. Public Personnel Management, 13, 355–359. Smith, J. E., & Hakel, M. D. (1979). Convergence among data sources, response bias, and reliability and validity of a structured job analysis questionnaire. Personnel Psychology, 32, 677–692. Society for Industrial and Organizational Psychology. (2003). Principles for the validation and use of personnel selection procedures (4th ed.). Bowling Green, OH: Author. Stevens, M. J., & Campion, M. A. (1994). The knowledge, skill, and ability requirements for teamwork: Implications for human resource management. Journal of Management, 20, 503–530. Stewart, G. L., Fulmer, I. S., & Barrick, M. R. (2005). An exploration of member roles as a multilevel linking mechanism for individual traits and team outcomes. Personnel Psychology, 58, 343–365. Strong, M. H., Jeanneret, P. R., McPhail, S. M., Blakley, B. R., & D’Egidio, E. L. (1999). Work context: Taxonomy and measurement of the work environment. In N. G. Peterson, M. D. Mumford, W. C. Borman, P. R. Jeanneret, & E. A. Fleishman (Eds.), An occupational information system for the 21st century: The development of O*NET (pp. 127–145). Washington, DC: American Psychological Association. Taylor, P. J., Li, W., Shi, K., & Borman, W. C. (2008). The transportability of job information across countries. Personnel Psychology, 61, 69–111.

Tett, R. P., & Guterman, H. A. (2000). Situation trait relevance, trait expression, and cross-situational consistency: Testing a principle of trait activation. Journal of Research in Personality, 34, 397–423. Toh, S. M., Morgeson, F. P., & Campion, M. A. (2008). Human resource configurations: Investigating fit with the organizational context. Journal of Applied Psychology, 93, 864–882. Tubre, T. C., & Collins, J. M. (2000). Jackson and Schuler (1985) revisited: A meta-analysis of the relationships between role ambiguity, role conflict, and job performance. Journal of Management, 26, 155–169. Tversky, A., & Kahneman, D. (1974, September 27). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131. Uhrbrock, R. S. (1922). The history of job analysis. Administration, 3, 164–168. Van Iddekinge, C. H., Putka, D. J., Raymark, P. H., & Eidson, C. E. (2005). Modeling error variance in job specification ratings: The influence of rater, job, and organization-level factors. Journal of Applied Psychology, 90, 323–334. Voskuijl, O. F., & van Sliedregt, T. (2002). Determinants of interrater reliability of job analysis: A meta-analysis. European Journal of Psychological Assessment, 18, 52–62. Wang, N. (2003). Examining reliability and validity of job analysis survey data. Journal of Applied Measurement, 4, 358–369. Wexley, K. N., & Silverman, S. B. (1978). An examination of differences between managerial effectiveness and response patterns on a structured job analysis questionnaire. Journal of Applied Psychology, 63, 646–649. Wilson, M. A. (2006). A history of job analysis. In L. Koppes (Ed.), Historical perspectives in industrial and organizational psychology (pp. 219–242). Mahwah, NJ: Erlbaum. Wilson, M. A., Harvey, R. J., & Macy, B. A. (1990). Repeating items to estimate the test–retest reliability of task inventory ratings. Journal of Applied Psychology, 75, 158–163. Woehr, D. J., & Huffcutt, A. I. (1994). Rater training for performance appraisal: A quantitative review. Journal of Occupational and Organizational Psychology, 67, 189–205.

41

CHAPTER 2

RECRUITMENT: A REVIEW OF RESEARCH AND EMERGING DIRECTIONS

Copyright American Psychological Association. Not for further distribution.

Brian R. Dineen and Scott M. Soltis

As the first decade of the 21st century comes to a close, companies are challenged as never before to attain the necessary human and social capital to develop, maintain, and increase their competitive advantage. Although organizations continue to wage the “war for talent,” as it has popularly been phrased (e.g., Lavelle, 2003), this contest has increased both in scope and complexity, with recruiters claiming difficulty in finding good workers and acquiring talent (see Ployhart, 2006). The result is a fascinating yet multiplex environment that holds great research potential. For example, the poaching of employees has become a more public phenomenon, especially in the technology sector, as start-ups frequently court employees at established firms such as Google and Microsoft (Delaney, 2007; see also Gardner, 2005). Even more apparent is the recent boom of social networking sites designed for domestic and international job seekers such as linkedin.com and doostang.com (McConnon, 2007). Recruiters interested in poaching passive job candidates might use these sites by first developing relationships with candidates before luring them away from competitors (Cappelli, 2001; Lievens & Harris, 2003). Thus Web technology and the increased pace of recruitment activities have leveled the information playing field and have gained importance recently as intriguing research topics. This chapter is devoted to reviewing the research that might inform this challenging landscape and also proposes new areas of research spawned by new developments and related anecdotal reports (e.g., Billsberry, 2007). Specifically, our goals are to

(a) present a detailed model of the recruitment process; (b) provide a selective review of recent research pertaining to the context, strategies, and processes associated with the stages depicted; and (c) suggest several future avenues for recruitment research. We view recruitment as a process (Rynes, 1991; Rynes & Cable, 2003) and thus define it as the actions organizations take to generate applicant pools, maintain viable applicants, and encourage desired candidates to join those organizations. This chapter is intended to complement reviews of the recruitment literature, both previous (Barber, 1998; Ehrhart & Ziegert, 2005; Rynes & Cable, 2003) and recent (Breaugh, 2008; Breaugh, Macan, & Grambow, 2008). A general theme of these reviews has been to recommend a more comprehensive examination of recruitment. Specifically, Barber (1998) and Rynes and Cable (2003) called for more integration of the context in which recruitment occurs. Breaugh et al. downplayed contextual issues and suggested exploring recruitment stages integratively, while also calling for more nuanced approaches to the study of certain topics. Viewing these excellent reviews as a foundation, we add our unique insights, cover additional studies, and call attention to new or expanded research areas. We focus mainly on research that has occurred since Rynes and Cable’s review and present their conclusions in Exhibit 2.1 as a means of placing our review in the context of previous research findings. As will become evident, these previous findings are organized thematically in

http://dx.doi.org/10.1037/12170-002 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

43

Dineen and Soltis

Exhibit 2.1 Summary of Findings Reported by Rynes and Cable (2003)

Copyright American Psychological Association. Not for further distribution.

Environmental and Contextual Considerations Firm ■ Location, size, and organizational image are important factors in job seekers’ application decisions. ■ Organizational reputation or image is highly correlated with organizational familiarity and moderately correlated with profitability and industry. ■ The most likely routes to improving organizational image are to improve familiarity and to increase the amount of information available to applicants. ■ Organizational image appears to be important to applicant decisions both because it sends signals about more specific job attributes and because it influences expected pride from membership (i.e., social identity). ■ Other organizational-level characteristics, particularly size and industry, are used to make inferences about more specific vacancy characteristics. ■ Some of the main determinants of perceived person–organization fit are the same as factors influencing perceived organizational image. ■ Applicants’ preinterview beliefs about organizations affect their interview performance and impressions. Applicants with positive preinterview exhibit more positive impression management behaviors, ask more positive confirmatory questions, and perceive recruiter behaviors more positively. ■ In general, affirmative action policies are perceived positively by those who might benefit from them and negatively by white males. ■ Negative reactions to affirmative action can be minimized by placing a strong emphasis on merit (e.g., affirmative action as tiebreaker policies) and explaining the reasons behind the policy. ■ Although there are some organizational characteristics that are widely favored by most job seekers (e.g., fairness, high pay), the strength—and sometimes direction—of preferences varies according to individual differences in values, personality, or beliefs.

Vacancy ■ Pay and benefits are of at least moderate importance in job choice. However, importance varies across individuals and market characteristics. ■ In general, college students prefer high pay levels, pay raises based on individual rather than team performance, fixed rather than variable pay, and flexible rather than fixed benefits. ■ Job challenge and interesting work appear to be particularly important to students who have exhibited high academic and social achievement. ■ High pay levels, strong promotion opportunities, and performance-based pay are relatively more important to students with high levels of social achievement (e.g., extracurricular activity, offices). ■ High academic achievers (i.e., those with high GPA and test scores) are more attracted by commitment-based employment philosophies than are high social achievers. ■ Organizations appear to modify vacancy characteristics in reactive rather than strategic fashion, thus limiting potential recruitment effectiveness. Labor Market ■ High-quality applicants (i.e., as assessed by grades and number of job offers) generally appear to be more critical of recruiting practices (e.g., recruiters, recruiting delays). However, those with greater work experience may be slightly more forgiving. Generating Viable Candidates Targeting Strategies ■ White males still have better access than other groups to informal sources of referral. ■ Social referrals are still unequal by race and gender, and they have effects on employment outcomes. ■ Job seekers’ social networks explain variance in job choices over and above general preferences and specific academic preparation.

Messaging Strategies ■ Results regarding recruitment source effects are inconsistent across studies. Even the strongest conclusion from research conducted before 1991—that informal sources are superior to formal ones in terms of posthire outcomes—appears to be open to question. ■ Sources differ in terms of the types of applicants they produce and the amount of information they appear to provide. However, the precise nature of these differences varies across studies.

44

Recruitment: A Review of Research and Emerging Directions

Exhibit 2.1 (Continued) ■

■

■ ■

Copyright American Psychological Association. Not for further distribution.

■

Individuals often use more than one source in locating and applying for jobs. The typical practice of coding only one source is problematic and can have a substantial effect on study results. The same source (e.g., the Internet) can be used in different ways by different employers. Thus, the types of applicants attracted and the amount of information associated with the same source can also vary dramatically across employers. Realistic job previews (RJPs) are associated with consistent, but small, increases in employee retention. RJPs do not appear to cause greater applicant self-selection out of the applicant process. The issue of whether different types of employees self-select as a result of RJPs remains unexamined. Applicants appear to go through two phases of job search, as follows: (a) a broad, exploratory phase in which general information is sought mostly through formal sources and (b) a more focused stage in which informal sources are increasingly used to gain detailed information about a small subset of identified alternatives.

Maintaining Status of Viable Applicants Screening Considerations ■ Applicant reactions to selection procedures can be explained largely in terms of perceived fairness or justice. ■ In general, applicants appear to accept the use of cognitive ability tests in selection. ■ Although there are sometimes differences in perceived test fairness across demographic groups, there is little evidence that the use of testing causes job seekers to drop out of applicant pools. ■ In campus recruiting contexts, delays between recruitment phases can cause significant dropout from applicant pools. Dropout will probably be most severe among applicants with the most opportunities. ■ In other types of labor markets, dropout due to delays may be heaviest among those who need immediate employment.

Interactions With Organizational Agents ■ Recruiters can make a difference to applicants’ job choices, particularly at the extremes of recruiter effectiveness. However, recruiter effects are typically overshadowed by job and organizational attributes. ■ Line recruiters and representatives met on site visits are more influential (in either direction) than staff recruiters and representatives met on campus. ■ Applicants regard trained recruiters as somewhat more effective than untrained ones, although the effects on job choices are probably not large. ■ Trained recruiters are more likely to follow a standardized protocol in interviews and to ask more screening-related questions. Thus, they are probably likely to produce more valid selection decisions. ■ Although applicants like recruiters who spend more time recruiting than selecting, attraction to the job itself may suffer if recruitment is overemphasized relative to selection. ■ Recruiter characteristics are often used to make inferences about organizational and job characteristics and likelihood of receiving an offer. ■ Recruiters and other organizational representatives are often mentioned as sources of applicant beliefs about person– organization fit. ■ Recruiter behaviors (particularly warmth) have a clear effect on applicant interview performance. Applicant behaviors have much less effect on recruiter behaviors, suggesting that recruiters have much more control over interview processes and outcomes than do applicants.

Note. From Handbook of Psychology: Industrial and Organizational Psychology (Vol. 12, p. 69), by W. C. Borman, D. R. Ilgen, and R. J. Klimoski (Eds.), 2003, Hoboken, NJ: Wiley. Copyright 2003 by Wiley. Adapted with permission.

Exhibit 2.1 according to the recruitment process framework we introduce next. FRAMEWORK OF THE RECRUITMENT PROCESS Keeping previous conclusions in mind, we present the framework shown in Figure 2.1 as a guide to the current review. Figure 2.1 integrates Barber’s (1998) and

Breaugh et al.’s (2008) sequential stage models and also integrates Rynes and Cable’s (2003) more explicit consideration of environmental and contextual issues as well as key process issues. Specifically, Figure 2.1 illustrates three primary recruitment stages: generating viable candidates, maintaining the status of viable candidates, and achieving closure in terms of encouraging desired candidates to accept job offers and join the organization. Two key decision points demarcate 45

Copyright American Psychological Association. Not for further distribution.

Dineen and Soltis

FIGURE 2.1. Framework of the recruitment process. PE = person–environment; ELM = elaboration likelihood model.

these three stages: the job seeker’s decision to formally enter the selection process (i.e., application decision) and the organization’s decision to formally invite the applicant to join the company (i.e., job offer). Specifically, Figure 2.1 shows that before job seekers decide to apply, organizations work to generate viable job candidates (Stage 1). Figure 2.1 suggests various targeting strategies related to timing and types of candidates to target, as well as messaging strategies such as diversity advertising and use of various recruitment sources such as the Web. Exhibit 2.1 shows research that had occurred within this stage as of the Rynes and Cable (2003) review, including limited work on targeting and messaging strategies that yielded little in the way of consistent results. Once job seekers actually apply, organizations must then focus on maintaining the status of the most viable applicants (Stage 2). At this stage, timeliness and perceived fairness of the selection process become important considerations, as do interactions with organizational agents such as recruiters (which can also occur before the job seeker applies, as indicated in Figure 2.1) and site visits (which can also occur after the job offer). Here, previous research 46

(Exhibit 2.1) indicates notable progress in assessing recruiter effects and somewhat limited progress in terms of the recruitment effects related to applicant screening processes. After organizations formally offer positions, they must persuade candidates to join the organization (Stage 3). For example, timing issues again become important in terms of windows of opportunity for persuading job seekers who have competing offers. Virtually no work had occurred within this stage at the time of Rynes and Cable’s (2003) review, as reflected in Exhibit 2.1. The bottom of Figure 2.1 also portrays key process variables that are particularly relevant at each stage. For example, at Stage 1, recruiters’ social networking might help identify viable candidates to target, and job seekers’ information processing determines how well various messaging strategies might work. At Stage 2, communication and signaling become more important as job seekers learn more details about the position. At Stage 3, processes such as job choice decision making and negotiation are likely to be key. Finally, overlaid across the entire recruitment process are environmental and contextual consider-

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

ations, as illustrated at the top of Figure 2.1. Specifically, we view certain firm-level characteristics as well as vacancy and labor market characteristics as potentially influencing core considerations at any of the three stages. For example, the labor market, firm brand equity, or the nature of the vacancy (e.g., whether an organization considers it to be core or peripheral to their mission) may drive the use of targeting or messaging strategies or exploding offers (i.e., job offers that are rescinded if not accepted within a predetermined time period). A fair amount of work had occurred in the area of environmental and contextual characteristics at the time of Rynes and Cable’s (2003) review (Exhibit 2.1), especially in terms of firm and specific vacancy characteristics. However, these scholars indicated the need for continued work along these lines. We recognize that not all of the categories in Figure 2.1 have necessarily received sufficient research attention either prior to or during the time covered in our review to draw definitive conclusions. In those cases we discuss the issue (e.g., recruiting older candidates) and discuss directions in which we would like to see the research move. In other areas that have received more attention (e.g., firm reputation, recruiter interactions), we selectively review pertinent findings to update the literature. We begin by reviewing the literature pertaining to key environmental and contextual considerations that potentially affect all three recruitment stages. ENVIRONMENTAL AND CONTEXTUAL CONSIDERATIONS Nearly 20 years ago, Rynes and Barber (1990) called for more focus on contingencies in the recruitment process, but progress has only recently been evident. As Figure 2.1 shows, we posit that recruitment activities are contingent on characteristics of the recruiting firm, the specific vacancy, and the prevailing labor market. Each potentially plays an important role throughout the stages of recruitment, and we discuss them in turn.

Firm Characteristics Following earlier work attempting to link the recruitment and marketing literatures (Maurer, Howe, &

Lee, 1992), studies in recent years have increasingly examined firm characteristics in the context of recruitment. Specifically, characteristics receiving attention have been brand image (Collins & Stevens, 2002), organizational image (Chapman, Uggerslev, Carroll, Piasentin, & Jones, 2005), reputation (Turban & Cable, 2003), firm personality (Slaughter, Zickar, Highhouse, & Mohr, 2004), and firm knowledge (Cable & Turban, 2001). Although earlier work reflected in Exhibit 2.1 also examined some of these characteristics, researchers have begun to take a more nuanced perspective on these topics by integrating marketing concepts to a greater degree. For example, Turban and Cable (2003), after controlling for industry, number of recruiters available to interview, and interview date, found links between firm reputation and applicant pool outcomes including increased numbers of applicants generated (ΔR2 = .02, a 5% increase above control variables) and higher-quality applicants and interviewees produced (in terms of grade point average; ΔR2 = .03, a 14% increase for applicants; and ΔR2 = .06, a 30% increase for interviewees). Cable and Turban (2003) found that job seekers recalled significantly more information about familiar firms (β = .20). They also found that applicants’ perceptions that a company had a favorable reputation increased the pride they anticipated from joining the organization (β = .28), which translated into lower salary requirements to work for higher-reputation firms (β = −.19). Collins and Stevens (2002) found applicant pool size and quality to be influenced by brand equity generated through sponsorships, wordof-mouth, publicity, and general recruitment advertising (and combinations thereof). Collins and Han (2004) found that all but sponsorship influenced brand image via general attitudes toward a company or perceived job attributes. More specifically, Collins and Han focused on differences in effects of lowinvolvement recruitment strategies (i.e., requiring little or no search and processing effort; e.g., visual stimuli, sponsorships of sporting events) versus high-involvement strategies (i.e., greater cognitive processing effort; e.g., detailed description of joborganization characteristics). They found that lowinvolvement strategies are beneficial primarily in terms of applicant pool size and quality when a firm 47

Copyright American Psychological Association. Not for further distribution.

Dineen and Soltis

has not already invested in corporate advertising or does not already have a good reputation. Conversely, a firm that uses high-involvement practices must have already established awareness in the minds of job seekers, either through a priori advertising or reputation-building efforts (Δ adjusted R2 interaction effects ranged from .09 to .13 for advertising and reputation, representing increases of between 24% and 37% over main effects and industry and company size control variables). Complementing these studies and once again moving beyond previous findings in Exhibit 2.1, Collins (2007) found that product awareness moderates the influence of high- or low-involvement recruitment practices, such that low-involvement practices enhance application intentions and actual application decisions but are more effective when product awareness is low. When product awareness is high, high-involvement practices are more effective in influencing these outcomes (ΔR2 for the block of interactions = .22 for application intentions, representing a 76% increase over main effects and control variables; Δ Cox and Snell R2 = .12 for application decisions, representing a 55% increase).

Vacancy Characteristics The nature of the vacancy itself also continues to be an important contextual factor (Rynes & Barber, 1990). Chapman et al. (2005) provided metaanalytic evidence that job-organizational factors were among the stronger predictors of recruitment outcomes, including particularly strong effects of type of work on pursuit intentions (ρ = .53). They also noted that pay (ρ = .15) and the combination of compensation and advancement (ρ = .14) predict pursuit intentions to a much lesser degree than many other vacancy characteristics. However, earlier work (Williams & Dreher, 1992) suggested that although pay does not influence the number of applicants, it could influence job acceptance rates. This importance is underscored by the increased effect pay has on acceptance intentions (ρ = .28; Chapman et al., 2005). Studies also have begun to consider the role of specific non-pay inducements such as work–life benefits, flexible work, and career paths on job pursuit intentions and perceptions of the organization. 48

Casper and Buffardi (2004) examined work–life benefits and flexible work schedules using a wide sample of job seekers and new hires recently starting a job. The authors found that flexible work schedules and dependent care assistance influenced both job pursuit intentions (β = .27, β = .21) and perceived organizational support (β = .28, β = .43). Carless and Wintle (2007) found that flexible (M = 4.07) or dual career paths (M = 3.69) were significantly more attractive than traditional career paths (M = 2.83, all on a 5-point scale). Rau and Hyland (2002) drew on boundary theory (Ashforth, Kreiner, & Fugate, 2000) and found that individuals with higher role conflict found flexible work schedules relatively more attractive, likely because they reduce the cost of role transitions and thus ease role conflict. However, those with lower role conflict found telecommuting relatively more attractive, likely because it increases boundary flexibility and reduces transition costs across role boundaries, whereas those experiencing greater role conflict found such blurring of roles to be unattractive. These findings challenge the assumption that job seekers universally desire telecommuting and flexible work arrangements. Several studies span the firm and vacancy contextual categories identified in Figure 2.1. Lievens and Highhouse (2003) and Slaughter et al. (2004) compared traditional instrumental factors related to the vacancy (e.g., pay, benefits) with symbolic characteristics (e.g., “personality”) of firms. In studies of students and bank employees, Lievens and Highhouse found that organization trait inferences predicted attraction above job-organizational factors (ΔR2 = .09, a 22% increase, and ΔR2 = .18, a 53% increase, in these respective samples). They also found more differentiation among organizations based on trait inferences (i.e., vs. job-organizational attributes), suggesting room for competitive advantage based on symbolic factors such as brand image. Slaughter et al. validated a five-dimensional measure of organizational personality (e.g., style, thrift, boy scout, dominance, innovativeness), finding all but the dominance factor to relate to attraction (R2 = .32 with thrift negatively related) when modeled simultaneously. That organizations are better able to distinguish themselves in terms of symbolic factors compared

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

with more instrumental vacancy-based factors makes sense considering the increased transparency of instrumental factors such as pay via sources such as the Web (e.g., http://www.salary.com). Indeed, researchers have suggested that the Web levels the playing field for job seekers and lowers costs of searching for comparative information about instrumental vacancy characteristics (Lievens & Harris, 2003; Rynes & Cable, 2003). In general, the tension between symbolic and instrumental factors highlights that recruitment research should incorporate multiple predictors as a way of gaining a more realistic picture of the overall process. In terms of specifically examining the impact of symbolic factors, considerable work remains. For example, much research has focused on comparing firms that have positive images with firms that have no discernable images but has failed to account for temporal effects related to image. Thus, research is needed that examines the short- and long-term effects of a negative image or reputation or the mechanisms that might explain changes in image or reputation perceptions over time. As an exception, Brooks, Highhouse, Russell, and Mohr (2003) showed that familiar companies engender more polarized reactions (in either positive or negative directions) than do unfamiliar companies, thus calling into question the overall relationship between familiarity and reputation or attraction past work has suggested (Barber, 1998). A pragmatic reason for researchers to focus on firms with low images or reputations is that such firms have greater need for prescriptive recruitment advice, whereas high-reputation firms often enjoy a steady stream of candidates, making selection rather than recruitment key to overall staffing utility. In general, research is needed regarding how to affect or leverage a firm’s reputation or image. For example, the research presented earlier might have implications for small firms’ recruitment strategies (Williamson, Cable, & Aldrich, 2002). A useful starting point might be Cable and Turban’s (2001) flow diagram for leveraging various levels of image, familiarity, and reputation. The diagram prescribes actions (e.g., maintain low profile, modify or correct employer image using experiential information sources) based on firm familiarity, reputation, and accuracy of understanding of employer image.

Overall, progress is evident over the past few years in examining vacancy characteristics and characteristics that span firm and vacancy categories, compared with the prior piecemeal conclusions shown in Exhibit 2.1. The Chapman et al. (2005) meta-analysis provided a much-needed synthesis of prior findings in these areas, and the recent consideration of flexible work and holistic examinations of symbolic and instrumental characteristics are encouraging yet still leave open the need for continued research attention.

Labor Market Characteristics Consistent with Rynes and Barber (1990), we believe the literature would benefit if researchers more extensively embraced labor market characteristics as a key contextual aspect of recruitment (Billsberry, 2007) that is likely to affect relationships across stages. Cappelli (2005) concluded that businesses cannot know whether a labor shortage is likely someday, but they certainly cannot expect a labor surplus in the foreseeable future. However, this prediction, combined with recent unemployment levels and fluctuations in the global job market, implies the need to reconsider and better customize general recruitment strategies to match labor’s current supply and demand. For example, levels of internal demand and external supply of candidates might dramatically alter recruitment strategies; job seekers who perceive that they have more or fewer choices are likely to react differently to recruitment stimuli. Also, the nature of the labor market has shifted toward what has been termed a free agent market (Rynes & Cable, 2003). This raises the interesting question of just how much effort people put into their job search when they expect that their tenure may be short. Finally, labor supply diversity has the potential to dramatically alter recruitment strategies and approaches in the United States (e.g., an aging workforce, more Hispanic workers). As was the case at the time of Rynes and Cable’s review (Exhibit 2.1), there continues to be gaps in our understanding in this area. Keeping these critical firm, vacancy, and labor market contextual factors in mind, we turn to a discussion of the three primary recruitment stages. We present key processes and outcomes that we believe 49

Dineen and Soltis

are most relevant to each stage while acknowledging that these processes and outcomes are not necessarily exhaustive or bound solely to a particular stage.

Copyright American Psychological Association. Not for further distribution.

GENERATING VIABLE CANDIDATES The generation of viable job candidates (Stage 1 in Figure 2.1) greatly determines the potential utility of the remainder of the staffing process (Barber, 1998). As will be seen as the remainder of our review unfolds, this stage of the recruitment process has received the greatest research attention. In accordance with Figure 2.1, we explore general targeting strategies (i.e., whom and when to target) and the messages embedded in recruitment (i.e., where and how to target). Key processes within this stage include social networking between recruiters and potential applicants and the type of information processing in which job seekers engage. Important outcomes at this stage should include high-quality and/or diverse applicant pools (Carlson, Connerley, & Mecham, 2002), along with building relationships with potential candidates. Thus, the ratio of viable candidates to total applicants is critical, as are other yield ratios (e.g., number of candidates ultimately hired to number of applicants). However, much previous and current research has focused on attraction as the key mechanism for generating an applicant pool. For example, Chapman et al. (2005) identified six antecedent categories that have been related to attraction in varying degrees (job and organizational attributes, ρ = .39; recruiter characteristics, ρ = .29; perceptions of the recruitment process, ρ = .42; perceived fit, ρ = .45; perceived alternatives, ρ = .16; and hiring expectancies, ρ = .33). Although this key outcome has not changed dramatically, research into what makes an organization attractive has begun to reflect contemporary trends by examining antecedents such as organizational ecological reputation (Aiman-Smith, Bauer, & Cable, 2001) and work–life benefits (Carless & Wintle, 2007). As discussed earlier, Carless and Wintle found a significant link between various work–life benefits and organizational attraction. Aiman-Smith et al. found that ecological rating more strongly affected attractiveness (β = .34) than did pay (β = .28), promotion opportunity (β = .23), or layoff policy (β = .29). 50

An outcome closely aligned yet distinct from attraction, job pursuit intentions captures the extent to which an individual will actively strive to join the organization. For the job seeker, attraction is more passive, but pursuit intentions indicate a more active mind-set with regard to vying for a position. Job seekers might be attracted to a company but may perceive that they are underqualified for the advertised position and decide not to pursue it. The Chapman et al. (2005) meta-analysis also suggested that most of the antecedents of attraction have received various levels of research attention and are also important for affecting job pursuit intentions (job and organizational attributes, ρ = .38; recruiter characteristics, ρ = .36; perceptions of the recruitment process, ρ = .27; perceived fit, ρ = .55; and hiring expectancies, ρ = .33). The previously cited studies of ecological reputation (Aiman-Smith et al., 2001) and work–life benefits (Casper & Buffardi, 2004) found significant (though generally weaker) effects on pursuit intentions in addition to attraction. In terms of the importance of diversity in organizations, Brown, Cober, Keeping, and Levy (2006) found that participants who were high in racial tolerance were more likely to pursue employment at organizations that emphasize diversity values (ΔR2 = .08, or a 40% increase over racial tolerance and diversity values condition main effects).

Targeting Strategies After organizations identify their desired outcomes but before they craft recruitment communications or choose a medium to disseminate messages, they must identify target audiences they wish to recruit. The potential is great for interplay between contextual considerations and targeting strategies identified in Figure 2.1. For example, whether the company considers the vacancy to be a core or peripheral position may drive their targeting or messaging strategies. Also, the number of vacancies for a given position relative to forecasted labor supply might affect the relative importance of applicant pool quality versus quantity outcomes, which in turn may influence strategies. Despite prior work in this area, research is lacking on many fronts. Broader targeting decisions. Previous work has focused on the targeting of various demographic

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

groups. For example, Rynes, Orlitzky, and Bretz (1997) found that experienced hires were evaluated more highly than new graduates on several skillbased characteristics, although new graduates were rated more highly on open-mindedness and willingness to learn. Use of experienced hires was also related to organizational growth but less dynamic business environments. However, companies also make broader decisions about their overall targeting strategies. Efforts might consist of person-to-person (i.e., one-to-one) communication, whereby a recruiter initiates contact with specifically qualified people, or an employee refers friends or acquaintances. Relational approaches might include maintaining contact with groups of former employees. The boomerang effect, typical of companies such as P&G (e.g., Horovitz, 2003), describes the recruitment of former employees. A more traditional approach is to recruit en masse by disseminating messages with broad appeal that do not target any single individual. Companies with exemplary reputations may choose not to recruit but rather to allow candidates to proactively approach them. Social identity theory suggests that organizational membership partly shapes self-concept (Ashforth & Mael, 1989). Thus, to enhance self-esteem and personal prominence, job seekers are likely to be attracted to firms that enjoy favorable reputations. Such firms may expect larger applicant pools (Turban & Cable, 2003). Scholars have recently raised the related issue of the firm as celebrity. It takes longer for a firm to build contextual factors such as reputation, brand image, or product awareness, but certain system shocks might bring windfalls to organizations in terms of recruitment. Rindova, Pollock, and Hayward (2006) described this as the celebrity effect, manifested when the mass media dramatizes an event related to a firm, and suddenly the general public views that firm as more attractive, regardless of actual performance metrics or longer-term proof of quality. For example, the Flutie effect is the “phenomenon of having a successful college sports team increase the exposure and prominence of a university” (“Flutie Effect,” 2008) and refers to Boston College’s 16% spike in applications for undergraduate admissions the year after the school’s quarterback, Doug Flutie, beat an archrival opponent with a miraculous touch-

down pass in the final seconds of a key football game. Although the duration of such an effect is unknown, the finding is generally consistent with Rindova, Williamson, Petkova, and Sever’s (2005) finding that the prominence dimension of reputation, and not the perceived product quality dimension, predicted price premiums organizations enjoyed. From a recruitment perspective, firms might try to capitalize on this phenomenon by strategically using the mass media to publicize their programs or accomplishments. Of course, as with firm reputation, a negative system shock might work against a firm. Also, the long-term sustainability of this approach is open to investigation. Timing issues. Another key targeting decision identified in Figure 2.1 involves when an organization should engage in recruitment activities (Rynes & Cable, 2003). For example, some job seekers (e.g., college graduates) and some jobs (e.g., holiday retail) operate in distinct cycles. When companies consider these cycles, questions arise: When in our particular recruitment cycle should we move? Can we be too early? Should we try to preempt the market? For example, the authors’ university basketball program recently made the national news when its former head coach offered a scholarship to an eighth grader. Soelberg’s (1967) implicit favorite model suggests possible benefits to firms that enter the recruitment market early. However, given that the practice is salient, unfortunately little research has been done in cycle timing since Rynes (1991) addressed it in her review. Internal recruitment. Also largely falling outside the realm of prior research and prior recruitment reviews (e.g., Breaugh et al., 2008; Taylor & Collins, 2000) is internal recruitment. We view this as a vital missing link (see also Billsberry, 2007). First, global staffing and expatriate assignments are receiving more attention, given the globalization of business in general. Gong (2003b) presented a model of the mix of parent-country nationals, host-country nationals, and third-country nationals optimal for global staffing purposes. This model informs the recruitment strategy of parent companies when trying to fill positions in foreign subsidiaries. Gong (2003a) found that when cultural distance is greater 51

Copyright American Psychological Association. Not for further distribution.

Dineen and Soltis

and companies use parent-company expatriates for staffing top management jobs, they increase their subsidiary labor productivity. Second, internal recruitment processes potentially produce turf wars over valued human capital. For example, Ling and Dineen (2005) used agency and stewardship theories to suggest that managers may either hinder or encourage efforts of valued employees to transfer internally. Specifically, agency theory predicts that managers will act in self-interest and tend to hinder transfers of valued employees. Stewardship theory predicts that managers will maximize firm and employee interests by encouraging deserving employees to transfer internally, even if such transfers will be detrimental to the manager. The authors suggested that success of these efforts depends on managers’ levels of social capital within the firm. Discovering ways to hinder managers’ hoarding behavior and encourage career building of valued employees (e.g., through incentives or other governance mechanisms) therefore seems important if companies are to use internal recruitment (see also SHRM, 2008). Somaya, Williamson, and Lorinkova (2008) recently offered a related and intriguing perspective: “letting go” of valued human capital, although seemingly detrimental, may actually be beneficial from the standpoint of creating social network ties to new areas of the business where transferees move (or in Somaya et al.’s case, to competitors). Finally, Ostroff and Clark (2001) found that various antecedent demographics (e.g., education, gender, children under the age of 15), job-related variables (e.g., information, attitudes, future employment), and community- and familyrelated variables predicted various internal mobility opportunities (e.g., lateral promotions, with or without relocation). Among the myriad of results this study offered was that lateral moves involving a career change were less appealing to those with smaller children, ostensibly because moving involved a greater potential disruption in family dynamics. Conversely, only moving concerns were related to willingness to accept promotions without corresponding career changes. Recruiting passive job candidates. One of the more interesting issues covered over the last few years 52

has been the recruiting of passive job candidates— employed individuals not actively searching for jobs but willing to consider outside opportunities. Termed poaching or talent raiding, this targeting strategy has become increasingly prevalent (Cappelli, 1999). For example, Rao and Drazin (2002) found that young or poorly connected mutual fund firms were more likely to recruit from competitors to increase innovation. In general, newer or less connected firms use poaching to gain entry into product markets when resources are more highly constrained. Poaching also allows newer firms to integrate more quickly to industry norms and avoid the mistakes that experienced employees of older firms have already learned. By contrast, wellconnected firms do not appear to gain as much from poaching, possibly because they gain only redundant talent exposure and may even be constrained by their level of connectedness. Of course, an unresolved issue is the threshold at which a valued employee will submit to poaching overtures and move to a lesser-connected, newer firm, and thus higher-risk career position. Related to this question is where firms should go to try to recruit passive job seekers. Dunford, Boudreau, and Boswell (2005) found that executives were more apt to search for jobs when their stock options had a market value below their exercise price (β = .13), a situation called underwater. Dunford, Oler, and Boudreau (2008) found that executives—especially CEOs—were more likely to turn over when faced with underwater options. This suggests a potential strategy of targeting employees of poorly performing firms; thus, the labor market generally seems to have direct implications for poaching strategies. The extensive job search literature also might offer recruiters insight into who might be more likely to be searching for new jobs (e.g., Kanfer, Wanberg, & Kantrowitz, 2001). Yet another factor associated with poaching is the competitive responses that poached firms might use. Gardner (2005) examined this issue among software industry competitors. He posited that poached firms may ignore the poaching or they may respond defensively (e.g., raising their inducement levels, requiring remaining employees to sign noncompete agreements) or retaliatorily/defensively (e.g., poaching talent from the firm that initiated the

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

poaching, severing business relationships). As the value of poached human capital increases, retaliatory/ defensive responses become more likely in comparison with purely defensive responses, and this effect is further enhanced as transferability of human capital increases. In general, these results suggest that firms losing human capital through poaching will retaliate if they see that the poaching is damaging their interests or heightening the competitiveness of the poaching firm. Finally, poaching may be a network-driven phenomenon. Williamson and Cable (2003), for example, found that firms were more likely to hire top management team members from organizations with which they had board of director interlocks (β = .59, SE = .23). Targeting nontraditional candidates. Rynes and Barber (1990) claimed that firms should increase their focus on nontraditional applicants to redress projected labor shortages. Very little work has considered the recruitment strategies that might be optimal for older workers, although it has been well documented that the working population is aging and a large percentage of older workers expect to continue working past retirement age (cf. Adams & Rau, 2004). Indeed, with some sources claiming considerable differences in work preferences of “Generation Y” versus baby-boomer generation workers (Armour, 2005) and some suggesting far fewer differences (e.g., Deal, 2007), rigorous research is needed in this area. Some work has begun to address predictors of bridge employment among older workers (i.e., employment between retirement from a full-time position and full retirement; e.g., Adams & Rau, 2004; Davis, 2003), and this may be a fruitful area in which to begin. Davis identified several key factors that lead retirees to participate in bridge employment (e.g., career pull opportunities, entrepreneurial orientations) or to avoid such employment (e.g., age, organizational tenure, clear retirement plans). Adams and Rau also found traditional constraints (e.g., inadequate transportation, poor health) to relate negatively to job seeking (incident rate ratio = .77, where the incident rate ratio refers to an increase [if the value is greater than one] or decrease [if less than one] in the rate of job seeking activity expected

with a one-unit change in a predictor). Surprisingly, older-worker constraints (e.g., perceived stereotypes against older workers) related positively to olderworker job search behavior (incident rate ratio = 1.41), although the authors surmised that olderworker constraint perceptions could be heightened because of job search experiences or experiences with rejection. To the extent that recruitment efforts can alleviate concerns over these constraints (e.g., Walmart’s image as an age-friendly workplace), companies may gain an advantage in attracting older workers. To recruit older workers, firms might also look to factors that engage older employees. Avery, McKay, and Wilson (2007) found that older workers are more engaged when they are satisfied with both younger and older coworkers (ΔR2 = .07, or a 39% increase over several other demographic controls). Thus, communications aimed at recruiting older workers might highlight how well potential older workers will likely fit with coworkers (i.e., person– group fit) rather than focusing solely on how well their abilities will fit the job demands or how their values will match the organization’s values. Rau and Adams (2005) also discovered that targeted equal employment opportunity statements, in combination with the opportunity to transfer knowledge through mentoring and flexible schedules, had more influence than any of these policies alone (partial η2 = .04). Beyond targeting older workers, other nontraditional applicant populations deserve research attention, and we review work related to diversity advertising in the following section. Thus, targeting strategy research progress has been less in terms of building on prior findings (which, as shown in Exhibit 2.1, were already limited) but more in terms of beginning to explore newer critical areas such as internal recruitment, poaching, and the recruitment of older workers. Although encouraging, researching a greater breadth of targeting strategies will need to be matched by attempts to provide richer, more in-depth investigations in these areas.

Messaging Strategies Figure 2.1 indicates that once a company determines its targeting strategy, it must develop and disseminate 53

Dineen and Soltis

Copyright American Psychological Association. Not for further distribution.

recruitment messages. Although potential messaging strategies have great breadth, we selectively review research related to message orientation and diversity messaging and some recent sourcing research. We further recognize that messages designed to influence person–environment fit are another crucial element of the applicant generation stage and we refer readers to Volume 3, chapter 1, this handbook, and KristofBrown, Zimmerman, and Johnson (2005) for detailed overviews of that literature. Message orientation. A key consideration in crafting a recruitment message is the message orientation (Frase-Blunt, 2003). Here, work has addressed the use of screening, recruiting, or dual-purpose orientations. For example, Williamson, Lepak, and King (2003) found that recruiting-oriented Web sites (i.e., those that try to “sell” the company to a recruit) influenced content usefulness perceptions to a greater extent than screening-oriented Web sites (i.e., those that provide information to allow job seekers to withdraw if they are a poor fit; β = .20). Usefulness perceptions then led to attraction (β = .41). Dual-purpose orientations exhibited a slightly lower but nonsignificant difference from recruiting orientation in terms of attraction. However, as noted more than 20 years ago (Rynes & Boudreau, 1986), we know little about how these strategic recruitment decisions are made. Dineen and Williamson (2008) provided preliminary evidence suggesting that recruiter compensation characteristics (i.e., whether a recruiter was compensated based on applicant pool quality rather than quantity; γ = .34) and higher firm reputation (γ = .22) influenced recruiter use of a screening orientation and that a screening orientation was linked to applicant pool quality (β = .19). Another key unresolved issue is when in the recruitment process an organization should provide recruitment-oriented messages and when it should provide screeningoriented messages. Considering once again contextual factors such as firms with negative reputations or particularly undesirable job features, it is interesting to consider messaging strategies that these firms might use. Ashforth, Kreiner, Clark, and Fugate (2007) provided a framework of approaches to avoid negative employee reactions to “dirty work” (i.e., undesirable 54

work offered to sanitation workers or tobacco company employees). Using an exploratory, semistructured interview format, these authors found that such companies used tactics such as the formation of occupational ideologies (e.g., “This work is valuable to society despite its negatives”) or social buffers (e.g., a pest-exterminator company might include exterminators’ testimonials about the satisfaction they have given their customers). Other possibilities are the use of humor and defensive tactics such as social comparison with others who are relatively worse off (e.g., one tobacco company comparing itself with another that has a worse public relations record). It would be interesting to see how these approaches could be applied to a recruitment context. Finally, we recognize that screening- versus recruiting-oriented approaches build on the welldocumented realistic job preview (RJP) tradition (e.g., Phillips, 1998; Wanous, 1992). Although we do not replicate Breaugh’s (2008) recent extensive review of the RJP literature, we reiterate a key point from that review in response to concerns about potential adverse self-selection effects (e.g., Bretz & Judge, 1998). That is, even if negative information leads highly qualified job seekers to drop out of a selection process, that result still seems better than having them become disillusioned after the recruitment process and quit soon after starting the job. Diversity advertising. Recent attention has been given to the way recruitment communications portray diversity. For example, it is increasingly recognized that to recruit minorities and women, firms must impress on these groups that the company values diversity (Avery & McKay, 2006) by signaling fairness and inclusion. Avery and McKay presented a framework of impression management techniques organizations can use to convey diversity images to potential minority applicants. They posited that firms could use ingratiation by portraying highly diverse ads, recruiting at traditional minority institutions, or placing recruiting ads in targeted media (e.g., traditional women’s magazines). An organization might also use promotion by presenting evidence of successful diversity management or exemplification by sponsoring minority events. Such efforts are thought to depend in part on (a) an organization’s

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

broader reputation for diversity (e.g., a defensive strategy might be optimal for a poor-reputation firm, whereas an assertive strategy might benefit a highreputation firm) and (b) the available pool of diverse applicants. These ideas should also be studied in the context of other diverse populations such as older or disabled workers and should be directly compared with other factors (e.g., what happens when a lessdiverse-friendly firm offers superior inducements). Given this general backdrop, two means of promoting a diversity image have been studied: the demographics of recruiters (not shown to have as much impact; ρ = −.05 to .03; Chapman et al., 2005) and recruitment advertisement diversity. In general, advertisements depicting diversity and Equal Employment Opportunity statements tend to be more attractive. For example, Avery, Hernandez, and Hebl (2004) found that Black and Hispanic participants were more attracted to companies when their recruitment advertisements used a Black or Hispanic representative instead of a Caucasian representative (d = 1.07 for Black participants, d = .78 for Hispanics). It is worth noting that the representative’s race did not affect Caucasian participants. The authors concluded that Black and Hispanic applicants were attracted because of a similarity mechanism, not because they perceived that the organization valued diversity. Cropanzano, Slaughter, and Bachiochi (2005) found that job seekers generally found preferential treatment plans to be unappealing, which suggests that minority applicants want to be perceived as having been treated fairly and not as receiving preferential treatment. Regarding the shaping of perceptions of organizational diversity, Kim and Gelfand (2003) examined the role that race and ethnic identity play in forming organizational inferences from diversity initiatives that are included in recruitment brochures. Ethnic identity and diversity initiative significantly affect socioemotional inferences, such as treatment of employees and relationships among employees. A recruitment brochure that included a diversity initiative also increased the likelihood that those who were high in ethnic identity would take the job offer, although this effect was small (ΔR2 = .01). Race did not exhibit a significant main effect on inferences or job offer acceptance. Martins and

Parsons (2007) corroborated this finding by demonstrating that individuals’ attitudes and beliefs about gender-related issues (e.g., gender identity, attitudes toward affirmative action) moderated the impact of gender diversity initiatives in recruitment literature on attraction. Finally, Avery (2003) found that openness to racial diversity moderated the effectiveness of diversity portrayals in recruitment advertisements and that such portrayals were useful only for supervisory positions. This further suggests that restricting diversity portrayals to lower-level employees may do more harm than good by raising cynicism among minority job seekers. Taken together, it seems important for firms to try to understand the mindset of their target audience and to use diversity material carefully. Recent recruitment source research. Companies must somehow disseminate the messaging strategies described earlier. In Figure 2.1 we note that Web recruitment and other sources play a crucial role in executing messaging strategies. We review the recruitment source literature and the critiques associated with it in less depth, given previous efforts (see Exhibit 2.1 and Breaugh, 2008). We do, however, note recent developments in three areas: social networks, word-of-mouth, and Web-based recruitment. First, studies have long implied that social networks have a role in recruitment either as conduits to job information (Granovetter, 1973) or as resources that shape individual decisions (Kilduff, 1990). More recently, Leung (2003) explicitly examined the use of social and business networks in the staffing of entrepreneurial ventures. This exploratory study suggested that social ties were used heavily to fill human resource needs in the start-up phase of a company with a shift toward business network ties in the growth phase. In addition, in a finding somewhat contrary to Granovetter’s famous strength of weak ties argument, Leung provided preliminary evidence that companies used strong, direct ties in both the start-up and growth phases when selecting new employees. Although this finding may be an artifact of the nature of the sample (i.e., entrepreneurial ventures often lack legitimacy and thus require a great deal of trust on behalf of new employees) and sample size (i.e., four 55

Copyright American Psychological Association. Not for further distribution.

Dineen and Soltis

organizations), it is important nonetheless in helping understand the role of networks as a recruiting source. The previously discussed work by Williamson and Cable (2003) also took a networks approach in examining the role of board interlocks in hiring decisions (see also Somaya et al., 2008). Recent work has also examined word-of-mouth communication that takes place within informal networks. For example, Van Hoye and Lievens (2007a) found that word-of-mouth (i.e., informal Web-based communication about companies) was viewed as more credible (partial η2 = .19) and associated with higher organizational attractiveness (partial η2 = .08) than Web-based testimonials. Van Hoye and Lievens (2007b) further discovered that negative word-ofmouth information could interfere with the effect of recruitment advertising. In general, nonrecruitmentrelated word-of-mouth communications might be perceived as more credible and lead to more accurate culture perceptions (Cable, Aiman-Smith, Mulvey, & Edwards, 2000) than recruitment-related communications (Kanar, Collins, & Bell, 2006), probably because, being from an external source, it is not perceived as trying to sell the organization (Fisher, Ilgen, & Hoyer, 1979). This is consistent with other work that has found that social network contacts influence job seekers (Kilduff, 1990). Wordof-mouth information from potential coworkers may be viewed as more credible because they are expected to have heightened expertise (Cable & Turban, 2001). Another recruitment source that has, not surprisingly, seen tremendous growth over the last decade is the World Wide Web; Ployhart (2006) described this growth as “nothing short of radical” (p. 875). Research has suggested that the Web is a powerful tool for sending messages to potential applicants, and scholars have suggested that the Web makes it easier for job seekers to find information about companies and apply for jobs (Lievens & Harris, 2003). However, the popular press has reported that the Web has increased extraneous application traffic from unqualified job seekers (Frase-Blunt, 2003; “Internet misuse,” 2003). Cober, Brown, Keeping, and Levy (2004) presented a model of the relationships between organizational Web sites and recruitment outcomes. In 56

general, they proposed that Web site façade relates to job seekers’ reactions but less strongly when job seekers have favorable prior attitudes toward the organization. These attitudes then influence job seekers’ perceptions of usability, Web site attitude, and actual search behavior, which in turn influence image, familiarity, and applicant attraction. One of Cober et al.’s more interesting propositions suggested that simply browsing a company’s Web site could alter organizational image. Building on Cober et al.’s (2004) work, Allen, Mahto, and Otondo (2007) examined objective factors (e.g., job-organization attributes), subjective factors (e.g., brand image and fit), and critical contact factors (e.g., nature of Web-site medium) on applicant attraction. Using a large sample mostly comprising job seekers browsing actual job ads, Allen et al. found that organizational image, but not mere familiarity, related to attitudes toward the organization (β = .32). They suggested that media richness perceptions also affected credibility and satisfaction, which related to attitudes toward the organization. Cable and Yu (2006) also found that, in general, richer and more credible mediums enhanced the correspondence between pre- and postviewing organizational image beliefs, even if such beliefs were already overestimated. Job seekers tended to hold underestimated perceptions of organizational images prior to exposure to career fair or Web-based sources (i.e., a Web page or electronic bulletin board). The aforementioned research further suggested that smaller or unfamiliar firms might be able to overcome barriers related to being unknown if they can first direct job seekers to their Web site. This renders the process of driving job seekers to a company Web site an important and overlooked research topic. Returning to the work of Collins and colleagues (e.g., Collins & Han, 2004), once a viewer is on a Web site, high- versus low-involvement recruitment strategies might be optimal depending on a firm’s brand image or reputation. Whereas Web-based recruitment research thus far has tended to be recruitment-oriented (i.e., attracting a maximum number of job seekers as the goal), some research has also examined the quality of applicants that might be generated by using features such as interactivity that are mostly available only

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

on the Web. Dineen and colleagues (Dineen, Ash, & Noe, 2002: Dineen, Ling, Ash, & DelVecchio, 2007; Dineen, & Noe, 2009) have shown that providing job seekers with customized information might encourage them to consider the information more carefully by making it more experiential or personally relevant (Cable & Turban, 2001), which might lead to better outcomes (e.g., well-fitting job seekers may be more attracted; poorly fitting job seekers might be less attracted). Specifically, Dineen and Noe found that customization is better at encouraging poorly fitting job seekers to eliminate themselves than at encouraging well-fitting job seekers to apply (odds ratios = 1.64 for person–organization fit and 2.13 for demands–abilities fit). Such a finding is consistent with image (Ordonez, Benson, & Beach, 1999) and prospect (Kahneman & Tversky, 1979) theories, which suggest that job seekers tend to screen-out poor options more than they screen-in good options. Furthermore, when a company narrows the initial pool of applicants, it decreases its legal liabilities because it must later reject fewer applicants. Related to Cober et al. (2004) and Allen et al. (2007), good aesthetic features may be necessary to trigger the benefits of customization (Dineen et al., 2007) and may be even more important for firms that have low familiarity (Collins & Han, 2004), such as the fictitious companies used in Dineen and colleagues’ research. Foundational to these findings is the notion of job seeker processing motivation (Breaugh & Starke, 2000). Specifically, it is likely that applicant pool quality is tied at least partly to the degree to which job seekers are willing and able to carefully scrutinize recruitment information. Recent work has begun to draw on the elaboration likelihood model (ELM) to address these issues (e.g., Cable & Turban, 2001; Dineen & Noe, 2009; Jones, Schultz, & Chapman, 2006; Roberson, Collins, & Oreg, 2005). Developed by Petty and Cacioppo (1986), the ELM suggests that people can be persuaded through a central route of high elaboration where information is given careful attention or through a more peripheral route where information is more passively processed without careful thought. For example, Jones et al. found that those exposed to a condition that encouraged peripheral processing of information chose ads con-

taining non-job-related features such as attractive fonts rather than those containing higher-quality arguments. In general, understanding prior job seeker cognitions and how they influence recruitment information processing at different recruitment stages is important in understanding this information’s impact on job seeker decision making. Finally, limited work has differentiated between various subsources in Web-based recruitment. For example, Jattuso and Sinar (2003) found that generalized job boards such as monster.com or careerbuilder.com attracted lower-quality applicants (in terms of educational qualifications, d = .67; and skills, d = .13) and applicants with a lesser degree of fit (d = 1.68) than industry/position specific job boards such as tvjobs.com or salesjobs.com. A replication of this study may be useful given that many of the more general job boards now have specialized subcomponents (e.g., engineering.careerbuilder.com). Future Web-based recruitment research should continue to address job seeker reactions to these different subsources in terms of usefulness and privacy concerns (Lievens & Harris, 2003). Investigations might also address the types of Web-based recruitment information that are used at different recruitment stages (e.g., job board information in the applicant generation phase, electronic bulletin boards that offer neutral testimonials about company culture in the post-offer stage), and the use of third-party recruitment firms as intermediaries between companies and job seekers. Even though considerable work remains, progress has been evident in the area of messaging strategies, relative to the prior work outlined in Exhibit 2.1. We are encouraged by the move away from attempts to examine general source effects to more nuanced investigations of specific sources such as the Web or word-of-mouth communications. Also, recent work has continued to draw on the RJP tradition but has expanded that concept to look at overall firm characteristics (e.g., negative reputations) and message frames (e.g., screening, dual-purpose). Finally, diversity advertising has received attention that has previously been largely absent. Organizations experience relative levels of success in generating viable candidates for their vacancies. Once an applicant pool is generated, organizations 57

Dineen and Soltis

must turn their attention to maintaining the status of their most viable applicants.

Copyright American Psychological Association. Not for further distribution.

MAINTAINING THE STATUS OF VIABLE APPLICANTS The second stage of the recruitment process comprises organizational efforts to ensure that viable, higher-quality applicants maintain interest in being considered for vacancies. As organizations form applicant pools or court candidates individually, screening considerations often take center stage in terms of workforce planning, but they also likely transfer to recruitment considerations, as shown in Figure 2.1. Indeed, often overlooked is the continuing need for effective, ongoing recruitment of candidates at this stage to maintain their status as potential employees until the company can tender job offers. As indicated in Figure 2.1, several processes are vital to this stage, including communication and rapport building as job applicants continue to interact with organization agents. Specifically, it is primarily at this stage that recruiters and those involved in site visits build rapport with applicants and signal organizational intentions. Indeed, as shown in Figure 2.1, interactions with organizational agents take place across all three stages, but are most prominent during applicant maintenance. Key outcomes at this stage include remaining competitive with other firms seeking similar job candidates and tendering offers in a timely manner to maximize the chances of employing valued applicants. Thus a parallel topic in this stage is the way in which applicants screen out companies and withdraw from the pool. Several chapters in this handbook address issues related to screening and selection and are not covered here (see in particular chap. 13, this volume). For an overview of one key recruitment consideration— selection process fairness perceptions—we refer the reader to chapter 12 of this volume. Another key consideration is selection process timeliness. Yet, since Rynes, Bretz, and Gerhart (1991) found that delays in the selection process can lead to attrition from that process, especially among more qualified applicants, little has been accomplished from a research perspective to build on these findings. Thus, for purposes of our review, we specifically 58

address two primary issues: recruiter interactions and site visits. As will become apparent by the shorter length of this section (and the next) in relation to previous sections of this chapter, opportunities for research are plentiful. Although, as with other topics, researchers have recently reviewed recruiter interactions (Breaugh, 2008), we address some key issues. Chapman et al. (2005) concluded that recruiter personableness exhibits a fairly strong relationship with pursuit intentions (ρ = .50). However, its effects are weaker for more distal outcomes such as job choice, and it is likely that applicants rely less on recruiter signals as more information about job and organizational characteristics becomes salient. Alternatively, recruitment initiatives such as recruiter behaviors likely signal job-organizational characteristics, and research has generally shown that job-organization factors mediate the effects of attraction on recruitment (Turban, 2001). Indeed, it appears that later in the process when job seekers decide whether to accept the job, they focus more on what their work environment will be like rather than on particular aspects of the recruitment process such as recruiter interactions, and earlier longitudinal work showed this to be true from the application phase through the job choice decision stage (Taylor & Bergmann, 1987). Some researchers have attempted to ground the role of recruiters in psychological theory. For example, Larsen and Phillips (2002) laid out a series of propositions regarding the propensity for recruiters to influence applicants based on the ELM. Specifically, they proposed that recruiter demographics and friendliness exert less influence on job applicant attraction when those applicants engage in central processing of organizational and job attributes. Job applicants’ use of central processing is more likely, for example, when they possess lower stress levels, or greater job and company knowledge, interview experience, financial need, or self-esteem. Alternatively, Chapman and Webster (2006) used expectancy theory and concepts of procedural justice and signaling to unpack the mechanisms underlying recruiter– applicant interactions. The authors found that recruiter friendliness was related to applicant perceptions of procedural justice of the process, post-interview organizational attractiveness, and expectancy of

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

receiving an offer (βs = .60, .21, and .26, respectively). Rynes et al. (1991) found, however, that signaling more greatly affected less-experienced job seekers. Despite these recent studies, progress has generally been limited in the area of recruiter interactions since the Rynes and Cable (2003) review (Exhibit 2.1). However, rather than indicating a need for additional work, we do not view recruiter interactions as having as high of a priority going forward as other topics covered in this review. Conversely, as reviewed by Breaugh et al. (2008) and indicated in Exhibit 6.1, researchers have generally neglected on-site visits, although job applicants often decide about organizations during those times. Previous work has shown that site-visit perceptions relate to job choice (R2 = .05; Turban, Campion, & Eyring, 1995). More recently, McKay and Avery (2006) provided a comprehensive model of the site-visit process in terms of racioethnic issues. They suggest that minority site visitors perceive a stronger link between organizational/ community diversity, vertical integration, and diversity climate perceptions, as well as between quality of site-visit/community interactions and diversity climate perceptions. Diversity climate relates to acceptance intentions and even more so when job opportunities are perceived to be high. This model, if verified, might have important implications for firms looking to move diverse candidates from job application to job choice. It seems especially important given that site visits likely provide a truer picture of a company’s diversity climate than applicants might glean from initial recruiting communications and failure to uphold those first communications might not only lead minority candidates to drop out but might also lead them to translate these misunderstood communications to other minority job seekers. Future research should address this issue as well as extend McKay and Avery’s propositions to older job seekers or other nontraditional populations. After organizations complete their selection processes, and assuming they have successfully retained viable applicants, they typically extend job offers to chosen candidates. From here, organizations must use closure processes to ensure that the candidates accept the offers. We now turn to this stage.

POSTOFFER CLOSURE After firms tender job offers to desired candidates, much work remains from a recruitment standpoint. Indeed, postoffer closure is often overlooked but can be a vital tipping point between a valued job applicant accepting an offer or going elsewhere, often to a competitor. Key processes likely to occur at this stage include the applicant’s decision making, offer negotiation, and the organization’s ability to recognize competitors’ overtures toward the candidate in terms of inducements and offer timing. Job choice continues to receive attention as a critical outcome at this stage. For example, in terms of the postoffer/prehire time period, Breaugh, Greising, Taggart, and Chen (2003) studied the effects of recruiting sources on the propensity to hire, finding that direct applicants (8.1%) and employee referrals (12.4%) were hired at greater rates than those recruited through newspapers (1.1%), colleges (1.3%), or job fairs (4.8%). Boswell, Roehling, LePine, and Moynihan (2003) longitudinally examined over several recruitment phases how 14 job and organization factors related to offer acceptance or rejection. They found that the mostmentioned factors influencing acceptance decisions were nature of work (37.6%), location (37.6%), company culture (36.5%), and advancement opportunities (25.8%). As shown in Figure 6.1, ongoing interactions with firm agents likely carry over in terms of importance in this final stage. In the study conducted by Boswell et al. (2003), respondents indicated that meeting with multiple constituents while on a site visit positively affected job choice decisions as did follow-up contacts from the company. The aforementioned studies represent the most proximal postoffer outcomes (e.g., hiring rate, acceptance intentions), yet continued attention is needed to additional distal outcomes such as turnover, performance, and satisfaction of new hires (Hoffman & Woehr, 2006; Meglino, Ravlin, & DeNisi, 2000). More subtle outcomes might include negotiated changes in salary level (i.e., How much does a company need to concede to secure the employment of a job candidate?) or maintenance of internal equity among job incumbents. The issue of external competitiveness (Milkovich & Newman, 59

Copyright American Psychological Association. Not for further distribution.

Dineen and Soltis

2008) suggests that companies are pressured to extend lucrative offers. However, they risk upsetting the balance or equity perceptions of current employees who hold similar positions within the organization. Although studied extensively under the umbrella of wage compression in the labor and economics literatures (e.g., Heyman, 2008), little research has been done regarding how potential wage compression issues might affect an organization’s approach to postoffer recruitment. Among the recruitment issues relevant at this stage, investigations of multiple and/or competing offers seem critical. The most valued recruits are also the most likely to have several job offers. This makes processes such as negotiation vital to recruitment all the way to the point of job acceptance. Furthermore, the issues of job choice timing and exploding offers seem crucial (Rynes & Cable, 2003; Schwab, Rynes, & Aldag, 1987). Yet, little research has been done on job choice timing, and the scant research that has been conducted on exploding offers has found no significant effects on job choice outcomes (Boswell et al., 2003). Schwab et al. addressed the issue of multiple offers using marginal utilities (i.e., by assessing the marginal utility of continuing to pursue additional offers vs. taking the current offer without knowing the value of potential offers). They also addressed simultaneous versus sequential evaluation of alternatives, which returns to the need for research investigations to include multiple jobs to enhance external validity. Horvath and Millard (2008) recently provided preliminary evidence suggesting that attraction–intention relationships varied in a nonlinear fashion across recruitment stages (e.g., pre/ postoffer) but also as a function of other vacancies simultaneously being considered. For example, the relationship between attraction and applicant intentions partially depended on where the candidate was with other companies (e.g., postoffer). However, it is important to note that the Chapman et al. (2005) meta-analysis did not find very strong effects for perceived alternatives on acceptance intentions (ρ = −.06) or actual job choice (ρ = .07), suggesting that simultaneous offers might be less important than at first glance. Chapman et al. speculated that quality rather than quantity of competing offers may be influential. 60

Other interesting issues might surface at this stage and deserve research attention. Given demographic shifts in nontraditional family arrangements (e.g., Conlin, 2003), work and family issues such as transferability to a new location and care for an elderly parent often occur on a case-by-case basis, making job offer and subsequent negotiation processes more challenging and complex. Another interesting issue is whether prior contextual characteristics such as brand equity and vacancy characteristics matter as much at this stage. Finally, a key process at this stage might be a firm’s engagement in competitive intelligence, or knowledge of competitors’ actions or likely actions. For example, knowing the window of opportunity competing offers have given job seekers would be valuable during salary negotiations, as would knowledge of specific packages competitors have offered. Network ties within an industry might help firms gain advantage here. In general, given the dearth of prior work in the postoffer closure area (as reflected in Exhibit 2.1), we view this is a new area ripe for fresh perspectives and rigorous work. FUTURE DIRECTIONS AND CONCLUSION With this review of the recruitment literature come recommendations for future research. In closing, we suggest several areas beyond those mentioned. First, we have identified several studies that examined key outcomes within specific phases. However, researchers have tried to build on earlier studies (e.g., Cable & Judge, 1996; Taylor & Bergmann, 1987) through longitudinal research capturing multiple phases. For example, Carless (2005) studied the effects of perceived person–job and person–organization fit over time on attraction and likelihood of accepting a job offer. Chapman and Webster (2006) collected data at multiple times to observe recruiter effects on placement, attraction, job choice, and job choice intentions. Chapman et al.’s (2005) meta-analytic evidence highlighted the importance of continuing to study recruitment relationships over multiple stages by suggesting that attraction is not directly related to job choice but rather is at least partially mediated by pursuit and acceptance intentions. It is also important to note that at later recruitment stages restriction in

Copyright American Psychological Association. Not for further distribution.

Recruitment: A Review of Research and Emerging Directions

range on predictors may be at issue. Chapman et al. noted that person–environment fit and job choice relationships might be restricted because poorly fitting individuals have already eliminated themselves at earlier stages (see Cable & Judge, 1996, for a discussion of this issue). Indeed, in Chapman et al., job choice was the weakest predicted outcome. Second, although little work has addressed unitor firm-level outcomes, it is incumbent on researchers to approach investigations in this way (e.g., firmlevel financial outcomes, unit-level turnover; e.g., Ployhart, 2006; Rynes & Cable, 2003; Taylor & Collins, 2000). This may require cross-level research designs spanning several companies. Ployhart provided a diagram outlining a process of linking micro and macro levels of analysis. Taylor and Collins encouraged researchers to assess recruitment issues within the purview of the resource-based view of the firm (Barney, 1991). Specifically, they encouraged researchers and recruitment specialists to evaluate recruitment efforts according to how valuable, rare, and/or inimitable they are, and thus assess their potential to provide a source of sustainable competitive advantage. Also important are continued assessments of applicant pool size and quality. It is encouraging that these are starting to receive attention from scholars (e.g., Collins & Han, 2004) and practitioners (e.g., Cascio & Boudreau, 2008; http://www.staffing.org, 2005), although quality remains ill defined and likely differs according to firm objectives (e.g., diversity, cognitive ability). Third, we have discussed the importance of using different metrics at different stages of the recruitment process. There is thus a need for better understanding of when a given recruitment metric is more or less optimal and why and to define success at each stage commensurate with the importance of the prevailing outcome. For example, at the stage of generating viable candidates, it seems vital to generate a large enough pool of applicants of acceptable quality. Yet many recruitment studies continue to rely on attitudinal outcomes such as attraction as the ultimate criterion at this phase, overlooking issues such as the number of candidates a company must reject. Beyond defining recruitment success differently at different stages, Barber, Wesson, Roberson, and Taylor (1999) also suggested that small and

large firms tend to define recruitment success differently (e.g., longer- and shorter-term focuses, respectively). Several online (e.g., staffing.org) and print (Cascio & Boudreau, 2008) sources exist to enable companies to assess the value of their recruiting function and practitioners should avail themselves of these (see also, Carlson et al., 2002). Finally, our review reveals that much more effort is needed to examine recruitment strategies and processes that occur after a job seeker applies for a position. We would be mistaken to assume that recruitment could end once a job seeker has submitted an application, yet the unbalanced state of the recruitment literature seems to reflect this logic. We hope that future literature reviews will show more progress in analyzing the stages that follow candidate generation. Recruitment continues to be critical to organizational functioning, and much research is needed to inform practitioners in the throes of the talent war. Compared with prior work reflected in Exhibit 6.1, we are encouraged by the increased breadth of topics more recently being addressed as well as the increased depth of studies in areas such as firm reputation and the integration of marketing and recruitment principles. Yet, we still see a need for richer, more in-depth examinations of other crucial areas such as site visits, the labor market, and recruitment timing. Although we have provided a selective rather than exhaustive review, we once again encourage researchers to follow recommendations of other recent reviews and to embrace the unique challenges and opportunities we have attempted to illustrate in this chapter. It is an exciting time for advanced inquiry in an area greatly needing scholarly input. We look forward to future literature reviews that will undoubtedly be needed to classify and describe this pending work.

References Adams, G., & Rau, B. (2004). Job seeking among retirees seeking bridge employment. Personnel Psychology, 57, 719–744. Aiman-Smith, L., Bauer, T., & Cable, D. (2001). Are you attracted? Do you intend to pursue? A recruiting policy-capturing study. Journal of Business and Psychology, 16, 219–237. 61

Dineen and Soltis

Allen, D. G., Mahto, R. V., & Otondo, R. F. (2007). Webbased recruitment: Effects of information, organizational brand, and attitudes toward a web site on applicant attraction. Journal of Applied Psychology, 92, 1696–1708. Armour, S. (2005, November 6). Generation Y: They’ve arrived at work with a new attitude. USA Today, p. B1.

Copyright American Psychological Association. Not for further distribution.

Ashforth, B. E., Kreiner, G. E., Clark, M. A., & Fugate, M. (2007). Normalizing dirty work: Managerial tactics for countering occupational taint. Academy of Management Journal, 50, 149–174. Ashforth, B. E., Kreiner, G. E., & Fugate, M. (2000). All in a day’s work: Boundaries and micro role transitions. Academy of Management Journal, 25, 472–491. Ashforth, B. E., & Mael, E. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20–39. Avery, D. R. (2003). Reactions to diversity in recruitment advertising—Are differences black and white? Journal of Applied Psychology, 88, 672–679. Avery, D. R., Hernandez, M., & Hebl, M. R. (2004). Who’s watching the race? Racial salience in recruitment advertising. Journal of Applied Social Psychology, 34, 146–161. Avery, D. R., & McKay, P. F. (2006). Target practice: An organizational impression management approach to attracting minority and female job applicants. Personnel Psychology, 59, 157–187. Avery, D. R., McKay, P. F., & Wilson, D. C. (2007). Engaging the aging workforce: The relationship between perceived age similarity, satisfaction with coworkers, and employee engagement. Journal of Applied Psychology, 92, 1542–1556. Barber, A. E. (1998). Recruiting employees: Individual and organizational perspectives. Thousand Oaks, CA: Sage. Barber, A. E., Wesson, M. J., Roberson, Q. M., & Taylor, M. S. (1999). A tale of two job markets: Organizational size and its effects on hiring practices and job search behavior. Personnel Psychology, 52, 841–868. Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 18, 99–120.

Breaugh, J. A. (2008). Employee recruitment: Current knowledge and important areas for future research. Human Resource Management Review, 18, 103–118. Breaugh, J. A., Greising, L. A., Taggart, J. W., & Chen, H. (2003). The relationship of recruiting sources and pre-hire outcomes: Examination of yield ratios and applicant quality. Journal of Applied Social Psychology, 33, 2267–2287. Breaugh, J. A., Macan, T. H., & Grambow, D. M. (2008). Employee recruitment: Current knowledge and directions for future research. In G. P. Hodgkinson & J. K. Ford (Eds.), International review of industrial and organizational psychology (Vol. 23, pp. 45–82). Hoboken, NJ: Wiley. Breaugh, J. A., & Starke, M. (2000). Research on employee recruitment: So many studies, so many remaining questions. Journal of Management, 26, 405–434. Bretz, R. D., & Judge, T. A. (1998). Realistic job previews: A test of the adverse self-selection hypothesis. Journal of Applied Psychology, 83, 330–337. Brooks, M., Highhouse, S., Russell, S., & Mohr, D. (2003). Familiarity, ambivalence, and firm reputation: Is corporate fame a double-edged sword?. Journal of Applied Psychology, 88, 904–914. Brown, D. J., Cober, R. T., Keeping, L. M., & Levy, P. E. (2006). Racial tolerance and reactions to diversity information in job advertisements. Journal of Applied Social Psychology, 36, 2048–2071. Cable, D., Aiman-Smith, L., Mulvey, P., & Edwards, J. (2000). The sources and accuracy of job applicants’ beliefs about organizational culture. Academy of Management Journal, 43, 1076–1085. Cable, D., & Judge, T. (1996). Person–organization fit, job choice decisions, and organizational entry. Organizational Behavior and Human Decision Processes, 67, 294–311. Cable, D., & Turban, D. (2001). Establishing the dimensions, sources, and value of job seekers’ employer knowledge during recruitment. In G. Ferris (Ed.), Research in personnel and human resource management (Vol. 20, pp. 115–163). Greenwich, CT: JAI Press.

Billsberry, J. (2007). Experiencing recruitment and selection. Hoboken, NJ: Wiley.

Cable, D., & Turban, D. (2003). The value of organizational reputation in the recruitment context: A brand equity perspective. Journal of Applied Social Psychology, 33, 2244–2266.

Borman, W. C., Ilgen, D. R., & Klimoski, R. J. (Eds.). (2003). Handbook of psychology: Industrial and organizational psychology (Vol. 12). Hoboken, NJ: Wiley.

Cable, D., & Yu, K. (2006). Managing job seekers’ organizational image beliefs: The role of media richness and media credibility. Journal of Applied Psychology, 91, 828–840.

Boswell, W. R., Roehling, M. V., LePine, M. A., & Moynihan, L. M. (2003). Individual job-choice decisions and the impact of job attributes and recruitment practices: A longitudinal field study. Human Resource Management, 42, 23–37.

Cappelli, P. (1999). The new deal at work: Managing the market-driven workforce. Boston: Harvard Business School Press.

62

Cappelli, P. (2001). Making the most of on-line recruiting. Harvard Business Review, 79, 139–146.

Recruitment: A Review of Research and Emerging Directions

Cappelli, P. (2005). Will there really be a labor shortage? Human Resource Management, 44, 143–149. Carless, S. A. (2005). Person–job fit versus person– organization fit as predictors of organizational attraction and job acceptance intentions: A longitudinal study. Journal of Occupational and Organizational Psychology, 78, 411–429.

Copyright American Psychological Association. Not for further distribution.

Carlson, K. D., Connerley, M. L., & Mecham, R. L., III (2002). Recruitment evaluation: The case for assessing the quality of applicants attracted. Personnel Psychology, 55, 461–490. Carless, S. A., & Wintle, J. (2007). Applicant attraction: The role of recruiter function, work–life balance policies and career salience. International Journal of Selection and Assessment, 15, 394–404. Cascio, W. F., & Boudreau, J. W. (2008). Investing in people. Upper Saddle River, NJ: Pearson. Casper, W. J., & Buffardi, L. C. (2004). Work–life benefits and job pursuit intentions: The role of anticipated organizational support. Journal of Vocational Behavior, 65, 391–410. Chapman, D. S., Uggerslev, K. L., Carroll, S. A., Piasentin, K. A., & Jones, D. A. (2005). Applicant attraction to organizations and job choice: A meta-analytic review of the correlates of recruiting outcomes. Journal of Applied Psychology, 90, 928–944. Chapman, D. S., & Webster, J. (2006). Toward an integrated model of applicant reactions and job choice. International Journal of Human Resource Management, 17, 1032–1057. Cober, R. T., Brown, D. J., Keeping, L. M., & Levy, P. E. (2004). Recruitment on the net: How do organizational web site characteristics influence applicant attraction? Journal of Management, 30, 623–646. Collins, C. J. (2007). The interactive effects of recruitment practices and product awareness in job seekers’ employer knowledge and applicant behaviors. Journal of Applied Psychology, 92, 180–190. Collins, C. J., & Han, J. (2004). Exploring applicant pool quantity and quality: The effects of early recruitment practices, corporate advertising, and firm reputation. Personnel Psychology, 57, 685–717. Collins, C. J., & Stevens, C. K. (2002). The relationship between early recruitment-related activities and the application decisions of new labor-market entrants: A brand equity approach to recruitment. Journal of Applied Psychology, 87, 1121–1133. Conlin, M. (2003, October 20). Unmarried America. Business Week, 106. Cropanzano, R., Slaughter, J., & Bachiochi, P. (2005). Organizational justice and black applicants’ reactions to affirmative action. Journal of Applied Psychology, 90, 1168–1184.

Davis, M. A. (2003). Factors related to bridge employment participation among private sector early retirees. Journal of Vocational Behavior, 63, 55–71. Deal, J. (2007). Retiring the generation gap. San Francisco: Wiley. Delaney, K. J. (2007, June 28). Start-ups make inroads with Google’s work force. Wall Street Journal, B1. Dineen, B. R., Ash, S. R., & Noe, R. A. (2002). A web of applicant attraction: Person–organization fit in the context of Web-based recruitment. Journal of Applied Psychology, 87, 723–734. Dineen, B. R., Ling, J., Ash, S. R., & DelVecchio, D. (2007). Aesthetic properties and message customization: Navigating the dark side of web recruitment. Journal of Applied Psychology, 92, 356–372. Dineen, B. R., & Noe, R. A. (2009). Effects of customization on application decisions and applicant pool characteristics in a Web-based recruitment context. Journal of Applied Psychology, 94, 224–234. Dineen, B. R., & Williamson, I. O. (2008). Effects of environmental and organizational attributes on recruitment message orientation. Paper presented at the meeting of the Society for Industrial/Organizational Psychology, San Francisco. Dunford, B., Boudreau, J., & Boswell, W. (2005). Out-ofthe-money: The impact of underwater stock options on executive job search. Personnel Psychology, 58, 67–101. Dunford, B., Oler, D., & Boudreau, J. (2008). Underwater stock options and voluntary executive turnover: A multidisciplinary perspective integrating behavioral and economic theories. Personnel Psychology, 61, 687–726. Ehrhart, K. H., & Ziegert, J. C. (2005). Why are individuals attracted to organizations? Journal of Management, 31, 901–919. Fisher, C. D., Ilgen, D. R., & Hoyer, W. D. (1979). Source credibility, information favorability, and job offer acceptance. Academy of Management Journal, 22, 94–103. Flutie effect. (2008). Retrieved August 2008 from http://en.wikipedia.org/wiki/Flutie_effect Frase-Blunt, M. (2003, August). Candidate glut. HR Magazine, 89–93. Gardner, T. M. (2005). Interfirm competition for human resources: Evidence from the software industry. Academy of Management Journal, 48, 237–256. Gong, Y. (2003a). Subsidiary staffing in multinational enterprises: Agency, resources, and performance. Academy of Management Journal, 46, 728–739. Gong, Y. (2003b). Toward a dynamic process model of staffing composition and subsidiary outcomes in 63

Dineen and Soltis

multinational enterprises. Journal of Management, 29, 259–280. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1380.

Lavelle, L. (2003, September 29). After the jobless recovery, a war for talent. BusinessWeek, 3851, 92–92.

Heyman, F. (2008). How wage compression affects job turnover. Journal of Labor Research, 29, 11–26.

Leung, A. (2003). Different ties for different needs: Recruitment practices of entrepreneurial firms at different developmental phases. Human Resource Management, 42, 303–320.

Hoffman, B. J., & Woehr, D. J. (2006). A quantitative review of the relationship between person– organization fit and behavioral outcomes. Journal of Vocational Behavior, 68, 389–399. Horovitz, B. (2003, April 23). At P & G’s historic reunion. USA Today, p. B1. Copyright American Psychological Association. Not for further distribution.

tion likelihood model. Journal of Business and Psychology, 16, 347–364.

Horvath, M., & Millard, M. R. (2008). Recruitment stage, organizational attraction, and applicant intentions. Paper presented at the Society for Industrial and Organizational Psychology Meeting, San Francisco. Internet misuse may contribute to long joblessness. (2003, February 18). Silicon Valley/San Jose Business Journal. Retrieved May 2005 from http://sanjose.bizjournals. com/sanjose/stories/2003/02/17/daily17.html Jattuso, M. L., & Sinar, E. F. (2003). Source effects in Internet-based screening procedures. International Journal of Selection and Assessment, 11, 137–140. Jones, D. A., Shultz, J. W., & Chapman, D. S. (2006). Recruiting through job advertisements: The effects of cognitive elaboration on decision making. International Journal of Selection and Assessment, 11, 167–179. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–293. Kanar, A. M., Collins, C., & Bell, B. (2006). Can an organization overcome a negative image? A longitudinal experimental examination of the effect. Paper presented at the meeting of the Academy of Management, Atlanta, GA. Kanfer, R., Wanberg, C. R., & Kantrowitz, T. M. (2001). Job search and employment: A personalitymotivational analysis and meta-analytic review. Journal of Applied Psychology, 86, 837–855. Kilduff, M. (1990). The interpersonal structure of decision making: A social comparison approach to organizational choice. Organizational Behavior and Human Decision Processes, 47, 270–288. Kim, S. S., & Gelfand, M. J. (2003). The influence of ethnic identity on perceptions of organizational recruitment. Journal of Vocational Behavior, 63, 396–416. Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. (2005). Consequences of individuals’ fit at work: A meta-analysis of person–job, person–organization, person–group, and person–supervisor fit. Personnel Psychology, 58, 281–342. Larsen, D. A., & Phillips, J. I. (2002). Effect of recruiter on attraction to the firm: Implications of the elabora64

Lievens, F., & Harris, M. M. (2003). Research on Internet recruitment and testing: Current status and future directions. In C. L. Cooper & I. T. Robertson (Eds.), The international review of industrial and organizational psychology, (Vol. 18, 131–165). Chichester, England: Wiley. Lievens, F., & Highhouse, S. (2003). The relation of instrumental and symbolic attributes to a company’s attractiveness as an employer. Personnel Psychology, 56, 75–102. Ling, J., & Dineen, B. R. (2005). Internal transfers: A tale of human capital, social capital, and the manager as agent or steward. Paper presented at the meeting of the Southern Management Association, Charleston, SC. Martins, L. L., & Parsons, C. K. (2007). Effects of gender diversity management on perceptions of organizational attractiveness: The role of individual differences in attitudes and beliefs. Journal of Applied Psychology, 92, 865–875. Maurer, S., Howe, V., & Lee, T. (1992). Organizational recruiting as marketing management: An interdisciplinary study of engineering graduates. Personnel Psychology, 45, 807–833. McConnon, A. (2007, September 10). Social networking is graduating—and hitting the job market: How do the online Rolodexes stack up? Business Week, 4049, 4. McKay, P. F., & Avery, D. R. (2006). What has race got to do with it? Unraveling the role of racioethnicity in job seekers’ reactions to site visits. Personnel Psychology, 59, 395–429. Meglino, B. M., Ravlin, E. C., & DeNisi, A. S. (2000). A meta-analytic examination of realistic job preview effectiveness: A test of three counterintuitive propositions. Human Resource Management Review, 10, 407–434. Milkovich, G. T., & Newman, J. M. (2008). Compensation. Boston: McGraw-Hill/Irwin. Ordonez, L., Benson, L., III, & Beach, L. R. (1999). Testing the compatibility test: How instructions, accountability, and anticipated regret affect prechoice screening options. Organizational Behavior and Human Decision Processes, 78, 63–80. Ostroff, C., & Clark, M. A. (2001). Maintaining an internal market: Antecedents of willingness to change jobs. Journal of Vocational Behavior, 59, 425–453.

Recruitment: A Review of Research and Emerging Directions

Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: Springer-Verlag. Phillips, J. M. (1998). Effects of realistic job previews on multiple organizational outcomes: A meta-analysis. Academy of Management Journal, 41, 673–690. Ployhart, R. E. (2006). Staffing in the 21st century: New challenges and strategic opportunities. Journal of Management, 32, 868–897.

Copyright American Psychological Association. Not for further distribution.

Rao, H., & Drazin, R. (2002). Overcoming resource constraints on product innovation by recruiting talent from rivals: A study of the mutual fund industry, 1986–94. Academy of Management Journal, 45, 491–507. Rau, B. L., & Adams, G. A. (2005). Attracting retirees to apply: Desired organizational characteristics of bridge employment. Journal of Organizational Behavior, 26, 649–660. Rau, B., & Hyland, M. (2002). Role conflict and flexible work arrangements: The effects on applicant attraction. Personnel Psychology, 55, 111–136. Rindova, V., Pollock, T., & Hayward, M. (2006). Celebrity firms: The social construction of market popularity. Academy of Management Review, 31, 50–71. Rindova, V. P., Williamson, I. O., Petkova, A. P., & Sever, J. M. (2005). Being good or being known: An empirical examination of the dimensions, antecedents, and consequences of organizational reputation. Academy of Management Journal, 48, 1033–1049. Roberson, Q. M., Collins, C. J., & Oreg, S. (2005). The effects of recruitment message specificity on applicant attraction to organizations. Journal of Business and Psychology, 19, 319–339. Rynes, S. (1991). Recruitment, job choice, and post-hire consequences: A call for new research directions. In M. Dunnette & L. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 2, pp. 399–444). Palo Alto, CA: Consulting Psychologists Press. Rynes, S. L., & Barber, A. E. (1990). Applicant attraction strategies: An organizational perspective. Academy of Management Review, 15, 286–310. Rynes, S. L., & Boudreau, J. W. (1986). College recruiting in large organizations: Practice, evaluation, and research implications. Personnel Psychology, 39, 729–757. Rynes, S. L., Bretz, R. D., Jr., & Gerhart, B. (1991). The importance of recruitment in job choice: A different way of looking. Personnel Psychology, 44, 487–521. Rynes, S. L., & Cable, D. M. (2003). Recruitment research in the 21st century. In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.), Handbook of psychology:

Industrial and organizational psychology (Vol. 12, pp. 55–76). Hoboken, NJ: Wiley. Rynes, S. L., Orlitzky, M. O., & Bretz, R. D. (1997). Experienced hiring versus college recruiting: Practices and emerging trends. Personnel Psychology, 50, 487–521. Schwab, D. P., Rynes, S. L., & Aldag, R. J. (1987). Theories and research on job search and choice. In K. M. Rowland & G. R. Ferris (Eds.), Research in personnel and human resources management (Vol. 5, pp. 129–166.) Greenwich, CT: JAI Press. SHRM. (Producer). (2008). Seeing forward: Succession planning at 3M [DVD]. (Available from http://www.shrm. org/about/foundation/products/Pages/SeeingForwardD VD.aspx) Slaughter, J. E., Zickar, M. J., Highhouse, S., & Mohr, D. C. (2004). Personality trait inferences about organizations: Development of a measure and assessment of construct validity. Journal of Applied Psychology, 89, 85–103. Soelberg, P. (1967). Unprogrammed decision making. Industrial Management Review, 8, 19–29. Somaya, D., Williamson, I. O., & Lorinkova, N. (2008). Gone but not lost: The different performance impacts of employee mobility between cooperators versus competitors. Academy of Management Journal, 51, 936–953. Staffing.org. (2005, June 22). How do you calculate the ROI of job boards? Retrieved July 2005 from http://www.staffing.org/library_ViewArticle.asp? ArticleID=165 Taylor, M., & Bergmann, T. (1987). Organizational recruitment activities and applicants’ reactions at different stages of the recruitment process. Personnel Psychology, 40, 261–285. Taylor, M. S., & Collins, C. J. (2000). Organizational recruitment: Enhancing the intersection of theory and practice. In C. L. Cooper & E. A. Locke (Eds.), Industrial and organizational psychology: Linking theory and practice (pp. 304–334). Oxford, England: Basil Blackwell. Turban, D. B. (2001). Organizational attractiveness as an employer on college campuses: An examination of the applicant population. Journal of Vocational Behavior, 58, 293–312. Turban, D. B., & Cable, D. M. (2003). Firm reputation and applicant pool characteristics. Journal of Organizational Behavior, 24, 733–751. Turban, D. B., Campion, J. E., & Eyring, A. R. (1995). Factors related to job acceptance decisions of college recruits. Journal of Vocational Behavior, 47, 193–213. 65

Dineen and Soltis

Van Hoye, G., & Lievens, F. (2007a). Investigating Webbased recruitment sources: Employee testimonials versus word-of-mouse. International Journal of Selection and Assessment, 15, 372–382.

Williamson, I. O., & Cable, D. M. (2003). Organizational hiring patterns, interfirm network ties, and interorganizational imitation. Academy of Management Journal, 46, 349–358.

Van Hoye, G., & Lievens, F. (2007b). Social influences on organizational attractiveness: Investigating if and when word of mouth matters. Journal of Applied Social Psychology, 37, 2024–2047.

Williamson, I. O., Cable, D. M., & Aldrich, H. (2002). Smaller but not necessarily weaker: How small businesses can overcome barriers to recruitment. In J. Katz & T. Welbourne (Eds.), Advances in entrepreneurship, firm emergence, and firm growth (Vol. 5, pp. 83–106). Greenwich, CT: JAI Press.

Wanous, J. P. (1992). Organizational entry: Recruitment, selection, orientation, and socialization of newcomers (2nd ed.). Reading, MA: Addison Wesley.

Copyright American Psychological Association. Not for further distribution.

Williams, M. L., & Dreher, G. F. (1992). Compensation system attributes and applicant pool characteristics. Academy of Management Journal, 35, 571–595.

66

Williamson, I. O., Lepak, D. P., & King, J. (2003). The effect of company recruitment web site orientation on individuals’ perceptions of organizational attractiveness. Journal of Vocational Behavior, 63, 242–263.

CHAPTER 3

CAREER ISSUES

Copyright American Psychological Association. Not for further distribution.

Yehuda Baruch and Nikos Bozionelos

It has not been simple to arrive at a universally acceptable scientific definition of career. Nevertheless, definitions (e.g., Arnold, 1997; Arthur, Hall, & Lawrence, 1989; Greenhaus, Callanan, & Godshalk, 2000; Kram, 1985) revolve around the notion of sequential employment-related experiences through time and across space. Despite difficulties with its definition, career is a salient element in each person’s working life and demands planning and management from both individuals and organizations. In this chapter, we focus on current career issues, with particular emphasis on the changing nature of careers and with a specific focus on the elements of time and space in their evolvement. We discuss new forms of employment and their impact on careers, the notion of career success and perspectives to approach it, as well as careers in the present era of globalization. We use the term career actor to denote the individual as an active accumulator of experiences that compose a career. THE TIME ELEMENT IN CAREERS Time is unidirectional and a common denominator for everyone. Hence, a substantial number of models conceptualize careers as sequences of distinct stages (see the seminal work of Levinson, 1978, and Super, 1957, as well as Baird & Kram, 1983; Dalton, Thompson, & Price, 1977; Evans, 1986; Form & Miller, 1949; Greenhaus, 1987; Hall & Nougaim, 1968; Schein, 1978). The principles and motives behind these models are not identical. For example, Levinson’s intention was to describe individual development throughout life, a central aspect

of which was career, whereas Super’s intention was to specifically describe career development through the life course. Nevertheless, they all contain the notion of career progression through time from a nonstochastic point of view; that is, they legitimately posit that each stage is partly built on the previous ones. Transitions through the stages are not always smooth or painless (Smart & Peterson, 1997). For example, individuals may realize that they have to retrain or acquire additional formal education if they are to be competitive in the labor market. There is some support for the notion of career stage. For example, as predicted, individuals in the early stages of their careers tend to report lower organizational commitment and job satisfaction and higher turnover intentions than individuals in more advanced career stages (Ornstein, Cron, & Slocum, 1989). Baruch (2004) integrated and expanded extant career stage models into a single one that takes into account changes that have occurred since their inception. These include changes in the society, the economy, business practices, science, and technology. The model suggests the following stages in career progression: 1. Foundation, which starts from the early years of life and extends through childhood and adolescence. This part of life does not normally contain those systematic sequential work experiences that could be considered as career (though there are exceptions, such as in the case of performers and athletes). Nevertheless, for a substantial proportion of the population this stage does contain irregular, seasonal, and occasionally regular work

http://dx.doi.org/10.1037/12170-003 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

67

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

2.

3.

4.

5.

68

experiences (for a short review, see Hartung, Porfeli, & Vondracek, 2005, pp. 388–389). Furthermore, individuals appear to understand the concept of career and to think in terms of their own career development as early as in preadolescent years (Hartung et al., 2005). Most important, this period of life is associated with the accumulation of basic knowledge as well as with the development of fundamental values, attitudes, and aspirations that serve as anchors that guide, constrain, stabilize, and reinforce subsequent career direction and progress (Schein, 1978; Watson & McMahon, 2005). Indeed, attitudes toward work, occupational aspirations, and occupational expectations that are formed during that part of life are predictive of occupational aspirations, expectations, and career outcomes in adulthood (Helwig, 2008). Career entry, which involves the acquisition of knowledge, skills, and qualifications (e.g., by means of tertiary education, apprenticeship, or on-the-job training) to enter a job or a profession. This can be either within an organization or independently (e.g., one’s own business). Advancement, which involves development of expertise in the job or profession as well as generally upward movements in the organizational hierarchy, or expansion of one’s own business. This stage may be characterized by moves between functions and organizations, or by failures in one’s own business endeavors. In some cases, and especially for those whose work lives evolve within organizations, this stage may be associated with reaching a plateau, that is, a level with no prospects, either objectively or subjectively evaluated, for further advancement (see Chao, 1990). Reevaluation, which involves appraisal of the extent to which one’s aspirations have been fulfilled. This may be triggered either by endogenous factors (e.g., lack of challenge in one’s job) or by exogenous factors (e.g., plateauing, redundancy, or realization that one’s skills or profession are facing obsolescence). Reinforcement, which involves implementation of decisions that have been made in the previous stage. This may be revealed in a number of ways, ranging from refocusing on one’s job or profes-

sion with reinvigorated enthusiasm to returning to formal education (e.g., training to qualify for a different job or working toward an advanced degree) for a career change. 6. Decline, a stage that is normally characterized by consideration of and preparation for withdrawal from working life. This withdrawal may be gradual. This is a stage in which the desire to pass on one’s knowledge and experience tends to be strong in many individuals (Westermeyer, 2004). 7. Retirement, which involves disengagement from the labor market. This does not necessarily mean that the individual completely disengages from work or from one’s job or profession. For example, one may continue to follow closely the developments in one’s field, to conduct volunteer work, or to engage in bridge employment, that is, paid work of reduced amount and intensity (e.g., Griffin & Hesketh, 2008). The model is general, in line with typical models that depict social processes. Therefore, the number and sequence of stages should be considered as a general indicator rather than as applying to every single individual. Furthermore, the model deliberately does not attach specific age boundaries to stages. As also acknowledged by later (e.g., Baird & Kram, 1983; Dalton et al., 1977) and, retrospectively, by early (Super, 1990) career models, the time of entry into and the timing of transitions between stages are not invariant. Indeed, empirical research indicates that the correspondence between chronological age and career stage is far from perfect. For example, Ornstein and her associates (Ornstein et al., 1989; Ornstein & Isabella, 1990) reported correlations of .32 and .26, for their predominantly male and female samples respectively, between age and Super’s career stages. The following are some reasons for the lack of absolute correspondence: 1. Variance across jobs and professions in qualification requirements and the time needed to obtain these qualifications. Some jobs (e.g., manual work, some cases of service work) require only minimal training and no qualifications or accreditation, whereas many professions (e.g., medical, law specialists) require lengthy and rigorous education and training as well as professional accredi-

Copyright American Psychological Association. Not for further distribution.

Career Issues

tation. Furthermore, there are differences across professions in the number of stages they involve (e.g., the military). These affect the point of entry into each career stage, as well as the duration of each stage. 2. Substantial variance among individuals in the way their lives evolve. As an illustration, volatile labor markets that require constant learning and flexibility, combined with state funding for support to the unemployed, have generated in recent times a new “breed” of individuals who live a significant proportion of their lives as unemployed or underemployed (e.g., Jensen & Slack, 2003; Marston & McDonald, 2008). This may happen, for example, when individuals are unable to find comparable jobs after redundancy and cannot adapt to an alternative career direction (see Feldman, 1996; Feldman & Leana, 2000; McQuaid & Lindsay, 2002). 3. A general, and ongoing, increase over time in educational qualifications that are required to enter the job market. Whereas in the early part of the 20th century people started working in their mid-teens and university education was reserved for the affluent or tiny minority of highly talented individuals, the 21st century witnesses a growing number of people attending university (i.e., up to 50% of the age group; e.g., Davis & Bauman, 2008), hence starting full-time work when they are about 10 years older than their counterparts in the relatively recent past. Therefore, it is unrealistic and inappropriate to adopt the model under a “one-size-fits-all” mentality (Baruch, 2004; Sullivan, 1999); rather, it should be treated as a generic frame of reference. Nevertheless, age boundaries may be more precisely attached to career stages within certain professions. For instance, it is unlikely to find many full-time academics in their mid-20s, meaning that, for academics, the advancement stage does not normally start before the late 20s.

Minicycles Within Stages It is important to view stages as involving minicycles that are composed of miniature stages themselves (Baruch, 2004; Hall, 2002; Hall & Mirvis, 1995; see also Smart & Peterson, 1997; Super, 1980, 1990;

Super, Savickas, & Super, 1996, for empirical evidence). Career actors often need to evaluate their accomplishments and prospects and make decisions to reinvigorate or redirect their careers. For example, specialized workers may realize that their skills are becoming obsolete (e.g., Pang, Chua, & Chu, 2008) or professionals (e.g., engineer, physician) may attempt a switch to management to reenergize their careers (e.g., Linney, 2001). As a consequence, these individuals may decide to pursue retraining or additional education as a means to maintain their employability and reanimate their careers, respectively (e.g., Beutell & O’Hare, 2006). Naturally, the redirected career needs time to become established. However, competencies and experience gained in earlier cycles are likely to facilitate the advancement stage of subsequent cycles, hence accelerating the cyclical process (Baruch & Quick, 2007). The frequency of minicycles has been increasing over the years. Continuous changes in the nature of work and skill requirements have been part of the cause. These changes have been effected by augmentation in competition and the accelerating pace of technological development (e.g., Iida & Morris, 2008; Landry, Mahesh, & Hartman, 2005), which has led to the decline or disappearance of industries with their attached jobs (e.g., the mining industry in the United Kingdom) along with the parallel creation of new industries. Another part of the cause has been the constant organizational transformation and the use of downsizing as part of the repertoire of organizational strategies (e.g., Tsai, Wu, Wang, & Huang, 2006); these factors have diminished the commitment to the employer (Baruch, 1998; Littler, Wiesner, & Dunford, 2003; Rubin & Brody, 2005). As a consequence, individuals are more likely to change jobs and work environments, either involuntarily (e.g., because of redundancy) or voluntarily (i.e., as a planned move for career enhancement; Farber, 2008). In either case, they have to reestablish themselves. Key capacities for individuals in the process of minicycles are the ability to learn and the motivation to repeat the learning process (Hall & Mirvis, 1995). Though learning has always been an important factor in the way careers evolve, its importance has become even more critical in the contemporary era. Donohue (2007), for example, found that individuals pursuing 69

Baruch and Bozionelos

career change are much more likely (odds ratio of 2.01) to engage in activities that involve learning, such as skills improvement by means of education.

Copyright American Psychological Association. Not for further distribution.

A Contemporary Career Stage Model Baruch’s model, with some further modifications, has been incorporated into a comprehensive model of careers, depicted in Figure 3.1. The model aims to depict the current state of knowledge on career progression through stages, as discussed earlier. It attaches age intervals to stages, but this should serve only as a general indicator. The model considers foundation, career entry, and advancement as applicable to all individuals. Advancement is viewed as universal because the first years after career entry always contain advancements in experience, knowledge, and skills, hard and soft. Though these are not always translated into increases in status, they certainly advance the marketability of career actors. Reinforcement is included because it applies to the substantial numbers of individuals who still engage in traditional careers and to those who may abandon large organizations to pursue noncorporate careers (discussed later). However, it is by no means applicable to everyone (this is denoted by the faded script in the figure).

No reevaluation stage is included, unlike in earlier models, because this has been associated with the alleged midlife crisis of the early to mid 40s (Levinson, 1978), for which there is no support (Lachman, 2004). Instead, reevaluation may occur at any point in life that significant events, especially work-related ones, occur (Lachman, 2004). This fact was incorporated into the model by means of recurring cycles of evaluation, refoundation, and reestablishment, which are characterized by constant learning (also suggested by Hall, 1993, as cited in Hall & Chandler, 2005). The last two stages have been labeled “gradual change of roles,” instead of “decline,” and “retirement(?)” The latter signifies that retirement is included only as a possibility, despite the fact that human lives are (still) finite, and their progression toward the end is associated with physical decline. This is an important theme for all parties involved— individuals, organizations, and societies; hence, it deserves some consideration. Decline and retirement(?) There is no dispute that physical abilities, such as strength and agility, decline with age (e.g., Miyamoto et al., 2008; Over & Thomas, 1995). Yet, only a small proportion of jobs or occupations are highly physically demand-

FIGURE 3.1. A contemporary model of career progression through time. 70

Copyright American Psychological Association. Not for further distribution.

Career Issues

ing, especially in the contemporary era when technological advances and the change in the nature of the economy have shifted the great majority of the workforce away from manual work (Department of Professional Employees, 2008). Hence, the decline in physical strength and agility is an issue for only a small proportion of career actors. When cognitive ability is considered, the picture becomes even more positive for aging career actors. From early adulthood until the age of 65, by which time most individuals in developed countries retire (e.g., Friedberg, 2007), there is a slight decline in general cognitive ability (d = −.20; Schaie, 2005; see also Schwartzman, Gold, Andres, Arbuckle, & Chaikelson, 1987).1 However, this change is not uniform across all aspects of cognitive ability, with certain domains showing decline (i.e., those that pertain mostly to fluid intelligence, such as perceptual speed), other domains demonstrating improvement (i.e., those that pertain mostly to crystallized intelligence, such as verbal ability and mechanical tasks), and many other domains showing no alteration (Arbuckle, Maag, Pushkar, & Chaikelson, 1998; Schaie, 2005; Schwartzman et al., 1987). For example, in one of the most comprehensive longitudinal studies, Schaie (2005) found a decrease in scores on numerical ability (d = −.29) and perceptual speed (d = −.26) from age 25 to age 67 but also gains in scores on verbal ability (d = .24) and no significant changes in spatial orientation and inductive reasoning. Any decreases in cognitive ability that can be considered of substance (i.e., of magnitude of d in the order of .5 and beyond; Cohen, 1988) appear in the 7th and 8th decades of life (Schaie, 2005).2 Importantly, however, that engagement in intellectual activities, such as reading and cognitive training (e.g., in the form of mentally stimulating activities), can prevent decline or can even improve cognitive ability in late adulthood (Arbuckle et al., 1998; Ball et al., 2002; Boron, Willis, & Schaie, 2007; Schaie, 2005). Furthermore, empirical research has reported no relationship between age and creativity, a quality that is in demand in today’s economy, from the age of the early 20s until 60 (Binnewies, Ohly, & Niessen, 2008). Finally, older

individuals can apparently be at least as comfortable in using new, sophisticated technology as their younger counterparts (Bozionelos, 2001). Moreover, improvements in the way of life and medical advances have been dramatically increasing life expectancy and delaying the onset of physical decline (e.g., Kinsella, 2005). For example, life expectancy in the United States has increased from 47 years to nearly 77 years within the past century (e.g., Sonnega, 2006), and similar gains have been witnessed around the globe (Kinsella & Phillips, 2005). This means that individuals are able, and will be more so in the future, to continue working and developing themselves far beyond what traditional retirement schemes and career stage models assumed at the time of their inception. In addition, increases and projected increases in life expectancy (United Nations, 2006) render the viability of pension programs highly problematic if people continue to retire at the ages that were the norm in previous generations (see Feldman & Ng, 2007). This is further exacerbated by the continuous decline in birth rates in most developed and fast developing economies (United Nations, 2008), which limits the numbers of younger individuals who enter the workforce. If people continue to retire by the “traditional” age of 65 and, even worse, at early retirement to avoid redundancies of younger employees, the workforce will be shrinking at levels that may not be able to sustain the economy (e.g., Kozhaya, 2007). It appears, therefore, that for most people decline and retirement, although still relevant, would come at much older ages than might have been anticipated. This has implications for individuals (e.g., for targets set in personal career planning), organizations (e.g., reconsideration of their internal labor markets), and states, where legislators will have to confront these realities. THE SPACE ELEMENT IN CAREERS The space element encompasses both the job or profession (i.e., roles and tasks the individual performs) and the environment (social, organizational) in

1

Effect sizes, d, in this section were calculated by the authors using relevant information provided by Schaie (2005).

2

At this point it is worth noting that the peak in cognitive ability is not reached in early adulthood but in middle adulthood between the ages of 40 and 60, depending on the faculty; that is, the relationship between adult lifetime and cognitive ability is of inverted U-form (see Schaie, 2005).

71

Baruch and Bozionelos

Copyright American Psychological Association. Not for further distribution.

which work takes place (Baruch, 2004; Baruch & Rosenstein, 1992). Hence, the space aspect of careers is multidimensional, because work environments differ substantially in their structures, cultures, sizes, and purposes. Before progressing further, the following points need to be made: 1. It is rather inappropriate to think of careers as always attached to particular jobs or professions or that they are composed of a natural evolution of roles within the same general occupational context (e.g., someone starts as an engineer in a corporation, and as he or she advances up the corporate ladder, he or she assumes managerial and executive responsibilities). That consideration had its roots in the economic and labor market structures that prevailed for a few decades before and after World War II (to be discussed later). The relative economic stability of that era meant that a substantial proportion of (e.g., Parkin, Powell, & Matthews, 2007), though by no means all, individuals were able to perform the same job or stay within the same profession for the totality of their working lives (e.g., Sullivan, 1999). However, as also implied earlier, this gradually changed in the last quarter of the previous century. The turbulent economic environment and the shift in the nature of western economies from manufacturing to services (Amsden, 2001; “The Great Jobs Switch,” 2005; Shambaugh, 2006) forced many individuals to change jobs because the jobs they had been trained for were not available anymore. Hence, nowadays, individual careers should not be considered as progressing in parallel with a specific job or profession or even a series of related jobs. 2. Experiences that pertain to work roles and environments are the central, but not the only, aspect of careers. Personal life, and its related roles and experiences, is also part of them. Careers are shaped by both work and personal experiences between which there is a constant interplay (Greenhaus & Beutell, 1985; Greenhaus & Powell, 2006; Super, 1980). 3. The actual physical space where work takes place has been changing over the years. Traditionally, most individuals would fulfill their work-related tasks and roles within particular physical locations (e.g., office, desk, factory, till) where they 72

would be physically placed. However, advances in technology (i.e., especially in telecommunications but also in transportation), societal changes (e.g., awareness of work–life balance; see, for example, Gregory & Milner, 2009), and changes in the ways business is conducted (e.g., globalization that places collaborators, clients, or suppliers within physical distance; see, for example, Moore et al., 2008) have led to changes in this respect. Hence, in the contemporary era, individuals may conduct their work-related activities at home (Baruch & Nicholson, 1997) or at multiple locations (i.e., teleworking; WorldatWork, 2009; Thatcher & Zhu, 2006) or may have physical bases that periodically change (e.g., individuals on expatriate missions). Therefore, the pattern in which careers move through space has shifted over the years as a consequence of changes in business practices and in society, and this change is reflected in the notions of the traditional, the boundaryless, the protean, and the postcorporate career.

The Traditional Career The notion of the traditional career reflects the view that careers are bound to single organizations or particular professions that individuals normally join at the entry stage of their careers. The consequent career stages evolve within that same organization or profession. Organizational and/or professional structures would support the advancement stage wherein individuals would gain promotions, moving up the organizational and/or professional ladder. These structures would provide help (e.g., by supporting acquisition of formal education, such as a graduate degree or additional professional accreditation), structures, and opportunities for the reinforcement stage so that individuals would be able to revamp their careers and assume more senior roles in the organization or in the profession. They would also assist in the final stages, offering, for example, retirement preparation programs (e.g., Peiperl & Baruch, 1997). Most career stage models discussed earlier made the implicit assumption that careers generally fit that pattern. An important aspect of the notion of the traditional career was that the career was managed by the organization (e.g., Arthur &

Copyright American Psychological Association. Not for further distribution.

Career Issues

Rousseau, 1996; Gutteridge, Leibowitz, & Shore, 1993). The necessity or the pressure on individuals to actively manage their careers was limited. The notion of the traditional career was shaped by the societal and work structures of the industrial and part of the postindustrial era and can be roughly placed as starting with the railroad and continuing until the 1970s. The growth of the economy during that era gave rise to organizational structures that abided by Weberian principles (i.e., large organizations with clearly delineated roles and hierarchies and clear prescriptions of rules for movement between roles and hierarchical layers). These structures supported the notion of the traditional career (Reitman & Schneer, 2008; Sullivan, 1999), especially for the white-collar workforce (Cappelli, 1999). Career systems were hierarchy-based (e.g., Whyte, 1956; Wilensky, 1961), and career progression was linear and mostly based on a series of contests or tournaments (Rosenbaum, 1979) the winners of which would progress in the hierarchy. Hughes (1937) considered that “a career consists, objectively, of a series of status and clearly defined offices . . . typical sequences of position, achievement, responsibility, and even adventure” (p. 409). It should be noted that the traditional notion of career does not exclude lateral moves, as far as these are part of a planned organizational career system and as far as employment within the organization is secure (Peiperl & Baruch, 1997). Many considered the traditional career as the ideal type of career or career archetype (e.g., Reitman & Schneer, 2008). However, it is erroneous to assume that during the time period that gave rise to that notion of career organizational structures were entirely stable and careers devoid of unforeseen events or that there were not substantial numbers of individuals who were working outside organizational boundaries as self-employed (e.g., Blanchflower & Shadford, 2007). Therefore, the traditional career should be viewed as a metaphor that generally fits the mentality of careers in a particular era rather than as the norm in that era.

The Boundaryless Career The notion of the boundaryless career was introduced in response to the changing conditions in the econ-

omy and in the corporate world that commenced in the 1970s and became evident in the 1980s. At that time, competition increased and the basis of the economy started shifting from process toward knowledge. In parallel, governments started weighing financial objectives, such as reduction in deficit and debt, more heavily; which meant reduced subsidies for problematic sectors (e.g., the manufacturing industry in the United Kingdom) and privatization (Baruch, 2004). As a reaction to these conditions, and to be financially efficient, organizations resorted to tactics that included restructuring and downsizing or moving operations abroad (e.g., Coucke, Pennings, & Sleuwaegen, 2007), which normally involved reshuffling and flattening of hierarchies as well as layoffs (e.g., Thornhill & Saunders, 1998). In addition, the partial shift in organizational priorities (e.g., toward maximization of shareholder value) motivated the acquisition of skills from outside rather than developing these internally (Cappelli, 1999, 2006). As a consequence, the stable internal labor markets that served as platform for the traditional career ceased to be the norm. Another, though less powerful, force toward the change in career patterns was the shift in values of (Western) societies toward individualism as a result of the affluence of the post-World War II era and the liberal movements of the 1960s (for the connection between societal affluence and individualism, see Kashima & Kashima, 2003). Core values of individualism are personal choice and engagement in activities, including work tasks that are meaningful and interesting for the individual (e.g., Triandis, 1995). Hence, career actors were more likely to move between organizations on their own initiative in order to advance their careers or to find personally fulfilling roles (see Reitman & Schneer, 2008). Resiliency and self-efficacy, which pertain to the ability to sustain and rebound from adversary events, became a crucial career competence under these conditions (e.g., Betz, 2007; London, 1983; Rickwood, Roberts, Batten, Marshall, & Massie, 2004; Waterman, Waterman, & Collard, 1994). In parallel, as we have seen, the ability and especially the motivation for constant learning also became critical factors (e.g., Hall & Mirvis, 1996). The idea of the boundaryless career, as initially understood, denotes the lack of being bound into a 73

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

single organization and the lack of a single direction in career movement (Arthur et al., 1989; DeFillippi & Arthur, 1994). However, despite the broadness of its meaning (see Arthur & Rousseau, 1996), the connotation that had been typically assigned to the term was quite narrow (Briscoe & Hall, 2006; Sullivan & Arthur, 2006). Simply equating boundarylessness to physical mobility across organizational boundaries was a rather simplistic way to justify the attention that the concept enjoyed. Therefore, the meaning of the term has been recalibrated and refined over time. Authors (Baruch, 2006; Sullivan & Arthur, 2006) also considered boundarylessness in terms of the mental preparedness of career actors to be resilient in their choices and expectations so they are able to adapt to environmental contingencies. Preparedness to move even in cases when this is not personally convenient is not the same as actual movement. Indeed, empirical research has implied a distinction between mental preparedness for mobility and actual physical mobility of career actors (Briscoe, Hall, & DeMuth, 2006). Therefore, the current dominant thinking is that the notion of the boundaryless career encompasses both physical and psychological mobility (Lazarova & Taylor, 2009; Sullivan & Arthur, 2006) and that career actors can be classified according to their position into these two, apparently orthogonal, dimensions (though there is still some debate; see, for example, Greenhaus, Callanan, & DiRenzo, 2008, who conclude that it is more appropriate to view boundarylessness exclusively in terms of physical mobility).

The Protean Career The protean idea (Hall, 1976) refers to careers as driven by the values of career actors (i.e., what individuals themselves consider important and worthy) and directed through personal choice rather than by external agents, such as the organization (Briscoe & Hall, 2006; Hall, 2004; Hall & Mirvis, 1996). (See also Vol. 3, chap. 4, this handbook.) Therefore, in a similar vein to the boundaryless career, the protean career can be conceived along two dimensions: the extent to which the career is driven by values and the extent to which it is self-directed. Briscoe et al. (2006) developed a scale measure of career propensity that assesses both values-drive (e.g., “In the past 74

I have sided with my own values when the company has asked me to do something I don’t agree with” and “What’s most important to me is how I feel about my career success, not how other people feel”) and self-directedness (e.g., “Ultimately, I depend on myself to move my career forward” and “I am responsible for my success or failure in my career”). Hence, career actors can be categorized according to their positions into these two dimensions (Briscoe & Hall, 2006). In a complementary manner, and to the extent that actors are able to conduct their careers according to their values and in the direction they wish, the protean career can also be viewed as a unidimensional attitude that reflects subjective career success (Baruch, Bell, & Gray, 2005; Baruch & Quick, 2007), a concept that will be reviewed later. The notion of the protean career does appear to capture the demands that the contemporary era poses to career actors. For example, in a study of retired navy admirals who had pursued second careers, Baruch and Quick (2007) found that protean career orientation was associated with less stress, less emotional exhaustion, and greater satisfaction with the career transition process (r = −.32, −.27, and .26, respectively); shorter time in making the transition (i.e., to find the new job, r = .13); as well as higher overall career satisfaction (β = .63). And De Vos and Soens (2008) studied individuals who had received career counseling and found that protean orientation was strongly related to career insight, which reflects insight of the self, skills, and aspirations (β = .87), and to proactive career enhancement behaviors (β = .76). At first glance, the notion of the protean career may appear to overlap with the notion of boundarylessness, and this is the view of certain authors (Granrose & Baccili, 2006). However, the two concepts are distinct. The concept of the protean career focuses on the orientation, both value-driven and behavioral, of the actor toward the career, whereas the boundarylessness concept is primarily concerned with the structure of the career itself. As an illustration, Briscoe and Hall (2006) described the “solid citizen” as the individual who is driven by personal values and who is in full control of his or her career, whilst at the same time choosing to pursue one’s career within a single organization because

Career Issues

Copyright American Psychological Association. Not for further distribution.

the organizational culture is in line with his or her personal values. The career of that individual is protean in orientation but within organizational boundaries from a structural perspective. In line with the conceptual distinction between the protean and the boundaryless career, Briscoe et al. (2006) found average correlations of .28 and .38 (.21 and .28 for their executive sample alone) between scores on value-driven and self-directed career orientations and scores on the boundaryless career mindset, respectively. The relationship between the boundaryless and the protean perspectives. The establishment of the conceptual distinction between the protean and the boundaryless career permits the contemplation to combine them. The combination of the two dimensions of boundaryless career, psychological and physical mobility, along with the two dimensions of protean career, value-drive and self-direction, yields 16 potential career profiles. Briscoe and Hall (2006) concluded that eight of these profiles are realistically likely to exist in contemporary career contexts, and they provided labels and descriptions of them. At the one extreme, there are individuals who score low on all four dimensions and who are described with the label trapped/lost. At the other extreme, there are those who score high on all four dimensions and who are labeled protean career architects. Exhibit 3.1 presents the 16 potential profiles along with the 8 profiles that were identified as realistic. Segers, Inceoglou, Vloeberghs, Bartram, and Henderickx (2008) were partly successful in identifying distinct relationships between some of these profiles and an array of motives. For instance, protean career architects were more likely to be motivated by personal principles (reflective of the value-drive dimension of the protean career), achievement and personal growth (reflective of the self-direction dimension of the protean career), autonomy and affiliation (reflective of the psychological mobility dimension of boundarylessness), and personal interests (pertinent to both the psychological and physical mobility dimensions of boundarylessness; Segers et al., 2008). Additional work is certainly needed to further solidify the foundation of the two concepts and their resulting career profiles. Furthermore, career scholars should be alert to iden-

tify career profiles that apparently do not exist in the present era but may emerge along with changes in the economic and social environment. Work on the further development and integration of these dominant perspectives will be of substantial benefit to the field of careers because it will provide the tools for a more accurate and precise understanding of the processes involved in career pathing of individuals and, hence, for a fine-grained diagnosis of career situations and guidance on career development.

The Postcorporate Career The term postcorporate career (Peiperl & Baruch, 1997) refers to careers that take place outside large organizations, the actors being individuals who have left such organizations, voluntarily or involuntarily, or individuals who are unable or unwilling to pursue corporate careers because of the uncertainty that is inherent in them. The postcorporate notion was epitomized by the situation at the end of the previous century when many large organizations consistently resorted to layoffs and outsourcing their functions (e.g., Cascio, 1995; Insinga & Werle, 2000; Leanna & Feldman, 1992), hence creating the need for smaller and agile organizations to provide outsourcing services (e.g., computer network maintenance, specialized consultancy) and compelling many individuals to work either alone (e.g., independent vendors) or outside large corporations (e.g., small consultancy firms, partnerships). The postcorporate career pattern cannot be easily located on the protean-boundaryless framework because it is difficult to identify with certainty the position of its actors on self-directedness, a protean dimension, and on the physical and psychological mobility dimensions of boundarylessness. That is, actors with different positions (i.e., high or low) on these dimensions could be equally likely to pursue postcorporate careers. Hence, it is considered separately. Postcorporate careers do not normally provide hierarchical rewards; rather, they provide intrinsic (e.g., achievement) and financial rewards to their actors (e.g., Cascio, 1995; Feldman & Bolino, 2000). The notion of the postcorporate career is still highly contemporary, for example, with the major waves of redundancy following the credit-crunch crisis that 75

76

“Fortressed” These career actors are driven, without self-direction, solely by personal values, and they lack ability to recognize opportunities. Hence, sense of success is achieved only under conditions that both match their personal values and are stable (e.g., within a financially healthy organization with traditional structure whose culture matches the values of the career actor). Such conditions may be difficult to find in the contemporary era. Not likely to encounter in contemporary career contexts.

“Trapped/lost” Because of low proactiveness and inability to see possibilities across boundaries, career actors are restricted to very narrow career possibilities over which they can exercise limited control. Success is more a function of luck than of direct control.

“Wanderer” Because personal values are of limited importance to them, these career actors generally adapt easily. However, their careers are mostly controlled by circumstances because they lack the ability to identify, evaluate, and actively drive themselves to opportunities across boundaries.

No

No

Yes

Yes No

No

No No

Psychological mobility

Values-driven Self-directed

Physical mobility

Boundarylessness

Protean orientation

Not likely to encounter in contemporary career contexts.

Not likely to encounter in contemporary career contexts.

No Yes

Not likely to encounter in contemporary career contexts.

Not likely to encounter in contemporary career contexts.

Yes Yes

Exhibit 3.1 Combining the Protean With the Boundaryless Careers Orientation: Emerging Career Profiles

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

Yes

Yes

“Solid citizen” Career actors are able to actively manage their careers according to their values. They have the ability to adopt a broad perspective across boundaries. Either because of values (e.g., belief that they should be committed to the employer) or because of circumstances (e.g., family or other constraints) there is limited ability to move. “Protean career architect” Driven by personal values, career actors self-direct their careers, being aware of possibilities across boundaries and willing to physically cross them. Presumably, few individuals fall into this category. A potential challenge for these career actors may be to find or maintain a balance (e.g., to temporarily resist their values in order to stay in a situation for as long as is required to establish themselves).

“Organization man/woman” Career actors generally tend to conduct their careers in ways that fit the needs of others (e.g., the employer) rather than their own needs. This is because their motivation to physically move is limited, though their preparedness to do so is high (hence, they may accept internal transfers); they are either unclear about or place low importance on their values and needs. “Hired gun/hired hand” Career actors are able to identify opportunities across boundaries, are psychologically prepared to move, and are willing to physically move toward the directions they decide. Hence, these career actors are generally adaptive in their careers. However, the fact that their career decisions are not directed by values (either because they are unclear about these or because these are of a lower priority to them) may lower their probabilities of finding an environment where they fit perfectly. They may also lack loyalty.

“Idealist” Career actors are able to identify opportunities across boundaries, but they generally lack the motive and ability to direct themselves towards these. Such individuals may thrive within environments that fit their values and allow and encourage information sharing.

Not likely to encounter in contemporary career contexts.

Not likely to encounter in contemporary career contexts.

Not likely to encounter in contemporary career contexts.

Note. Based on Briscoe and Hall (2006).

Yes

No

Copyright American Psychological Association. Not for further distribution.

Career Issues

77

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

commenced in 2008 (e.g., Holman, 2009). However, postcorporate careers have not become the norm, as some had predicted (e.g., Gray, 2001). The traditional, corporate-bound, career has proved quite robust and has survived through the pressures of competition and rationalization.

in the contemporary environment (i.e., traditional, boundaryless, and postcorporate) or actors’ mentalities (i.e., the protean and partly the boundaryless). As seen, there have been recent attempts to combine some of these concepts, and further systematic attempts in this direction will be of great benefit.

Is the Traditional Career Still Alive?

Holland’s Theory: Struggling Through Time to Find the Ideal Space

As seen, the boundaryless and the postcorporate career notions were developed to describe changing career structures, and they were rather successful in their mission. However, has the traditional career vanished? The answer to that question is unequivocally negative, despite popular views (e.g., Hansen, 2009). This is what, for example, was demonstrated by a recent large-scale study (Moss, Salzman, & Tilly, 2008) that focused on call centers in the financial and retail industries and reconstructed the ways in which their organizational structures and internal labor markets have evolved since the early 1980s. Call centers epitomize the new economy; hence, they provided a most appropriate setting to investigate whether the traditional career has been inexorably disappearing. Findings suggested that the principles of the traditional career, with long-term employment and upward moves in the hierarchy being its trademarks, were clearly evident in the call center industry. Furthermore, far from disappearing, vertical career paths and additional hierarchical layers were often introduced to improve effectiveness and motivate and retain able employees, roles that simple monetary rewards were unable to fulfill (Moss et al., 2008). Findings that are suggestive of the resilience of the traditional career have also been reported in more traditional industries, such as in the public sector (Koskina, 2008; McDonald, Brown, & Bradley, 2005) as well as in the Japanese private sector (Iida & Morris, 2008). Therefore, authors (Baruch, 2006; Lips-Wiersma & Hall, 2007) have called for balancing traditional and contemporary viewpoints: The traditional career is present, but the boundaryless and the postcorporate career also describe contemporary patterns, which, however, are not the norm. It should be finally noted that the four notions of career described here are not mutually exclusive. They either represent career structures that are found 78

Holland (1959, 1997) advanced the idea that individuals are inclined to enter work environments whose characteristics (e.g., demands, opportunities for cultivation and expression of talent, prestige) fit their personality and interests. He categorized individuals into six types: realistic, investigative, artistic, social, enterprising, and conventional. These types fall on a circumplex that is determined by two axes: working with things (e.g., materials, machines, tools) versus with people (e.g., care), and working with data (e.g., facts) versus with ideas (e.g., insights, concepts; Prediger & Vansickle, 1992). Each of Holland’s types reflect particular interests and values and, hence, preferences for activities that are attached to particular jobs, professions, or occupations. There is general support for the theory (Holland, 1996; Tracey & Rounds, 1993), which has enjoyed substantial popularity in practice (e.g., career counseling; Rayman & Atanasoff, 1999). More recently, cognitive abilities have been added as a parallel dimension to interest types; that is, individuals are more likely to pursue careers within those jobs or professions whose cognitive demands (e.g., task complexity) they have the ability to handle (see Armstrong, Day, McVay, & Rounds, 2008). Holland’s idea, therefore, implies that careers contain a “struggle” for career actors to find occupational environments that fit their interests and values and whose demands they can meet. This is relevant to the consideration of both the time and the space element of careers. For example, it can assist in the understanding of career-related decisions and mobility at the foundation and career entry stages because individuals are apparently more likely to choose educational subjects and move toward jobs that are in line with their interest types (see McLaughlin & Tiedeman, 1974). Furthermore, the theory appears compatible with the concepts of the protean and

Copyright American Psychological Association. Not for further distribution.

Career Issues

boundaryless career because the former emphasizes value-drive and self-directedness, and the latter emphasizes mobility across borders (until the individual finds an environment that fits his or her interests and values). Nevertheless, the theory needs further development. For example, there is lack of clarity regarding what exactly the “types” represent. Holland apparently considered interests and values as functionally identical to personality, but these are distinct, though related, constructs (e.g., Hogan & Blake, 1999); current thinking about the theory considers personality, interests, and abilities as distinct dimensions (Armstrong et al., 2008; see also chap. 5, this volume). Furthermore, changes in the nature of the economy have caused the shrinkage of environments that fitted certain interest types. For example, realistic types (who tend to be attracted to specialist blue-collar jobs) flourished in the era in which the economy was based on manual work and manufacturing, but may have difficulties in finding appropriate work environments in today’s service-driven and largely knowledge-based economy (see also Holland, 1996). In this respect, the theory may need to be developed to capture the influence of economic and social changes on occupational environments. This could be accompanied by work, which is lacking so far, to directly link Holland’s theory with space notions of careers. For example, the degree of congruence between individuals’ career interests and the characteristics of their work environment should be inversely related to their psychological boundarylessness (representing propensity to move to a different environment), with the relationship being moderated by the value-drive dimension of career proteity. A SHORT INTERLUDE TO COMMENT ON THE UNIVERSALITY OF CAREER MODELS Accounts of the development of the previously discussed notions of careers take the perspective of what we would refer to as the “Western developed world,” which generally includes North America, Western Europe, Oceania (i.e., Australia and New Zealand), Japan, as well as some special areas (e.g., Hong Kong, Singapore). The perspective of this part of the world was adopted because its economy, albeit with some

variance across regions, has been functioning according to the principles of capitalism for most of the 20th century (and especially after the second World War); political systems in this developed world generally allowed individuals to make choices according to their personal needs, attitudes, and values (though within norms influenced by the national cultures). Whereas the rest of the world was functioning under different political and economic conditions, the previously discussed notions of careers are sufficient to describe career patterns around most parts of the globe. To illustrate, it appears that the notion of the traditional career captures well careers in the Soviet Union (which included Russia and nearby countries) and the countries that belonged to its sphere of direct control. Since the establishment of the communist regime in the early 1920s, and especially after the end of the second World War, unemployment was kept at zero levels and virtually everyone was directly or indirectly employed by the state. Employment was permanent and movement across enterprises was minimal, because it was strongly discouraged (Brand, 1991). Career ladders were rigid, financial compensation was tied to hierarchical position, and promotions were based on seniority and on connections with the communist party. Thus, the traditional career was both the archetype and the norm in the Soviet era (though apparently few there would consider their careers as “ideal”). A difference from the Western notion was that the largest proportion of the workforce (according to some estimates, up to two-thirds; Maslova, 1991) was employed in manual work, including agriculture. The economic and social conditions that gave rise to protean, boundaryless, and postcorporate careers were absent in the Soviet era. Similar conditions to those in the Soviet Union were prevailing in China until the 1980s; with rigid state control over productivity targets, organizational structures, and hiring, as well as complete dissociation between individual performance and organizational rewards (e.g., Ding & Warner, 2001). The apparent difference in culture between the Soviet Union and China (with the Soviet Union being substantially more individualistic as a society than the collectivistic China; see Trompenaars, 1994) did not have much of an impact on the structure and ideology of careers because, in contrast to 79

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

the West, the prevailing structures allowed minimal expression of thinking or individual choice. The notions of the protean and boundaryless career became especially relevant to China in the second period of economic reforms that commenced in 1993 and are still ongoing, because since that point in time the economy has been largely, though not entirely, operating according to capitalist principles (though to assume equality in the treatment of all employees in China in the preeconomic reform era is an erroneous oversimplification; see White, 1989). In the past 2 decades, the Chinese economy seems to have gone, albeit at much faster rate, through the same stages that gave rise to the notions of the protean, boundaryless, and postcorporate careers in Western economies. For example, the restructuring of the Chinese economy was accompanied by substantial layoffs in state-owned enterprises (e.g., Hassard, Morris, Sheehan, & Yuxin, 2006). Back to the question of universality of career models, it should be noted that a large proportion of the world population (mostly in developing countries) might not have a cohesive set of work experiences that fits our idea of career and may not even have a concept of it. The lives, work and nonwork, of many people are still quite similar to the lives of human generations before the industrial revolution, the stage in history that gave rise to those sequential systematic work experiences we today conceive as career (Baruch, 2006). CAREERS AND THE PSYCHOLOGICAL CONTRACT The study of careers cannot be conducted in isolation from the psychological contract. (See also Vol. 3, chap. 5, this handbook.) As seen, careers are associated with particular mentalities that include expectations and obligations (e.g., the traditional career is accompanied by the expectation that the organization provides employment in exchange for fair effort). The psychological contract refers to perceptions regarding reciprocal obligations between the employee and the employer (Argyris, 1960; Rousseau, 1995). Recent meta-analytic work (Zhao, Wayne, Glibkowski, & Bravo, 2007) indicates that the extent to which employees perceive the psychological contract as 80

honored or breached is related to key outcomes, which include work attitudes, such as job satisfaction, organizational commitment and turnover intentions (absolute ρ = .54, .38, and 42, respectively), and actual performance, such as in-role and extra-role behaviors (absolute ρ = .24 and .14, respectively). The evolution in the nature of careers has been accompanied by changes in the nature of the career actor’s psychological contract. Psychological contracts can be conceptualized as falling into a continuum, one end of which consists of relationships of a purely transactional nature and the other end of which consists of relationships that include transactional and relational elements (e.g., De Cuyper, Rigotti, De Witte, & Mohr, 2008). Contracts of a transactional nature are characterized by high reciprocity, normally based in exchange obligations of a monetary nature, such as pay-for-performance, and no expectations for long-term relationship. Relational psychological contracts do not contain limits regarding relationship duration and include emotional elements, such as commitment and loyalty (De Cuyper et al., 2008). Clearly, the notion of the traditional career is compatible with a psychological contract that includes relational elements. Employment insecurity, unclear career paths, and involuntary employer changes have largely stripped the psychological contract from its relational elements (Baruch, 1998; Bozionelos, 2003a; Brown, 2005). For example, Allen, Freeman, Russell, Reizenstein, and Rentz (2001) found that reductions in downsizing survivors’ feelings of employment security were associated with lowered organizational commitment (β = .31) and heightened turnover intentions (β = −.18), with the decrease in organizational commitment persisting 1 year after the downsizing. Krause, Stadil, and Bunke (2003) reported similar findings from a study with employees who simply witnessed layoffs in another organization rather than experiencing layoffs within their own organization. This change in the psychological contract of career actors feeds back to career patterns: Career actors have low loyalty to their employers or are reluctant to commit to corporate careers, factors that reinforce modern forms of careers, such as the boundaryless and the postcorporate. This cycle may

Career Issues

be difficult to break, especially considering the current economic climate and the fact that the new psychological contract may have been embedded in the mentality of the workforce.

Copyright American Psychological Association. Not for further distribution.

Employability Instead of Security and Vertical Advancement? The new psychological contract, which parallels the notions of boundaryless and postcorporate career, places emphasis on employability (Scholarios et al., 2008), which refers to work-centered adaptability that enhances individuals’ ability to identify and seize career opportunities by means of facilitating intra- and interorganizational job movement (Fugate, Kinicki, & Ashforth, 2004; Van der Heijde & Van der Heijden, 2006). In the present era, employability is fundamental for most actors to maintain a career (Fugate et al., 2004) or at least an uninterrupted career. For example, employability has been found to relate strongly to intensity of job search (β = .71) and probabilities of reemployment (β = .49) in cases of job loss (McArdle, Waters, Briscoe, & Hall, 2007). The fact that career actors may operate according to the new psychological contract, which places emphasis on employability instead of employment security and steady vertical advancement, does not necessarily mean that they welcome it (Baruch, 2004). For example, employees highly value promotions, a trademark of the traditional career and its attached psychological contract, because these represent increases in status and formal acknowledgement of service (e.g., Moss et al., 2008; Nicholson & De Waal-Andrews, 2005). Individuals also value highly employment security. For example, in a recent cross-generational study, Dries, Pepermans, and De Kerpel (2008) found that over 80% of participants attached prime importance to employment security, with no difference between those born in the years from 1925 to 1945, who presumably conducted their careers in an era dominated by the traditional career notion, and those born after 1981. Furthermore, it is rather questionable whether enhancement of employability is always feasible or counterbalances the lack of security and steady advancement (Baruch, 2001, 2004; De Cuyper, Bernhard-Oettel, Berntson, De Witte, &

Alarco, 2008). To illustrate, Taylor, Audia, and Gupta (1996) found no evidence that increased job responsibility could offset the negative effects of reduced promotion opportunities on voluntary turnover. In addition, the new career with its quest for employability is associated with elevated stress levels and health risks for career actors (e.g., Cooper, 2006), whilst it does not appear advantageous in terms of monetary rewards or subjective feelings of success (e.g., Reitman & Schneer, 2003). Therefore, a stage may have been reached that is in the best interests of neither career actors nor organizations; as the new psychological contract with its emphasis on employability means reduced commitment on the part of the workforce (see, for example, D’Amato & Herzfeldt, 2008). Some grounds for optimism? As a note of optimism, empirical work implies that career actors are able to appreciate those employers who offer them career environments that incorporate traditional career elements. Ng and Feldman (2008) found that employees who perceived that their employers provided them with better career deals (i.e., with respect to financial rewards, job security, advancement prospects, training opportunities) than the market expressed considerably greater affective and normative commitment (β = .62 and .54, respectively), which can be translated into a psychological contract with more relational elements. This may be an avenue for change. If a sufficient proportion of employers realize the benefits of offering good career deals to employees, even under conditions of pressure, the cycle may slow down or even reverse; with organizations trying to offer environments that simulate the traditional career and career actors adding relational elements to their mentalities. CAREER SUCCESS The notion of success or failure is inherent in the conceptualization of careers (e.g., Hall, 1976; see also Nicholson & De Waal-Andrews, 2005), and the identification of factors that govern the successful progression of individuals within the space element is a key issue in the study of careers. Major perspectives to career success include the structural, the human capital, and, more recently, the social capital view. 81

Baruch and Bozionelos

Copyright American Psychological Association. Not for further distribution.

Objective and Subjective Career Success Traditionally, career success was considered solely in terms of attainment of rewards that are valued by the society. This refers to the objective or extrinsic view, which judges careers via the prism of the external observer and uses objectively verifiable criteria to evaluate success (Heslin, 2005; Nicholson & De Waal-Andrews, 2005; see also Khapova & Korotov, 2007). Such criteria typically include status attained in the organization or the profession, the pace of vertical advancement in the corporate, professional or occupational hierarchy, occupational prestige, and financial attainment (e.g., Heslin). However, and in line with the notion of the protean career, career actors also conduct their own evaluations of their careers, using their personal values, beliefs, and aspirations in their judgments. In these judgments individuals use both self (e.g., levels of experienced satisfaction, perceived security, opportunities to be creative) and other (e.g., advancement and performance in comparison with others, recognition by others) reference points (Dries, Pepermans, & Carlier, 2008; Heslin) and consider past as well as prospective accomplishments (Gattiker & Larwood, 1986, 1988; Nicholson & De Waal-Andrews, 2005). This represents the subjective or intrinsic notion of career success (Gattiker & Larwood, 1986; Van Maanen, 1977). In their meta-analysis, Ng, Eby, Sorensen, and Feldman (2005) found corrected correlations between measures of objective and subjective career success in the range of .18 to .30, a finding that indicates that these are related yet clearly distinct constructs. Subjective success acquires particular substantive importance when careers are characterized by deceleration or stagnation in objective terms, which appears more likely in the present era. For example, reduced likelihood for upwards promotion relates to poorer affective commitment and greater probability of voluntary turnover (Taylor et al., 1996). Furthermore, subjective evaluations of success can serve as compasses in career decisions and as motivators to reach career-related goals, hence eventually feeding back to objective career outcomes (see Hall & Chandler, 2005). Therefore, both perspectives need to be employed in the consideration of career success. 82

The Human Capital View of Career Success The human capital view (Becker, 1964, 1975) suggests that societal or organizational rewards, such as hierarchical and income progression, are distributed according to relevant competencies which, de facto, contribute to organizational performance or to the functioning of the society. Such competencies can be acquired through a number of means, including education, training, general and jobspecific work experience, and tenure with the organization. The human capital approach is in line with the contest mobility view of intraorganizational career progression, which advocates that individuals compete with each other for organizational rewards in open and fair contests where actors are judged solely on the basis of their credentials and contributions (Turner, 1960). The concept of know-how that has been advanced by Arthur and associates (Arthur, Claman, & DeFillippi, 1995; DeFillippi & Arthur, 1994) bears substantial relevance to human capital. The idea of know-how, however, adopts specifically a career perspective because it refers to the total repertoire of knowledge, skills, and other talents that career actors have at their disposal in their quest for success. Melamed (1996a) offered a comprehensive classification of human capital attributes into three categories: General human capital, which encompasses attributes that should facilitate performance in the vast majority of jobs. These include, for example, educational attainment, cognitive ability, and the personality traits of conscientiousness and emotional stability, all of which have been found to relate to job performance (Barrick, Mount, & Judge, 2001; Ng & Feldman, 2009; Salgado et al., 2003). Job-specific human capital, which refers to attributes that relate to performance only in certain jobs. For example, job-specific knowledge and skills or personality traits that facilitate performance in particular jobs, such as extraversion for managers (Barrick et al., 2001). Empirical evidence suggests that the distinction between general and job-specific human capital is valid and highly pertinent to the modern economy (Poletaev & Robinson, 2008). Finally, Melamed also identified job-irrelevant human capital, which is considered later.

Copyright American Psychological Association. Not for further distribution.

Career Issues

Empirical research provides support for the human capital approach with respect to objective career success. Ng et al. (2005) found that an array of human capital variables was related to both number of promotions and financial attainment (average corrected correlations of .06 and .21, respectively). And in a longitudinal investigation, Judge, Higgins, Thoresen, and Barrick (1999) found that measures of cognitive ability, conscientiousness, and emotional stability obtained in the career foundation stage were related to occupational prestige in the reinforcement stage and to financial attainment on the edge of the retirement stage (r = .53, .41, and .34, respectively). Education, and its related attributes such as training, is probably the archetypical human capital factor. Indeed, human capital is frequently operationalized solely in terms of educational attainment (e.g., McArdle et al., 2007). This is partly justifiable because in Ng et al.’s meta-analysis, education, along with general work experience, demonstrated the strongest relationship (corrected r = .29 in both cases) of all human capital factors with financial attainment. The pervasiveness of the idea of human capital is illustrated by the fact that it is considered a core aspect of employability (Fugate et al., 2004). And practices seem to espouse the approach: Employers normally hire for cognitive ability, conscientiousness, and other personality characteristics that they believe add value to their businesses (e.g., Tracey, Sturman, & Tews, 2007), and they invest in internal development or external acquisition of education, skills, and experience (e.g., Lepak & Snell, 1999). States also invest in the human capital of their citizens (e.g., by financing educational institutions and by providing training incentives to organizations; Lanzi, 2007). And, finally, individuals, as career actors, invest in their own human capital (e.g., undergraduate and graduate education, seeking international assignments for the acquisition of relevant experience). Human capital generally fares less well in accounting for individual differences in subjective career success. For example, Ng et al. (2005) found very limited relationship of educational attainment, job, and organizational tenure with career satisfaction (corrected r = .03, −.02, and .02, respectively). However, stable

human capital characteristics fare better. Ng et al. identified corrected correlations of .14 and .36 for conscientiousness and emotional stability with subjective career success. And in their lifelong longitudinal study, Judge et al. (1999) reported a correlation of .30 between cognitive ability and subjective career success. The route via which these stable human capital attributes relate to subjective evaluations of success, however, is not clear. As cognitive ability, conscientiousness, and emotional stability relate to job performance, which should in turn be translated to objective career achievements, the relationship may be spurious; taking into account that, as seen, objective success partly influences subjective evaluations. By the same token, however, education, which also relates to job performance, should also relate to subjective career success, but it essentially does not. An alternative account invokes nature. Personality is to a substantial extent rooted in the genetic makeup (e.g., Bouchard & Loehlin, 2001; Jang et al., 2006) and is largely invariant during adulthood (e.g., Judge et al., 1999), and certain personality traits, including emotional stability and extraversion, are linked to the physiological mechanisms for emotion (see Eysenck, 1992; Hooker, Verosky, Miyakawa, Knight, & D’Esposito, 2008). This can account for the fact that certain personality traits appear to predispose individuals to consistently positive or negative outlooks of situations (Chien, Ko, & Wu, 2007; DeNeve & Cooper, 1998) and can also account for the relevant finding that subjective career success remains stable over substantial time intervals, whereas objective career success changes during those same intervals (Wolff & Moser, 2009). The implication is that certain individuals are predisposed to evaluate their careers favorably or unfavorably and, hence, that subjective career success is partly outside individual and organizational control. This means that availability of opportunities, such as education, training, and challenging assignments, to enhance human capital is not necessarily accompanied by improvements in employees’ subjective feelings of success, at least not analogously to corresponding gains in objective success. Career success and job performance. Readers may have noticed that in our discussion objective 83

Baruch and Bozionelos

Copyright American Psychological Association. Not for further distribution.

career success is not treated as tantamount to job performance. There are a number of reasons for this: 1. Objective career success is determined by the joint effect of multiple work-related outcomes, not all of which are under the control of career actors (e.g., a recession with its consequent contraction of the job market, as to be seen later). Furthermore, objective career outcomes are largely influenced by career decisions. These decisions range from the choice of employer at the career-entry stage to whether to contemplate an interorganizational move or to pursue additional education at a later stage. For example, a career actor who is an excellent performer may choose to stay with an employer instead of accepting a competitor’s offer; however, the employer may subsequently restructure and downsize, which may restrict promotion opportunities or even leave the career actor temporary unemployed. In this case, the career success outcome may not be commensurate with job performance. Careers, and their attached success, have much deeper and broader determinants than how well career actors perform in particular work roles. 2. Consideration of the relationship between job performance and career success would be meaningful only within the same job or profession and within the same environment. Otherwise, comparability issues can seriously compromise the validity of any conclusions. To demonstrate, criteria for the allocation of organizational rewards, such as promotions and money, vary across sectors. For example, not-for-profit organizations are less likely to use job performance as a criterion for promotion (Devaro & Brookshire, 2007). Furthermore, criteria for career success vary widely across jobs and professions as well as across societal and organizational contexts, which means that objective evaluations may be conflicting depending on the criterion used. As an illustration, the success of a plateaued middle manager, a semiskilled worker who has always been able to find employment, and a heart surgeon with a poor survival record for his or her patients would be judged differently according to 84

occupational prestige, employability, and monetary compensation. 3. Even in cases, however, in which the relationship has been tested with standardized instruments and with controls for structural factors, nonperfect associations have been identified. A recent study (Carmeli, Shalom, & Weisberg, 2007) of individuals pursuing corporate careers in both the service and the manufacturing sector found that an array of performance indices, including overall job performance, withdrawal behaviors (e.g., absenteeism), amount of overtime work, and organizational citizenship, accounted for 12% of the variance in objective career success; a significant but clearly not impressive amount. Similar findings were reported by Van Scotter, Motowidlo, and Cross (2000) in a study of air force mechanics, in which task and contextual performance combined together accounted for less than 10% of the variance in promotions. In fact, there is research that suggests an even more limited relationship. For example, Cannings and Montmarquette (1988) found no link between performance ratings and probabilities of promotion among male managers. One potential explanation for the rather weak link between job performance and objective career success lies in the imperfections of organizational systems in evaluating employee performance (e.g., Latham & Mann, 2006) as well as in the error involved in the operationalization and measurement of job performance by researchers. Another explanation, however, involves the presence of factors that are not directly relevant to job performance but nevertheless affect career prospects. These include job-irrelevant human capital and social capital, which will be discussed next. Job-irrelevant human capital. Job-irrelevant human capital refers to individual characteristics that logically should not bear a relationship to job performance but which nevertheless appear to influence career success (Melamed, 1996a). This category of human capital deserves some special attention because it has gained recognition only recently. Physical characteristics of individuals, reflected by physical appearance, are most representative of

Copyright American Psychological Association. Not for further distribution.

Career Issues

job-irrelevant human capital. These include height and physical attractiveness. These features may indeed offer performance advantages in certain jobs (e.g., modeling, certain sports). However, in the majority of jobs or professions these characteristics should not present performance advantages or disadvantages, especially in the modern economy that relies heavily on cognitive rather than manual work. Hence, from a logical point of view these should bear no relationship to career success. However, Hosoda, Stone-Romero, and Coats’s (2003) metaanalysis of 27 field and laboratory experiments indicated an advantage in promotion and hiring decisions for more physically attractive individuals (combined d = .37). Success in being hired may influence the timing of career entry, probabilities of upward interorganizational moves, or staying in employment in the case of lay-offs; hence, it is highly pertinent to career success. Furthermore, meta-analytic and field research conducted by Judge and Cable (2004) revealed a relationship of height with hierarchical status (ρ = .24) and financial attainment (average β = .26). Hence, empirical evidence suggests a clear link between observable physical characteristics and objective career success. One account for the observed relationships invokes implicit personality theory (Bruner & Tagiuri, 1954) and posits that others form impressions of competence, social skill, and stature on the basis of individuals’ appearance (e.g., Ashmore, 1981). These impressions, in turn, affect decisions on rewards (e.g., job offers, promotions; see Hosoda et al., 2003; Judge & Cable, 2004). According to this explanation, decision makers (e.g., line managers, promotion and hiring panels) advantage individuals with certain physical characteristics without being consciously aware of doing so. A second account, advanced herein, is that individuals with certain physical characteristics are consciously given an advantage (or disadvantage) by decision makers, who function under the belief that physical appearance has direct value adding properties for their units. This is especially the case in interactive service industries (e.g., hospitality) in which employers generally assume that physical attractiveness of service providers is related to customer satisfaction (Warhurst, Nickson, Witz, & Cullen, 2000). Both

of these explanations imply passivity on the part of individual career actors whose physical characteristics simply evoke reactions that influence their careers. Yet a third account implies that visible physical characteristics do not relate to career success simply because of how individuals “look” but because of what individuals “do.” This account suggests that physical attributes, such as height, are directly linked to human personality ( Judge & Cable, 2004) or cognitive ability (Case & Paxson, 2008), which in turn impact job performance and career success. However, Melamed and Bozionelos (1992) found that the significant relationship between height and promotion rate of British civil service managers remained virtually unchanged after imposing statistical control for personality. The corresponding finding of Judge and Cable (2004) for cognitive ability was analogous. Systematic research on the mechanisms that link job-irrelevant human capital with career success is important because the magnitude of relationships is comparable with those of other forms of human capital. For example, the average correlation between physical height and financial attainment in Judge and Cable’s (2004) studies was .29, whereas the corresponding correlations for education and conscientiousness, forms of general human capital, in Ng et al.’s (2005) meta-analysis were .29 and .07, respectively. States invest enormous amounts in educational programs, as do individuals for their personal education and development, and organizations invest considerable resources in training, development, and selection systems with the expectation of having educated and dependable individuals in their ranks. But, ironically, factors that are out of their control seem able to offset these investments. Identification of the mechanisms of the link should assist in the further development of hiring and award allocation procedures, including awareness and training of decision makers. This should reduce the likelihood of biased decisions, hence improving fairness and outcomes at individual, organizational, and societal level. Fluid versus crystallized human capital. Using the analogy of fluid versus crystallized intelligence (e.g., Cattell, 1941; Horn & Cattell, 1966), it is proposed that human capital can be categorized in terms 85

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

of attributes that can be developed or enhanced, such as skills and education, and attributes that are generally stable and cannot be easily modified or developed, such as personality traits, cognitive ability, and physical characteristics; the former labeled as crystallized human capital and the latter as fluid human capital. This typology should be seen as orthogonal to Melamed’s (1996a) typology. The combination of the two typologies yields six categories of human capital, according to relevance to job performance (ranging from general to job-irrelevant) and degree of modifiability. Such categorization can assist career decision making and planning. For example, specialists or career actors will be able to decide in favor of careers that fit best their crystallized human capital and avoid those directions that require fluid human capital in which they have serious deficiencies. Subsequently, they can identify which crystallized human capital characteristics they need to further cultivate to achieve success in their chosen career paths.

The Structural View of Career Success The structural view, sometimes referred to as the opportunistic approach, places emphasis on the influence of structural factors on career progression (e.g., Spilerman, 1977). These can be of an organizational, environmental, or societal nature. Organizational factors include organizational structure, size, type of ownership, criteria for allocation of organizational rewards, and characteristics of internal labor markets (e.g., Sonnenfeld, Peiperl, & Kotter, 1988). Gunz (1988) commented on the distinctive “career logic” of every organization, which denotes the fact that career paths and rules to navigate through these paths are determined by the unique structural and cultural features of each organization. Evidently, organizational factors concern only individuals who pursue traditional careers in the totality or part of their working lives. Environmental factors, or labor market forces, include the type of industry (e.g., service vs. manufacturing, degree of regulation, profit vs. non-profit) and the economic circumstances (e.g., period of war, recession) under which careers take place (e.g., Long & Link, 1983). Finally, societal factors refer to the structure of the society (e.g., Melamed, 1995a). These include, for example, the 86

educational system and employment legislation, which can influence skill availability in the labor market and opportunities or constraints to career mobility, respectively. There is empirical evidence for the veracity of the structural perspective. In a recent study of MBA graduates from elite schools, Oyer (2008) found that both the timing of their entry into investment banking careers and the financial success of these careers were related to the conditions in the stock market while they were in graduate school. MBAs were more likely to go directly into investment banking upon graduation (β = .14) if stock market returns were positive (i.e., the index increased) for the duration of the program, and probability of pursuing a career in investment banking was increased by 75% for those who went directly into the industry, in comparison with those who had an interest to do so but did not enter the industry directly upon graduation. And in a 20-year longitudinal study with a representative sample of the workforce with the United States in the career entry stage, Shin (2007) found that upward and downward intra- and interorganizational movements were affected by the industry’s structural changes (e.g., contraction, expansion) and density of mergers. The human capital and the structural approach account for the contribution of factors of different nature to career success; hence, they should be seen as complementary. To illustrate, in a study in a German engineering company, Bruderl, Preisendorfer, and Ziegler (1993) found that human capital affected probabilities of promotion, but the relationship was moderated by the opportunity structure of the organization, which was a function of the number of workers in each hierarchical level and of the number of promotions made at each level in a specific year. Number of promotions was affected by expansion or contraction of the company, which, in turn, was affected by the conditions in the industry and the economy in general. At this point, it should be noted that it is erroneous to view career actors as passive followers of a script or as pawns for forces represented by, primarily, structural and, secondarily, human capital factors. Instead, it is more appropriate to view individuals as improvising actors who navigate within the space element of careers. The environment undoubtedly

Copyright American Psychological Association. Not for further distribution.

Career Issues

imposes constraints (see also Ozbilgin, Kusku, & Erdogmus, 2005, p. 2001), but individuals can optimize their career options within these constraints or even overcome them in some cases. For example, active job search behaviors, such as preparation and updating of one’s resume, have been found to predict success in finding employment in the transition from university to the labor market (β = .54; Brown, Cober, Kane, Levy, & Shalhoop, 2006), as well as in the speed of the transition from the armed forces to the civilian labor market (r = .23; Baruch & Quick, 2007). And individual career choices account for career success over and above the contribution of structural and human capital factors (Melamed, 1996b). Therefore, career progression and success should be regarded as the outcome of a constant interplay between extraneous forces, such as structural variables, and internal forces that are manifested in the actions of career actors. In fact, not only do individuals navigate through the structural element, but they may also shape it in an interactive manner; as the processes through which individual careers progress mold organizations and affect their probabilities of success (Feldman, 1985; Lazarova & Taylor, 2009).

The Social Capital View of Career Success The social capital perspective has been consolidated rather recently (Seibert, Kramer, & Liden, 2001), though the principles behind it had been in place for a substantial amount of time. This approach asserts that factors going beyond structure and human capital should be taken into cognizance in order to develop an exhaustive description of what determines career success. The interest in social capital is evidenced by the fact that it is seen as a component of employability (Fugate et al., 2004). The social capital view attests that informal interpersonal processes, which are captured by the concept of social capital, play an important role in career success. The term social capital was popularized by Coleman (1988),3 though the notion, and the term itself, is found in earlier writings (see Portes, 1998), Bourdieu’s (1980, 1986) being the most notable. Social capital can be considered at macro level (e.g., 3

Inkpen & Tsang, 2005) or micro level. However, the present work exclusively focuses on the latter because this refers to the individual career actor. Social capital signifies resources (i.e., information, influence, solidarity) that an individual has at one’s disposal by means of his or her relationship ties with other individuals within a particular social structure, such as the work organization, the profession, or the society in general (Adler & Kwon, 2002; Bourdieu, 1986; Portes, 1998). These resources impact social outcomes, such as career success, by means of two properties: substitutability and appropriability (Adler & Kwon). Substitutability refers to the capacity to substitute for or complement qualities or resources that the individual may be lacking (e.g., credentials, performance, direct access to information, position power). Appropriability captures the fact that relationship ties of a certain type can be utilized for multiple purposes (e.g., friends in the workplace may offer emotional support, performance feedback, information on internal job openings, and access to organizational decision makers). Through these properties, social capital makes “possible the achievement of certain ends that in its absence would not be possible” (Coleman, 1988, p. S98). It should be noted that there are not only the relationship ties of which the individual is aware, and which compose his or her social capital, but also those ties of which the individual is not consciously aware (i.e., the ties of the ties) that influence outcomes (see Higgins & Kram, 2001). Elements of the social capital perspective are found in the know-whom factor of Arthur and associates’ view of career enhancing strategies. The know-whom factor refers to the accumulation and mobilization of interpersonal ties in the pursuit of career success (Arthur et al., 1995; DeFillippi & Arthur, 1994). Empirical literature provides support for the link between social capital and objective career success. Social capital appears to assist in the career entry and in later career stages. For example, Jokisaari and Nurmi (2005) found that the social capital of final year university students was related to the probability of having a full-time job commensurate with their educational credentials 6 months later (odds ratios

The impact of Coleman’s article becomes evident from a search of the Thompson Scientific database that covers the period from 1945 until present, using social capital as key word. Before 1988, when Coleman’s article was published, there were 16 articles using the term, whereas after 1988 this number well exceeds 3,000.

87

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

1.89 and .65 for the two measures of social capital used, respectively). Similarly, Combes, Linnemer, and Visser (2008) found that social capital was able to overcompensate for deficiencies in number and quality of publications in the hiring of economics faculty in French academia: The combination of having one’s doctoral advisor in the selection panel and having earned one’s doctorate from a university with which another member of the panel was affiliated nearly tripled (i.e., increased from 17% to 47%) the probability of success from the baseline condition, whereas a shift from the 10th to the 90th percentile in quality of publications increased the probability of success over the baseline condition by only about half (i.e., from 14% to 21%; Combes et al.). Finally, social capital has also been found to relate to subjective, along with objective, career success (Seibert et al., 2001). Mentoring and network resources. An operationalization of social capital that facilitates the development of more fine-grained advice on tactics and strategies for individual career management can be expressed in terms of primary mentoring relationships and network ties (e.g., Bozionelos, 2003b; see also Seibert et al., 2001). Mentoring in its classical or traditional form refers to an intensive developmental relationship between two persons of unequal status or power, the mentor and the protégé, in which the former provides a variety of functions for the latter (e.g., Kram, 1985; see also chap. 17, this volume). These functions fall into two dimensions: careerinstrumental (e.g., protection, assignment of challenging tasks, exposure and visibility) and socioemotional (e.g., acceptance and confirmation, role modeling; Kram; Tepper, Shaffer, & Tepper, 1996). The mentoring relationship develops over a long period of time and may transcend organizational boundaries (Baugh & Fagenson-Eland, 2005; Kram). The function of mentoring as a career-enhancing platform is in agreement with the sponsored mobility view, which asserts that certain individuals have an advantage in the allocation of organizational and societal rewards because they enjoy the sponsorship and support of powerful members of the organization or of the social system (Turner, 1960). It should be borne in mind that mentoring relationships vary with respect to the number and qual88

ity of mentoring functions provided within them (Fletcher & Ragins, 2007; Ragins, Cotton, & Miller, 2000) and that individuals may be involved in none, one, or many mentoring relationships in their careers (for a mini-review of relevant evidence, see Kirchmeyer, 2005, p. 654). Network ties encompass the totality of the individual’s interpersonal ties, excluding any traditional mentoring relationships. Network ties vary in terms of strength, which encompasses longevity, intensity, intimacy, and reciprocity; range, which signifies the social distance that a particular tie can reach (see Marsden, 1990); diversity, which refers to the social domain (e.g., work, professional association, nonwork) of the tie’s origin; and direction, as ties may be formed with individuals of lower, equal, or greater power or experience than the career actor (see Higgins & Kram, 2001). Mentoring relationships with peers (see McManus & Russell, 2007) fall into the category of network ties. In a direct analogy to the functions of mentoring, the functionality of network ties is also divided into career-instrumental and socioemotional (Fombrun, 1982; Ibarra, 1993; see also Saint-Charles & Mongeau, 2009), with most ties serving in varying proportions both instrumental and emotional functions (Brass, 1992; Kram & Isabella, 1985; Saint-Charles & Mongeau). Meta-analyses (Allen, Eby, Poteet, Lentz, & Lima, 2004; Eby, Allen, Evans, Ng, & DuBois, 2008; Kammeyer-Mueller & Judge, 2008) of the abundant empirical research on the relationship between mentoring receipt and career success confirm the link between the two. Allen et al., for example, found relationships of mentoring received with monetary compensation and number of promotions (sample-weighted r = .12 and .31, respectively) as well as with career satisfaction and expectations for advancement (sample-weighted r = .21 and .26, respectively). Furthermore, empirical evidence is emerging that indicates that providing mentoring to others is linked with mentors’ career success (Bozionelos, 2004; Eby, Durley, Evans, & Ragins, 2006). As a result, organizations have been implementing formal mentoring systems, which involve mentor–protégé relationships that are formally arranged, recognized, and supported; though the benefits of formal mentoring pertain mostly to

Copyright American Psychological Association. Not for further distribution.

Career Issues

subjective career outcomes (see Baugh & FagensonEland, 2007). A lesser amount of systematic empirical work has been dedicated exclusively to the relationship of network resources to career success, though relevant research is accumulating. Wolff and Moser (2009) investigated the relationship of six networking behaviors (i.e., building, maintaining, and using internal and external organizational contacts) to career success. Most of these behaviors were related, though only moderately at best, to current monetary compensation (mean partial r = .10, range of the four significant partial rs = 11–.20) and current subjective career success (mean partial r = .17, range of the four significant partial rs = .16–.32) over and above a range of human capital and structural factors. Two of these networking behaviors, maintaining and using internal contacts, were also related, albeit weakly (partial r = .05 and .03, respectively), to growth in monetary compensation over a 2-year period. And Higgins (2005), looking at the labor market as a wider system, described what she termed a “career ecosystem” by presenting the case of Baxter, a major company in the biochemical industry, and the way that their managers were indoctrinated into a network of relationships where entrepreneurship was a key factor to success. A substantial number of Baxter’s former executives came to dominate the start-up bioindustry and a large proportion of its former managers became the CEOs of such start-ups. Their success can be attributed to their access to a common pool of knowledge, information, and solidarity that was available through their common links. Mentoring or network ties? An issue that has implications for career tactics and strategies is whether mentors and network ties relate independently to career success or whether they can substitute for each other. For example, authors have suggested that in the present era career actors need multiple developmental relationships because professional and learning demands are greater and cannot be fulfilled by a single traditional mentoring relationship (de Janasz & Sullivan, 2004; Higgins & Kram, 2001) that best fits the demands of the traditional career. Some research findings are suggestive of overlapping contributions of mentoring and network ties to

career success, the implication being that network ties can substitute for mentoring and vice versa (Bozionelos, 2003b; 2008). However, Seibert et al.’s (2001) work appears to suggest that their effects are additive (i.e., independent), whereas yet other research suggests that the effects are additive only for subjective career success (Blickle, Witzki, & Schneider, 2009). Yet, and as a compromise, there is some evidence that the benefits of mentoring and network ties are realized at different career stages, with network ties being more beneficial in the longterm (Higgins & Thomas, 2001). However, to add to the complexity, recent empirical research (Blickle et al., 2009) implies that mentoring and network ties are linked in the sense that mentoring is partly responsible for the development of network ties (e.g., the protégé models the mentor’s behaviors of ties acquisition), which in turn influences career success. Finally, perplexity is further augmented by the fact that the instrumental and the socioemotional functions of mentoring and network ties appear to relate differentially to objective and subjective career outcomes. For example, Allen et al. (2004) found that socioemotional mentoring was substantially more strongly related to subjective (sample-weighted r = .25) than to objective career success (mean sample-weighted r = .05), a pattern that was partly repeated for career-instrumental mentoring (sample-weighted r = .29 and mean sample-weighted r = .12, respectively). On the other hand, Bozionelos (2003b) found that only the instrumental (β = .12), but not the socioemotional, network resources were associated with objective career success, whereas the pattern was more complicated with respect to subjective career success, with instrumental and socioemotional networking being related to different facets of it. Additional research is necessary, taking into account that career actors’ psychological resources (e.g., persistence, proactivity) and time are not limitless; hence, it is important to direct these correctly in their effort to build social capital. The mechanism of the social capital effect: Politics or job performance? As we have seen, there is evidence that supports the social capital approach to career success. The mechanism for the link, however, 89

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

remains to be unveiled. Potential explanations, not necessarily mutually exclusive, include job performance and political processes. The former posits that social capital enhances the performance of those who possess it, which, in turn, impacts career success (e.g., income attainment). For example, challenging assignments or role modeling that mentors provide increase protégés’ job expertise and, subsequently, their performance, which is rewarded by means of organizational or societal rewards (e.g., more clients, hence, greater income). However, the political approach suggests that it is by means of political processes that social capital impacts career success. For example, a mentor or another powerful social tie may directly influence a promotion decision in favor of the career actor. This is in line with the political view of organizational life (e.g., Mintzberg, 1985), which posits that, on many occasions, decisions that influence careers (e.g., hiring, promotions, transfers, training) are based on political processes and motives rather than on transparent and merit-based processes and criteria (e.g., Bozionelos, 2005; Ferris & Judge, 1991; Ferris & King, 1991; Koskina, 2008; Pfeffer, 1989). Tharenou (1997), for example, posited that “promotion . . . is [also] determined by networks and politics” (p. 83) and that “politics influences who advances in management” (p. 83). There is some evidence that social capital is linked with enhanced job performance. For example, Thompson (2005) found a relationship between network-building behaviors and supervisory performance ratings. However, Green and Bauer (1995) found no relationship between mentoring and objective measures of performance. Supervisory performance ratings may be influenced by the social capital of career actors (e.g., Bozionelos & Wang, 2007). Furthermore, in a study that deliberately compared the political and the performance approaches using objective performance criteria, Kirchmeyer (2005) found evidence that was more in support of the political rather than the performance view. Additional research is crucial because the issue is of critical importance, especially for organizations that invest in formal mentoring systems or are advised to foster cultures that favor the development of informal mentoring relationships and social capital (e.g., Pastoriza, Arino, & Ricart, 2008; Singh, Bains, 90

& Vinnicombe, 2002). Organizations’ primary interest is the performance, rather than the career success, of their employees. If social capital operates mainly via the political route, which can be detrimental for both organizational performance and employee morale (e.g., Bozionelos, 2005; Vigoda, 2000; and for some points on the dangers of social capital, see Portes, 1998, pp. 15–18), organizational policies may need to be reconsidered or systems be developed that filter out the political elements of social capital. DIVERSITY AND CAREERS Diversity refers to heterogeneity among members of the workforce according to one or more salient characteristics (e.g., Bell, 2007; Milliken & Martins, 1996; see also Vol. 1, chap. 20, this handbook). Hence, a diverse workforce is composed of two or more subgroups. Diversity is pertinent to careers because it apparently relates to variance in employment patterns and opportunities. Individuals may pursue different career goals or be offered different career options, depending on the group they belong to. Traditionally, diversity has been viewed in terms of surface-level diversity (Harrison, Price, & Bell, 1998), which signifies differences in overtly identifiable features that are usually rooted in biological or genetic differences. The most prominent of these features include sex, race, ethnicity, and age (Tsui, Egan, & O’Reilly, 1992). These characteristics are readily observable, are not susceptible to alteration, and are operationalizable in terms of demographic factors. Religion could also be added to these, as could sexual orientation (if revealed) and certain physical features, such as body weight and forms of disability. These features are either permanent or difficult to alter and are readily revealed in social contact. Finally, special types of family arrangements, such as being a single parent or a part of a dual-career couple, are other forms of surface-level diversity. Though not permanent, these conditions normally last for substantial portions of worklife and tend to manifest themselves early in the course of social interaction. Evidently, various forms of surface-level diversity may coexist (e.g., an overweight single mother), giving rise to hybrid diversity (Baruch, 2004). It should also be mentioned that some surface-level diversity charac-

Copyright American Psychological Association. Not for further distribution.

Career Issues

teristics, such as body weight, sex, and race, could also be seen under the prism of job-irrelevant human capital because these may influence career success independently of job performance. Apart from the surface level, however, diversity also operates at “deep level.” Deep-level diversity refers to heterogeneity or separation in attitudes, beliefs, values, or even skills, knowledge, and experiences (Harrison et al., 1998; Jackson, May, & Whitney, 1995; Milliken & Martins, 1996; see also Harrison & Klein, 2007). These characteristics are constructs; hence, they can only be inferred via substantial interpersonal interaction and long observation. Furthermore, unlike surface-level features, deep-level diversity characteristics are susceptible to change ( Jackson et al., 1995). Both surface- and deep-level diversity relate to outcomes in the work environment (Harrison et al., 1998; Harrison, Price, Gavin, & Florey, 2002). However, from a careers perspective, it is surfacelevel diversity that is of greatest relevance. Similarity at surface level leads to expectations for similarity at deep level (Philips & Loyd, 2006). And surface-level characteristics are more likely to evoke immediate reactions that pertain to bias and prejudice (Milliken & Martins, 1996), which give rise to discrimination. Discrimination refers to systematic disparity in opportunities and outcomes between groups with comparable potential (Dipboye & Halverson, 2004) and is linked to careers by its association with differentiation in career opportunities for diverse groups, such as probabilities of obtaining a job, being promoted, or offered a favorable transfer. The similarity-attraction paradigm (Byrne, 1971) and the self-categorization and social identity perspective (Tajfel & Turner, 1986) explain the link between diversity and discrimination: Individuals categorize themselves and others according to salient attributes, such as surface-level characteristics, and identify with those whom they perceive as belonging to the same categories as themselves, and to whom they are, in turn, attracted and with whom they prefer to associate. Next, we focus on gender, the diversity form that has attracted most attention from the perspective of careers (Baruch, 2006). However, other forms of 4

surface-level diversity are mentioned too, to the extent that these exhibit similarities with gender.

Gender and Careers Women today comprise nearly half of the working population in developed countries (United Nations, 2007). It appears, however, that their careers lag behind those of men in terms of objective criteria for success. For example, Ng et al.’s (2005) meta-analysis indicated significant disadvantages for women in both number of promotions and financial attainment, with women lagging behind men one tenth and one fifth of a standard deviation, respectively (corresponding d = .45 and .18, respectively4). This finding concurs with the official statistics on earnings differences between women and men, which is in the ratio of 4 to 5, in the general population of the United States (U.S. Department of Labor, 2008). The term glass ceiling has been coined to describe the presence of an invisible barrier that prevents women and other groups from advancing into powerful positions (Morrison, White, & Van Velsor, 1987) in any environment where careers are pursued (e.g., Achkar, 2008). As to be seen below, there is some evidence of cracks in the glass ceiling, though this may not necessarily reflect elimination of disparity in career opportunities. A number of accounts, which are not mutually exclusive, have been offered for the observed gender differences in objective career success: 1. Different expectations and priorities, rooted partly in socialization. Women, for example, are likely to be expected to prioritize family or their partner’s careers over their own (e.g., by taking career breaks to look after their children, working less overtime, taking the burden of domestic labor, changing their jobs to trail their partners who relocate; Nieva, 1985; Russo, 1985). Indeed, in a study of women in dual-career couples, Valcour and Ladge (2008) found that prioritization of their partner’s careers over their own was negatively related to their financial attainment (β = −.32). And in a study with university graduates in the United Kingdom, Chevalier (2007) found that women were substantially more likely than

d values calculated by the authors.

91

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

their male counterparts to expect a career break for family reasons; which, in turn, accounted for 10% of the gender gap in financial attainment. Work–family conflict, the conflict between those work and family roles that are mutually incompatible (Greenhaus & Beutell, 1985), is also pertinent. Work–family conflict appears more detrimental for women than for men. For example, Mayrhofer, Meyer, Schiffinger, and Schmidt (2008) found that family responsibilities were negatively related to work centrality, which was, in turn, related to career success for women but not for men. 2. Human capital and career path differences. According to this view, women follow different educational paths and, hence, accumulate less or different crystallized human capital, which, in turn, handicaps their career attainment. Women, for example, are more likely to major in “soft” subjects whereas men tend to major in sciences, engineering, and business-related disciplines (e.g., Bradley, 2000; Dahlmann, Elsner, Jeschke, Natho, & Schroder, 2008) that are more likely to lead to corporate careers, line instead of staff positions, and higher earning occupations. Indeed, Penner (2008) found that occupational differences accounted for approximately three-quarters of the gender gap in earnings among newly hired employees in a large U.S. firm (for updated data on earnings gaps across occupations in the United States, see Bureau of Labor Statistics, 2007). Even nowadays, females with secondary education are less likely than their male counterparts to express intentions to follow a career in computer science or information technology (Anderson, Lankshear, Timms, & Courtney, 2008; Papastergiou, 2008). At this point, it should be mentioned that males tend to score higher in aspects of cognitive ability, such as mental rotation and mechanical reasoning (see Hyde, 2005), that facilitate entry into scienceand engineering-related degree programs; whereas women tend to score higher in verbal ability (Hyde), which is associated with “softer” disciplines that are less likely to lead to careers with strong objective success prospects. However, although these explanations have received empirical support, in many cases gender differences in objective career success persist even after 92

controlling for the factors attached to these accounts. To illustrate, in a 10-year longitudinal study with a representative sample of university graduates, Bobbitt-Zeher (2007) found that around one-third of the gender difference in financial attainment remained unaccounted for after controlling for an array of background factors (e.g., socioeconomic background), personal values (e.g., importance attached to money), family situation, human capital (i.e., cognitive ability, undergraduate major, GPA), structural factors (i.e., profit vs. nonprofit sector, industry type, occupation) and work involvement (i.e., hours worked per week). This, therefore, provides an additional explanation for the gender gap in objective career success, which draws on social capital and direct discrimination. It is argued, for example, that women possess less social capital than men or use it suboptimally in the work environment, a fact that undermines their career advancement (e.g., Tharenou, 1997). Concurring with this view, a recent study with a representative sample of the U.K. adult population indicated that women have fewer network ties than their male counterparts (Li, Savage, & Warde, 2008). Explanations for women’s deficit in social capital include the following: 1. Women’s lower capacity, as relative newcomers in organizational life, to create career-instrumental social capital, as well as their alleged tendency to underestimate its importance for career advancement and to rely more on formal, merit-based, procedures and human capital in their quest for success. There is evidence to support this view. Van Emmerik (2006), in a study in academia, found that males were more able in creating career-instrumental network ties than were women. Ng et al.’s (2005) meta-analysis indicated that educational attainment and amount of working hours, two variables that are representative of human capital and work effort, respectively, were more strongly related to financial attainment for women than for men; and Forret and Dougherty (2004) found that networking behaviors accounted for greater variance in men’s than in women’s objective career success. In a qualitative study, Ozbilgin and Healy (2004) described how augmentation in the attention paid to human capital in hiring and promotion

Copyright American Psychological Association. Not for further distribution.

Career Issues

increased the proportion of female representation in academic ranks in Turkey. Finally, in a study with women managers in Australian banks, Metz and Tharenou (2001) found that human capital accounted for seven and three times more variance than social capital in their objective career success at lower-middle and upper management levels, respectively. 2. High degree of homophily in women’s social capital ties (Ibarra, 1993; Morrison & Von Glinow, 1990). This means that women tend to form mentoring and network ties primarily with other women and do not integrate into male networks, which carry most career-enhancing information and influence because of the traditional male dominance of organizational power structures. Indeed, Ibarra’s (1992) findings on homophily differences between men and women’s networks implied that that it was necessary for women to form career-instrumental ties with men, with similar findings reported by Stackman and Pinder (1999). And in a study that focused on career experiences of financial professionals in Wall Street firms, Roth (2004) found evidence for homophily in women’s mentoring and network ties, which was, in turn, related to reduced likelihood to work in the most highly paid areas. The theoretical underpinnings of women’s deficit in social capital include the similarity-attraction paradigm, referred to earlier (Tsui & O’Reilly, 1989); the perception that formation of ties with women, and other career disadvantaged groups, are of lower value (Ibarra, 1993; Ragins & Sundstrom, 1989); and identity preservation, which suggests that women consciously choose not to integrate into male dominated networks in order to demonstrate loyalty to their gender (Ibarra; Kanter, 1977; Kram, 1985). Indeed, in line with the last point, Davey (2008) found that women in male-dominated professions detested the fact that they had to disguise their feminine values and characteristics, such as sensitivity, in order to acquire the social capital necessary to advance their careers. A supplementary account invokes personality. For example, women score lower than men on the assertiveness aspect of extraversion (d = .51; Hyde, 2005), a personality trait that is associated with

social capital (Okun, Pugliese, & Rook, 2007; Wolff & Moser, 2006). However, the role of personality should be secondary, taking into account that genders do not differ in the other aspects of extraversion (see Hyde, 2005). Apart from being disadvantaged in the social capital arena, however, women may also face direct discrimination because of general beliefs held by key organizational players: that women are less committed to, less competent, and less suitable for tasks that pertain to organizational life (e.g., Adler, 1993; Eagly & Carli, 2007, p. 64). This view is partly represented by the status expectation theory (Berger, Fisek, Norman, & Zelditch, 1977), which suggests that those groups that have traditionally enjoyed higher status, such as men, are also implicitly assumed to be more competent than traditionally lower status groups, such as women. In line with this view, Greenhaus and Parasuraman (1993) found that women’s successful performance was more likely than that of men’s to be attributed to contingencies rather than abilities, a finding that was corroborated by those of Roth (2004). And Lyness and Heilman’s (2006) finding that women managers who were promoted had substantially better performance ratings than their male counterparts (β = .29) implies that women may have to prove themselves more in terms of performance to receive organizational rewards. These accounts, fully or in part, apply to a number of other diversity groups that are allegedly discriminated against, ranging from racial and ethnic minorities (see Ibarra, 1993) to overweight individuals (e.g., Rudolph, Wells, Weller, & Baltes, 2009). For example, Li et al. (2008) found that along with women, ethnic minorities were also disadvantaged in terms of social capital; and Penner (2008) found similar effects of occupational segregation on starting salaries of ethnic minorities as were found for women. Furthermore, any adverse effects of diversity on careers are likely to be exacerbated in cases of hybrid diversity (Baruch, 2004).

The Paradox of Subjective Career Success and the Kaleidoscope Career Women’s apparent disadvantage in objective career success is not reflected in their subjective career evaluations. Indeed, women do not lag behind their 93

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

male counterparts in this dimension. In their metaanalysis, Ng et al. (2005) found no gender difference in subjective career success. And in a study with business owners that was conducted since the metaanalysis, Powell and Eddleston (2008) reported exactly analogous findings, with female business owners reporting the same levels of subjective success as their male counterparts despite the fact that their businesses were lagging behind in objective criteria. This paradox can be accounted for in terms of differential career expectations for men and women (Powell & Eddleston, 2008) as well as in terms of alternative domains (e.g., family, work–family balance, social relations, contribution to the society) from which women can draw a sense of achievement and satisfaction; in contrast to men who appear to derive such sense mostly from work-related experiences that pertain to attainment of power and material rewards (Dyke & Murphy, 2006; see also Lirio et al., 2007). Therefore, it appears that women’s careers have more protean elements than men’s because women seem to use alternative criteria, apart from rigid definitions of success or failure, to guide and evaluate their workrelated experiences (see Mainiero & Sullivan, 2005; Valcour & Ladge, 2008). Mainiero and Sullivan’s (2005, 2006) kaleidoscope career model builds on the subjective element of careers with the purpose of developing an explanation for the differential gender perspective. Using the kaleidoscope metaphor (see Inkson, 2007, for the use of metaphors in the study of careers), Mainiero and Sullivan (2005) suggested the following: Like a kaleidoscope that produces changing patterns when the tube is rotated and its glass chips fall into new arrangements, women shift the pattern of their careers by rotating different aspects of their lives to arrange roles and relationships in new ways. (p. 111) Relationships are pivotal in women’s careers, and the kaleidoscope contains three “glass chips” through which they evaluate and guide their careers: authenticity, or being true to oneself in the midst of the constant interplay of personal development with work and nonwork issues; balance, defined as making decisions in a way that work and nonwork aspects of life 94

form a coherent whole; and challenge, or engagement in activities that demand responsibility and allow learning and growth. A problem and challenge, however, is that organizational structures and mentalities are not yet attuned to the female perspective and career needs, despite the fact that females comprise around half of the workforce (O’Neil, Hopkins, & Bilimoria, 2008).

A Final Note on Diversity and Careers As seen, there is evidence that certain groups are disadvantaged in terms of objective career success, and there are theory-based accounts for these disadvantages. However, there is contrasting evidence as well. For example, using human resource records over a 9-year period from a large U.S. service sector company, Petersen and Saporta (2004) found no gender differences in promotion rates or salary attainment for administrative, professional, and managerial employees. In fact, at higher hierarchical levels, women had faster promotion rates than their male counterparts. Females were disadvantaged only in hierarchical level and salary on hiring, but this could be accounted for by a male superiority in previous work experience, that is, human capital (for a short review of literature with similar findings see Petersen & Saporta, 2004, p. 897). Furthermore, there is some evidence that the female, and group minority, disadvantage in objective career outcomes is constantly falling as women and minorities are becoming established in the work environment. For example, in their meta-analysis, Ng et al. (2005) found that the strength of the relationship between gender and financial attainment was attenuated as a function of the publication year of the study, with more recent studies reporting weaker relationships. In line with this, in their narrative review, Altman and Shortland (2008) found some evidence that both organizational views of women’s fitness for international assignments, which have traditionally been viewed as prestigious and sought after, and women’s proportional representation in such assignments have been improving over the years. Finally, Beckman and Phillips (2005) found evidence that law firms were more likely to promote women to top positions when their clients also had women in such positions. This implies a “snowball” or “drag” effect: The more women advance into high positions, the

Copyright American Psychological Association. Not for further distribution.

Career Issues

more likely it is for women in general to break the glass ceiling. Whether the apparent disadvantages of certain groups in objective career success will not be an issue in the future remains to be seen. However, it should also be kept in mind that ascendance to positions of power by individuals belonging to traditionally disadvantaged groups does not necessarily mean that the issue of differential opportunities has been resolved. This may simply reflect window dressing (e.g., appointing a token woman to the board, frequently in a “feminine” role, such as HR director), conquering empty castles (i.e., gaining positions that are not attractive for the dominant majority anymore; Hofbauer & Fischlmayr, 2004), or, even worse, cases of glass cliff (i.e., where minority members are deliberately appointed to senior positions that bear a greater risk of failure or criticism; Ryan & Haslam, 2007; Wilson-Kovacs, Ryan, Haslam, & Rabinovich, 2008). Therefore, the issue of diversity and careers probably goes beyond the achievement of statistical nonsignificance in financial or status attainment. Legislation may assist in the striving for equality in opportunities for career success (e.g., Leck, 2002; Weichselbaumer & Winter-Ebmer, 2007). However, this may also be the cause of undesired side effects. For example, achievements and contributions of traditionally disadvantaged groups may be doubted because these may be attributed to legislation instead of merit (e.g., Heilman & Welle, 2006); and there may be majority (e.g., male) backlash (Leck, 2002). In addition, the partial move toward postcorporate careers and less rigid organizational structures with their attached work arrangements (e.g., temporary employment, freelancing) may in fact reduce career opportunities for disadvantaged groups (e.g., Holgate & McKay, 2009). Hence, abiding by legislation should not be the primary motivator for employers and key organizational players to work toward equality in career opportunities. The main motive should be enhancement of organizational performance (and of a harmonious society) rather than political correctness, an issue that has arguably plagued the struggle for equality. This is because perceptions of fairness in the allocation of career rewards are associ5

ated with organizational commitment and job performance (e.g., Parker & Kohlmeyer, 2005; Ramaswami & Singh, 2003). Furthermore, impartial allocation of career rewards (e.g., promoting the most able individuals) enables employers to exploit the talent of the totality of their workforce (Harel, Tzafrir, & Baruch, 2002) and enhance bottom-line performance, as empirical research attests (McKay, Avery, & Morris, 2008). GLOBALIZATION AND CAREERS Globalization has brought a flow of people, knowledge, information, ideas, and products across national borders and has added to the issues in the study of careers (Tams & Arthur, 2007). We now address the issue of careers across national cultures and the management of expatriation.

Careers Across Cultures Our accumulated knowledge and models of careers have been primarily developed on the basis of work in North America and other highly economically developed parts of the Western world (as discussed earlier). The extent to which, and under what conditions, these are applicable across national cultural settings needs to be investigated because, for example, there is evidence that career orientations and priorities vary across cultures. To illustrate, in a study of MBA students from seven culturally distinct countries, Malach-Pines and Kaspi-Baruch (2008) found differences among participants from all countries in factors that included the meaning of work, protean and traditional career orientations, and expectations from careers in management. And Segers et al. (2008) identified relationships between dimensions of national culture and protean and boundaryless career orientations (e.g., r = −.32 between masculinity and the value-drive dimension of career proteity). Finally, in their meta-analysis, Tracey and Rounds (1993) found greater support for the structure of Holland’s model in the United States than with research conducted outside the United States (mean correspondence indices of .69 and .46, respectively, d = .81,5 calculated as the number of

d value calculated by the authors.

95

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

model predictions that are in agreement with the data minus the number of model predictions that are not in agreement with the data divided by the total number of model predictions). On the one hand, globalization has certainly increased the power and relevance of some already existing concepts, such as the boundaryless and the protean career (e.g., see Cappellen & Janssens, 2005; Hall, 2004; Thomas, Lazarova, & Inkson, 2005). The former can be viewed from the additional perspective of psychological preparedness of career actors to cross national boundaries and of actual movement across borders, and the latter can be viewed from the perspective of the extent to which career actors value cross-border work experiences and take initiatives in directing their careers toward such experiences. On the other hand, concepts (and this may include the concept of career itself) that have been developed with economically developed societies as reference points may have limited or no applicability in certain areas or populations of the globe. For instance, the notion of the traditional career may be of limited relevance in many societies (e.g., see International Labour Organization, 2009) where the majority of the workforce is self-employed or employed only as temporary or seasonal workers (e.g., in farms) or in small businesses that have limited or no hierarchical structures. Ironically, in such societies, where work lives for the majority are similar to work lives before the industrial revolution, the notion of the postcorporate career may be most relevant as a career model, despite the fact that they have never experienced the traditional, corporate-bound career. Another issue that emerged along with the movement of career actors across national boundaries is whether factors that are associated with career benefits within particular national contexts provide similar benefits in different contexts. As an illustration, as discussed earlier, mentoring has been established as an experience that relates to both objective and subjective career success. However, relevant empirical research has been conducted mainly in the AngloSaxon work environment. Research in non-AngloSaxon cultural environments yields results that are not always supportive of the career enhancing properties of mentoring. For example, Bozionelos (2006) 96

found that informal mentoring was related to objective but not to subjective career success of protégés in Hellas (Greece) and subsequently developed a culture-dependent explanation, which suggests that receiving traditional mentoring may have detrimental effects on subjective evaluations and workplace adjustment in particular cultural environments (e.g., when there is no trust of authority and there is a clear separation between the ingroup and the outgroup). Furthermore, although mentoring appears to be a universal phenomenon (e.g., Gentry, Weber, & Sadri, 2008), there are some indications that its exact meaning, functions, and prevalence vary across cultures. For example, a study of information technology professionals from four European countries with distinct cultures indicated considerable variance in the prevalence of informal mentoring, with a range from 9.9% for Norway to 42.9% for Italy (Bozionelos, 2007). A complementary study reported a mentoring prevalence of 72.6% among white-collar workers in China (Bozionelos & Wang, 2006), suggesting that the national culture probably influences the meaning attached to developmental relationships, which, in turn, affects the prevalence of mentoring in the workplace. In addition, careers that are pursued partly or thoroughly across national boundaries are likely to impose demands for new skills or qualities. For example, Ng et al. (2005) considered international experience as part of human capital, a view that is endorsed in the international management literature (e.g., Takeuchi, Tesluk, & Marinova, 2006). Other human capital attributes that may be essential in cross-border career movements include ability to learn languages, cultural empathy, cultural intelligence, and the personality trait of openness (e.g., Earley, 2002; Peltokorpi, 2008; Tarique & Schuler, 2008). We suggest using caution before adopting competencies or traits that are narrow and exclusive to cross-cultural adaptation (e.g., see Berry & Ward, 2006). There may be a need first to enrich or broaden the constructs we have developed for understanding careers to account for career progression and success in a globalized environment. Finally, the movement of labor across national borders has increased the prevalence of another form of diversity, cultural diversity (e.g., Chope, 2008;

Career Issues

Sippola & Smale, 2007). Because culture represents beliefs, values, and assumptions, cultural diversity operates mostly at the deep level. This imposes additional challenges for organizations in terms of managing it with respect to their career systems, taking into account that, as seen, the effects of deep-level diversity tend to be realized in the longer term.

Copyright American Psychological Association. Not for further distribution.

Expatriation Expatriation refers to the pursuit and accumulation of work experiences outside the country in which an individual has been reared and whose cultural values the individual espouses. Expatriation and its natural consequence, repatriation, are becoming more common (e.g., Mercer Human Resource Consulting, 2008). This is due to the increasing numbers of organizations that cross national borders and the facilitation in labor movement across such borders (e.g., World Bank, 2009). Expatriation may be distinguished as either corporate-sponsored or self-initiated. Corporate-sponsored expatriation. The term corporate-sponsored expatriation applies to individuals who are sponsored by a parent-country organization to go to a mission abroad. Traditionally, expatriation has been viewed from this perspective. Sponsoring employees in expatriate missions has necessitated the management of their careers during their stay in the host country and upon their return as repatriates (Baruch & Altman, 2002). Expatriate and repatriate career management imposes a number of challenges. For example, expatriates or their families may face cultural adaptation problems that may impact their performance (e.g., Harrison & Shaffer, 2005), which in turn may have repercussions for their careers. Furthermore, although expatriation is usually a choice made with the expectation of career enhancement (e.g., Doherty & Dickmann, 2009; Stahl, Miller, & Tung, 2002) it may fail to deliver its anticipated career benefits. In a recent matched-sample study within a single organization, Benson and Pattie (2008) found that repatriates had received fewer intraorganizational promotions than their counterparts who did not acquire expatriate experience and were not more likely to have been contacted by headhunters for potential interorganizational career

moves. Kraimer, Shaffer, and Bolino (2009) found that the proportions of recent repatriates who reported that they had been demoted and promoted in their employing organizations were similar (15% and 17%, respectively). No career advantages of expatriation renders it not unusual for repatriates to express dissatisfaction with their intraorganizational career prospects and to express intentions to leave (Bossard & Peterson, 2005) or to actually leave the organization (e.g., Baruch, Steele, & Quantrill, 2002; Black & Gregersen, 1999); which makes intraorganizational career success on repatriation of interest to organizations. Indeed, repatriate turnover is apparently double the rate of the rest of the workforce (e.g., see GMAC Global Relocation Services, 2008). Organizations, nonetheless, can enhance career prospects for their expatriates upon their return if they use systems such as predeparture career planning, systematic career revision during the assignment, and allocation of a formal mentor based in the parent country (e.g., Bolino, 2007; Lazarova & Cerdin, 2007). At this point, however, it should be noted that extant research has been methodologically constrained by the fact that, presumably because of accessibility issues, most samples are composed of individuals who are either still in their expatriate assignment or are repatriates and still employed with the parent organization; most samples do not include repatriates who have left the parent company. Hence, information on career success and motives for leaving of those expatriates who have actually left the parent organization are largely unaccounted for. Therefore, a complementary account for the effects of expatriation on career success, which draws on contemporary notions of careers as protean and boundaryless, has been developed (Lazarova & Cerdin, 2007; see also Suutari & Brewster, 2003). This explanation posits that expatriation has mostly positive effects on career success but not necessarily as this is viewed from the traditional corporate-bound career perspective. That is, career actors may, upon return, opt to “cash” the human capital gained by the expatriate experience by means of an interorganizational career move that is seen as providing greater career gains. Research provides some support for this view (Lazarova & 97

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

Cerdin; Stahl, Chua, Caligiuri, Cerdin, & Taniguchi, 2009; Suutari & Brewster). For example, Lazarova and Cerdin found that active engagement in tactics that should facilitate interorganizational career movement explained 19% of the variance in repatriate turnover intentions over and above the variance accounted for by their satisfaction with organizational support, whereas, the latter was not able to explain variance over and above the former. The implication is that organizations may not have full control over retaining their repatriates, regardless of the quality of their expatriation and repatriation management systems. Whether pursuing intra- or interorganizational career success, a key aspect for successful career outcomes of expatriates upon repatriation appears to be proactivity. Activities that should enhance prospects for intraorganizational career success upon repatriation include the following (e.g., Bolino, 2007; Kraimer et al., 2009; O’Sullivan, 2002): seeking expatriate assignments with developmental value or in subsidiaries with strategic importance; actively maintaining and using one’s social capital in the parent-country, which can provide information on developments back home and can prevent the “out of sight, out of mind” syndrome; and seeking a repatriate position that allows use and recognition of the skills and experience acquired while abroad. On the other hand, actively engaging one’s social capital for career-related information and support and scanning the job market should enhance the prospects of interorganizational career moves (see Lazarova & Cerdin, 2007). Finally, corporate-sponsored expatriation includes the cases of host country–origin individuals who are sent to the parent country (inpatriates) as well as cases of third-country origin individuals whom organizations use with increasing frequency (e.g., Kiessling & Harvey, 2006; Tarique, Schuler, & Gong, 2006). These individuals may encounter different or additional career issues than those faced by the traditional parent-country expatriates (e.g., an inpatriate may face the extra challenge of being treated as a minority member in the parent country; Harvey, Novicevic, Buckley, & Fung, 2005), which poses an additional challenge for organizations and for future research. 98

Self-initiated expatriation. Self-initiated expatriation involves cases of individuals who take on employment in a foreign country without sponsorship from a corporation in their home country. These individuals differ substantially from corporatesponsored expatriates, as they cannot rely on the resources of a parent-country corporation in their endeavor (Bozionelos, 2009; see also Suutari & Brewster, 2000). It is surprising that this type of expatriation has received limited attention so far. Self-initiated expatriation is a phenomenon that has existed for a long time (e.g., Inkson, Arthur, Pringle, & Barry, 1997; Inkson & Myers, 2003) but is becoming more common for a number of partly interrelated reasons, including (a) increasing facilitation in the movement of labor across national borders (e.g., the European Union), (b) the development of technology (especially the World Wide Web) that provides information on job openings anywhere in the globe, and (c) dramatic shortages in skilled or unskilled labor that lead many countries to resort to hiring from abroad (e.g., Harry, 2007; Harvey, Hartnell, & Novicevic, 2004; Martin, 2003). The growing frequency of self-initiated expatriation is in line with the increasing tendency for individuals to take ownership of their own careers, as discussed earlier in the chapter. The demographic synthesis of the population of self-initiated expatriates is apparently different from that of corporate-sponsored expatriates. The latter tend to be executives, managers, and high-profile professionals (e.g., scientists, engineers), whereas self-initiated expatriates come from every occupational and societal stratum. They can be high school or university graduates in the career exploration stage (e.g., Inkson & Myers, 2003), higher (Richardson & McKenna, 2002; Suutari & Brewster, 2000) or lower (e.g., Bozionelos, 2009) profile professionals or managers, or simply unskilled or semiskilled workers seeking employment. This last category is likely to be the largest. For example, in the United Kingdom alone, there are currently more than 2 million foreign nationals in employment, most of them unskilled or semiskilled workers, especially in the manufacturing, construction, and distribution sectors (Bowcott & Booth, 2009). Therefore, an additional reason for studying self-initiated expatriates is their difference in

Copyright American Psychological Association. Not for further distribution.

Career Issues

occupational terms from their corporate-sponsored counterparts. Another distinguishing feature of self-initiated expatriation pertains to the motive behind the decision to expatriate. Corporate-sponsored expatriation is normally a conscious career enhancement move. In contrast, self-initiated expatriation is more likely to be either a spontaneous move at the career exploration stage without calculated career expectations (e.g., Inkson & Myers, 2003), or the only available move for survival (see Bozionelos, 2009), with career enhancement (e.g., increase in status, learning) not even remotely considered in the initial decision. Self-initiated expatriation naturally has career instrumental value (e.g., Inkson & Myers, 2003; Jokinen, Brewster, & Suutari, 2008; Vance, 2005). However, the way it influences the career journey and career success may differ from corporate-sponsored expatriation. For example, objective career benefits may be realized in the longer term. Furthermore, in contrast to the majority of their corporate-sponsored counterparts, many self-initiated expatriates may never return to their home countries, staying abroad for the totality of their work lives instead and essentially becoming immigrants; though their initial plan may have been to return to their home countries after an interval abroad. There is still limited knowledge of the factors that contribute to the success of selfinitiated expatriation and of the extent to which these factors overlap with the corresponding factors in corporate-sponsored expatriation. In addition, the notion of career success for self-initiated expatriates may differ. Traditional indexes of objective success (e.g., promotions) may not be applicable, and operationalizations of subjective success that are fully or partly anchored on conventional objective achievements (e.g., satisfaction with advancement and prospects for advancement) may equally be inappropriate. Instead, outcomes that pertain either to learning and accumulation of experience or to the assurance of a living for themselves and their offspring may be more pertinent to the concept of career success for the majority of self-initiated expatriates. Understanding the perspectives and needs of these expatriates is important because the extent to which these are met is likely to affect the performance of their host organizations or nations (e.g., Lee, 2005).

Virtual Global Careers—Nonphysical Crossing of Borders Finally, the fact that information technology has enabled human interaction regardless of physical proximity has led to cases of individuals who regularly interact with colleagues or clients across borders without physically leaving their home countries (e.g., researchers collaborating in projects, employees in outsourced call center operations). This imposes additional challenges in the study of careers (Forret, 2007; Tams & Arthur, 2007). For example, these career actors are exposed to other cultures and may need to accommodate certain degrees of cultural adaptation (e.g., to adapt their behaviors to the styles of customers). However, it is uncertain whether, or to what extent, their careers can be characterized as global, considering that their experiences are probably different from those of expatriates in the traditional physical sense. What is also of importance is how their virtual cross-cultural experiences affect their own views of their work and personal lives. This is an issue of interest in the study of careers, taking into account that globalization has been boosted mostly by technology and that the latter is unlikely to reverse or come to a standstill; meaning that the cases of individuals who regularly engage in work-related interactions across borders without being physically mobile should be constantly increasing. SUMMARY AND DIRECTIONS In this chapter, we covered most major themes along with recent developments in the study of careers. Common axes in our consideration were the elements of time and space. These elements are common across individual careers, which are, however, highly differentiated by the timing and pace of movement through the space element, the content of that element, and the way career actors evaluate workrelated experiences. Career progression and success is determined by both human capital and environmental factors, and are facilitated by processes that involve the social capital of career actors. The extent to which individuals actively direct their careers according to their values in a protean way, as well as their willingness and ability to learn, has acquired increasing importance over time; which has witnessed 99

Copyright American Psychological Association. Not for further distribution.

Baruch and Bozionelos

a shift from organizational toward individual or collaborative responsibility for career management. The changing nature of careers has also been epitomized by the establishment of boundaryless and postcorporate careers. These, however, have not displaced the traditional career, and neither have they changed the fact that most individuals are still most comfortable with that archetypal career. The importance of diversity as a factor in career progression was also discussed, taking mainly the perspective of gender. The conclusion was that although there is an apparent reduction in career opportunities mismatch between diverse groups, actual equality may go beyond lack of significant results in pertinent studies. Finally, the impact of globalization on careers was considered, focusing on whether concepts and career-enhancing factors are applicable across geographical boundaries and on the movement, physical or virtual, of career actors across borders. Despite the substantial knowledge accumulated, however, understanding is still far from complete, especially as careers are constantly evolving: The space element is embedded into societal structures and technology, which are interacting and constantly changing at an increasing pace. In addition, technology appears to influence the time element as well (i.e., extending the limits of work life). Hence, there is a requirement for continuous research that needs to focus on issues that include the following: ■

■

■

The influence of the prolongation of life, and especially of the active part of life, on careers. Our models of the time element may need substantial updating. Further development and refinement of the notions of the protean and boundaryless careers, and their combination in order to understand individual career patterns and improve career advice. Extant notions should also be extended to the domain of cross-border work experiences. In addition, they could be integrated with theories of vocational choice such as Holland’s. Documentation of new career patterns to which changes in the economic and social environment may give rise (i.e., in the same manner as it gave rise to the boundaryless and postcorporate notions).

100

■

■

■

■

■

■

■

Further understanding of the relationship between organizational practices and the psychological contract of career actors and especially whether the conditions under which its relational elements can be cultivated. Enhancement of our knowledge of social capital and job-irrelevant human capital and the processes by which these influence careers. The monitoring of discrepancies in career progression among diverse groups and the development of understanding of the causes of any shifts in these discrepancies (e.g., apparent closure of differences in objective career success). Furthermore, attention may need to be directed at diversity dimensions (e.g., body weight) that have been neglected in the past. How globalization, and the familiarization of various cultures, races, and ethnicities with each other may be influencing the impact of surfacelevel diversity on careers. The meaning of work and life experiences for that substantial part of the world population for whom the concept of career as understood in the past 150 years does not seem to apply. The comprehension of how national culture influences careers and their related constructs (e.g., mentoring). The meaning of virtual global careers and their impact on career actors.

Careers are of critical importance to individuals, employers, society, and the economy, and we hope and believe that this chapter sheds light on the present state of the art in their study.

References Achkar, E. (2008). Will women ever break the glass ceiling in medicine? American Journal of Gastroenterology, 103, 1587–1588. Adler, N. J. (1993). An international perspective on the barriers to the advancement of women managers. Applied Psychology: An International Review, 42, 289–300. Adler, P. S., & Kwon, S-W. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27, 17–40. Allen, T. A., Freeman, D. M., Russell, J. E. A., Reizenstein, R. C., & Rentz, J. O. (2001). Survivor reactions to organizational downsizing: Does time ease the pain?

Career Issues

Journal of Occupational and Organizational Psychology, 74, 145–164. Allen, T. D., Eby, L. T., Poteet, M. L., Lentz, E., & Lima, L. (2004). Career benefits associated with mentoring for protégés: A meta-analysis. Journal of Applied Psychology, 89, 127–136. Altman, Y., & Shortland, S. (2008). Women and international assignments: Taking stock—A 25-year review. Human Resource Management, 47, 199–216.

Copyright American Psychological Association. Not for further distribution.

Amsden, A. H. (2001). The rise of “the rest”: Challenges to the West from late-industrializating economies. New York: Oxford University Press. Anderson, N., Lankshear, C., Timms, C., & Courtney, L. (2008). “Because it’s boring, irrelevant, and I don’t like computers”: Why high school girls avoid professionally oriented ICT subjects. Computers and Education, 50, 1304–1318. Arbuckle, T. Y., Maag, U., Pushkar, D., & Chaikelson, J. S. (1998). Individual differences in trajectory of intellectual development. Psychology and Aging, 13, 663–675. Argyris, C. (1960). Understanding organizational behavior. Homewood, IL: Dorsey. Armstrong, P. I., Day, S. X., McVay, J. P., & Rounds, J. (2008). Holland’s RIASEC model as an integrative framework for individual differences. Journal of Counseling Psychology, 55, 1–18.

Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and job performance at the beginning of the new millennium: What do we know and where do we go next. International Journal of Selection and Assessment, 9, 9–30. Baruch, Y. (1998). The rise and fall of organizational commitment. Human Systems Management, 17, 135–143. Baruch, Y. (2001). Employability: A substitute for loyalty? Human Resource Development International, 4, 543–566. Baruch, Y. (2004). Managing careers: Theory and practice. London: Prentice Hall. Baruch, Y. (2006). Career development in organizations and beyond: Balancing traditional and contemporary viewpoints. Human Resource Management Review, 16, 125–138. Baruch, Y., & Altman, Y. (2002). Expatriation and repatriation in MNCs: A taxonomy. Human Resource Management, 41, 239–259. Baruch, Y., Bell, M., & Gray, D. (2005). Generalist and specialist graduate business degrees: Tangible and intangible value. Journal of Vocational Behavior, 67, 51–68. Baruch, Y., & Nicholson, N. (1997). Home, sweet work: Requirements for effective home working. Journal of General Management, 23(2), 15–30.

Arnold, J. (1997). Managing careers into the 21st century. London: Chapman.

Baruch, Y., & Quick, J. C. (2007). Understanding second careers: Lessons from a study of U.S. Navy admirals. Human Resource Management, 46, 471–491.

Arthur, M. B., Claman, P. H., & DeFillippi, R. J. (1995). Intelligent enterprise, intelligent careers. Academy of Management Executive, 9(4), 7–20.

Baruch, Y., & Rosenstein, E. (1992). Career planning and managing in high tech organizations. International Journal of Human Resource Management, 3, 477–496.

Arthur, M. B., Hall, D. T., & Lawrence, B. S. (1989). Generating new directions in career theory: The case for a transdisciplinary approach. In M. B. Arthur, D. T. Hall, & B. S. Lawrence (Eds.), Handbook of career theory (pp. 7–25). Cambridge, England: Cambridge University Press.

Baruch, Y., Steele, D. J., & Quantrill, G. A. (2002). Management of expatriation and repatriation for novice global player. International Journal of Manpower, 23, 659–671.

Arthur, M. B., & Rousseau, D. M. (1996). The boundaryless career as a new employment principle. In M. G. Arthur & D. M. Rousseau (Eds.), The boundaryless career (pp. 3–20). New York: Oxford University Press. Ashmore, R. D. (1981). Sex stereotypes and implicit personality theory. In D. L. Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 37–81). Hillsdale, NJ: Erlbaum. Baird, L., & Kram, K. (1983). Career dynamics: Managing the supervisor/subordinate relationship. Organizational Dynamics, 11(4), 46–64. Ball, K., Berch, D. B., Helmers, K. F., Jobe, J. B., Leveck, M. D., Marsiske, M., et al. (2002). Effects of cognitive training interventions with older adults: A randomized controlled trial. JAMA, 288, 2271–2281.

Baugh, S. G., & Fagenson-Eland, E. A. (2005). Boundaryless mentoring: An exploratory study of the functions provided by internal versus external organizational mentors. Journal of Applied Social Psychology, 35, 939–955. Baugh, S. G., & Fagenson-Eland, E. A. (2007). Formal mentoring programs: A “poor cousin” to informal relationships? In B. R. Ragins & K. E. Kram (Eds.), The handbook of mentoring at work (pp. 249–271). Thousand Oaks, CA: Sage. Becker, G. S. (1964). Human capital. New York: National Bureau of Economic Research. Becker, G. S. (1975). Human capital. Chicago, IL: Chicago University Press. Beckman, C. M., & Phillips, D. J. (2005). Interorganizational determinants of promotion: Client leadership and the attainment of women attorneys. American Sociological Review, 70, 678–701. 101

Baruch and Bozionelos

Bell, M. P. (2007). Diversity in organizations. Mason, OH: South-Western. Benson, G. S., & Pattie, M. (2008). Is expatriation good for my career? The impact of expatriate assignments on perceived and actual career outcomes. International Journal of Human Resource Management, 19, 1636–1653. Berger, J., Fisek, M. H., Norman, R. Z., & Zelditch, M., Jr. (1977). Status characteristics and social interaction: An expectations-states approach. New York: Elsevier.

Copyright American Psychological Association. Not for further distribution.

Berry, J. W., & Ward, C. (2006). Commentary on “redefining interactions across cultures and organizations.” Group and Organization Management, 31, 64–77. Betz, N. E. (2007). Career self-efficacy: Exemplary recent research and emerging directions. Journal of Career Assessment, 15, 403–422. Beutell, N. J., & O’Hare, M. M. (2006). Career pathfinders: A qualitative study on career development. Psychological Reports, 98, 517–528. Binnewies, C., Ohly, S., & Niessen, C. (2008). Age and creativity at work: The interplay between job resources, age and idea creativity. Journal of Managerial Psychology, 23, 438–457.

Bourdieu, P. (1986). The forms of capital. In J. E. Richardson (Ed.), Handbook of theory for research in the sociology of education (pp. 241–258). New York: Greenwood Press. Bowcott, O., & Booth, R. (2009, February 4). Increase in foreign workers over past three years. The Guardian. Retrieved March 23, 2009, from http://www. guardian.co.uk/uk/2009/feb/04/immigrationforeign-workers Bozionelos, N. (2001). Computer anxiety: Relationship with computer experience and prevalence. Computers in Human Behavior, 17, 213–224. Bozionelos, N. (2003a). Career perceptions and the psychological contract in organizational downsizing. In M. A. Rahim, R. T. Golembiewski, & K. D. Mackenzie (Eds.), Current topics in management (Vol. 8, pp. 295–312). New Brunswick, NJ: Transaction Publishers. Bozionelos, N. (2003b). Intraorganizational network resources: Relation to career success and personality. International Journal of Organizational Analysis, 11, 41–66. Bozionelos, N. (2004). Mentoring provided: Relation to mentor’s career success, personality, and mentoring received. Journal of Vocational Behavior, 64, 24–46.

Black, J. S., & Gregersen, H. B. (1999). The right way to manage expats. Harvard Business Review, 77(2), 52–63.

Bozionelos, N. (2005). When the inferior candidate is offered the job: The selection interview as a political and power game. Human Relations, 58, 1605–1631.

Blanchflower, D. G., & Shadforth, C. (2007). Entrepreneurship in the UK. Foundations and Trends in Entrepreneurship, 3, 257–364.

Bozionelos, N. (2006). Mentoring and expressive network resources: Their relationship with career success and emotional exhaustion among Hellenes employees involved in emotion work. International Journal of Human Resource Management, 17, 362–378.

Blickle, G., Witzki, A. H., & Schneider, P. B. (2009). Mentoring support and power: A three-year predictive field study on protégé networking and career success. Journal of Vocational Behavior, 74, 181–189. Bobbitt-Zeher, D. (2007). The gender income gap and the role of education. Sociology of Education, 80, 1–22. Bolino, M. C. (2007). Expatriate assignments and intraorganizational career success: Implications for individuals and organizations. Journal of International Business Studies, 38, 819–835. Boron, J. B., Willis, S. L., & Schaie, K. W. (2007). Cognitive training gain as a predictor of mental status. Journal of Gerontology, 62B, P45–P52. Bossard, A. B., & Peterson, R. B. (2005). The repatriate experience as seen by American expatriates. Journal of World Business, 40, 9–28. Bouchard, T. J., Jr., & Loehlin, J. C. (2001). Genes, evolution, and personality. Behavior Genetics, 31, 243–273. Bourdieu, P. (1980). Le capital social: Notes provisoires [Social capital: Provisional notes]. Actes de la Recherche en Sciences Sociales, 31, 2–3. 102

Bozionelos, N. (2007, May). A comparative study on mentoring in four European countries. In L. T. Eby & S. C. Evans (Chairs), Cross-cultural perspectives on mentoring research. Symposium conducted at the meeting of the Society for Industrial and Organizational Psychology, New York. Bozionelos, N. (2008). Intraorganizational network resources: How they relate to career success and organizational commitment. Personnel Review, 37, 349–363. Bozionelos, N. (2009). Expatriation outside the boundaries of the multinational corporation: A study with expatriate nurses in Saudi Arabia. Human Resource Management, 48, 111–134. Bozionelos, N., & Wang, L. (2006). The relationship of mentoring and network resources with career success in the Chinese organizational environment. International Journal of Human Resource Management, 17, 1531–1546. Bozionelos, N., & Wang, L. (2007). An investigation on the attitudes of Chinese workers towards individually-

Career Issues

based performance related reward systems. International Journal of Human Resource Management, 18, 284–302. Bradley, K. (2000). The incorporation of women into higher education: Paradoxical outcomes. Sociology of Education, 73, 1–18.

Cascio, W. F. (1995). Whither industrial and organizational psychology in a changing world of work? American Psychologist, 11, 928–939.

Brand, H. (1991). Perestroika and its impact on the Soviet labor market. Monthly Labor Review, 114(12), 38–45.

Case, A., & Paxson, C. (2008). Stature and status: Height, ability, and labor market outcomes. Journal of Political Economy, 116, 499–532.

Brass, D. J. (1992). Power in organizations: A social network perspective. In G. Moore & J. A. Whitt (Eds.), Research in politics and society: The political consequences of social networks (Vol. 4, pp. 295–323). Greenwich, CT: JAI Press. Copyright American Psychological Association. Not for further distribution.

Carmeli, A., Shalom, R., & Weisberg, J. (2007). Considerations in organizational career advancement: What really matters. Personnel Review, 36, 190–205.

Briscoe, J. P., & Hall, D. T. (2006). The interplay of boundaryless and protean careers: Combinations and implications. Journal of Vocational Behavior, 69, 4–18. Briscoe, J. P., & Hall, D. T., & DeMuth, R. L. F. (2006). Protean and boundaryless careers: An empirical exploration. Journal of Vocational Behavior, 69, 30–47. Brown, D. J., Cober, R. T., Kane, K., Levy, P. E., & Shalhoop, J. (2006). Proactive personality and the successful job search: A field investigation with college graduates. Journal of Applied Psychology, 91, 717–726. Brown, W. S. (2005). The new employment contract and the “at risk” worker. Journal of Business Ethics, 58, 195–201. Bruderl, J., Preisendorfer, P., & Ziegler, R. (1993). Upward mobility in organizations: The effects of hierarchy and opportunity structure. European Sociological Review, 9, 173–188. Bruner, J. S., & Tagiuri, R. (1954). The perception of people. In G. Lindzey (Ed.), Handbook of social psychology (Vol. 2, pp. 634–654). Cambridge, MA: AddisonWesley. Bureau of Labor Statistics. (2007, May). May 2007 National occupational employment and wage estimates. Retrieved February 27, 2009, from http://www.bls.gov/oes/current/oes_nat.htm

Cattell, R. B. (1941). Some theoretical issues in adult intelligence testing. Psychological Bulletin, 38, 592. Chao, G. T. (1990). Exploration of the conceptualization and measurement of career plateau: A comparative analysis. Journal of Management, 16, 181–193. Chevalier, A. (2007). Education, occupation, and career expectations: Determinants of the gender pay gap for UK graduates. Oxford Bulletin of Economics and Statistics, 69, 819–842. Chien, L. L., Ko, H. C., & Wu, J. Y. W. (2007). The fivefactor model of personality and depressive symptoms: One-year follow-up. Personality and Individual Differences, 43, 1013–1023. Chope, R. C. (2008). Practice and research in career counseling and development—2007. Career Development Quarterly, 57, 98–173. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S95–S120. Combes, P-P., Linnemer, L., & Visser, M. (2008). Publish or peer-ish? The role of skills and networks in hiring economics professors. Labour Economics, 15, 423–441.

Byrne, D. E. (1971). The attraction paradigm. New York: Academic Press.

Cooper, C. L. (2006). The changing nature of work: The new psychological contract and associated stressors. In A. M. Rossi, P. L. Perrewe, & S. L. Sauter (Eds.), Stress and quality of working life: Current perspectives in occupational health (pp. 1–7). Greenwich, CT: Information Age Publishing

Cannings, K., & Montmarquette, C. (1991). Managerial momentum: A simultaneous model of the career progress of male and female managers. Industrial and Labor Relations Review, 44, 212–228.

Coucke, K., Pennings, E., & Sleuwaegen, L. (2007). Employee layoff under different modes of restructuring: Exit, downsizing or relocation. Industrial and Corporate Change, 16, 161–182.

Cappellen, T., & Janssens, M. (2005). Career paths of global managers: Towards future research. Journal of World Business, 40, 348–360. Cappelli, P. (1999). Career jobs are dead. California Management Review, 42(1), 146–167.

Dahlmann, N., Elsner, M., Jeschke, S., Natho, N., & Schroder, C. (2008, June). Gender gap in technological disciplines: Societal causes and consequences. Paper presented at the IEEE International Symposium on Technology and Society, Fredericton, Canada.

Cappelli, P. (2006). Churning of jobs. In J. H. Greenhaus & G. A. Callanan (Eds.), Encyclopedia of Career Development: Vol. 1 (pp. 165–167). Thousand Oaks, CA: Sage.

Dalton, G., Thompson, M. P., & Price, P. (1977). The four stages of professional careers: A new look at performance by professionals. Organizational Dynamics, 6(1), 19–42. 103

Baruch and Bozionelos

D’Amato, A., & Herzfeldt, R. (2008). Learning orientation, organizational commitment, and talent retention across generations: A study of European managers. Journal of Managerial Psychology, 23, 929–953. Davey, K. M. (2008). Women’s accounts of organizational politics as a gendering process. Gender, Work, and Organization, 15, 650–671.

Copyright American Psychological Association. Not for further distribution.

Davis, J. W., & Bauman, K. J. (2008). School enrolment in the United States: 2006. Washington, DC: U.S. Census Bureau. De Cuyper, N., Bernhard-Oettel, C., Berntson, E., De Witte, H., & Alarco, B. (2008). Employability and employees’ well-being: Mediation by job insecurity. Applied Psychology: An International Review, 57, 488–509. De Cuyper, N., Rigotti, T., De Witte, H., & Mohr, G. (2008). Balancing psychological contracts: Validation of a typology. International Journal of Human Resource Management, 19, 543–561. DeFillippi, R. J., & Arthur, M. B. (1994). The boundaryless career: A competency-based perspective. Journal of Organizational Behavior, 15, 307–324. de Janasz, S. C., & Sullivan, S. E. (2004). Multiple mentoring in academe: Developing the professional network. Journal of Vocational Behavior, 64, 263–283. DeNeve, K. M., & Cooper, H. (1998). The happy personality: A meta-analysis of 137 traits and subjective well-being. Psychological Bulletin, 124, 197–229. Department of Professional Employees. (2008). Vital workforce statistics: Fact sheet 2008. Retrieved February 19, 2009, from http://www.dpeaflcio.org/ programs/factsheets/fs_2008_workforce.pdf Devaro, J., & Brookshire, D. (2007). Promotions and incentives in nonprofit and for-profit organizations. Industrial and Labor Relations Review, 60, 311–339. De Vos, A., & Soens, N. (2008). Protean attitude and career success: The mediating role of self-management. Journal of Vocational Behavior, 73, 449–456. Ding, D. Z., & Warner, M. (2001). China’s labourmanagement system reforms: Breaking the “three old irons” (1978–1999). Asia Pacific Journal of Management, 18, 315–334. Dipboye, R. L., & Halverson, S. K. (2004). Subtle (and not so subtle) discrimination in organizations. In R. W. Griffin & A. M. O’Leary-Kelly (Eds.), The dark side of organizational behavior (pp. 131–158). San Francisco: Jossey-Bass. Doherty, N., & Dickmann, M. (2009). Exposing the symbolic capital of international assignments. International Journal of Human Resource Management, 20, 301–320. Donohue, R. (2007). Examining career persistence and career change intent using the career attitudes and 104

strategies inventory. Journal of Vocational Behavior, 70, 259–276. Dries, N., Pepermans, R., & Carlier, O. (2008). Career success: Constructing a multidimensional model. Journal of Vocational Behavior, 73, 254–267. Dries, N., Pepermans, R., & De Kerpel, E. (2008). Exploring four generations’ beliefs about career. Journal of Managerial Psychology, 23, 907–928. Dyke, L. S., & Murphy, S. A. (2006). How we define success: A quantitative study of what matters most to women and men. Sex Roles, 55, 357–371. Eagly, A. H., & Carli, L. L. (2007). Women and the labyrinth of leadership. Harvard Business Review, 85(9), 62–71. Earley, P. C. (2002). Redefining interactions across cultures and organizations: Moving forward with cultural intelligence. Research in Organizational Behavior, 24, 271–299. Eby, L. T., Allen, T. D., Evans, S. C., Ng, T., & DuBois, D. L. (2008). Does mentoring matter? A multidisciplinary meta-analysis comparing mentored and nonmentored individuals. Journal of Vocational Behavior, 72, 254–267. Eby, L. T., Durley, J. R., Evans, S. C., & Ragins, B. R. (2006). The relationship between short-term mentoring benefits and long-term mentor outcomes. Journal of Vocational Behavior, 69, 424–444. Evans, P. (1986). New directions in career management. Personnel Management, 18(12), 26–30. Eysenck, H. J. (1992). Four ways five factors are not basic. Personality and Individual Differences, 13, 667–673. Farber, H. S. (2008). Job loss and the decline in job security in the United States. Manuscript in preparation. Feldman, D. C. (1985). The new careerism: Origins, tenets, and consequences. The Industrial Psychologist, 22, 39–44. Feldman, D. C. (1996). The nature, antecedents, and consequences of underemployment. Journal of Management, 22, 385–407. Feldman, D. C., & Bolino, M. C. (2000). Career patterns of the self-employed: Career motivations and career outcomes. Journal of Small Business Management, 38, 53–67. Feldman, D. C., & Leana, C. R. (2000). Whatever happened to laid-off executives: A study of reemployment challenges after downsizing. Organizational Dynamics, 29(1), 64–75. Feldman, D. C., & Ng, T. W. H. (2007). Careers: Mobility, embeddedness, and success. Journal of Management, 33, 350–377. Ferris, G., & Judge, T. (1991). Personnel/human resource management: A political influence perspective. Journal of Management, 17, 477–488.

Career Issues

Ferris, G. R., & King, T. R. (1991). Politics in human resources decisions: A walk on the dark side. Organizational Dynamics, 20(2), 59–71. Fletcher, J. K., & Ragins, B. R. (2007). Stone center relational cultural theory: A window on relational mentoring. In B. R. Ragins & K. E. Kram (Eds.), The handbook of mentoring at work (pp. 373–399). Thousand Oaks, CA: Sage. Fombrun, C. J. (1982). Strategies for network research in organizations. Academy of Management Review, 7, 280–291.

Copyright American Psychological Association. Not for further distribution.

Form, W. H., & Miller, D. C. (1949). Occupational career patterns as a sociological instrument. American Journal of Sociology, 54, 317–329. Forret, M. L. (2007, June). Verbal comment. In N. Bozionelos, M. Crocitto, & H. van Emmerik (Chairs), Shaping the future: How to accommodate the global career of tomorrow? Round Table Discussion conducted at the International Conference of the Eastern Academy of Management, Amsterdam, The Netherlands. Forret, M. L., & Dougherty, T. W. (2004). Networking behaviors and career outcomes: Differences for men and women? Journal of Organizational Behavior, 25, 419–437. Friedberg, L. (2007, March). The recent trend towards later retirement. Work Opportunities for Older Americans, 9, 1–7. Fugate, M., Kinicki, A. J., & Ashforth, B. E. (2004). Employability: A psycho-social construct, its dimensions, and applications. Journal of Vocational Behavior, 65, 14–38. Gattiker, U. E., & Larwood, L. (1986). Subjective career success: A study of managers and support personnel. Journal of Business and Psychology, 1, 78–94. Gattiker, U. E., & Larwood, L. (1988). Predictors for managers’ career mobility, success, and satisfaction. Human Relations, 41, 569–591. Gentry, W. A., Weber, T. J., & Sadri, G. (2008). Examining career-related mentoring and managerial performance across cultures: A multilevel analysis. Journal of Vocational Behavior, 72, 241–253. GMAC Global Relocation Services. (2008). Global relocation trends: 2008 survey report. Oak Brook, IL: GMAC Global Relocation Services. Granrose, C. S., & Baccili, P. A. (2006). Do psychological contacts include boundaryless or protean careers? Career Development International, 11, 163–182.

Green, S. G., & Bauer, T. N. (1995). Supervisory mentoring by advisers: Relationship with doctoral student potential, productivity and commitment. Personnel Psychology, 48, 537–561. Greenhaus, J. H. (1987). Career management. New York: Dryden Press. Greenhaus, J. H., & Beutell, N. J. (1985). Sources of conflict between work and family roles. Academy of Management Review, 10, 76–88. Greenhaus, J. H., Callanan, G. A., & DiRenzo, M. (2008). A boundaryless perspective on careers. In J. Barling & C. L. Cooper (Eds.), The SAGE handbook of organizational behavior: Vol. 1. Micro approaches (pp. 277–299). London: Sage. Greenhaus, J. H., Callanan, G. A., & Godshalk, V. M. (2000). Career management (3rd ed.). Fort Worth, TX: Dryden Press. Greenhaus, J., & Parasuraman, S. (1993). Job performance attributions and career advancement prospects: An examination of gender and race effects. Organizational Behavior and Human Decision Processes, 55, 273–297. Greenhaus, J., & Powell, G. (2006). When work and family are allies: Theory of work-family enrichment. Academy of Management Review, 31, 77–92. Gregory, A., & Milner, S. (2009). Editorial: Work–life balance: A matter of choice? Gender, Work, and Organization, 16, 1–13. Griffin, B., & Hesketh, B. (2008). Post-retirement work: The individual determinants of paid and volunteer work. Journal of Occupational and Organizational Psychology, 81, 101–121. Gunz, H. P. (1988). Organizational logics of managerial careers. Organization Studies, 9, 529–554. Gutteridge, T. G., Leibowitz, Z. B., & Shore, J. E. (1993). Organizational career development. San Francisco: Jossey-Bass. Hall, D. T. (1976). Careers in organizations. Pacific Palisades, CA: Goodyear. Hall, D. T. (2002). Careers in and out organizations. Thousand Oaks, CA: Sage. Hall, D. T. (2004). The protean career: A quarter-century journey. Journal of Vocational Behavior, 65, 1–13. Hall, D. T., & Chandler, D. E. (2005). Psychological success: When the career is a calling. Journal of Organizational Behavior, 26, 155–176.

Gray, J. (2001). The end of career. Communications of the ACM, 44(11), 65–69.

Hall, D. T., & Mirvis, P. H. (1995). The new career contract: Developing the whole person at midlife and beyond. Journal of Vocational Behavior, 47, 269–289.

The great jobs switch. (2005, September 29). The Economist. Retrieved February 23, 2009, from http://www.economist.com/opinion/displaystory.cfm ?story_id=4458528

Hall, D. T., & Mirvis, P. H. (1996). The new protean career: Psychological success and the path with a heart. In D. T. Hall (Ed.), The career is dead—long live the career (pp. 15–45). San Francisco: Jossey-Bass. 105

Baruch and Bozionelos

Hall, D. T., & Nougaim, K. (1968). An examination of Maslow’s need hierarchy in an organizational setting. Organizational Behavior and Human Performance, 3, 12–35. Hansen, R. S. (2009). Portfolio careers: Creating a career of multiple part-time jobs. Quintessential Careers. Retrieved February 22, 2009, from http://www. quintcareers.com/portfolio_careers.html

Copyright American Psychological Association. Not for further distribution.

Harel, G. H., Tzafrir, S. S., & Baruch, Y. (2003). Achieving organizational effectiveness through promotion of women into managerial positions: HRM practice focus. International Journal of Human Resource Management, 14, 247–263. Harrison, D. A., & Klein, K. J. (2007). What’s the difference? Diversity constructs as separation, variety, or disparity in organizations. Academy of Management Review, 32, 1199–1228. Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond relational demography: Time and the effects of surfaceand deep-level diversity on work group cohesion. Academy of Management Journal, 41, 96–107. Harrison, D. A., Price, K. H., Gavin, J. H., & Florey, A. T. (2002). Time, teams, and task performance: Changing effects of surface- and deep-level diversity on group functioning. Academy of Management Journal, 45, 1029–1045. Harrison, D. A., & Shaffer, M. A. (2005). Mapping the criterion space for expatriate success: Task- and relationship-based performance, effort and adaptation. International Journal of Human Resource Management, 16, 1454–1474. Harry, W. (2007). Employment creation and localization: The crucial human resource issues for the GCC. International Journal of Human Resource Management, 18, 132–146. Hartung, P. J., Porfeli, E. J., & Vondracek, F. W. (2005). Child vocational development: A review and reconsideration. Journal of Vocational Behavior, 66, 385–419. Harvey, M., Hartnell, C., & Novicevic, M. (2004). The inpatriation of foreign healthcare workers: A potential remedy for the chronic shortage of professional staff. International Journal of Intercultural Relations, 28, 127–150. Harvey, M., Novicevic, M. M., Buckley, M. R., & Fung, H. (2005). Reducing inpatriate managers’ “liability of foreignness” by addressing stigmatization and stereotype threats. Journal of World Business, 40, 267–280. Hassard, J., Morris, J., Sheehan, J., & Yuxin, X. (2006). Downsizing the danwei: Chinese state-enterprise reform and the surplus labour question. International Journal of Human Resource Management, 17, 1441–1455. Heilman, M. E., & Welle, B. (2006). Disadvantaged by diversity? The effects of diversity goals on compe106

tence perceptions. Journal of Applied Social Psychology, 36, 1291–1319. Helwig, A. A. (2008). From childhood to adulthood: A 15-year longitudinal career development study. The Career Development Quarterly, 57, 38–50. Heslin, P. A. (2005). Conceptualizing and evaluating career success. Journal of Organizational Behavior, 26, 113–136. Higgins, M. (2005). Career imprints: Creating leaders across an industry. San Francisco: Jossey-Bass. Higgins, M. C., & Kram, K. E. (2001). Reconceptualizing mentoring at work: A developmental network perspective. Academy of Management Review, 26, 264–288. Higgins, M. C., & Thomas, D. A. (2001). Constellations and careers: Toward understanding the effects of multiple developmental relationships. Journal of Organizational Behavior, 22, 223–247. Hofbauer, J., & Fischlmayr, I. C. (2004). Feminization of international assignments: Conquering empty castles? International Studies of Management and Organization, 34(3), 46–67. Hogan, R., & Blake, R. (1999). John Holland’s vocational typology and personality theory. Journal of Vocational Behavior, 55, 41–56. Holgate, J., & McKay, S. (2009). Equal opportunities policies: How effective are they in increasing diversity in the audio-visual industries’ freelance labour market? Media, Culture and Society, 31, 151–163. Holland, J. L. (1959). A theory of occupational choice. Journal of Counseling Psychology, 6, 35–45. Holland, J. L. (1996). Exploring careers with a typology: What we have learned and some new directions. American Psychologist, 51, 397–406. Holland, J. L. (1997). Making vocational choices: A theory of vocational personalities and work environments (3rd ed.). Odessa, FL: Psychological Assessment Resources. Holman, K. (2009). Downsized. Mergers and Acquisitions: The Dealmaker’s Journal, 44(2), 30–31. Hooker, C. I., Verosky, S. C., Miyakawa, A., Knight, R. T., & D’Esposito, M. (2008). The influence of personality on neural mechanisms of observational fear and reward learning. Neuropsychologia, 46, 2709–2724. Horn, J. L., & Cattell, R. B. (1966). Refinement and test of the theory of fluid and crystallized general intelligences. Journal of Educational Psychology, 57, 253–270. Hosoda, M., Stone-Romero, E. F., & Coats, G. (2003). The effects of physical attractiveness on job-related outcomes: A meta-analysis of experimental studies. Personnel Psychology, 56, 431–462.

Career Issues

Copyright American Psychological Association. Not for further distribution.

Hughes, E. C. (1937). Institutional office and the person. American Journal of Sociology, 43, 404–413. Hyde, J. S. (2005). The gender similarities hypothesis. American Psychologist, 60, 581–592.

Judge, T. A., & Cable, D. M. (2004). The effect of physical height on workplace success and income: Preliminary test of a theoretical model. Journal of Applied Psychology, 89, 428–441.

Ibarra, H. (1992). Homophily and differential returns: Sex differences in network structure and access in an advertising firm. Administrative Science Quarterly, 37, 422–447.

Judge, T. A., Higgins, C. A., Thoresen, C. J., & Barrick, M. R. (1999). The big five personality traits, general mental ability, and career success across the life span. Personnel Psychology, 52, 621–652.

Ibarra, H. (1993). Personal networks of women and minorities in management: A conceptual framework. Academy of Management Review, 18, 56–87.

Kammeyer-Mueller, J. D., & Judge, T. A. (2008). A quantitative review of mentoring research: Test of a model. Journal of Vocational Behavior, 72, 269–283.

Iida, T., & Morris, J. (2008). Farewell to the salaryman? The changing roles and work of middle managers in Japan. International Journal of Human Resource Management, 19, 1072–1087.

Kanter, R. M. (1977). Men and women of the corporation. New York: Basic Books.

Inkpen, A. C., & Tsang, E. W. K. (2005). Social capital, networks, and knowledge transfer. Academy of Management Review, 30, 146–165.

Kashima, Y., & Kashima, E. S. (2003). Individualism, GNP, climate, and pronoun drop: Is individualism determined by affluence and climate, or does language use play a role? Journal of Cross-Cultural Psychology, 34, 125–134.

Inkson, K. (2007). Understanding careers: The metaphors of working lives. Thousand Oaks, CA: Sage.

Khapova, S. N., & Korotov, K. (2007). Dynamics of western career attributes in the Russian context. Career Development International, 12, 68–85.

Inkson, K., Arthur, M. B., Pringle, J., & Barry, S. (1997). Expatriate assignment v. overseas experience: Contrasting models of international human resource development. Journal of World Business, 32, 351–368.

Kiessling, T., & Harvey, M. (2006). Global organizational control: A new role by inpatriates. Multinational Business Review, 14(2), 1–27.

Inkson, K., & Myers, B. A. (2003). “The big OE”: Selfdirected travel and career development. Career Development International, 8, 170–181.

Kinsella, K. G. (2005). Future longevity: Demographic concerns and consequences. Journal of the American Geriatrics Society, 53, S299–S303.

Insinga, R. C., & Werle, M. J. (2000). Linking outsourcing to business strategy. Academy of Management Executive, 14(4), 58–70. International Labour Organization. (2009). Laborsta: Database of labour statistics. Retrieved March 10, 2009, from http://laborsta.ilo.org Jackson, S. E., May, K. E., & Whitney, K. (1995). Understanding the dynamics of diversity in decisionmaking teams. In R. A. Guzzo & E. Salas (Eds.), Team decision-making effectiveness in organizations (pp. 204–261). San Francisco: Jossey-Bass.

Kinsella, K., & Phillips, D. R. (2005). Global aging: The challenge of success. Population Bulletin, 60(1). Kirchmeyer, C. (2005). The effects of mentoring on academic careers over time: Testing performance and political perspectives. Human relations, 58, 637–669. Koskina, A. (2008). The “pros” and “cons” of career development in the Greek public sector. Personnel Review, 37, 264–279. Kozhaya, N. (2007). The retirement age in Quebec: A worrying situation. Montreal, Canada: Montreal Economic Institute.

Jang, K. L., Livesley, W. J., Ando, J., Yamagata, S., Suzuki, A., Angleitner, A., et al. (2006). Behavioral genetics of the higher-order factors of the Big Five. Personality and Individual Differences, 41, 261–272.

Kraimer, M. L., Shaffer, M. A., & Bolino, M. (2009). The influence of expatriate and repatriate experiences on career advancement and repatriate retention. Human Resource Management, 48, 27–47.

Jensen, L., & Slack, T. (2003). Underemployment in America: Measurement and evidence. American Journal of Community Psychology, 32, 21–31.

Kram, K. E. (1985). Mentoring in the workplace: Development relationships in organizational life. Glenview, IL: Scott, Foresman.

Jokinen, T., Brewster, C., & Suutari, V. (2008). Career capital during international work experiences: Contrasting self-initiated expatriate experiences and assigned expatriation. The International Journal of Human Resource Management, 19, 979–998.

Kram, K. E., & Isabella, L. A. (1985). Mentoring alternatives: The role of peer relationships in career development. Academy of Management Journal, 28, 110–132.

Jokisaari, M., & Nurmi, J-E. (2005). Company matters: Goal-related social capital in the transition to working life. Journal of Vocational Behavior, 67, 413–428.

Krause, A., Stadil, T., & Bunke, J. (2003). Auswirkungen von downsizing-maßnahmen auf das organisationale commitment der verbleibenden mitarbeiter: Ein vorher-nachher-vergleich [Impact of downsizing 107

Baruch and Bozionelos

measures on the organizational commitment of remaining employees: Before-and-after-implementation]. Gruppendynamik und Organisationsberatung, 34, 355–372. Lachman, M. E. (2004). Development in midlife. Annual Review of Psychology, 55, 305–331. Landry, B. J. L., Mahesh, S., & Hartman, S. (2005). The changing nature of work in the age of e-business. Journal of Organizational Change Management, 18, 132–144.

Copyright American Psychological Association. Not for further distribution.

Lanzi, D. (2007). Capabilities, human capital, and education. The Journal of Socio-Economics, 36, 424–435. Latham, G. P., & Mann, S. (2006). Advances in the science of performance appraisal: Implications for practice. International Review of Industrial and Organizational Psychology, 21, 295–337. Lazarova, M. B., & Cerdin, J-L. (2007). Revisiting repatriation concerns: Organizational support versus career and contextual influence. Journal of International Business Studies, 38, 404–429. Lazarova, M., & Taylor, S. (2009). Boundaryless careers, social capital, and knowledge management: Implications for organizational performance. Journal of Organizational Behavior, 30, 119–139. Leanna, C. R., & Feldman, D. C. (1992). Coping with job loss: How individuals, organizations, and communities respond to layoffs. New York: Lexington Books. Leck, J. D. (2002). Making employment equity programs work for women. Canadian Public Policy/Analyse de Politiques, 28, S85–S100. Lee, C. H. (2005). A study of underemployment among self-initiated expatriates. Journal of World Business, 40, 172–187. Lepak, D. P., & Snell, S. A. (1999). The human capital resource architecture: Toward a theory of human capital allocation and development. Academy of Management Review, 24, 31–48. Levinson, D. (1978). Seasons of a man’s life. New York: Knopf. Li, Y., Savage, M., & Warde, A. (2008). Social mobility and social capital in contemporary Britain. British Journal of Sociology, 59, 391–411. Linney, G. (2001). Preparing for a career in medical management: 8 steps to help you evaluate your career path. Physician Executive, 27(4), 14–17. Lips-Wiersma, M., & Hall, D. T. (2007). Organizational career development is not dead: A case study on managing the new career during organizational change. Journal of Organizational Behavior, 28, 771–792. Lirio, P., Lituchy, T. R., Monserrat, S. I., Olivas-Lujan, M. R., Duffy, J. A., Fox, S., et al. (2007). Exploring career-life success and family social support of suc108

cessful women in Canada, Argentina, and Mexico. Career Development International, 12, 28–50. Littler, C. R., Wiesner, R., Dunford, R. (2003). The dynamics of delayering: Changing management structures in three countries. Journal of Management Studies, 40, 225–256. London, M. (1983). Toward a theory of career motivation. Academy of Management Review, 8, 620–630. Long, J. E., & Link, A. N. (1983). The impact of market structure on wages, fringe benefits, and turnover. Industrial and Labor Relations Review, 36, 239–250. Lyness, K. S., & Heilman, M. E. (2006). When fit is fundamental: Performance evaluations and promotions of upper-level female and male managers. Journal of Applied Psychology, 91, 777–785. Mainiero, L. A., & Sullivan, S. E. (2005). Kaleidoscope careers: An alternative explanation for the opt-out evolution. Academy of Management Executive, 19(1), 106–123. Mainiero, L. A., & Sullivan, S. E. (2006). The opt-out revolt: Why people are leaving companies to create kaleidoscope careers. Palo Alto, CA: Davies-Black. Malach-Pines, A., & Kaspi-Baruch, O. (2008). The role of culture and gender in the choice of a career in management. Career Development International, 13, 306–319. Marsden, P. V. (1990). Network data and measurement. Annual Review of Sociology, 16, 435–463. Marston, G., & McDonald, C. (2008). Feeling motivated yet? Long-term unemployed people’s perspectives on the implementation of workfare in Australia. Australian Journal of Social Issues, 43, 255–269. Martin, P. L. (2003). Managing labor migration: Temporary worker programs for the 21st century. Geneva: International Institute for Labour Studies. Maslova, I. (1991). State employment programmes in the light of the new law of employment. In G. Standing (Ed.), In search of flexibility: The new Soviet labour market (pp. 121–144). Geneva: International Labour Organization. Mayrhofer, W., Meyer, M., Schiffinger, M., & Schmidt, A. (2008). The influence of family responsibilities, career fields, and gender on career success. Journal of Managerial Psychology, 23, 292–323. McArdle, S., Waters, L., Briscoe, J. P., & Hall D. T. (2007). Employability during unemployment: Adaptability, career identity, and human and social capital. Journal of Vocational Behavior, 71, 247–264. McDonald, P., Brown, K., & Bradley, L. (2005). Have traditional career paths given way to protean ones? Evidence from senior managers in the Australian public sector. Career Development International, 10, 109–129.

Career Issues

McKay, P. F., Avery, D. R., & Morris, M. A. (2008). Mean racial-ethnic differences in employee sales performance: The moderating role of diversity climate. Personnel Psychology, 61, 349–374.

Morrison, A. M., White, R. P., & Van Velsor, E. (1987). Breaking the glass ceiling: Can women reach the top of America’s largest corporations? Reading, MA: AddisonWesley.

McLaughlin, D. H., & Tiedeman, D. V. (1974). Elevenyear career stability and changes as reflected in project talent data through the Flanagan, Holland, and Roe occupational classification systems. Journal of Vocational Behavior, 5, 177–196.

Moss, P., Salzman, H., & Tilly, C. (2008). Under construction: The continuing evolution of job structures in call centers. Industrial Relations, 47, 173–208.

Copyright American Psychological Association. Not for further distribution.

McManus, S. E., & Russell, J. E. A. (2007). Peer mentoring relationships. In B. R. Ragins & K. E. Kram (Eds.), The Handbook of mentoring at work (pp. 273–297). Thousand Oaks, CA: Sage. McQuaid, R. W., & Lindsay, C. (2002). The “employability gap”: Long-term unemployment and barriers to work in buoyant labour markets. Environment and Planning C: Government and Policy, 20, 613–628. Melamed, T. (1995a). Career success: The moderating effect of gender. Journal of Vocational Behavior, 47, 35–60. Melamed, T. (1996a). Career success: An assessment of a gender-specific model. Journal of Occupational and Organisational Psychology, 69, 217–242. Melamed, T. (1996b). Validation of a stage model of career success. Applied Psychology: An International Review, 45, 35–65. Melamed, T., & Bozionelos, N. (1992). Managerial promotion and height. Psychological Reports, 71, 587–593. Mercer Human Resource Consulting. (2008). Benefits survey for expatriates and globally mobile employees. Geneva: Mercer LLC. Metz, I., & Tharenou, P. (2001). Women’s career advancement: The relative contribution of human and social capital. Group and Organization Management, 26, 312–342. Milliken, F. J., & Martins, L. L. (1996). Searching for common threads: Understanding the multiple effects of diversity in organizational groups. Academy of Management Review, 21, 402–433. Mintzberg, H. (1985). The organization as political arena. Journal of Management Studies, 22, 133–154. Miyamoto, K., Takebayashi, H., Takimoto, K., Miyamoto, S., Morioka, S., & Yagi, F. (2008). A new simple performance test focused on agility in elderly people: The ten-step test. Gerontology, 54, 365–372. Moore, P. R., Ng, A. H. C., Yeo, S. H., Sundberg, M., Wong, C. B., & De Vin, L. J. (2008). Advanced machine service support using internet-enabled three-dimensionalbased virtual engineering. International Journal of Production Research, 46, 4215–4235. Morrison, A. M., & Von Glinow, M. A. (1990). Women and minorities in management. American Psychologist, 45, 200–208.

Ng, T. W. H., Eby, L. T., Sorensen, K. L., & Feldman, D. C. (2005). Predictors of objective and subjective career success: A meta-analysis. Personnel Psychology, 58, 367–408. Ng, T. W. H., & Feldman, D. C. (2008). Can you get a better deal elsewhere? The effects of psychological contract replicability on organizational commitment over time. Journal of Vocational Behavior, 73, 268–277. Ng, T. W. H., & Feldman, D. C. (2009). How broadly does education contribute to job performance? Personnel Psychology, 62, 89–134. Nicholson, N., & De Waal-Andrews, W. (2005). Playing to win: Biological imperatives, self-regulation, and trade-offs in the game of career success. Journal of Organizational Behavior, 26, 137–154. Nieva, V. F. (1985). Work and family linkages. In L. Larwood, A. H. Stromberg, & B. A. Gutek (Eds.), Women and work (Vol. 1, pp. 162–190). Beverly Hills, CA: Sage. Okun, M. A., Pugliese, J., & Rook, K. S. (2007). Unpacking the relation between extraversion and volunteering in later life: The role of social capital. Personality and Individual Differences, 42, 1467–1477. O’Neil, D. A., Hopkins, M. M., & Bilimoria, D. (2008). Women’s careers at the start of the 21st century: Patterns and paradoxes. Journal of Business Ethics, 80, 727–743. Ornstein, S., Cron, W. L., & Slocum, J. W. (1989). Life stage versus career stage: A comparative test of the theories of Levinson and Super. Journal of Organizational Behavior, 10, 117–133. Ornstein, S., & Isabella, L. (1990). Age v. stage models of career attitudes of women: A partial replication and extension. Journal of Vocational Behavior, 36, 1–19. O’Sullivan, S. L. (2002). The protean approach to managing repatriation transitions. International Journal of Manpower, 23, 597–616. Over, R., & Thomas, P. (1995). Age and skilled psychomotor performance: A comparison of younger and older golfers. International Journal of Aging and Human Development, 41, 1–12. Oyer, P. (2008). The making of an investment banker: Stock market shocks, career choice, and lifetime income. The Journal of Finance, 63, 2601–2628. Ozbilgin, M., & Healy, G. (2004). The gendered nature of career development of university professors: The 109

Baruch and Bozionelos

case of Turkey. Journal of Vocational Behavior, 64, 358–371. Ozbilgin, M., Kusku, F., & Erdogmus, N. (2005). Explaining influences on career “choice”: The case of MBA students in comparative perspective. International Journal of Human Resource Management, 16, 2000–2028.

Copyright American Psychological Association. Not for further distribution.

Pang, M., Chua, B-L., & Chu, C. W. L. (2008). Learning to stay ahead in an uncertain environment. International Journal of Human Resource Management, 19, 1383–1394. Papastergiou, M. (2008). Are computer science and information technology still masculine fields? High school students’ perceptions and career choices. Computers and Education, 51, 594–608. Parker, R. J., & Kohlmeyer, J. M. (2005). Organizational justice and turnover in public accounting firms: A research note. Accounting, Organizations and Society, 30, 357–369. Parkin, M., Powell, M., & Matthews, K. (2007). Economics (7th ed.). Harlow, England: Pearson. Pastoriza, D., Arino, M. A., & Ricart, J. E. (2008). Ethical managerial behaviour as an antecedent of organizational social capital. Journal of Business Ethics, 78, 329–341. Peiperl, M., & Baruch, Y. (1997). Back to square zero: The postcorporate career. Organizational Dynamics, 25(4), 7–22. Peltokorpi, V. (2008). Cross-cultural adjustment of expatriates in Japan. International Journal of Human Resource Management, 19, 1588–1606.

Powell, G. N., & Eddleston, K. A. (2008). The paradox of the contented female business owner. Journal of Vocational Behavior, 73, 24–36. Prediger, D. J., & Vansickle, T. R. (1992). Locating occupations on Holland’s hexagon: Beyond RIASEC. Journal of Vocational Behavior, 40, 111–128. Ragins, B. R., Cotton, J. L., & Miller, J. S. (2000). Marginal mentoring: The effects of type of mentor, quality of relationship, and program design on work and career attitudes. Academy of Management Journal, 43, 1177–1194. Ragins, B. R., & Sundstrom, E. (1989). Gender and power in organizations. Psychological Bulletin, 105, 51–88. Ramaswami, S. N., & Singh, J. (2003). Antecedents and consequences of merit pay fairness for industrial salespeople. Journal of Marketing, 67, 46–66. Rayman, J., & Atanasoff, L. (1999). Holland’s theory and career intervention: The power of the hexagon. Journal of Vocational Behavior, 55, 114–116. Reitman, F., & Schneer, J. A. (2003). The promised path: A longitudinal study of managerial careers. Journal of Managerial Psychology, 18, 60–75. Reitman, F., & Schneer, J. A. (2008). Enabling the new careers of the 21st century. Organization Management Journal, 5, 17–28. Richardson, J., & McKenna, S. (2002). Leaving and experiencing: Why academics expatriate and how they experience expatriation. Career Development International, 7, 67–78.

Penner, A. M. (2008). Race and gender differences in wages: The role of occupational sorting at the point of hire. The Sociological Quarterly, 49, 597–614.

Rickwood, R. R., Roberts, J., Batten, S., Marshall, A., & Massie, K. (2004). Empowering high-risk clients to attain a better quality of life: A career resiliency framework. Journal of Employment Counseling, 41, 98–104.

Petersen, T., & Saporta, I. (2004). The opportunity structure for discrimination. American Journal of Sociology, 109, 852–901.

Rosenbaum, J. L. (1979). Tournament mobility: Career patterns in a corporation. Administrative Science Quarterly, 24, 221–241.

Pfeffer, J. (1989). A political perspective on careers: Interests, networks, and environments. In M. B. Arthur, D. T. Hall, & B. S. Lawrence (Eds.), Handbook of career theory (pp. 380–396). Cambridge, England: Cambridge University Press.

Roth, L. M. (2004). The social psychology of tokenism: Status and homophily processes on Wall Street. Sociological Perspectives, 47, 189–214.

Phillips, K. W., & Loyd, D. L. (2006). When surface- and deep-level diversity collide: The effects on dissenting group members. Organizational Behavior and Human Decision Processes, 99, 143–160.

Rousseau, D. M. (1995). Psychological contracts in organizations: Understanding written and unwritten agreements. Thousand Oaks, CA: Sage. Rubin, B. A., & Brody, C. J. (2005). Contradictions of commitment in the new economy: Insecurity, time, and technology. Social Science Research, 34, 843–861.

Poletaev, M., & Robinson, C. (2008). Human capital specificity: Evidence from the dictionary of occupational titles and displaced worker surveys, 1984–2000. Journal of Labor Economics, 26, 387–420.

Rudolph, C. W., Wells, C. L., Weller, M. D., & Baltes, B. B. (2009). A meta-analysis of empirical studies of weightbased bias in the workplace. Journal of Vocational Behavior, 74, 1–10.

Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24, 1–24.

Russo, N. F. (1985). Sex-role stereotyping, socialization, and sexism. In A. G. Sargent (Ed.), Beyond sex roles (pp. 150–187). St. Paul, MN: West Publishing.

110

Career Issues

Ryan, M. K., & Haslam, S. A. (2007). The glass cliff: Exploring the dynamics surrounding the appointment of women to precarious leadership positions. Academy of Management Review, 32, 549–572. Saint-Charles, J., Mongeau, P. (2009). Different relationships for coping with ambiguity and uncertainty in organizations. Social Networks, 31, 33–39.

Copyright American Psychological Association. Not for further distribution.

Salgado, J. F., Anderson, N., Moscoso, S., Bertua, C., de Fruyt, F., & Rolland, J. P. (2003). A meta-analytic study of general mental ability validity for different occupations in the European Community. Journal of Applied Psychology, 88, 1068–1081. Schaie, K. W. (2005). Developmental influences on adult intelligence: The Seattle longitudinal study. New York: Cambridge University Press. Schein, E. H. (1978). Career dynamics: Matching individual and organizational needs. Reading, MA: Addison Wesley. Scholarios, D., Van der Heijden, B. I. J. M., Van der Schoot, E., Bozionelos, N., Epitropaki, O., Jedrzejowicz, P., et al. (2008). Employability and the psychological contract in European ICT sector SMEs. International Journal of Human Resource Management, 19, 1035–1055. Schwartzman, A. E., Gold, D., Andres, D., Arbuckle, T. Y., & Chaikelson, J. (1987). Stability of intelligence: A 40-year follow up. Canadian Journal of Psychology, 41, 244–256. Segers, J., Inceoglou, I., Vloeberghs, D., Bartram, D., & Henderickx, E. (2008). Protean and boundaryless careers: A study on potential motivators. Journal of Vocational Behavior, 73, 212–230. Seibert, S. E., Kraimer, M. L., & Liden, R. C. (2001). A social capital theory of career success. Academy of Management Journal, 44, 219–237. Shambaugh, D. (Ed.). (2006). Power shift: China and Asia’s new dynamics. Berkeley, CA: University of California Press. Shin, T-J. (2007). The impact of structural dynamics on job mobility rates in the United States. Social Science Research, 36, 1301–1327. Singh, V., Bains, D., & Vinnicombe, S. (2002). Informal mentoring as an organizational resource. Long Range Planning, 35, 389–405. Sippola, A., & Smale, A. (2007). The global integration of diversity management: A longitudinal case study. International Journal of Human Resource Management, 18, 1895–1916.

the limit? Research Highlights in the Demography and Economics of Aging, 8, 1–4. Sonnenfeld, J. A., Peiperl, M. A., & Kotter, J. P. (1988). Strategic determinants of managerial labor markets: A career systems view. Human Resource Management, 27, 369–388. Spilerman, S. (1977). Careers, labor market structure, and socioeconomic achievement. American Journal of Sociology, 83, 551–593. Stackman, R. W., & Pinder, C. C. (1999). Context and sex effects on personal work networks. Journal of Social and Personal Relationships, 16, 39–64. Stahl, G. K., Chua, C. H., Caligiuri, P., Cerdin, J-L., & Taniguchi, M. (2009). Predictors of turnover intentions in learning-driven and demand-driven international assignments: The role of repatriation concerns, satisfaction with company support, and perceived career advancement opportunities. Human Resource Management, 48, 89–109. Stahl, G. K., Miller, E. L., & Tung, R. L. (2002). Toward the boundaryless career: A closer look at the expatriate career concept and the perceived implications of an international assignment. Journal of World Business, 37, 216–227. Sullivan, S. E. (1999). The changing nature of careers: A review and research agenda. Journal of Management, 25, 457–484. Sullivan, S. E., & Arthur, M. B. (2006). The evolution of the boundaryless career concept: Examining physical and psychological mobility. Journal of Vocational Behavior, 69, 19–29. Super, D. E. (1957). The psychology of careers. New York: Harper & Row. Super, D. E. (1980). A life-span, life-space approach to career development. Journal of Vocational Behavior, 16, 282–298. Super, D. E. (1990). A life-span, life-space approach to career development. In D. Brown & L. Brooks (Eds.), Career choice and development (2nd ed., pp. 197–261). San Francisco: Jossey-Bass. Super, D. E., Savickas, M. L., & Super, C. M. (1996). The life-span, life-space approach to careers. In D. Brown & L. Brooks (Eds.), Career choice and development (3rd ed., pp. 121–178). San Francisco: Jossey-Bass. Suutari, V., & Brewster, C. (2000). Making their own way: International experience through self-initiated foreign assignments. Journal of World Business, 35, 417–436.

Smart, R., & Peterson, C. (1997). Super’s career stages and the decision to change careers. Journal of Vocational Behavior, 51, 358–374.

Suutari, V., & Brewster, C. (2003). Repatriation: Empirical evidence from a longitudinal study of careers and expectations among Finnish expatriates. International Journal of Human Resource Management, 14, 1132–1151.

Sonnega, A. (2006, March). The future of human life expectancy: Have we reached the ceiling or is the sky

Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel & 111

Baruch and Bozionelos

Copyright American Psychological Association. Not for further distribution.

W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24). Chicago: Nelson-Hall.

Trompenaars, F. (1994). Riding the waves of culture. New York: Irwin.

Takeuchi, R., Tesluk, P. E., & Marinova, S. V. (2006). Role of international experiences in the development of cultural intelligence. In S. Dey & V. N. Posa (Eds.), Cultural intelligence: An introduction (pp. 56–91). Hyderabad, India: ICFAI University Press.

Tsai, C-F., Wu, S-L., Wang, H-K., & Huang, I-C. (2006). An empirical research on the institutional theory of downsizing: Evidence from MNC’s subsidiary companies in Taiwan. Total Quality Management, 17, 633–654.

Tams, S., & Arthur, M. B. (2007). Studying careers across cultures: Distinguishing international, cross-cultural, and globalization perspectives. Career Development International, 12, 86–98.

Tsui, A. S., Egan, T. D., & O’Reilly, C. A., III. (1992). Being different: Relational demography and organizational attachment. Administrative Science Quarterly, 37, 549–579.

Tarique, I., & Schuler, R. (2008). Emerging issues and challenges in global staffing: A North American perspective. The International Journal of Human Resource Management, 19, 1397–1415.

Tsui, A. S., & O’Reilly, C. A., III. (1989). Beyond simple demographic effects: The importance of relational demography in superior-subordinate dyads. Academy of Management Journal, 32, 402–423.

Tarique, I., Schuler, R., & Gong, Y. (2006). A model of multinational enterprise subsidiary staffing composition. The International Journal of Human Resource Management, 17, 207–224.

Turner, R. J. (1960). Sponsored and contest mobility and the school system. American Sociological Review, 25, 855–867.

Taylor, M. S., Audia, G., & Gupta, A. K. (1996). The effect of lengthening job tenure on managers’ organizational commitment and turnover. Organization Science, 7, 632–648. Tepper, K., Shaffer, B. C., & Tepper, B. J. (1996). Latent structure of mentoring function scales. Educational and Psychological Measurement, 56, 848–857. Tharenou, P. (1997). Managerial career advancement. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (Vol. 12, pp. 39–93). New York: Wiley. Thatcher, S. M. B., & Zhu, X. M. (2006). Changing identities in a changing workplace: Identification, identity enactment, self-verification, and telecommuting. Academy of Management Review, 31, 1076–1088. Thomas, D. C., Lazarova, M. B., Inkson, K. (2005). Global careers: New phenomenon or new perspectives? Journal of World Business, 40, 340–347. Thompson, J. A. (2005). Proactive personality and job performance: A social capital perspective. Journal of Applied Psychology, 90, 1011–1017. Thornhill, A., & Saunders, N. K. (1998). The meanings, consequences, and implications of the management of downsizing and redundancy: A review. Personnel Review, 27, 271–295. Tracey, T. J., & Rounds, J. (1993). Evaluating Holland’s and Gati’s vocational-interest models: A structural meta-analysis. Psychological Bulletin, 113, 229–246. Tracey, J. B., Sturman, M. C., & Tews, M. J. (2007). Ability versus personality. Cornell Hotel and Restaurant Administration Quarterly, 48, 313–322. Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO: Westview Press. 112

United Nations. (2006). Life expectancy at birth, both sexes combined (years). Retrieved January 23, 2009, from http://data.un.org/Data.aspx?d=PopDiv&f=variable ID%3a68#PopDiv United Nations. (2007). Women’s share of labour force. Retrieved January 23, 2009, from http://data.un.org/ Data.aspx?q=labour&d=GenderStat&f=inID%3a107 United Nations. (2008). Average annual rate of reduction of total fertility rate. Retrieved January 23, 2009, from http://data.un.org/Data.aspx?q=fertility+rate&d= SOWC&f=inID%3a80#SOWC U.S. Department of Labor. (2008, October). Highlights of women’s earnings in 2007 (Report 1008). Retrieved February 27, 2009, from http://www.bls.gov/cps/ cpswom2007.pdf Valcour, M., & Ladge, J. J. (2008). Family and career path characteristics as predictors of women’s objective and subjective career success: Integrating traditional and protean career explanations. Journal of Vocational Behavior, 73, 300–309. Vance, C. M. (2005). The personal quest for building global competence: A taxonomy of self-initiating career path strategies for gaining business experience abroad. Journal of World Business, 40, 374–385. Van der Heijde, C. M., & Van der Heijden, B. I. J. M. (2006). A competence-based and multidimensional operationalization and measurement of employability. Human Resource Management, 45, 449–476. Van Emmerik, I. J. H. (2006). Gender differences in the creation of different types of social capital: A multilevel study. Social Networks, 28, 24–37. Van Maanen, J. (1977) Towards a theory of career. In J. Van Maanen (Ed.), Organizational careers: Some new perspectives (pp. 161–179). New York: Wiley.

Career Issues

Van Scotter, J. R., Motowidlo, S. J., & Cross, T. C. (2000). Effects of task performance and contextual performance on systemic rewards. Journal of Applied Psychology, 85, 526–535. Vigoda, E. (2000). The relationship between organizational politics, job attitudes, and work outcomes: Exploration and implications for the public sector, Journal of Vocational Behavior, 57, 326–47.

Copyright American Psychological Association. Not for further distribution.

Warhurst, C., Nickson, D., Witz, A., & Cullen, A. M. (2000). Aesthetic labour in interactive service work: Some case study evidence from the “new” Glasgow. The Service Industries Journal, 20(3), 1–18.

Whyte, W. H. (1956). The organization man. New York: Simon & Shuster. Wilensky, H. L. (1961). Careers, lifestyles, and social integration. International Social Science Journal, 12, 553–558. Wilson-Kovacs, D., Ryan, M. K., Haslam, S. A., & Rabinovich, A. (2008). “Just because you can get a wheelchair in the building doesn’t necessarily mean that you can still participate”: Barriers to the career advancement of disabled professionals. Disability and Society, 23, 705–717.

Waterman, R. H., Jr., Waterman, J. A., & Collard, B. A. (1994). Toward a career-resilient workforce. Harvard Business Review, 72(4), 87–95.

Wolff, H-G., & Moser, K. (2006). Entwicklung und validierung einer networkingskala [Development and validation of a networking scale]. Diagnostica, 52, 161–180.

Watson, M., & McMahon, M. (2005). Children’s career development: A research review from a learning perspective. Journal of Vocational Behavior, 67, 119–132.

Wolff, H-G., & Moser, K. (2009). Effects of networking on career success: A longitudinal study. Journal of Applied Psychology, 94, 196–206.

Weichselbaumer, D., & Winter-Ebmer, R. (2007). The effects of competition and equal treatment laws on gender wage differentials. Economic Policy, 50, 235–287.

WorldatWork. (2009, February 18). Telework revs up as more employers offer work flexibility. Retrieved February 19, 2009, from http://www.world atwork.org/waw/adimComment?id=31331

Westermeyer, J. F. (2004). Predictors and characteristics of Erikson’s life cycle among men: A 32-year longitudinal study. International Journal of Aging and Human Development, 58, 29–48.

World Bank. (2009). World development report 2009: Reshaping economic geography. Washington, DC: The World Bank.

White, L. T., III (1989). Policies of chaos: The organizational causes of violence in China’s cultural revolution. Princeton, NJ: Princeton University Press.

Zhao, H., Wayne, S. J., Glibkowski, B. C., & Bravo, J. (2007). The impact of psychological contract breach on work-related outcomes: A meta-analysis. Personnel Psychology, 60, 647–680.

113

CHAPTER 4

INDIVIDUAL DIFFERENCES: THEIR MEASUREMENT AND VALIDITY

Copyright American Psychological Association. Not for further distribution.

Oleksandr S. Chernyshenko, Stephen Stark, and Fritz Drasgow

Individual differences can be defined broadly as the dimensions along which people vary. There are a great many individual difference dimensions, including how straight one can hit a golf ball, how much one likes to organize his or her work desk, or how fast one can read and comprehend a word passage. In domains of differential and industrial and organizational (I/O) psychology, however, the term individual differences is usually reserved to describe individuals’ basic tendencies, capacities, and dispositions that influence the observed range and frequency of their behavior (Motowidlo, Borman, & Schmit, 1997). Traditionally, individual difference variables include cognitive and psychomotor abilities or skills, personality traits, motives, values, and interests. Although researchers may still debate the extent to which these tendencies are inherited and shaped by environmental factors early in life, the general consensus is that most individual difference variables are relatively stable throughout adulthood and, in the absence of strong situational constraints, play a significant role in guiding behavior. Individual differences are an essential part of the person–situation interaction. Whereas the fields of experimental psychology and organizational behavior are devoted to the study of how situations influence behavior, differential psychology is devoted to the study of how individual differences influence behavior and performance. Mischel (1968) notwithstanding, both situations and individual differences are important and account for roughly equal amounts of variance (Funder & Ozer, 1983). In their widely cited article, Schmidt and Hunter (1998) provided

estimates of correlations between job performance and various individual difference measures under optimal conditions (i.e., when job performance is measured without error and there is no restriction of range on the individual difference variable). Their Table 1 shows that cognitive ability can be expected to correlate in excess of .50 with performance and several other measures have correlations greater than .35. Importantly, in selection contexts, these criterion correlations are essentially effect size measures, because they directly indicate the expected performance improvement (in z score units) when the average applicants’ scores on the predictor increase by one z score unit (or 1 SD). For example, a .50 correlation means that, on average, selection of applicants who are 1 SD higher on the predictor would result in a .50 SD improvement in performance. Clearly, individual differences are powerful predictors of behavior and performance in work organizations. Interestingly, the person-by-situation interaction has been studied far less often and frequently accounts for much less variance than the main effects of people and situations. The relatively weak support for the person-by-situation interaction is one of the main reasons that I/O psychologists have shown an acute interest in using individual difference variables in theories of work behavior. To illustrate, consider Figure 4.1, in which we present a conceptual model for the role of individual differences in predicting job performance. Here, we have combined the ideas of Hunter (1983); Ackerman, Kanfer, and Goff (1995); Campbell (1990); McCrae

http://dx.doi.org/10.1037/12170-004 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

117

Chernyshenko, Stark, and Drasgow

Individual Differences (Dispositions and Personal History):

Cognitive and Psychomotor Abilities, Personality, Motives, Values, and Interests, Narratives . . .

Mediators: Training, Satisfaction, Self-Efficacy, Organizational Characteristics . . .

Performance Determinants (Characteristic Adaptations): Declarative Knowledge, Procedural Skills, Motivation . . .

Job Performance: Task Performance, Citizenship, Counterproductivity, Leadership Effectiveness, Turnover, Cultural Adjustment . . .

Copyright American Psychological Association. Not for further distribution.

FIGURE 4.1. Conceptual model of the role of individual differences in predicting job performance.

and Costa (1996); Motowidlo et al. (1997); Chan and Drasgow (2001); Credé, Chernyshenko, Stark, Dalal, and Bashur (2007); and many others, to posit how the main classes of individual differences (i.e., abilities, personality, values) exert a direct or indirect influence on performance determinants (aka characteristic adaptations), which, in turn, directly influence employee performance. Of course, the strength of the path coefficients and the relevant mediators and outcomes will vary depending on whether one looks at cognitive ability, personality, or values, but the basic theoretical mechanisms are the same. For example, in the case of cognitive ability, researchers have shown that the relationship between general cognitive ability and task performance is partially or fully mediated by declarative knowledge, procedural skills, motivational factors, or all three (Ackerman et al., 1995; Hunter 1983; McCloy, Campbell, & Cudeck, 1994). Additionally, Ackerman et al. (1995) found that ability self-concept (aka self-efficacy) partially mediates the relationship between general cognitive ability and motivation. Chan and Drasgow (2001) showed that relationships between several personality variables and motivation to lead, a performance determinant in their model, were partially mediated by leadership experiences and self-efficacy, whereas motivation to lead variables predicted leadership potential ratings obtained from both assessment centers and 360-degree feedback. Arthur, Bell, Villado, and Doverspike (2006) also showed that the relationship between person–organization (P-O) fit variables and job performance were partially mediated by turnover intentions and organizational commitment, both of which can be seen as tapping motivational aspects of performance determinants. These and many other studies have illustrated how indi118

vidual differences exert their influence on employee performance. From the personnel selection perspective, these research findings are particularly useful because many mediator variables (training, experience, organizational characteristics) tend to be uniformly applied to all organizational members (i.e., every newcomer receives the same training course); so, in the absence of random work events or moderators, the rank order of individuals on many important performance dimensions is determined by their rank order on distal individual difference variables. The strength of direct effects may vary depending on which individual difference and performance variables are used, but many effects are appreciable enough to make a real difference to an organization’s bottom line. CHAPTER OUTLINE In this chapter, we focus primarily on three broad domains of individual differences: (a) cognitive and psychomotor abilities, (b) personality, and (c) motives, values, and interests. This grouping follows B. W. Roberts’s (2006) neosocioanalytic model of personality, which also includes narratives in the three classes of individual differences, a concept akin to what I/O psychologists call personal history or biographical data. Further, although narratives are clearly acquired throughout one’s life, the endurance of personal histories and their role in influencing subsequent behavior places them within the individual differences domain, rather than within situational mediators or immediate performance determinants. Recognizing that a detailed account of research in each of these domains is beyond the scope of this

Copyright American Psychological Association. Not for further distribution.

Individual Differences

chapter, we only briefly review the predominant taxonomies and highlight the similarities and differences of various classification systems. We focus instead on approaches to measuring individual differences and using the test scores for selection and classification. Key findings related to validity, adverse impact, measurement bias, and faking are reviewed, as are emerging issues connected to Web-based testing and computerized adaptive testing (CAT), which is gradually propagating from large educational testing programs into I/O settings. The advantages and challenges of implementing and maintaining this new technology are discussed using examples from the noncognitive domain, in recognition of a broadened perspective on the performance criterion that now includes factors such as citizenship performance, counterproductivity, organizational adjustment, and adaptability. We address personality assessment, in particular, because there are so many unresolved psychometric and substantive questions and a growing body of evidence suggesting that personality factors predict performance and retention well beyond general cognitive ability and with far less adverse impact. (See also chap. 5, this volume.) Because each individual difference domain can be seen as organized hierarchically from broad to narrow (B. W. Roberts, 2006), we begin each section with a brief overview of existing trait taxonomies, and we then discuss salient issues in measurement and scoring that may be domain specific. We begin with cognitive ability and then move to personality and vocational interests. In the cognitive ability domain, the broadest construct is, of course, general cognitive ability, often referred to as g. The medium level of specificity includes fluid and crystallized intelligences, Gf and Gc, which subsume narrower domains such as word knowledge, reading comprehension, mathematical reasoning, inductive reasoning, and deductive reasoning. We start with Spearman’s original (1904) model of intelligence, which hypothesized g and a specific ability for each domain. We then contrast Thurstone’s (1938) findings of correlated primary abilities but no general ability. The resolution of the debate between these two camps is described via Schmid and Leiman’s (1957) transformation and

its modern alternative, the bifactor model. The most widely accepted current model of intelligence, Carroll’s (1993) three-stratum theory, is then described. We conclude this section with a review of issues surrounding the use of cognitive ability test scores for job selection and licensure and certification. In the personality domain, there is still considerable debate concerning the structure of personality at both the broad (global) factor and narrow (lower order) factor levels. We therefore review the most predominant higher order (global) factor trait taxonomy, the Big Five (Costa & McCrae, 1988; Goldberg, 1990, 1993) and then explore Ashton et al.’s (2004) HEXACO model, and the seven-factor models proposed by Almagor, Tellegen, and Waller (1995) and Saucier (2003). We then focus on lower order factors, which are sometimes referred to as facets, and discuss reasons why the measurement of facets is becoming more common in organizational research and practice. Finally, we draw attention to the numerous challenges associated with the use of personality measures in applied settings. Specifically, we discuss new psychometric models for scaling personality statements, formats designed to reduce response distortion, and how advances in measurement technology are being used to customize and increase the efficiency of personality assessment. The third broad domain of individual differences is values. B. W. Roberts (2006) defined this construct as consisting of “all of the things that people feel are desirable—that is, what people want to do or would like to have in their life” (p. 7). As an illustration, we review Schwartz’s (1992) model of human values and discuss how this and other models are used in the rapidly growing research area of P-O fit. We place vocational interests at the medium level of generality in the values domain and describe Holland’s (1985) RIASEC model (realistic [R], investigative [I], artistic [A], social [S], enterprising [E], and conventional [C]). We then discuss some alternative ways of representing vocational interests (Tracey & Rounds, 1996) and the potential ramifications for using interest scores for vocational counseling, selection, and classification. (See also Vol. 3, chaps. 1 and 4, this handbook.) 119

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

The final trait category in the neo-socioanalytic model is narratives, that is, the stories people use to understand and explain themselves, their environments, and the actions of others. B. W. Roberts (2006) described narratives as helping people “create meaning and purpose in their lives and predictability in the events they observe and experience, and provide explanations of how people have come to be in their present circumstances” (p. 9). In a broad sense, information provided by narratives is already used in many selection systems. Information in personal statements, letters of recommendation, and unstructured interviews is often examined qualitatively and is used to make or support selection and promotion decisions. When this information is collected systematically using biodata forms or structured interviews, it is more readily quantified and has potentially higher validity. This section summarizes important research findings involving narrative data and highlights some unresolved questions about measurement and verification. COGNITIVE ABILITY The modern study of intelligence began with Spearman’s seminal article in 1904 in which he introduced his “two-factor” theory of intelligence. Spearman labeled the first factor g, or general ability, which he assumed influenced performance in all domains. The second factor actually refers to a category, rather than a particular ability. Spearman hypothesized that performance in any domain was influenced by g as well as a specific factor that was unique to the domain. Thus, performance across domains was correlated because of the shared influence of g, but the correlation was less than perfect because each domain was influenced by an independent specific factor. Performance in physics, for example, is influenced by g and a specific ability in physics. If data from a set of tests administered to a broad sample of individuals are factor analyzed, the first factor ordinarily provides a surprisingly good fit (see Drasgow, 2003, p. 110). Nonetheless, clear failures of the model can be seen in terms of large residuals unless the set of tests to be analyzed is carefully preselected. This led Spearman’s students and junior 120

colleagues to develop hierarchical models. These models have g at the apex as a single, broad trait that affects performance in all areas. Vernon (1950), for example, included two middle-level traits: v:ed, which refers to verbal–educational ability and was used to explain relations among reading comprehension, logical reasoning, and arithmetic reasoning, and k:m, which refers to spatial–mechanical ability. Hierarchical models with g influencing performance on all tests and midlevel traits affecting clusters of tests provide better fits to data than Spearman’s original model. Thurstone (1938) took a different approach to the study of human intelligence. In a landmark study, 218 University of Chicago students completed a battery of 56 tests. Thurstone extracted a dozen factors, and, after rotation, seven primary mental abilities were identified: spatial, perceptual, numerical, verbal relations, word fluency, memory, and inductive reasoning. To analyze his data, Thurstone had developed multiple factor analysis (Thurstone, 1947), with correlated factors defined by rotation to simple structure. Using his new methods, Thurstone did not find a general factor, only multiple correlated factors. This led to an acrimonious debate between Thurstone and the American group, on the one hand, and Spearman and the British group, on the other, concerning the existence of the general factor g. This debate was resolved in 1957 when Schmid and Leiman showed the mathematical equivalence of models with correlated factors and higher order factor models with second-order general factors and orthogonal first-order factors. Conceptually, the Schmid and Leiman analysis can be seen as subjecting the factor correlation matrices obtained by Thurstone and others for batteries of ability tests to factor analysis. The resulting factor or factors are termed second-order factors. When the factor correlation matrix obtained from a battery of tests is subjected to this analysis, a single, strong second-order factor is often obtained. Carroll (1993) described the results of a remarkably extensive literature review and reanalysis of factor analytic studies of cognitive ability. He identified 461 correlation matrices suitable for reanalysis (i.e., data from a broad sample were collected, at least three measures were included for each factor

Copyright American Psychological Association. Not for further distribution.

Individual Differences

hypothesized, and a reasonably large number of factors were included). Carroll’s findings were summarized in his three stratum model of intelligence, which is shown in Figure 4.2. In the three-stratum model, as in Spearman’s two-factor model and Vernon’s hierarchical model, g sits at the apex and explains the most variance. Carroll found many second-level factors (Figure 4.2 contains eight of the most important), including fluid intelligence Gf and crystallized intelligence Gc. The length of the lines from g to each of the second stratum abilities is inversely related to the strength of the relationship, so that Gf and Gc are the most strongly related to g. Stratum I factors lie beneath the second stratum and refer to narrower traits. For example, deduction, induction, and quantitative reasoning are some examples of first stratum traits beneath Gf and verbal ability, reading comprehension, and lexical knowledge are examples of Stratum I traits beneath Gc. Yung, Thissen, and McLeod (1999) explored the relation between Schmid and Leiman’s (1957) higher order factor model and the increasingly popular hierarchical factor model, which includes the bifactor model as a special case. In the bifactor model, every test is allowed to load on a general factor and then sets of tests load on additional factors. Each test is allowed to load on exactly two factors (the

general factor and one additional specific factor) and all factors are constrained to be uncorrelated. Yung et al. showed that the Schmid-Leiman higher order factor model is equivalent to the hierarchical factor model with additional constraints imposed. For example, they showed that the nonzero factor loadings of columns 1 and 2 of the factor loading matrix in their Table 1 have a 4:3 ratio. Here, Test 1 has a loading of .512 on Factor 1, so it has a loading of .384 (= .512 × 3⁄4) on Factor 2 and Test 2 has a loading of .576 on Factor 1 and therefore a loading of .432 (= .576 × 3⁄4) on Factor 2. There does not appear to be any substantive meaning to these proportionality constraints; they are simply a result of the “factor analyze a factor correlation matrix” strategy taken by Schmid and Leiman. Thus, with modern structural equations software, the more general hierarchical model seems preferable to higher order factor models. It has fewer constraints and therefore should fit even better than the models Carroll (1993) fit to find support for his three-stratum model. Today, there appears to be a scientific consensus that human cognitive abilities are organized hierarchically. At the top of the hierarchy is the general factor; at the level below are fluid and crystallized intelligence factors (also sometimes represented as verbal, quantitative, and abstract reasoning factors) and several memory, perception, and cognitive speed

Stratum III g

Stratum II

Gf

Gc

memory

visual auditory perception perception

retrieval

cognitive speed

processing speed

Stratum I: First-order common factors

FIGURE 4.2. Carroll’s three-stratum model. From Handbook of Psychology: Industrial and Organizational Psychology (p. 115), by W. C. Borman and D. R. Ilger (Eds.), 2003, New York: Wiley. Reprinted with permission. 121

factors; finally, at the third level are narrow factors representing fairly specific cognitive performances associated with higher order factors. This hierarchy is useful because it provides ways to account for variation among existing cognitive ability tests and performances and often serves as a basis for future test development efforts. It also allows researchers and practitioners to appreciate fully the strength of the general factor in terms of predicting both organizational and personal level outcomes. As many research studies have shown, unless the criterion being measured is very specific, there is usually little incremental validity beyond the general cognitive ability factor in predicting job performance (Ree, Earles, & Teachout, 1994) and training performance (Ree & Earles, 1991), and so a single score from a general intelligence measure is usually sufficient to make most selection decisions.

Assessment: Dominance Models: Classical Test Theory and Item Response Theory Reliable and valid assessment of cognitive abilities and intellectual achievement is immensely important in many fields of psychology. Regardless of the psychometric framework within which one works, most models for cognitive ability data assume a dominance (Coombs, 1964) relationship between trait level and the probability of obtaining a correct

response or a particular test score. Specifically, the more of the trait one possesses, the better the examinee’s chance of answering the item correctly or earning a high test score. Classical test theory (CTT) and linear factor analysis assume the relationship between trait level and expected score is linear, whereas item response theory models typically allow for nonlinear but monotonically increasing relationships, making them particularly useful for binary data. If we were to plot the probability of observing a correct response as a function of examinee ability level for any item in a cognitive ability scale, the two psychometric frameworks would produce trace lines, known as item response functions (IRFs), similar to those presented in Figure 4.3. The straight line curve represents CTT and linear factor analytic models. The monotonically increasing S-shaped curve is characteristic of item response theory (IRT) models such as the three-parameter logistic model (3PL; Birnbaum, 1968), which posits that each item has a discrimination parameter, a difficulty parameter, and lower asymptote (aka guessing) parameter (for details, see Hulin, Drasgow, & Parsons, 1983). Note that unlike the linear IRF, the 3PL IRF does not assume that the item is equally discriminating across all trait levels, as indicated by the flattening of the slope at the high and low ends of the trait continuum. In fact, the response probabilities show rapid change over the

1.0 Probability of Correct Response

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

0.8

0.6

0.4

0.2

0.0 -3

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

Ability Level

FIGURE 4.3. Example item response functions for linear and nonlinear dominance models. 122

3

Copyright American Psychological Association. Not for further distribution.

Individual Differences

interval +.20 to +1.20, which indicates that the item measures best for examinees located in that trait range and provides relatively little information about examinees located beyond. That nonlinear models can be used to examine how item information changes across the trait continuum makes them particularly suitable as the basis for CAT. In essence, CAT algorithms seek to identify and administer items or subsets of items that are most informative for each examinee, thus tailoring tests to increase score accuracy and reduce measurement error with the fewest number of items. The simplest example would be a test of mathematical achievement. In IRT, items involving simple algebraic operations are typically “easy” and would allow differentiation between examinees with very low or low ability scores (e.g., third graders vs. fifth graders), whereas items involving differential calculus are “difficult” and thus would be administered to differentiate among high-ability individuals. It would make little sense, however, to give calculus items to third graders or simple algebra items to examinees with degrees in statistics, because neither item group would differentiate within the inappropriate ability group. Linear models would not make sense here because they assume that an item is equally informative for all examinees, so the same set of items would be seen as appropriate for all examinees. Classical test theory methods treat all items as equally important, and for that reason total score is often a simple sum of item responses; factor scores from linear factor analysis might weight items by their factor loadings, but the value of an item does not change across examinees with different standings on the latent trait. A wide variety of cognitive ability measures are used in organizational settings. Because assessment time is often at a premium, most tests are relatively short and focus mainly on general cognitive ability. Perhaps the three best examples are the Wonderlic Personnel Test (WPT; Wonderlic Personnel Test, Inc., 1992), Raven’s Progressive Matrices (Raven, Raven, & Court, 1998), and the Watson–Glaser Test of Critical Reasoning (Watson & Glaser, 1980). The WPT takes only 12 minutes to administer; contains a combination of vocabulary, arithmetic reasoning, and spatial reasoning items (Grubb, Whetzel, & McDaniel, 2004); and yields a single score that

reflects an examinee’s general cognitive ability. Several parallel forms of the WPT test are available to allow for retesting. The Raven’s test, in contrast, consists exclusively of abstract reasoning items, which increase in difficulty as the test progresses. There are standard and advanced forms, both of which have multiple parallel versions and, like the Wonderlic, the Raven’s yields a single score reflecting general cognitive ability, although arguably it assesses fluid intelligence. The Watson–Glaser Test contains 80 reading passages aimed at assessing several critical reasoning skills (i.e., drawing inferences, recognizing assumptions) that are mostly part of the crystallized intelligence factor. The test is designed to be fairly difficult and thus is used mainly to assess applicants’ managerial potential. Parallel and shorter forms of the test are also available to allow retesting or to accommodate limited testing times. Manuals for these three tests, as well as others, usually provide extensive information with regard to their construct and predictive validity, availability of norms, and test administration procedures. Tests such as these are usually constructed on the basis of CTT principles and are scored simply as “number correct.” Most are available in both paperand-pencil and computerized formats, although the latter are typically just electronic “page turners” (Drasgow & Olson-Buchanan, 1999) that present the same items in the same order as the paper-and-pencil forms. The continuing reliance on CTT does not necessarily mean that test developers are unaware of IRT and its benefits. Rather, it is a testament to the astounding success of such relatively straightforward measures for selecting the best performing employees. As long as a test contains a mix of easy, medium, and difficult items, the resulting CTT-based scores should provide a fairly good rank order of examinees in terms of their cognitive abilities, so any gains in precision that might be achieved by implementing more sophisticated and costly methods of item selection and scoring might be difficult to justify. On the other side of the cognitive ability testing spectrum are standardized selection and classification tests. These measures take full advantage of the factor analytic research on the latent structure of cognitive ability and often use IRT techniques to evaluate and administer items and score examinees. 123

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

The most prominent example is the Armed Services Vocational Aptitude Battery (ASVAB), which is used to select and classify hundreds of thousands of recruits into hundreds of military occupational specialties (Campbell & Knapp, 2001). This is accomplished, in part, by using nine subtests (General Science, Arithmetic Reasoning, Word Knowledge, Paragraph Comprehension, Auto and Shop Information, Mathematics Knowledge, Mechanical Comprehension, Electronics Information, and Assembling Objects) to measure an array of specific skills, knowledge, and abilities rather than just one or a few broad dimensions. The primary difference with respect to general cognitive ability tests, such as those just described, is that the ASVAB has a strong knowledge component (e.g., Electronics Information, Auto and Shop Information) that enhances its usefulness for predicting performance in technical jobs. The second difference is that rather than yielding only one ability score, the ASVAB yields nine subtest scores, which can be combined to form many composites for classification decisions. Today, the ASVAB is administered in both paper-andpencil and computerized formats. The computerized version (i.e., CAT-ASVAB; Segall & Moreno, 1999) selects items adaptively and thus provides high measurement precision with fewer items. Scores on various “forms” are then equated using statistical methods commonly applied in large-scale educational testing programs (see Kolen & Brennan, 2004).

Using Cognitive Ability Measures in Selection: Validity, Effects of Range Restriction and Attenuation, and Adverse Impact The meta-analytic estimate of the correlation between general cognitive ability and job performance is .51, and the estimate of the correlation with training performance is .53 (Schmidt & Hunter, 1998). Moreover, job complexity appears to moderate the relationship between ability and performance, with the relationship being even stronger than .51 for highly complex jobs (Hunter, 1980; Schmidt & Hunter, 2004). Such findings are not limited to studies conducted in U.S. work settings. Salgado et al. (2003), for example, reported similar validities for job and training performance in European Union 124

Countries and also found that job complexity acted as a moderator. It is important to note that these validity estimates are based on meta-analyses involving hundreds if not thousands of studies. In any particular organization, however, it is unlikely that one would observe validity coefficients of such magnitude. This is because measurement artifacts such as criterion unreliability and predictor range restriction can severely reduce the observed correlations between predictors and criteria. Range restriction is particularly problematic when assessing the usefulness of a test because almost any organizational sample has been subjected to direct or indirect range restriction, self-selection effects, or both, which reduce the variance of trait scores and thus the correlations with criterion measures. To say it differently, because performance data can only be collected from those already selected into the organization, one can never know what the true validity of a predictor is, unless selection is at random or every applicant has been selected. The best known example of how range restriction can reduce observed correlation coefficients was given by Thorndike (1949). Because of a severe shortage of pilots during World War II, the U.S. military had to take everyone in a group of 1,036 volunteers, even though only 136 of these men would have ordinarily qualified for pilot training (i.e., received a passing score on a cognitive ability composite). The observed correlation between the cognitive composite and pilot training performance was .64 for the full sample but would have been only .18 if the selection process had followed the normal procedure (i.e., if only the 136 men with passing cognitive composite scores were actually selected). Unreliability of the criterion can also severely reduce (aka attenuate) observed predictor–criterion relationships because it is not uncommon for supervisor ratings to show reliabilities in only the .5 to .6 range (Viswesvaran, Ones, & Schmidt, 1996). This happens because unreliability increases the proportion of random variance contained in the criterion score, thus reducing the proportion of explainable variance. Consequently, correlation coefficients appear small, unless we remove or partial out the random variance component. For these reasons, various corrections for attenuation and

Copyright American Psychological Association. Not for further distribution.

Individual Differences

range restriction have been developed (Gulliksen, 1950) and are used to gauge more accurately the usefulness of a test for a specific group of applicants. The validity results just described notwithstanding, the use of cognitive ability tests for selection and promotion continue to raise concerns about fairness because of their impact on selection rates for members of protected groups, as defined by the Civil Rights Act of 1964. Criticisms of cognitive ability tests are fueled by differences in test score means across majority and minority groups, and the popular belief is that these differences result from measurement bias (i.e., a psychometric problem with the instruments). However, most studies suggest that these differences do not result from bias but rather from impact, which is defined as a difference in the underlying trait distributions at the time of testing (Stark, Chernyshenko, & Drasgow, 2004). Thus, a more fruitful path for future research might be to focus on the motivational and educational factors that influence test performance, rather than endlessly searching for fundamental flaws in current ability measures. Another avenue would be to augment selection batteries with measures of noncognitive constructs, such as personality or situational judgment tests, because they seem to provide incremental validity with minimal, if any, adverse impact (Lievens, Buyse, & Sackett, 2005; Olson-Buchanan et al., 1998). Internet or computer-based delivery of selection tests has grown rapidly in recent years as advances in computer technology have rapidly expanded the options available for psychological assessment. In fact, computerized assessments have many distinct advantages over paper-and-pencil tests, such as the introduction of new item types (e.g., video-based scenarios, in-basket simulations), immediacy of scoring and feedback, and improved test security. When used in conjunction with adaptive testing algorithms, computerized assessments provide increased testing accuracy and efficiency because items are selected to match an examinee’s skill level. Hence, adaptive tests do not contain items that are too easy or too difficult for a particular examinee. Proctored administration, such as provided by commercial computer-based testing organizations, can be obtained for a fairly hefty per hour of seat-time fee. Alternatively, many organizations have turned

to unproctored Internet testing in which job applicants can take assessments at times and places of their choosing. Of course, with unproctored testing, there is no assurance that the job applicant is the person who actually answered the test items. This has led to controversy (Tippins et al., 2006) about the efficacy and ethicality of this form of assessment. One suggestion is to require individuals passing the unproctored Internet test to take a proctored confirmation test before receiving a job offer (Segall, 2001); another option is to use computer video technology to capture an applicant’s image while he or she is taking the test for verification purposes. Research to date, however, has not found evidence of test compromise (Nye, Do, Drasgow, & Fine, 2008). Nonetheless, if applicants believe that the stakes are high enough, some will cheat. More work is needed to delineate the conditions under which applicants will, or will not, be tempted to increase their scores by seeking improper aid.

A Note About Physical and Psychomotor Abilities Tests of physical and psychomotor abilities are often administered together with various cognitive ability measures to select police officers, firefighters, military personnel, clerical workers, or professional athletes. In fact, motor coordination and finger dexterity subtests have long been a part of many civil service exams administered in the U.S. Employment Service (e.g., the General Aptitude Test Battery). However, as Carroll (1993) pointed out, most physical and psychomotor tests have low correlations with cognitive ability tests, and thus should be considered as part of a distinct domain of individual differences. Most commonly measured factors in this domain are the following: static strength, body equilibrium, reaction time, speed and coordination of limb– wrist–finger movements, finger–manual dexterity, arm–hand steadiness, and aim (Caroll, 1993; Fleishman, 1964; see also Hogan’s [1991] comprehensive review of the literature on physical abilities). In jobs in which excellent vision or hearings are required, various components of these sensory modes are also commonly measured. Although many physical and psychomotor factors appear to be relevant and even necessary to perform 125

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

certain jobs, the ways these factors have been measured and passing standards set have attracted substantial employment litigation (see Hogan & Quigley, 1986). For example, minimum height or weight requirements have often been arbitrarily set (perhaps as proxy measures of physical strength), resulting in a disproportional underselection of female applicants (Medows v. Ford Motor Co.) or Spanish Americans (Guardians Association v. Civil Service Commission). In the majority of such cases, the courts ruled in favor of the plaintiff and prohibited further use of such tests. Hence, organizations interested in using physical and psychomotor ability measures for selection and classification must ensure that these tests are nondiscriminatory or, if not possible, are job related and equally valid nondiscriminatory alternatives are unavailable. PERSONALITY

Broad and Narrow Taxonomic Structures The Big Five theory brought order to the somewhat chaotic research literature on the structure of personality. Before the Big Five, there was little agreement about the basic dimensions of normal personality, so many instruments and scales conceptualized the trait domain in their own unique ways (e.g., Gough, 1957). One consequence of this was that early reviews attempting to combine validities of various personality instruments found near zero correlations with important work outcomes (e.g., Guion & Gottier, 1965). These grim empirical findings were accompanied by the theoretical arguments of Mischel (1968, 1969, 1973) and his colleagues, who contended that the behavior of individuals was not sufficiently consistent across time and situations to allow valid predictions by means of personality measures (they cited the fact that observed validity coefficients for personality have rarely exceeded .30 and most are in the .10–.20 range). Fortunately, since the early 1990s, personality researchers have reached a consensus that the Big Five personality constructs, Extraversion, Agreeableness, Conscientiousness, Emotional Stability, and Openness to Experience, are sufficient to adequately describe the basic dimensions of normal personality (see Borkenau & Ostendorf, 1990; Costa & McCrae, 1988; Digman, 1990; Goldberg, 126

1990, 1993; McCrae & Costa, 1987). It is important that numerous studies have been conducted to map existing inventories onto the Big Five structure (e.g., Chernyshenko, Stark, & Chan, 2001; McCrae, Costa, & Piedmont, 1993), which facilitates integration of the vast empirical literature. Today, most personality test manuals include a section detailing how their scales and/or scale composites relate to the Big Five (e.g., Conn & Reike, 1994). Of course, as with any taxonomic model, there is some debate whether the proposed factors are indeed sufficient to describe basic personality. In their HEXACO model, Ashton et al. (2004) argued for a six-factor solution. Almagor, Tellegen, and Waller (1995) and Saucier (2003), in contrast, have proposed even more elaborate seven-factor structures. All these authors have suggested that one or two extra evaluative dimensions (e.g., positive evaluation, negative evaluation, honesty) are needed in addition to the Big Five (B. W. Roberts, 2006). Similarly, although the Big Five model has been fairly well replicated in many countries and cultures, some cross-cultural researchers have suggested that additional dimensions may be beneficial (e.g., F. M. Cheung et al., 2001). These disagreements are not surprising given that the researchers are attempting to partition a large and heterogeneous set of behaviors, feelings, and thoughts using factor analytic methods. The number of factors extracted depends on what variables are included in the analyses, their psychometric properties, and the degree of heterogeneity researchers are willing to tolerate in their solutions. The hierarchical nature of the personality domain effectively ensures that one could continue extracting factors as long as desired, but each successive iteration would reconfigure the available factor space into increasingly narrow parts. Although research into basic personality structures continues to be important, we want to draw attention to emerging research on narrower personality dimensions known as facets. Facets can be seen as more contextualized manifestations of broad personality factors (B. W. Roberts, 2006). For example, Conscientiousness can be seen as an overall tendency to exercise a certain degree of control over one’s internal or external environment and thus includes a range of behavioral patterns and thoughts such as industriousness, responsibility, orderliness, self-

Copyright American Psychological Association. Not for further distribution.

Individual Differences

control, virtue, and rule following (B. W. Roberts, Chernyshenko, Stark, & Goldberg, 2005). Order is particular to external features of one’s immediate environment and reflects the extent to which one organizes and plans things, whereas Industriousness, which refers to one’s drive to achieve goals and other outcomes, also reflects an attempt to control one’s environment, but the emphasis is on more distal or far-reaching contextual features. One reason facets are important is that employee performance behaviors occur in specific organizational contexts, such as power relations (e.g., leadership), social exchanges (e.g., citizenship), or evaluations of one’s costs and rewards (turnover). Hence, there is often a closer match between behaviors involved in a particular performance dimension and those pertaining to one or more narrow facets than with broad personality factors, which results in facets having higher predictive validities. Paunonen (1998) correlated several Big Five factor and facet measures with various behavioral criteria and concluded that “aggregating personality traits into their underlying personality factors could result in decreased predictive accuracy due to the loss of trait-specific but criterion-valid variance” (p. 538); other researchers have reached similar conclusions (e.g., Ashton, 1998; Mershon & Gorsuch, 1988). In the Ashton (1998) study, for example, the correlation between counterproductive work behaviors and scores on the Responsibility facet was −.40, but a correlation of only −.22 was obtained when a broad measure of Conscientiousness was used in place of the facet measure. From a theory-building perspective, using facets translates into stronger theory because more precise hypotheses about employee behavior can be generated and tested. For example, Moon (2001) found that it is the Order facet of Conscientiousness that drives early job performance because organized employees are often more effective in dealing with the overwhelming amount of information that accompanies new jobs. In the long run, however, it is the Industriousness facet of Conscientiousness that drives performance because people who score high on this facet tend to set more difficult goals and work harder to achieve them. Another important reason for studying facets is that it will help to clarify the conceptualization of

broader factors (Briggs, 1989; Saucier & Ostendorf, 1999). Currently, there are inconsistencies in the way broad factors are defined. For example, in the widely popular NEO Personality Inventory (NEO-PI) model, Costa and McCrae (1994) place the Warmth facet, which includes emotionally supportive behaviors and displays of unconditional positive regard for others, within their Extraversion factor, whereas the abridged Big Five dimensional circumplex model (AB5C; Hofstee, De Raad, & Goldberg, 1992) locates it within Agreeableness. Better clarity in broad factor definitions is particularly important for meta-analyses of personality–performance relationships because the misplacement of facets could bias validity estimates or be a source of disagreement among various research groups. Finally, from an applied perspective, narrow traits offer much higher fidelity in terms of personality description and thus enhance the diagnostic value of assessment. This is especially beneficial for respondents who fall in the middle of the distribution on a measure of a broad factor, because such scores can be obtained in many ways. Unlike extreme scores on a broad factor, which suggest that an individual is generally high or low on all subcomponents, middle scores can be attained by being average on all lower order components or high on some and low on others. A number of narrow trait taxonomies have been proposed to date. Among the most widely known are the 45-facet structure of the AB5C model (Hofstee, De Raad, & Goldberg, 1992) and the 30-facet structure of the NEO-PI (Costa, McCrae, & Dye, 1991). In both cases, the researchers used a combination of prior empirical studies, theoretical justifications, and intuitions to divide each Big Five factor into an equal number of facets. Another way to establish narrow trait taxonomies is to adopt a purely empirical stance and to conduct a series of factor analyses using examinees’ responses to a diverse array of personality indicators (e.g., adjectives, behavioral statements, or scales). A good example of such an effort is the 22-facet narrow-order taxonomy of the Tailored Adaptive Personality Assessment System (TAPAS; Stark, Drasgow, & Chernyshenko, 2008), which is currently being tried for use with the ASVAB to facilitate military personnel selection and classification decisions. The TAPAS taxonomy, shown in Table 4.1, combines the results of Saucier 127

Chernyshenko, Stark, and Drasgow

TABLE 4.1 Tailored Adaptive Personality Assessment System Facet Taxonomy Derived From Lexical and Questionnaire-Based Factor Analytic Studies Lower order facets Questionnaire-based

Copyright American Psychological Association. Not for further distribution.

studies

Known facet markers Lexical study

Adjectives

Existing scales

Big Five factor: Conscientiousness Ambitious, industrious, aimless, NEO Competence, NEO Achievement Striving, AB5C Purposefulness

Achievement

Industriousness

Order

Order

Organized, neat, disorganized, sloppy

AB5C Orderliness, NEO Order, 16PF Perfectionism, JPI Organization

Self-Control

Decisiveness

Decisive, firm, controlled, deliberate, inconsistent

AB5C Cautiousness, NEO Deliberation, MPQ Self-Control, HPI Impulse Control

Responsibility

Reliability

Dependable, responsible, prompt, unreliable

CPI Responsibility, CPI Achievement via Conformance, CPI Socialization, JPI Responsibility

Nondelinquency

Traditional, lawful, delinquent, law-abiding

MPQ Traditionalism, JPI Traditional, 16PF Rule Consciousness

Virtue

Honest, truthful, frank, honorable, deceitful,

CPI Good Impression, CPI Self-Control, HPI Virtuous

Big Five factor: Openness to Experience Intelligent, analytical, HPI Education, HPI Good Memory, CPI knowledgeable, Intellectual Efficiency, AB5C Intellect

Intellectual efficiency

Intellect

Ingenuity

Imagination

Creative, inventive, unimaginative

AB5C Ingenuity, HPI Generates Ideas, JPI Innovation

Curiosity

Perceptiveness

Curious, perceptive, insightful, unobservant,

16PF Sensitivity, HPI Curiosity, HPI Science Ability

Aesthetics

Aesthetic, artistic, musical, unsophisticated, unrefined

NEO Aesthetics, AB5C Reflection, MPQ Absorption, NEO Feelings, HPI Culture, AB5C Imagination, JPI Breadth

Tolerance

Tolerant, broadminded, biased

CPI Flexibility, NEO Values, JPI Tolerant

Depth

Introspective, reflective, philosophical, shallow

16PF Abstractness, AB5C Depth, AB5C Introspection, JPI Complexity

Big Five factor: Extraversion Assertive, direct, bold, weak, submissive, feeble

CPI Dominance, HPI Leadership, NEO Assertiveness, MPQ Social Potency

Dominance

Assertiveness

Sociability

Sociability

Sociable, gregarious, talkative, withdrawn

CPI Sociability, CPI Social Presence, 16PF Social Boldness, HPI No Social Anxiety

Excitement seeking

Unrestraint

Loud, loquacious, entertaining, dull, unexciting, shy

16PF Liveliness, NEO Excitement Seeking, HPI Exhibitionistic, HPI Likes Crowds, HPI Entertaining

Physical condition

Adventurousness

Active, physical, adventurous, outdoorsy

JPI Energy, NEO Activity

128

Individual Differences

TABLE 4.1 (Continued)

Lower order facets

Copyright American Psychological Association. Not for further distribution.

Questionnaire-based

Known facet markers

studies

Lexical study

Adjectives Existing scales Big Five factor: Agreeableness Affectionate, compassionate, AB5C Warmth,16PF Warmth, NEO Warmth, warm, cold, insensitive 16PF Self-Reliance, MPQ Social Closeness, NEO Positive Emotions

Warmth

Warmth-Affectionate

Generosity

Generosity

Charitable, helpful, generous, greedy, stingy, and selfish

CPI Femininity/Masculinity, 16PF Sensitivity, AB5C Sympathy, AB5C Tenderness, AB5C Understanding, AB5C Empathy

Cooperation

Gentleness

Agreeable, cordial, trusting, uncooperative, combative

AB5C Pleasantness, NEO Altruism, NEO Trust, HPI Easy to Live With

Modesty

Modest, humble, conceited, snobbish, egocentric Big Five factor: Emotional Stability 16PF Apprehensive, JPI anxiety, NEO Anxiety, Insecure, apprehensive, HPI Not Anxious, MPQ Stress Reaction, nervous, relaxed, certain JPI Cooperativeness,

No anxiety

Insecurity

Even temper

Irritability

Moody, hot-headed, calm, composed, temperamental

AB5C Calmness, AB5C Stability, AB5C Tranquility, NEO Hostility, HPI Even Tempered

Well-being

Emotionality

Happy, cheerful, optimistic, depressed, dejected

NEO Depression, AB5C Happiness, MPQ WellBeing, HPI No Guilt, 16PF Emotional Stability, CPI Well-Being

Note. Empty cells indicate that the dimensions do not appear in the lexical study or have equivalents. NEO = NEO Personality Inventory; AB5C = Big Five dimensional circumplex model; 16PF = Sixteen Personality Factor Questionnaire; JPI = Jackson Personality Inventory; MPQ = Multidimensional Personality Questionnaire; HPI = Hogan Personality Inventory; CPI = California Personality Inventory.

and Ostendorf’s (1999) analyses of 312 vectors of responses to 500 temperament adjectives with a series of factor analyses of questionnaire data from several major personality inventories (see B. W. Roberts et al., 2005, for the Conscientiousness factor). Importantly, lexical and questionnaire-based studies produced remarkably similar facets structures, suggesting a reasonably good convergence across methods and different samples. Minor differences were observed in the Openness to Experience representation, which was most likely due to the lack of adjectives describing one’s openness to nonintellectual stimuli (i.e., artistic, tolerant, or reflective) in the Saucier and Ostendorf (1999) study. There was also an Agreeableness facet, Modesty,

produced by the lexical study, that failed to emerge in the questionnaire-based study. Interestingly, one of the main departures of Ashton and Lee’s HEXACO model from the Big Five is the Honesty factor, which seems to be a combination of Modesty and Virtue and is a part of the Conscientiousness factor in the questionnaire-based study. Table 4.1 presents the TAPAS narrow facet taxonomy derived from lexical and questionnairebased studies, and, to facilitate interpretation of each facet, we list some adjectives and existing scale markers from such well-established inventories as the NEO-PI, the California Personality Inventory (CPI; Gough, 1957), the Sixteen Personality Factor Questionnaire (16PF; Conn & Rieke, 1994), the 129

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

Multidimensional Personality Questionnaire (MPQ; Tellegen,1982), the Jackson Personality Inventory ( JPI; Jackson, 1994), the Hogan Personality Inventory (HPI; Hogan & Hogan, 1992), and the AB5C scales from the International Personality Item Pool (AB5CIPIP; Goldberg, 1997). The table is organized around the Big Five factors, although this facet structure can be reconfigured to fit other broad factor taxonomies (i.e., the Virtue and Modesty facets can be combined into a separate Honesty factor). Alternatively, facets can be aggregated upward into somewhat broader dimensions, as was done by DeYoung, Quilty, and Peterson (2007).

Personality Assessment There is no shortage of personality inventories measuring both broad and narrow Big Five traits. Among the most widely used for selection are the NEO PI, 16PF, HPI, and CPI. All these inventories were developed using classical test theory methods and comprise scales with 10 to 15 items that ask respondents about their typical behavior. The number of response categories varies for different inventories, but all require negatively worded items to be reversed scored before the total score computations. Further, although many of these measures have served well in research contexts, it is important to remember that they were not designed with high-stakes selection applications in mind (by “high stakes,” we mean that personality scores are actually used to make decisions that carry significant monetary gains for an applicant). In a research context, the main goal is to examine patterns of covariation among variables, so a short scale with easy-to-understand item content and a straightforward response format is more than sufficient to provide a rough rank ordering of participants on the trait continua (as Drasgow & Kang, 1984, showed, correlations are robust). In a selection context, however, precision of measurement across a wide range of trait levels, test–retest capabilities, and at least some resistance to response distortions are needed. Recent research indicates that, to be useful in selection, not only considerable enhancements to current measures but, possibly, a change in the assessment approaches may be needed (White, Young, Hunter, & Rumsey, 2008). We dis130

cuss some of these assessment innovations in the next section. Dominance versus ideal point models. The first issue is that CTT and Likert’s (1932) approaches to scale construction may be inappropriate for personality items because they are based on dominance response process assumptions. Consider the following personality statement measuring Order: “When it comes to being clean and tidy, I am about average.” Under the dominance model, respondents with increasingly higher scores on the Order facet should dominate this item and, thus, answer it Agree or Strongly Agree. Clearly, in the context of ability testing, dominance models make sense. For example, if respondents were asked to find an antonym to the word dubious among four alternatives, those with higher verbal aptitude would have higher probabilities of answering correctly. However, research indicates that the process for the Order item above is different: Respondents with average orderliness tend to choose the Agree response, but those with either low or high orderliness tend to choose Disagree (Chernyshenko, Stark, Drasgow, & Roberts, 2007; Stark, Chernyshenko, Drasgow, & Williams, 2006). Coombs (1964) used the term ideal point to describe a response process in which individuals are more likely to endorse an item when it is close to them. The idea for ideal point models can be traced back to a series of remarkable papers by Louis Thurstone (1927, 1928, 1929) that developed theory and methods for measuring attitudes. However, Thurstone’s methods for scaling attitude stimuli were laborious, and they were quickly replaced by Likert’s techniques. Likert suggested identifying one end of the attitude continuum as positive, reverse-scoring items at the other end of the attitude continuum, and then computing a person’s score as the sum of his or her item scores. He also suggested using item-total correlations to evaluate the quality of individual items. Interestingly, Likert (1932) found that items tending toward neutrality (aka intermediate items) had low item-total correlations, and thus he felt they should be eliminated. Thurstone, in contrast, deliberately included such items as he found them to be located in the middle of the scale, whereas items

usually retained by Likert’s approach were located either at the positive or the negative end of the scale. How does this distinction between ideal point and dominance models affect personality assessment? First, the use of dominance models unnecessarily eliminates items located in the middle of the scale from consideration in the scale construction process. Items such as “My social skills are about average” or “I do a standard maintenance of my property and possessions, but rarely anything more” are not found in personality scales constructed under dominance process assumptions. That is because such items have IRF similar to the one shown in Figure 4.4. From the graph, it is clear that individuals having trait scores between −1 and +1 have a higher probability of endorsing the item than individuals at either extreme. Consequently, if one were to correlate the dichotomous item score with the total scale score, obtained by summing over the remaining items, the resulting item-total correlation would be small, thus suggesting that the item is virtually nondiscriminating and should be eliminated from consideration. However, the rapidly changing response probabilities between trait levels of −2 and −1 and between +1 and +2 indicate that the item dis-

criminates well and provides high information for examinees located in those regions. Therefore, the commonly used yardsticks for judging the quality of items can be misleading if an ideal point response process applies. Second, for sophisticated measurement applications, such as computerized adaptive testing and detection of differential item functioning across groups, it is important for one’s psychometric model to provide a valid representation of the data. In such applications, a misspecified model can easily lead to incorrect results and conclusions. Third, this distinction becomes important for preferential choice items (i.e., pairwise preferences or tetrads or pentads), because one must explicitly specify how the choice for a given set of stimuli is made. Preferential choice items contain two or more personality statements from the same or different dimensions, and respondents are asked to choose statements that are most or least like them (or both). Will individuals select stimuli that are most descriptive of them (i.e., an ideal point process) or those that they most dominate (i.e., a dominance process)? Finally, important theoretical issues concerning the structure of personality hinge on psychometric analyses. For example, Ashton and Lee (2007) have argued for a six-dimensional model

1.0 Probability of Correct Response

Copyright American Psychological Association. Not for further distribution.

Individual Differences

0.8

0.6

0.4

0.2

0.0 -3

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

3

Ability Level

FIGURE 4.4. Example item response functions (IRF) for an intermediate item assuming an ideal point response process. 131

Chernyshenko, Stark, and Drasgow

Copyright American Psychological Association. Not for further distribution.

of personality on the basis of results from a dominance model analysis (i.e., factor analysis). If the model is misspecified, conclusions from such analyses should be drawn with less certainty. Resistance to faking. Research comparing responses of applicants and nonapplicants on traditional personality measures clearly shows that applicants can increase their scores in socially desirable directions (Rosse, Stecher, Miller, & Levin, 1998; Stark, Chernyshenko, Chan, Lee, & Drasgow, 2001). Although there is some debate over the extent to which applicants actually do so, the experiences of military researchers show that the faking problem “has been one of the greatest challenges to the Army’s ability to implement and sustain the operational, large-scale use of self-report personality measures, especially in high-stakes testing situations” (White et al., 2008, p. 291) and “High levels of faking . . . can lead to highly inflated test scores that have little or no criterion-related validity” (p. 292). Furthermore, approaches to detecting and correcting for faking that involve social desirability or impression management scales are only modestly effective because the scales misclassify too many honest individuals as fakers (Zickar & Drasgow, 1996). Because faking impairs the quality of hiring decisions by changing the rank order of applicants (Christiansen, Goffin, Johnston, & Rothstein, 1994), personality measures must be made more fake resistant before they can be successfully used in selection. One of the promising avenues for making personality assessment more fake resistant is the multidimensional forced-choice (MFC) item format (Chernyshenko et al., 2009; Christiansen, Burns, & Montgomery, 2005; Heggestad, Morrison, Reeve, & McCloy, 2006; Jackson, Wrobleski, & Ashton, 2000; McCloy, Heggestad, & Reeve, 2005; Stark, 2002; Stark, Chernyshenko, & Drasgow, 2005; Vasilopoulos, Cucina, Dyomina, Morewitz, & Reilly, 2006; White & Young, 1998). In the MFC format, which actually dates back to the late 1940s and 1950s (Sisson, 1948), respondents are presented with a choice of two or more statements of similar desirability but different dimensions and asked to indicate which statement describes them more accurately. U.S. Army researchers recently implemented a varia132

tion of the forced-choice format in the Assessment of Individual Motivation (AIM; White & Young, 1998) inventory. A prevalent MFC format has four statements (aka a tetrad) from which respondents are asked to select one statement that is “most like me” and another that is “least like me.” By varying the number of statements representing each dimension and by assigning intermediate scores (1s) for statements that are not selected and 0s or 2s for selected statements, enough variation is introduced into the test scores to overcome the well-known problem of ipsativity and to recover normative information. Ipsative scores contain only information about the rank order of traits for a specific individual (i.e., trait profile) but do not indicate where an individual is actually located on those traits. Thus, ipsative scores can only be used for intraindividual comparisons (i.e., personal development feedback) but not selection or classification decisions requiring interindividual comparisons. Although several studies have shown criterion validity for such MFC measures and mixed results with respect to fakeability (Christiansen et al., 2005; Heggestad et al., 2006; Jackson et al, 2000; White & Young, 1998), formal measurement models for tetrad approaches have not been developed to date. Moreover, because of the complexity of the processes underlying respondents’ judgments involving four dimensions simultaneously, we suspect that any measurement model would be quite complex. Thus, it appears that many measurement analyses and applications, such as determination of reliability, computer-adaptive testing, differential item and test functioning, and person fit analysis, would be difficult to devise and implement. A more tractable MFC format is the pairwise preference judgment. Andrich (1989); Borman et al. (2001); Chernyshenko et al. (2009); Stark (2002); Stark and Drasgow (2002); and Stark et al. (2005) and proposed and evaluated several IRT models for both unidimensional and multidimensional pairwise preferences in the context of personality, performance, and attitude assessment. Stark’s (2002) Multidimensional Pairwise Preference (MDPP) model, which was designed specifically for multidimensional pairwise preference data, seems to be particularly

Copyright American Psychological Association. Not for further distribution.

Individual Differences

attractive for high-stakes personality assessment. The MDPP model assumes that respondents consider whether each statement in a two-alternative forced choice judgment accurately describes them. The generalized graded unfolding model (J. S. Roberts, Donoghue, & Laughlin, 2000) is used to characterize the process of evaluating the individual statements. Ultimately, the respondents pick the one statement in a pair that better describes them. In the MDPP model, each statement in a pool of personality statements can be paired with many others to form pairwise preference items. If there are no “enemies” (i.e., pairs of statements that cannot be allowed together in an item), a set of 20 statements ⎛ 20⎞ 20 i19 = 190 items. A pool would generate ⎜ ⎟ = ⎝ 2⎠ 2 of 1,000 statements would generate close to a half a million unique pairs, so even if additional constraints were introduced for pairing statements (e.g., requiring similar social desirabilities), tens of thousands of items could still be formed. This is particularly attractive for CAT applications because job applicants could be allowed to retest without concerns that they would see the same items. Moreover, different applicants would see different items, and concerns about test compromise would be reduced.

In the new TAPAS inventory (Stark, Drasgow, & Chernyshenko, 2008), for example, which measures up to 22 personality facets using a pool of more than 1,400 statements, the MDPP model is used to dynamically create pairwise preference items and score responses using multidimensional Bayes modal estimation. The use of CAT principles allows for the construction of individualized tests having high measurement precision with far fewer items than traditional nonadaptive personality assessments. Although the algorithm and computations are complex, TAPAS, like any other adaptive test, tries to find the subset of items that are most informative for a particular examinee. Item information is computed at the examinee’s trait location(s) using item response functions, but for items involving statements that represent different facets, the response functions are actually three-dimensional surfaces. An example item response surface involving personality statements representing Sociability (“My social skills are about average”) and Order (“Usually, my notes are so jumbled, even I have a hard time reading them”) is shown in Figure 4.5. In the figure, values along the vertical axis indicate the probability of preferring the Sociability statement (i.e., stimulus s) to the Order statement (i.e., stimulus t) given a

FIGURE 4.5. Item response surface for a MDPP item. From “Normative Scoring of Multidimensional Pairwise Preference Personality Scales Using IRT: Empirical Comparisons With Other Formats,” by O. S. Chernyshenko, S. Stark, M. S. Prewett, A. A. Gray, F. R. Stilson, & M. D. Tuttle, 2009, Human Performance, 22, p. 113. Copyright 2009 by Taylor & Francis. Reprinted with permission. 133

Chernyshenko, Stark, and Drasgow

respondent’s standing on the respective dimensions and each statement’s parameters (for more details, see Stark et al., 2005). As with the unidimensional item response function shown previously, MDPP information will be higher where the item response surface is steeper, or, in other words, where the probabilities change rapidly.

Copyright American Psychological Association. Not for further distribution.

Validity and the Use of Personality Test Scores for Selection On the basis of the Big Five framework, researchers have conducted several meta-analyses to examine which personality factors can be useful for predicting work performance (e.g., Hough, Eaton, Dunnette, Kamp, & McCloy, 1990; Ones, Viswesvaran, & Schmidt, 1993; Tett, Jackson, & Rothstein, 1991; see also chap. 5, this volume). Results have consistently shown that Conscientiousness is a valid predictor of job performance across nearly all occupations studied, whereas the other four broad factors predict success in specific occupations or relate to specific criteria. Barrick, Mount, and Judge (2001), for example, conducted a review of several meta-analyses involving personality dimensions and showed that the average corrected correlations between Conscientiousness and several performance criteria ranged between .19 and .26. In contrast, Openness to Experience had a .05 correlation with supervisory performance ratings but a .24 correlation with training performance. More recently, a number of studies have suggested that measures of narrow factors or facets may further increase the predictive power of personality variables (e.g., Dudley, Orvis, Lebiecki, & Cortina, 2006; Paunonen 1998). As B. W. Roberts et al. (2005) pointed out, this can happen if facets of the same broad factor have differential relationships with a criterion; essentially, aggregation into a single, broad factor can result in the loss of valid trait specific variance and reduce observed validities. For example, in their study, the correlation between work dedication scores and conscientiousness was zero when facet scores were summed into a single composite, but a multiple correlation of .26 was obtained when facets were entered separately into a regression equation. Further, although researchers continue to debate the actual magnitudes of personality-performance 134

validities and whether certain meta-analytic corrections should or should not be performed (see Morgeson et al., 2007), it is clear that personality variables can add to the prediction of job performance and even more so to the prediction of many discretionary performance behaviors such as citizenship (Borman, Penner, Allen, & Motowidlo, 2001; Organ & Ryan, 1995), counterproductivity (Berry, Ones, & Sackett, 2007), and leadership emergence (Judge, Bono, Ilies, & Gerhardt, 2002). The withinperson variance in behaviors emphasized by Mischel represents a limitation on the validity of any set of static personality predictors (i.e., the 0.3 “ceiling” on validity), but the consistency of behaviors allows valid predictions of various performance dimensions across situations. In addition to predicting performance across a wide range of occupations, personality measures have shown little, if any, impact against protected groups. For example, Foldes, Duehr, and Ones (2008) conducted a large-scale meta-analysis of racial group differences on various personality scales and found only small mean differences that would be unlikely to cause an adverse impact in selection, particularly if aggregated to form composites, because the differences varied in directionality. Similar results have been found for gender groups. With the exceptions of Agreeableness, on which females generally score higher, and Dominance, a facet of Extraversion, on which males typically score higher, the extent of gender differences is small (Feingold, 1994). Thus, when personality measures are used in conjunction with cognitive ability measures in a compensatory selection system, adverse impact against protected groups will be reduced (Hough, Oswald, & Ployhart, 2001). In summary, the outlook for personality measurement in selection contexts seems bright. Legal and societal concerns about adverse impact associated with the use of cognitive ability tests, as well as the increasing emphasis on citizenship performance, adaptability, and retention, should encourage organizations to use personality test scores for selection and classification decisions. As the predictive validity database for broad and narrow personality factors grows, organizations may want to tailor assessments to meet their specific needs in terms of the numbers

Individual Differences

of traits measured, the higher level composites formed from the primary trait scores, and how the scores can be used to enhance prediction across job families. To do that, future measures should have large pools of statements that can be used to measure broad and narrow factors, as well as the capability to assemble and administer multiple forms, in accordance with user specifications, and, if faking is likely, to present items in fake-resistant formats.

Copyright American Psychological Association. Not for further distribution.

Compound Personality Variables and Their Use Thus far, we have focused on broad and narrow personality traits consistent with the prevailing Big Five taxonomy. However, considerable research has been conducted on so-called compound personality variables, which are combinations of multiple personality traits. Examples of such traits include integrity (Ones et al., 1993), core self-evaluations (Erez & Judge, 2001), emotional intelligence (Goleman, 1995), and many others. The main advantage of these variables lies in their higher validities (which are obtained by virtue of combining several useful dimensions into an overall composite). The disadvantage is that they may carry a somewhat limited theoretical value because it is often unclear which part of the composite is driving validity. This situation seems analogous to the use of difference scores in the P-O fit literature. Edwards (2002) argued persuasively that much information is lost when the composite score (i.e., the difference) is created and that person and organization variables should be entered individually into regression equations to better understand fit. By entering the variables separately, their individual contributions can be clearly identified. Edwards noted that higher order relations can also be examined by entering quadratic and interaction terms. Similarly, with personality predictors of job performance, possible conjunctive and interactive effects could be better studied and tested with polynomial regression using unidimensional facet scores. One example of a compound variable receiving substantial attention in organizational literature is Integrity. In the past 30 years, Paul Sackett and his colleagues have written several comprehensive reviews of the integrity testing literature (the latest

being Berry, Sackett, & Wiemann, 2007). Whether “integrity” is measured with “overt” tests that explicitly ask respondents about their attitudes toward theft or with “covert” or personality-based tests that ask seemingly innocuous questions about the respondent’s typical behaviors, the underlying construct being measured is, essentially, a personality composite consisting of responsibility, virtue, rule following, excitement seeking, angry hostility, self-control, and social conformity. Hence, a large proportion of the variance in integrity scale scores seems to be rooted in the Big Five factor of Conscientiousness, with some additional variance contributed by facets from Extraversion, Emotional Stability, and Agreeableness. According to Ones et al.’s (1993) meta-analysis of 665 validity studies, integrity tests are valid predictors of job performance as well as counterproductive work behaviors (validity estimates range between .20 and .41), which is not surprising given the heterogeneity of this variable. Because most personality facets comprising integrity do not correlate with cognitive ability, integrity tests when used in conjunction with cognitive ability tests have the potential to significantly increase the validity of selection decisions. We deliberately use the word potential here because applicants can fairly easily fake integrity tests. As was shown by Hurtz and Alliger (2002), the difference between honest and fake-good conditions for both overt and covert integrity tests is about .70 SD, which is a sizable difference to ignore in the context of selection decisions. Another frequently researched compound variable is Core Self Evaluations (CSE; Erez & Judge, 2001). In the development of the CSE scale, Judge, Erez, Bono, and Thoresen (2002) explicitly conceptualized the construct as a broad composite of selfesteem, generalized self-efficacy (Locke, McClear, & Knight, 1996); emotional stability, and locus of control (Rotter, 1966). In the Big Five nomenclature, this translates into a combination of Emotional Stability, Conscientiousness, and Extraversion (or, more specifically, well-being, industriousness, and dominance facets). In fact, in their latest study, Judge, Hurst, and Simon (2009) reported the 12-item CSE scale to correlate the highest with Emotional Stability (.43), Extraversion (.39), and Conscientiousness (.29). Although seemingly rooted in Big Five, CSE 135

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

appears to be a useful construct from both theoretical and practical standpoints. Theoretically, it allows for a different factorial representation of the personality domain, in which achievement orientation, assertiveness, and optimistic tendencies are grouped into a single broad factor. Our own experience with many personality data sets supports such an alternative representation as well-being, achievement, and dominance facets tend to correlate more highly with each other than with any other Big Five facets, at least in samples of managers and U.S. military personnel. Practically, CSE has been shown to be related to a variety of outcomes including job and life satisfaction, job performance, and even better adjustment to foreign assignments with most validity coefficients ranging between .20 and .40 (for the latest review, see Judge, 2009). Finally, we note that Emotional Intelligence (EI) is probably one of the most contentious compound individual difference variables in recent years. Initially popularized by Daniel Goleman (1995, 1998), emotional intelligence was defined as a set of conceptually related psychological skills involving processing of affective information and hailed as a panacea for modern businesses that neglected the “critical ingredient of success” (i.e., affect). More specifically, EI was said to comprise several related dimensions: the verbal and nonverbal appraisal and expression of emotion in oneself, appraisal and recognition of emotion in others, the regulation of emotion in oneself, and the use of emotion to facilitate thought and performance (Davies, Stankov, & Roberts, 1998). Unfortunately, from the onset, the EI field was hindered by ill-defined operationalizations of these dimensions and inflated validity claims, so it received skeptical scrutiny from personnel selection researchers (e.g., Landy, 2005, Matthews, Roberts, & Zeidner, 2004). Most initial EI studies espoused what was later called a mixed model— a blend of competencies and dispositions loosely related to emotion that could be measured using personality-like questionnaires (e.g., self-awareness, empathy, impulse control, assertiveness, stress tolerance, and social skills). To no surprise, measures based on mixed models showed little evidence of discriminant validity with respect to Emotional Stability, Extraversion, and Agreeableness (e.g., 136

Davies et al., 1998), which is why we classify EI as a compound personality trait. We note, however, that more recent work on EI has focused on “mental ability models” (Mayer, Caruso, & Salovey, 1999) that explicitly define EI as an individual’s aptitude for processing affective information and can be measured in ways similar to cognitive ability (i.e., identification of emotions in pictures). This area of research is promising, although not without its psychometric challenges (e.g., scoring is usually based on expert or consensus judgments, which tend to be unreliable). Moreover, it is not yet clear whether EI should be conceptualized as an ability or whether it would be more beneficial to treat it as a learned set of competencies (i.e., procedural skills). This latter approach is a route that has been taken by the emerging domain of Cultural Intelligence (Ang & Van Dyne, 2008; Oolders, Chernyshenko, & Stark, 2008). VALUES AND VOCATIONAL INTERESTS

Value Taxonomies Values and motives reflect stable individual differences in what people want to do or want to have in their lives. They influence how people interpret characteristics of their external environments and are linked to various affective and motivational systems. They play an important role in forming one’s specific goals and, to some extent, define how those goals are pursued and with what intensity. Research on human values and their role in goal-oriented behavior is, of course, quite broad (see also Vol. 3, chap. 4, this handbook). Our focus here is on the structure and hierarchy of values and how that information can be used to predict employee behavior. Note that by hierarchy and structure, we do not imply priorities (e.g., Maslow, 1968) but rather how values can be aggregated into increasingly broader groupings. In the domain of basic human values, Schwartz’s (1992) taxonomy is one of the most widely researched. There are 10 motivationally distinct values groupings: power, achievement, hedonism, stimulation, self-direction, universalism, benevolence, tradition, conformity, and security. These values can be arranged into a circumplex having two orthogonal axes (openness–conservation and

Copyright American Psychological Association. Not for further distribution.

Individual Differences

enhancement–transcendence) by using multidimensional scaling, as shown in Figure 4.6. Alternatively, the 10 value groupings can be organized hierarchically into four distinct higher order factors, because proximal values in Figure 4.6 tend to correlate higher with each other than with distal values (e.g., achievement and power would form the SelfEnhancement broad factor, whereas security, conformity, and tradition would form the Conservation broad factor). Schwartz and Boehnke (2004) recently showed that these structures replicate in numerous samples and countries. A comparable taxonomy to Schwartz’s, designed specifically for work settings, is the work value taxonomy from the Minnesota Importance Questionnaire (MIQ; Rounds, Henly, Dawis, Lofquist, & Weiss, 1981). Developed as part of the Theory of Work Adjustment, the MIQ taxonomy comprises 20 basic work values or needs that are aggregated into six broader factors: achievement, comfort, status, altruism, safety, and autonomy. Note that this is very similar to Schwartz’s four-factor model, but Schwartz’s Self-Enhancement factor is split into separate achievement and power factors, whereas the Conservation factor is split into a comfort factor, which is con-

cerned specifically with one’s working conditions, and a safety factor that deals primarily with company policies and employee relations. There are no hedonism themes in the MIQ taxonomy, probably because work is rarely perceived as something really fun or pleasurable. One of the most interesting features of the MIQ taxonomy is that it was explicitly developed to help individuals make vocational choices. The Theory of Work Adjustment (Dawis & Lofquist, 1984) postulates that the correspondence between what an employee desires (basic needs and values) and what an organization actually provides (called occupational reinforcers) determines that employee’s subsequent job satisfaction and, possibly, turnover decision. Hence, individuals should be advised about value profiles for various occupations to avoid poor vocational choices. The importance of person–environment correspondence is now being recognized widely in the organizational behavior literature, and the term P-O fit has been coined to refer specifically to organizational contexts (Chatman, 1989). Many recent studies involving various value dimensions have shown that P-O fit predicts not only job satisfaction but also

FIGURE 4.6. Schwartz’s Value Taxonomy. From “Evaluating the Structure of Human Values With Confirmatory Factor Analysis,” by S. H. Schwartz and K. Boehnke, 2004, Journal of Research in Personality, 38, p. 233. Copyright 2004 by Elsevier. Reprinted with permission. 137

Chernyshenko, Stark, and Drasgow

Copyright American Psychological Association. Not for further distribution.

contextual performance, organizational commitment, turnover intentions, and even overall performance (for a recent meta-analysis, see Kristof-Brown, Zimmerman, & Johnson, 2005). For example, Chatman (1991) found that, for newly hired employees, P-O fit scores predicted job satisfaction and intent to stay with an organization 1 year later, and actual turnover 2.5 years later. Given these and similar findings, many researchers have argued to use P-O fit when selecting employees for long-term commitments (Bowen, Ledford, & Nathan, 1991).

Measurement Challenges Involving Use of Values for Selection Although research results concerning the use of values to enhance personnel decisions are compelling, there are many measurement and methodological challenges that must be overcome for operational use. First, traditional Likert-type measures just do not seem to work well. The problem is that most values, by definition, represent something desirable, so it is not uncommon to see uniformly high endorsement rates for nearly all values appearing in a questionnaire, even though Schwartz’s model and others hypothesize that some values are in conflict with each other (i.e., they lie on opposite sides of the circumplex). Recognizing this limitation, many values measures use rankings or Q-sorts rather than ratings. For example, the Rokeach Value survey (RVS; Rokeach, 1973), the Organization Culture Profile (OCP; O’Reilly, Chatman, & Caldwell, 1991), the Comparative Emphasis Scale (CES; Ravlin & Meglino, 1987), and an earlier version of the Schwartz Value Survey require respondents to rank specific values or characteristics in order of their importance, thus generating a preference profile for each individual. Although this may help to overcome social desirability response biases, because one cannot like or dislike everything, the resulting ranks are ipsative and do not necessarily indicate value strength. For example, two people could exhibit the same pattern of scores for three values but differ dramatically in their levels of preference. An individual who is very high on Autonomy, high on Variety, and average on Status would have the same ranking profile as one who is average on Autonomy, low on Variety, and very low on Status. Hence, using ranking data to 138

make normative comparisons or as components in a P-O fit equation is problematic. To be applicable to selection, value measurement approaches more akin to the MIQ, which uses a combination of pairwise preference and absolute judgment items to anchor the metric, might be needed (Rounds et al., 1981). Another issue, which is specific to the use of values measures in the P-O fit context, involves the need to combine person and organization scores to form correspondence indices. One popular index is a difference score, which is simple to compute and shows both the direction and degree of correspondence. However, difference scores have been criticized widely in the P-O fit literature for reasons that include conceptual ambiguity and unreliability (Edwards, 1993). However, using both person and organization scores in a polynomial regression framework (Edwards & Parry, 1993) is quite complicated and may not be well suited for diagnostic feedback and selection purposes. A second well-known correspondence index is the correlation between person and organization profiles. Such correlations reflect the rank-order similarity of person and organization values, but referring to them as measures of fit is somewhat misleading because they do not actually capture the distance between profiles. More important, it is unclear how this index can be used in selection, because an individual’s value profile could be perfectly correlated with an organization’s profile, but his or her value strengths could be extremely different. Conversely, one’s value profile could have a near zero correlation with the organization’s profile, yet the distances between his or her values and those of the organization could be very small, making the person a relatively good fit for a position. In addition, the reliability of the correspondence index is not known, and, as far as we know, no guidelines exist about what constitutes acceptable fit. These may be some of the reasons why values and congruence indices are not often used in selection contexts. Clearly, more research is needed in this area. As things currently stand, direct fit approaches, which operationalize P-O fit as the perceived value match between person and environment, appear to hold more promise for organizational applications than indirect approaches relying on a combination

Copyright American Psychological Association. Not for further distribution.

Individual Differences

of organizational and personal value scores. There are several kinds of direct fit measurement strategies (Edwards, Cable, Williamson, Lambert, & Shipp, 2006). Some approaches ask respondents to indicate explicitly the degree of fit (e.g., Posner, Kouzes, & Schmidt, 1985), whereas others use an implicit strategy and ask respondents to select statements that most closely describe their current situation with respect to a value being measured (Chernyshenko, Stark, & Williams, 2009). All direct fit measures are relatively easy to complete from a test taker’s perspective, do not require organizational scores (and hence additional correspondence indices), and yield a single fit score that may withstand potential legal scrutiny (e.g., its psychometric properties can be readily examined). However, such direct fit scores may not be seen as strictly measures of individual differences, because they reflect a dynamic distance between one’s current and desired states. Moreover, direct fit measures are not readily suitable for selection applications because they assume that individuals have been with an organization for some time period or are at least familiar with the job environment that they are likely to encounter.

Taxonomic Structures of Vocational Interests—RIASEC; People–Things, Data–Ideas, and Prestige In the classification of vocational interests, Holland’s (1959, 1985) theory of vocational personalities and work environments has been one of the most influential (see also chap. 3, this volume). Holland’s theory states that vocational interests and work environments can be meaningfully classified into six types, as noted earlier: realistic (R), investigative (I), artistic (A), social (S), enterprising (E), and conventional (C; RIASEC). These types can be arranged into a hexagonal model, which graphically describes the relationship of each interest type to another (see Figure 4.7). The Vocational Preference Inventory (VPI, Holland, 1985), the Strong Interest Inventory (Harmon, Hansen, Borgen, & Hammer, 1994), and the O*NET Interest Profile (O*NET Resource Center, 2003) are just a few examples of measures currently using the RIASEC model for interest assessment and job matching.

R C

I

E

A S

FIGURE 4.7. Holland’s hexagonal model of vocational interests and occupations. R = realistic; I = investigative; A = artistic; S = social; E = enterprising; C = conventional.

The structural implication of the RIASEC model is that the adjacent interest types are more similar than those further apart. As with the values framework, it is proposed that the congruence between an individual’s vocational interests and the work environment leads to higher satisfaction and performance. The other implication is that all individuals lie somewhere within the hexagon, and those lying closer to the periphery have more clearly defined interests. This assumption may be true to a limited extent. From a measurement perspective, vocational interest items that are near the perimeter would be the most informative, thereby providing reliable estimates of individual vocational interests. However, if there are vocational interest items that sit in the middle of the hexagon, measurement of individuals further from the periphery should be relatively stable as well. Although there is substantial empirical support for the RIASEC structure, this evidence is found primarily with measures specifically designed to assess the six types (Tracey & Rounds, 1993). However, when Holland (1985) developed the VPI, only a small number of occupational titles were chosen to represent the six types. RIASEC codes for the remaining occupations have been generated under the implicit assumption that the model generalizes to the full range of U.S. occupations. Recognizing this shortcoming, Deng, Armstrong, and Rounds (2007) recently conducted a large empirical study involving ratings of interests in occupations representing approximately 85% of the U.S. labor market for which RIASEC codes were also available. Their multidimensional scaling (MDS) results showed that 139

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

RIASEC types and the associated hexagonal structure were not sufficient to represent the full range of occupations; instead a three-dimensional solution similar to the one proposed in the spherical model of job interests (Tracey & Rounds, 1996) was obtained with occupations located throughout the space, not just on the periphery (Deng et al., 2007). This suggests that current interest measures may present an overly restricted range of career choices because they are mainly limited to occupations showing the hexagonal structure in multidimensional scaling studies. In light of the Deng et al. (2007) study, a new approach to vocational interest measurement is needed that allows vocational interests to be identified at one or more locations anywhere in the threedimensional space. Moreover, the new approach should go beyond simply classifying interests by a vocational theme (e.g., R) because occupations with realistic characteristics may lie near the periphery of the three-dimensional space or near the center. The use of such an alternative model to represent occupational interest structures is likely to have profound consequences for how individuals are matched to jobs. To locate individuals in a two or higher dimensional space, it may be necessary to use a set of reference axes such as the ones proposed by Hogan (1983) and Prediger (1982). Hogan (1983) postulated that there are two main factors: conformity (conventional–artistic items) and sociability. In contrast, Prediger (1982) had a different orientation to the axes and labeled them as people–things (realistic– social items) and data–ideas. Tracey and Rounds (1996) showed that these dimensions were not unique and could be generalizable across subject variables and RIASEC inventories. They also suggested adding a third dimension, prestige, to describe all jobs (see Sodano & Tracey, 2008). It is important to note that items falling close to these dimensional axes should be fairly unidimensional, a necessary assumption that allows one to use item response theory to measure individuals on the axes of interest. From an item response theory perspective, the key issue is the underlying process by which individuals respond to RIASEC items. Tay, Drasgow, Rounds, and Williams (2009) recently presented 140

evidence that an ideal point response process underlies answers to vocational interest items. Intuitively, when individuals indicate how much they would like a particular job, their judgment is based on how similar they are to the occupation on the underlying dimension, rather than how far they are above it. Thus, someone who has very high social interests would be just as unlikely to want a job offering only moderate social interaction as someone having very low social interests. Because many of the occupations studied by Deng et al. (2007) lie inside the hexagon, we should expect item response functions to exhibit folding (and this is what Tay et al., 2009, found), so the difference between ideal point models and dominance models in this context is large and, it appears, important.

Use of Vocational Interests in Selection and Classification To date, the majority of interest measures have been used for classification rather than selection purposes. Essentially, after completing an interest inventory, an individual is assigned a three-letter RIASEC code (e.g., ASC) based on his or her highest interest scores. The individual’s code is then compared with those for various occupations catalogued, for example, in the Dictionary of Occupational Titles (U.S. Department of Labor, 1991) or O*NET (O*NET Resource Center, 2003), and occupations matching the individual’s interests are identified. Because not all codes have corresponding job types, congruence indices have been developed to assist with matching. One example is the Brown and Gore (1994) C-Index, which compares a person’s three-letter RIASEC code with a set of occupations and yields a score between 0 and 18, with higher numbers indicating more congruence. Validity studies involving interests can be divided into two general categories: those predicting occupational choice and those predicting the extent to which interest congruence influences job satisfaction and other outcomes in the person–environment fit frameworks. In general, studies find that interest scores predict people’s future occupational choices reasonably well. For example, longitudinal studies by Zytowski (1974), Hansen and Swanson (1983), Hansen and Dik (2005) have found that approxi-

Copyright American Psychological Association. Not for further distribution.

Individual Differences

mately 50% to 60% of respondents subsequently worked in occupations consistent with their vocational interests. However, studies testing effects of congruence on job satisfaction have been less positive. Meta-analyses by Assouline and Meir (1987) and Tranberg, Slane, and Ekeberg (1993) found that the average correlation between interest congruence and job satisfaction was at best .20, which prompted concern about the “congruence problem.” Tinsley (2000) even went as far as to conclude that Holland’s theory is invalid. Arnold (2004), however, wrote an insightful commentary pointing out a number of measurement and methodological issues surrounding congruence studies that need to be resolved before the usefulness of interests is dismissed. He listed 14 potential problems for interest congruence studies and classified them as shortcomings of Holland’s theory (e.g., Holland’s framework may not provide an adequate characterization of the environment), as problems with research designs (e.g., using crosssectional and between-subjects designs to test longitudinal hypotheses, overreliance on job satisfaction as an outcome), or as issues related to external validity (e.g., Holland’s theory views a job’s characteristics as fixed, but the requirements of many jobs change over time). Clearly, more research is needed to address these problems adequately. In their widely cited article, Schmidt and Hunter (1998) stated that interests correlate .10 with job performance (see their Table 1) and .18 with performance in training. These correlations were originally obtained by Hunter and Hunter (1984) and bear some scrutiny. The validity estimate for job performance is based on just one inventory (the Strong Interest Inventory) and just three studies. The findings for training are again based on the Strong Interest Inventory, and just two studies with a total N of 383. Consequently, these relations should be interpreted cautiously. Certainly, more research is needed. What type of relationship should be expected between vocational interest–job characteristics congruence and job performance? If the technical difficulties noted by Arnold (2004) and others can be resolved, it seems likely that a reasonably strong relation will be found between congruence and job satisfaction. Job satisfaction, however, has an interesting relationship with job performance. Judge,

Thoresen, Bono, and Patton (2001) found that job satisfaction has a moderately strong relationship with performance for individuals in jobs of low to medium complexity, with an estimated mean validity of .29. However, for high-complexity jobs, a very strong relationship was found (mean validity of .52). Therefore, it seems reasonable to predict a modest relationship between vocational interest–job characteristic congruence and job performance for low- to medium-complexity jobs but a much stronger relationship for highly complex jobs. BEHAVIORAL NARRATIVES In his neo-socioanalytic model, B. W. Roberts (2006) placed narratives within the domain of individual differences, although personal history is clearly acquired throughout an individual’s life. The reason is the endurance of personal histories, because they play an important role in forming one’s identity and social reputation. Simply put, there are certain critical events in everyone’s life requiring immediate actions, and the memories of how these events were handled shape how an individual perceives himself or herself, how he or she is perceived by others, and how he or she would likely respond to similar events in the future. Further, although I/O psychologists have some theories that take into account critical events and life histories (e.g., the developmental–integrative model by Owens and Schoenefeldt [1979] or affective events theory by Weiss and Cropanzano [1996]), this potentially fruitful research area appears to be well suited for incorporating ideas from other areas of psychology. For example, Mumford and colleagues drew from findings in developmental psychology to create their ecology model (Mumford & Stokes, 1992), and Mael (1991) also incorporated themes from social identity theory. In a broad sense, information provided by narratives is already used in many selection systems, including personal statements, letters of recommendation, interviews, and biodata blanks. However, many of these measures also contain information about other domains of individual differences (i.e., abilities, personality, values, and interests), so the incremental validity of narrative measures largely 141

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, Stark, and Drasgow

depends on how much unique information is actually captured and whether it can be measured reliably. As noted by Mael (1991), in the discussion of biodata measures, the defining characteristic of narratives is that they are historical, so one’s current intentions, views, or responses to hypothetical situations should not be considered part of that domain. It is no surprise that there are conflicting accounts about the incremental validity of selection tools involving narrative information. Schmidt and Hunter (1998) noted, for example, that biodata measures provide little incremental validity beyond cognitive ability, whereas others have found significant incremental validities beyond both cognitive ability or personality tests (e.g., McManus & Kelly, 1999; Mount, Witt, & Barrick, 2002; Stokes, Toth, Searcy, Stroupe, & Carter, 1999). Furthermore, Bauer, McAdams, and Sakaeda (2005) found that narratives involving memories of personal growth events predict outcomes such as life satisfaction beyond the Big Five personality factors. Setting aside the issue of incremental validity, many studies have shown that narratives provide useful information for selection. Typical empirically keyed biodata measures have validities ranging from .25 to .40. For example, Bobko, Roth, and Potosky (1999) reported a correlation of .28 between biodata and job performance, computed by taking a weighted average from four previous meta-analyses. Validities of interviews and letters of recommendations are typically a bit lower, however, perhaps due partly to difficulties associated with coding qualitative data. Research on scoring biodata has mainly focused on methods for keying, rather than developing psychometric measurement models. For example, Stokes and Searcy (1999) discussed keying based on expert judgments, empirical keying based on correlations of the criterion with each option of each item, and internal consistency keying in which homogeneous clusters of items are identified by principal components or cluster analysis. There has been relatively little work on evaluating the fit of various IRT models to biodata or other measures of narratives, and we know of no research examining dominance versus ideal point models. Because narratives appear so important for understanding people’s behavior, the development of formal measurement models for 142

this domain should be a priority for future research. Findings from cognitive psychology, especially with regard to autobiographical memories, may hold particular promise in this endeavor. CONCLUSIONS In this chapter, we have reviewed several classes of individual differences that are important for understanding performance in organizational settings. Cognitive ability, personality, values, vocational interests, and narratives help us understand why some people find satisfaction in their jobs and perform well and others are dissatisfied and perform poorly. By no means are we dismissive of the context in which people work. Certainly good leadership that fosters employee engagement, appropriate compensation, and a safe workplace enhance performance, satisfaction, and commitment. Nonetheless, it is individual differences that explain why one individual enjoys using differential equations to solve engineering problems whereas another person finds satisfaction in helping others address their personal problems. Research on individual differences has progressed in different ways. For cognitive ability, there have been thousands of studies addressing a wide range of topics. We know that cognitive ability is a powerful predictor of performance in jobs that are complex, and the three-parameter logistic model provides an excellent fit to many tests. There has been much work on sophisticated measurement applications such as differential item functioning, computerized adaptive testing, and person fit. Carroll’s three-stratum model provides a comprehensive framework for the structure of cognitive abilities. There has been so much work on cognitive ability that it is difficult to identify significant topics that need further research. Barrick and Mount’s (1991) seminal meta-analysis revivified research on personality in organizational settings. The role of personality in understanding organizational behavior seems intrinsically more complex than cognitive ability. With cognitive ability, “more is better” and, in unrestricted populations, different abilities are so highly correlated that it is difficult to show incremental validity (Ree, Earles, & Teachout, 1994). However, the effectiveness of per-

Copyright American Psychological Association. Not for further distribution.

Individual Differences

sonality dimensions seems much more context dependent: Extraversion is important for sales personnel but not for accountants. Moreover, it is not clear that the Big Five framework is optimal for predicting job performance. The facets underlying each Big Five dimension are not as highly correlated as cognitive abilities and may provide incremental validity (B. W. Roberts et al., 2005). Psychometric models for the measurement of personality have not received much attention, and instead researchers have simply borrowed from the cognitive domain. As we have noted, dominance models are not well suited for personality, whereas ideal point models fit better. Much more work on the psychometrics of personality measurement, especially with respect to resistance to faking, seems to be needed. The burgeoning literature on P-O fit is a testament to the importance of values. As we noted, the measurement of values is a challenge because almost all values are socially desirable and consequently an agree–disagree format is unlikely to provide much information. Rankings and Q-sorts have been used, but these bring to bear another set of challenges related to ipsativity. More work on the measurement of values as well as on the ways of quantifying the congruence between people and organizations is needed (e.g., G. W. Cheung, 2009). Vocational interests have been studied for more than a half century, but much remains to be examined. For example, psychometric theory for assessing vocational interests needs more development, as empirical keying and heuristic methods, such as multidimensional scaling, have been used primarily to date. Similarly, biodata has been criticized for its “dust-bowl empiricism,” and thus framing biodata as an approach to studying life narratives may provide a useful theoretical conceptualization (e.g., Mael, 1991). We have not discussed genetic testing or drug testing in this chapter. There are claims that attending to individual differences in job applicants’ genes could improve their health and save lives by screening out people who are at risk from exposure to toxic substances. However, the Genetic Information Nondiscrimination Act of 2008 outlawed discrimination based on genetic information in employment and medical insurance, and consequently it is unlikely

that employers will consider genetic testing. However, many employers (more than half according to Heneman & Judge, 2009) now use drug testing. The American Council for Drug Education (ACDE; 2007) claimed that substance abusers are 33% less productive and 10 times more likely to be absent from work. Although the ACDE asserted that one worker in four, ages 18 to 34, used drugs in the previous year, it is interesting that Quest Diagnostics (2008) found that only 3.8% of job applicants tested positive. In sum, the efficacy of drug testing for improving performance, broadly defined, in the workplace appears to be a topic that needs to be studied empirically. In conclusion, we believe that individual difference variables will continue to be of vital importance for enhancing the effectiveness of selection, training, and performance management initiatives. Early research focused heavily on cognitive ability and its relationships with organizational criteria, but the past 20 years have shown a definite shift toward questions involving noncognitive predictors. In fact, many individual difference theories of job performance now explicitly incorporate one or more noncognitive variables (Chan & Drasgow, 2001; Motowidlo et al., 1997). Clearly, societal factors associated with test use have catalyzed this move toward broader assessments, but the benefits to both research and practice are unequivocal. The challenge now is to make noncognitive assessments work better in operational settings where testing time is a premium, response distortion is a serious concern, and the number of predictors that must be considered for decision making has increased dramatically.

References Ackerman, P. L., Kanfer, R., & Goff, M. (1995). Cognitive and noncognitive determinants and consequences of complex skill acquisition. Journal of Experimental Psychology: Applied, 1, 270–304. doi:10.1037/1076898X.1.4.270 Almagor, M., Tellegen, A., & Waller, N. (1995). The Big Seven model: A cross-cultural replication and further exploration of the basic dimensions of natural language of trait descriptions. Journal of Personality and Social Psychology, 69, 300–307. doi:10.1037/00223514.69.2.300 American Council for Drug Education. (2007). Why worry about drugs and alcohol in the workplace? 143

Chernyshenko, Stark, and Drasgow

Copyright American Psychological Association. Not for further distribution.

Retrieved July 10, 2009, from http://www.acde.org/ employer/DAwork.htm

Journal of Applied Psychology, 92, 410–424. doi:10.1037/0021-9010.92.2.410

Andrich, D. (1989). A probabilistic IRT model for unfolding preference data. Applied Psychological Measurement, 13, 193–216. doi:10.1177/014662168901300211

Berry, C. M., Sackett, P. R., & Wiemann, S. (2007). A review of recent developments in integrity test research. Personnel Psychology, 60, 271–301.

Ang, S., & Van Dyne, L. (2008). Conceptualization of cultural intelligence: Definition, distinctiveness, and nomological network. In S. Ang & L. Van Dyne (Eds.), Handbook of cultural intelligence: Theory, measurement, and applications (pp. 3–15). New York: Sharpe.

Birnbaum, A. (1968). Some latent trait models and their use in inferring an examinee’s ability. In F. M. Lord & M. R. Novick (Eds.), Statistical theories of mental test scores (pp. 397–472). Reading, MA: AddisonWesley.

Arnold, J. (2004). The congruence problem in John Holland’s theory of vocational decisions. Journal of Occupational and Organizational Psychology, 77, 95–113. doi:10.1348/096317904322915937

Bobko, P., Roth, P. L., & Potosky, D. (1999). Derivation and implications of a meta-analytic matrix incorporating cognitive ability, alternative predictor and job performance. Personnel Psychology, 52, 561–589. doi:10.1111/j.1744-6570.1999.tb00172.x

Arthur, W., Jr., Bell, S. T., Villado, A. J., & Doverspike, D. (2006). The use of person-organization fit in employment decision making: An assessment of its criterionrelated validity. The Journal of Applied Psychology, 91, 786–801. doi:10.1037/0021-9010.91.4.786 Ashton, M. C. (1998). Personality and job performance: The importance of narrow traits. Journal of Organizational Behavior, 19, 289–303. doi:10.1002/ (SICI)1099-1379(199805)19:3 3.0.CO;2-C Ashton, M. C., & Lee, K. (2007). Empirical, theoretical, and practical advantages of the HEXACO model of personality structure. Personality and Social Psychology Review, 11, 150–166. doi:10.1177/1088868306294907 Ashton, M. C., Lee, K., Perugini, M., Szarota, P., De Vries, R. E., DiBlas, L., et al. (2004). A six-factor structure of personality descriptive adjectives: Solutions from psycholexical studies in seven languages. Journal of Personality and Social Psychology, 86, 356–366. doi:10.1037/0022-3514.86.2.356

Borkenau, P., & Ostendorf, F. (1990). Comparing exploratory and confirmatory factor analysis: A study on the 5-factor model of personality. Personality and Individual Differences, 11, 515–524. doi:10.1016/ 0191-8869(90)90065-Y Borman, W. C., Buck, D. E., Hanson, M. A., Motowidlo, S. J., Stark, S., & Drasgow, F. (2001). An examination of the comparative reliability, validity, and accuracy of performance ratings made using computerized adaptive rating scales. Journal of Applied Psychology, 86, 965–973. doi:10.1037/00219010.86.5.965 Bowen, D. E., Ledford, G. E., & Nathan, B. R. (1991). Hiring for the organisation, not the job. Academy of Management Executive, 4, 35–51. Briggs, S. R. (1989). The optimal level of measurement for personality constructs. In D. M. Buss & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 246–260). New York: Springer.

Assouline, M., & Meir, E. I. (1987). Meta-analysis of the relationship between congruence and well-being measures. Journal of Vocational Behavior, 31, 319–332. doi:10.1016/0001-8791(87)90046-7

Brown, S. D., & Gore, P. A., Jr. (1994). An evaluation of interest congruence indices: Distribution characteristics and measurement properties. Journal of Vocational Behavior, 45, 310–327. doi:10.1006/jvbe.1994.1038

Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A metaanalysis. Personnel Psychology, 44, 1–26. doi:10.1111/ j.1744-6570.1991.tb00688.x

Campbell, J. P. (1990). Modeling the performance prediction problem in industrial and organizational psychology. In M. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 1, pp. 687–731). Palo Alto, CA: Consulting Psychologists Press.

Barrick, M., Mount, M. K., & Judge, T. A. (2001). Personality and performance at the beginning of the new millennium: What do we know and where do we go next? International Journal of Selection and Assessment, 9, 9–30. doi:10.1111/1468-2389.00160 Bauer, J. J., McAdams, D. P., & Sakaeda, A. R. (2005). Interpreting the good life: Growth memories in the lives of mature, happy people. Journal of Personality and Social Psychology, 88, 203–217. Berry, C. M., Ones, D. S., & Sackett, P. R. (2007). Interpersonal deviance, organizational deviance, and their common correlates: A review and meta-analysis. 144

Campbell, J. P., & Knapp, D. J. (Eds.). (2001). Exploring the limits in personnel selection and classification. Mahwah, NJ: Erlbaum. Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies. New York: Cambridge University Press. Chan, K. Y., & Drasgow, F. (2001). Toward a theory of individual differences and leadership: Understanding the motivation to lead. Journal of Applied Psychology, 86, 481–498. doi:10.1037/0021-9010.86.3.481

Individual Differences

Chatman, J. A. (1989). Improving interactional organizational research: A model of person–organization fit. Academy of Management Review, 14, 333–349. doi:10.2307/258171 Chatman, J. A. (1991). Matching people and organizations: Selection and socialization in public accounting firms. Administrative Science Quarterly, 36, 459–484. doi:10.2307/2393204

Copyright American Psychological Association. Not for further distribution.

Chernyshenko, O. S., Stark, S., & Chan, K. (2001). Investigating the hierarchical factor structure of the Fifth Edition of the 16PF: An application of the Schmid–Leiman orthogonalization procedure. Educational and Psychological Measurement, 61, 290–302. doi:10.1177/00131640121971248 Chernyshenko, O. S., Stark, S., Drasgow, F., & Roberts, B. W. (2007). Constructing personality scales under the assumptions of an ideal point response process: Toward increasing the flexibility of personality measures. Psychological Assessment, 19, 88–106. doi:10.1037/1040-3590.19.1.88 Chernyshenko, O. S., Stark, S., Prewett, M. S., Gray, A. A., Stilson, F. R., & Tuttle, M. D. (2009). Normative scoring of multidimensional pairwise preference personality scales using IRT: Empirical comparisons with other formats. Human Performance, 22, 105–127. Chernyshenko, O. S., & Stark, S., & Williams, A. (2009). Latent trait theory approach to measuring person– organization fit: Conceptual rationale and empirical evaluation. International Journal of Testing, 9, 358–380. Cheung, F. M., Leung, K., Zhang, J., Sun, H., Gan, Y., Song, W., & Xie, D. Indigenous Chinese personality constructs: Is the Five-Factor model complete? Journal of Cross-Cultural Psychology, 32, 407–433. Cheung, G. W. (2009). Introducing the latent congruence model for improving the assessment of similarity, agreement, and fit in organizational research. Organizational Research Methods, 12, 6–33. doi:10.1177/1094428107308914 Christiansen, N. D., Burns, G. N., & Montgomery, G. E. (2005). Reconsidering forced-choice item formats for applicant personality assessment. Human Performance, 18, 267–307. doi:10.1207/s15327043hup1803_4 Christiansen, N. D., Goffin, R. D., Johnston, N. G., & Rothstein, M. G. (1994). Correcting the 16PF for faking: Effects on criterion-related validity and individual hiring decisions. Personnel Psychology, 47, 847–860. doi:10.1111/j.1744-6570.1994. tb01581.x

model. Journal of Personality and Social Psychology, 55, 258–265. doi:10.1037/0022-3514.55.2.258 Costa, P. T., Jr., & McCrae, R. R. (1994). Revised NEO Personality Inventory (NEO-PI-R) and NEO-FiveFactor Inventory (NEO-FFI) professional manual. Odessa, FL: Psychological Assessment Resources. Costa, P. T., Jr., McCrae, R. R., & Dye, D. A. (1991). Facet scales for agreeableness and conscientiousness: A revision of the NEO Personality Inventory. Personality and Individual Differences, 12, 887–898. Credé, M., Chernyshenko, O. S., Stark, S., Bashshur, M., & Dalal, R. S. (2007). Job satisfaction as mediator: An assessment of job satisfaction’s position within the nomological network. Journal of Occupational and Organizational Psychology, 80, 515–538. doi:10.1348/096317906X136180 Davies, M., Stankov, L., & Roberts, R. D. (1998). Emotional intelligence: In search of an elusive construct. Journal of Personality and Social Psychology, 75, 989–1015. doi:10.1037/0022-3514.75.4.989 Dawis, R. V., & Lofquist, L. H. (1984). A psychological theory of work adjustment. Minneapolis: University of Minnesota Press. Deng, C.-P., Armstrong, P. I., & Rounds, J. (2007). The fit of Holland’s RIASEC model to US occupations. Journal of Vocational Behavior, 71, 1–22. doi:10.1016/ j.jvb.2007.04.002 DeYoung, C. G., Quilty, L. C., & Peterson, J. B. (2007). Between facets and domains: 10 aspects of the Big Five. Journal of Personality and Social Psychology, 93, 880–896. doi:10.1037/0022-3514.93.5.880 Digman, J. M. (1990). Personality structure: Emergence of the five-factor model. Annual Review of Psychology, 41, 417–440. Drasgow, F. (2003). Intelligence and the workplace. In W. C. Borman & D. R. Ilgen (Eds.), Handbook of psychology: Industrial and organizational psychology (pp. 107–130). New York: Wiley. Drasgow, F., & Kang, T. (1984). Statistical power of differential validity and differential prediction analyses for detecting measurement non-equivalence. Journal of Applied Psychology, 69, 498–508. doi:10.1037/ 0021-9010.69.3.498 Drasgow, F., & Olson-Buchanan, J. B. (Eds.). (1999). Innovations in computerized assessment. Mahwah, NJ: Erlbaum.

Coombs, C. H. (1964). A theory of data. New York: Wiley.

Dudley, N. M., Orvis, K. A., Lebiecki, J. E., & Cortina, J. M. (2006). A meta-analytic investigation of conscientiousness in the prediction of job performance: Examining the intercorrelations and the incremental validity of narrow traits. Journal of Applied Psychology, 91, 40–57. doi:10.1037/0021-9010.91.1.40

Costa, P. T., Jr., & McCrae, R. R. (1988). From catalog to classification: Murray’s needs and the five-factor

Edwards, J. R. (1993). Problems with these of profile similarity indices in the study of congruence in

Conn, S., & Reike, M. L. (Eds.). (1994). The 16PF fifth edition technical manual. Champaign, IL: Institute for Personality and Ability Testing.

145

Chernyshenko, Stark, and Drasgow

organizational research. Personnel Psychology, 46, 641–665. doi:10.1111/j.1744-6570.1993.tb00889.x Edwards, J. R. (2002). Alternatives to difference scores: Polynomial regression analysis and response surface methodology. In F. Drasgow & N. Schmitt (Eds.), Measuring and analyzing behavior in organizations (pp. 350–400). San Francisco: Jossey-Bass.

Copyright American Psychological Association. Not for further distribution.

Edwards, J. R., Cable, D. M., Williamson, I. O., Lambert, L. S., & Shipp, A. J. (2006). The phenomenology of fit: Linking the person and environment to the subjective experience of person-environment fit. Journal of Applied Psychology, 91, 802–827. doi:10.1037/ 0021-9010.91.4.802 Edwards, J. R., & Parry, M. E. (1993). On the use of polynomial regression equations as an alternative to difference scores in organizational research. Academy of Management Journal, 36, 1577–1613. doi:10.2307/256822 Erez, A., & Judge, A. (2001). Relationship of core self-evaluations to goal setting, motivation, and performance. Journal of Applied Psychology, 86, 1270–1279. Feingold, A. (1994). Gender differences in personality: A meta-analysis. Psychological Bulletin, 116, 429–456. doi:10.1037/0033-2909.116.3.429 Fleishman, E. A. (1964). Structure and measurement of physical fitness. Englewood Cliffs, NJ: Prentice-Hall. Foldes, H., Duehr, E. E., & Ones, D. S. (2008). Group difference in personality: Meta-analyses comparing five US racial groups. Personnel Psychology, 61, 579–616. doi:10.1111/j.1744-6570.2008.00123.x Funder, D. C., & Ozer, D. J. (1983). Behavior as a function of the situation. Journal of Personality and Social Psychology, 44, 107–112. doi:10.1037/00223514.44.1.107 Goldberg, L. R. (1990). An alternative “description of personality: “ The Big Five factor structure. Journal of Personality and Social Psychology, 59, 1216–1229. doi:10.1037/0022-3514.59.6.1216 Goldberg, L. R. (1993). The structure of phenotypic personality traits. The American Psychologist, 48, 26–34. doi:10.1037/0003-066X.48.1.26 Goldberg, L. R. (1997). A broad-bandwidth, publicdomain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe (Vol. 7, pp. 7–28). Tilburg, the Netherlands: Tilburg University Press.

Gough, H. G. (1957). Manual for the California Psychological Inventory. Palo Alto, CA: Consulting Psychologists Press. Grubb, W., Whetzel, D. L., & McDaniel, M. A. (2004). General mental ability tests in industry. In J. Thomas (Ed.), Comprehensive handbook of psychological assessment, Vol. 4: Industrial and organizational assessment (pp. 7–20). Hoboken, NJ: Wiley. Guion, R. M., & Gottier, R. F. (1965). Validity of personality measures in personnel selection. Personnel Psychology, 18, 135–164. doi:10.1111/ j.1744-6570.1965.tb00273.x Gulliksen, H. (1950). Theory of mental tests. New York: Wiley. Hansen, J. C., & Dik, B. J. (2005). Evidence of 12-year predictive and concurrent validity for SII Occupational Scale scores. Journal of Vocational Behavior, 67, 365–378. doi:10.1016/j.jvb.2004.08.001 Hansen, J. C., & Swanson, J. L. (1983). Stability of interests and the predictive and concurrent validity of the 1981 Strong–Campbell Interest Inventory for college majors. Journal of Counseling Psychology, 30, 194–201. doi:10.1037/0022-0167.30.2.194 Harmon, L. W., Hansen, J. C., Borgen, F. H., & Hammer, A. L. (1994). Strong Interest Inventory Applications and Technical Guide. Stanford, CA: Stanford University Press. Heggestad, E. D., Morrison, M., Reeve, C. L., & McCloy, R. A. (2006). Forced-choice assessments of personality for selection: Evaluating issues of normative assessment and faking resistance. The Journal of Applied Psychology, 91, 9–24. doi:10.1037/00219010.91.1.9 Heneman, H. G., III, & Judge, T. A. (2009). Staffing organizations (6th ed.). Middleton, WI: McGraw-Hill Irwin. Hofstee, W. K., De Raad, B., & Goldberg, L. R. (1992). Integration of the Big Five and circumplex approaches to trait structure. Journal of Personality and Social Psychology, 63, 146–163. doi:10.1037/0022-3514.63.1.146 Hogan, J. C. (1991). Physical abilities. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 2, pp. 753–831). Palo Alto, CA: Consulting Psychologists Press. Hogan, J., & Quigley, A. M. (1986). Physical standards for employment and the courts. American Psychologist, 41, 1193–1217.

Goleman, D. (1995). Emotional intelligence. New York: Bantam.

Hogan, R. (1983). A socioanalytic theory of personality. In Nebraska Symposium on Motivation (pp. 55–89). Lincoln: University of Nebraska Press.

Goleman, D. (1998). Working with emotional intelligence. New York: Bantam.

Hogan, R., & Hogan, J. (1992). Hogan Personality Inventory manual. Tulsa, OK: Hogan Assessment Systems.

146

Individual Differences

Holland, J. L. (1959). A theory of occupational choice. Journal of Counseling Psychology, 6, 35–45. doi:10.1037/h0040767 Holland, J. L. (1985). Making vocational choices: A theory of vocational personalities and work environments. Englewood Cliffs, NJ: Prentice Hall.

Copyright American Psychological Association. Not for further distribution.

Hough, L. M., Eaton, N. K., Dunnette, M. D., Kamp, J. D., & McCloy, R. A. (1990). Criterion related validities of personality constructs and the effect of response distortion on those validities. The Journal of Applied Psychology, 75, 581–595. doi:10.1037/00219010.75.5.581 Hough, L. M., Oswald, F. L., & Ployhart, R. E. (2001). Determinants, detection and amelioration of adverse impact in personnel selection procedures: Issues, evidence, and lessons learned. International Journal of Selection and Assessment, 9, 152–194. doi:10.1111/ 1468-2389.00171 Hulin, C. L., Drasgow, F., & Parsons C. K. (1983). Item response theory: Applications to psychological measurement. Homewood, IL: Dow Jones-Irwin. Hunter, J. E. (1980). Test validation for 12,000 jobs: An application of synthetic validity and validity generalization to the General Aptitude Test Battery (GATB). Washington. DC: U.S. Department of Labor. Hunter, J. E. (1983). A causal analysis of cognitive ability, job knowledge, job performance, and supervisory ratings. In F. Landy, S. Zedeck, & J. Cleveland (Eds.), Performance measurement and theory (pp. 257–266). Hillsdale, NJ: Erlbaum. Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternative predictors of job performance. Psychological Bulletin, 96, 72–98. doi:10.1037/ 0033-2909.96.1.72 Hurtz, G. M., & Alliger, G. M. (2002). Influence of coaching on integrity test performance and unlikely virtue scale scores. Human Performance, 15, 255–273. doi:10.1207/S15327043HUP1503_02 Jackson, D. N. (1994). Jackson Personality Inventory— revised manual. Port Huron, MI: Sigma Assessment Systems. Jackson, D. N., Wrobleski, V. R., & Ashton, M. C. (2000). The impact of faking on employment tests: Does forced-choice offer a solution? Human Performance, 13, 371–388. doi:10.1207/ S15327043HUP1304_3

Judge, T. A., Erez, A., Bono, J. E., & Thoresen, C. (2002). Are measures of self-esteem, neuroticism, locus of control, and generalized self-efficacy indicators of a common core construct? Journal of Personality and Social Psychology, 83, 693–710. Judge, T. A., Hurst, C., & Simon, L. S. (2009). Does it pay to be smart, attractive, or confident (or all three)? Relationships among general mental ability, physical attractiveness, core self-evaluations, and income. Journal of Applied Psychology, 94, 742–755. Judge, T. A., Thoresen, C. J., Bono, J. E., & Patton, G. K. (2001). The job satisfaction-job performance relationship: A qualitative and quantitative review. Psychological Bulletin, 127, 376–407. doi:10.1037/ 0033-2909.127.3.376 Kolen, M. J., & Brennan, R. L. (2004). Test equating, scaling, and linking: Methods and practices (2nd ed.). New York: Springer-Verlag. Kristof-Brown, A. L., Zimmerman R. D., & Johnson, E. C. (2005). Consequences of an individual’s fit at work: A meta-analysis of person-job, person-organization, person-group, and person-supervisor fit. Personnel Psychology, 58, 281–342. doi:10.1111/j.1744-6570. 2005.00672.x Landy, F. J. (2005). Some historical and scientific issues related to research on emotional intelligence. Journal of Organizational Behavior, 26, 411–424. doi:10.1002/ job.317 Lievens, F., Buyse, T., & Sackett, P. R. (2005). The operational validity of a video-based situational judgment test for medical college admissions: illustrating the importance of matching predictor and criterion construct domains. Journal of Applied Psychology, 90, 442–452. doi:10.1037/00219010.90.3.442 Likert, R. (1932). A technique for the measurement of attitudes. Archives of Psychology, 22(140), 1–55. Locke, E. A., McClear, K., & Knight, D. (1996). Selfesteem and work. International Review of Industrial and Organizational Psychology, 11, 1–32. Mael, F. A. (1991). A conceptual rationale for the domain and attributes of bio-data items. Personnel Psychology, 44, 763–792. Maslow, A. H. (1968). Toward a psychology of being (2nd ed.). Oxford, England: Van Nostrand.

Judge, T. A. (2009). Core self-evaluations and work success. Current Directions in Psychological Science, 18, 58–62. doi:10.1111/j.1467-8721.2009.01606.x

Matthews, G., Roberts, R. D., & Zeidner, M. (2004). Seven myths about emotional intelligence. Psychological Inquiry, 15, 179–196. doi:10.1207/ s15327965pli1503_01

Judge, T. A., Bono, J. E., Ilies, R., & Gerhardt, M. W. (2002). Personality and leadership: A qualitative and quantitative review. Journal of Applied Psychology, 87, 765–780. Medline doi:10.1037/0021-9010.87.4.765

Mayer, J. D., Caruso, D. R., & Salovey, P. (1999). Emotional intelligence meets traditional standards for an intelligence. Intelligence, 27, 267–298. doi:10.1016/S0160-2896(99)00016-1

147

Chernyshenko, Stark, and Drasgow

McCloy, R. A., Campbell, J. P., & Cudeck, R. (1994). A confirmatory test of a model of performance determinants. Journal of Applied Psychology, 79, 493–505. doi:10.1037/0021-9010.79.4.493 McCloy, R. A., Heggestad, E. D., & Reeve, C. L. (2005). A silk purse from the sow’s ear: Retrieving normative information from multidimensional forced-choice items. Organizational Research Methods, 8, 222–248. doi:10.1177/1094428105275374

Copyright American Psychological Association. Not for further distribution.

McCrae, R. R., & Costa, P. T., Jr. (1987). Validation of the five-factor model of personality across instruments and observers. Journal of Personality and Social Psychology, 57, 17–40. McCrae, R. R., & Costa, P. T. (1996). Toward a new generation of personality theories: Theoretical contexts for the five-factor model. In J. S. Wiggins (Ed.), The five-factor model of personality (pp. 51–87). New York: Guilford Press. McCrae, R. R., Costa, P. T., Jr., & Piedmont, R. L. (1993). Folk concepts, natural language, and psychological constructs: The California Psychological Inventory and the five-factor model. Journal of Personality, 61, 1–26. doi:10.1111/j.1467-6494.1993.tb00276.x McManus, M. A., & Kelly, M. L. (1999). Personality measures and bio-data: Evidence regarding their incremental predictive value in the life insurance industry. Personnel Psychology, 52, 137–148. doi:10.1111/j.1744-6570.1999.tb01817.x Mershon, B., & Gorsuch, R. L. (1988). Number of factors in personality sphere: Does increase in factors increase predictability of real life criteria? Journal of Personality and Social Psychology, 55, 675–680. doi:10.1037/0022-3514.55.4.675 Mischel, W. (1968). Personality and assessment. New York: Wiley. Mischel, W. (1969). Continuity and change in personality. American Psychologist, 24, 1012–1018. doi:10.1037/ h0028886 Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Psychological Review, 80, 252–283. doi:10.1037/h0035002 Moon, H. (2001). The two faces of conscientiousness: Duty and achievement striving in escalation of commitment dilemmas. Journal of Applied Psychology, 86, 533–540. doi:10.1037/0021-9010.86.3.535 Morgeson, F. P., Campion, M. A., Dipboye, R. L., Hollenbeck, J. R., Murphy, K., & Schmitt, N. (2007). Are we getting fooled again? Coming to terms with limitations in the use of personality tests for personnel selection. Personnel Psychology, 60, 1029–1049. doi:10.1111/j.1744-6570.2007. 00100.x Motowidlo, S. J., Borman, W. C., & Schmit, M. J. (1997). A theory of individual differences in task and contex148

tual performance. Human Performance, 10, 71–83. doi:10.1207/s15327043hup1002_1 Mount, M. K., Witt, L. A., & Barrick, M. R. (2000). Incremental validity of empirically keyed biodata scales over GMA and the five factor personality constructs. Personnel Psychology, 53, 299–323. Mumford, M. D., & Stokes, G. S. (1992). Developmental determinants of individual action: Theory and practice in applying background measures. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology (pp. 61–138). Palo Alto, CA: Consulting Psychologists Press. Nye, C. D., Do, B.-R., Drasgow, F., & Fine, S. (2008). Two-step testing in employee selection: Is score inflation a problem? International Journal of Selection and Assessment, 16, 112–120. doi:10.1111/j.14682389.2008.00416.x Olson-Buchanan, J. B., Drasgow, F., Moberg, P. J., Mead, A. D., Keenan, P. A., & Donovan, M. (1998). Interactive video assessment of conflict resolution skills. Personnel Psychology, 51, 1-24. Ones, D. S., Viswesvaran, C., & Schmidt, F. L. (1993). Comprehensive meta-analysis of integrity test validities: Findings and implications for personnel selection and theories of job performance. Journal of Applied Psychology, 78, 679-703. O*NET Resource Center. (2003). The O*NET analyst database. O*NET Consortium. Retrieved July 10, 2009, from http://www.onetcenter.org/database. html#archive Oolders, T., Chernyshenko, O.S., & Stark, S. (2008). Cultural intelligence as a mediator of relationships between openness to experience and adaptive performance. In S. Ang & L. Van Dyne (Eds.), Handbook of cultural intelligence: Theory, measurement, and applications (pp. 145–158). New York: Sharpe. O’Reilly, C. A., Chatman, J., & Caldwell, D. F. (1991). People and organizational culture: A profile comparison approach to assessing person-organization fit. Academy of Management Journal, 34, 487–516. Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional predictors of organizational citizenship behavior. Personnel Psychology, 48, 775–802. Owens, W. A., & Schoenfeldt, L. F. (1979). Toward a classification of persons. Journal of Applied Psychology, 65, 569–607. Paunonen, S. V. (1998). Hierarchical organization of personality and prediction of behavior. Journal of Personality and Social Psychology, 74, 538–556. Posner, B. Z., Kouzes, J. M., & Schmidt, W. H. (1985). Shared values make a difference: An empirical test of corporate culture. Human Resource Management, 24, 293–309.

Individual Differences

Pratt, M. G. (2000). The good, the bad, and the ambivalent: Managing identification among Amway distributors. Administrative Science Quarterly, 45, 456–493.

Saucier, G. (2003). An alternative multi-language structure for personality attributes. European Journal of Personality, 17, 179–205.

Prediger, D. J. (1982). Dimensions underlying Holland’s hexagon: Missing link between interests and occupations? Journal of Vocational Behavior, 21, 259–287.

Saucier, G., & Ostendorf, F. (1999). Hierarchical subcomponents of the Big Five personality factors: A cross-language replication. Journal of Personality and Social Psychology, 76, 613–627.

Quest Diagnostics. (2008). Drug testing index. Retrieved July 10, 2009, from http://www.questdiagnostics. com/employersolutions/dti/2008_03/dti_index.html

Copyright American Psychological Association. Not for further distribution.

Raven, J., Raven, J. C., & Court, J. H. (1998). Raven Manual. 1998 edition. Oxford, England: Oxford Psychologists Press. Ravlin, E. C., & Meglino, B. M. (1987). Effect of values on perception and decision making: A study of alternative work values measures. Journal of Applied Psychology, 72, 666–673. Ree, M. J., & Earles, J. A. (1991). Predicting training success: Not much more than g. Personnel Psychology, 44, 321–332.

Schmid, J., & Leiman, J. M. (1957). The development of hierarchical factor solutions. Psychometrika, 22, 53–61. Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124, 262–274. Schmidt, F. L., & Hunter, J. E. (2004). General mental ability in the world of work: Occupational attainment and job performance. Journal of Personality and Social Psychology, 86, 162–173.

Ree, M. J., Earles, J. A., & Teachout, M. (1994). Predicting job performance: Not much more than g. Journal of Applied Psychology, 79, 518–524.

Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 1–65). Orlando, FL: Academic Press.

Roberts, B. W. (2006). Personality development and organizational behavior. In B. M. Staw (Ed.), Research on organizational behavior (pp. 1–41). New York: Elsevier Science/JAI Press.

Schwartz, S. H., & Boehnke, K. (2004). Evaluating the structure of human values with confirmatory factor analysis. Journal of Research in Personality, 38, 230–255.

Roberts, B. W., Chernyshenko, O., Stark, S., & Goldberg, L. (2005). The structure of conscientiousness: An empirical investigation based on seven major personality questionnaires. Personnel Psychology, 58, 103–139.

Segall, D.O. (2001, April). Detecting test compromise in high-stakes computerized adaptive testing: A verification testing approach. Paper presented at the Annual Meeting of the National Council on Measurement in Education, Seattle, WA.

Roberts, J. S., Donoghue, J. R., & Laughlin, J. E. (2000). A general model for unfolding unidimensional polytomous responses using item response theory. Applied Psychological Measurement, 24, 2–32. Rokeach, M. (1973). The nature of human values. New York: Free Press.

Segall, D. O., & Moreno, K. E. (1999). Development of the CAT-ASVAB. In F. Drasgow & J. B. OlsonBuchanan (Eds.), Innovations in computerized assessment (pp. 35–65). Hillsdale, NJ: Erlbaum. Sisson, E. D. (1948). Forced choice—the new Army rating. Personnel Psychology, 1, 365–381.

Rosse, J. G., Stecher, M., Miller, J. L., & Levin, R. A. (1998). The impact of response distortion on preemployment personality testing and hiring decisions. Journal of Applied Psychology, 83, 634–644.

Sodano, S. M., & Tracey, T. J. (2008). Prestige in interest activity assessment. Journal of Vocational Behavior, 73, 310–317.

Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 80 (1, Whole No. 609).

Spearman, C. (1904). “General intelligence” objectively determined and measured. American Journal of Psychology, 15, 201–293.

Rounds, J. B., Henly, G A., Dawis, R. V., Lofquist, L. H., & Weiss, D. J. (1981). Manual for the Minnesota Importance Questionnaire. Minneapolis: University of Minnesota at Minneapolis, Vocational Psychology Research.

Stark, S. (2002). A new IRT approach to test construction and scoring designed to reduce the effects of faking in personality assessment. Unpublished doctoral dissertation, University of Illinois at Urbana–Champaign.

Salgado, J. F., Anderson, N., Moscoso, S., Bertua, C., de Fruyt, F., & Rolland, J. P. (2003). A meta-analytic study of general mental ability validity for different occupations in the European Community. Journal of Applied Psychology, 88, 1068–1081.

Stark, S., Chernyshenko, O. S., Chan, K. Y., Lee, W. C., & Drasgow, F. (2001). Effects of the testing situation on item responding: Cause for concern. Journal of Applied Psychology, 86, 943–953. Stark, S., Chernyshenko, O. S., & Drasgow, F. (2004). Examining the effects of differential item/test 149

Chernyshenko, Stark, and Drasgow

functioning (DIF/DTF) on selection decisions: When are statistically significant effects practically important? Journal of Applied Psychology, 89, 497–508.

Copyright American Psychological Association. Not for further distribution.

Stark, S., Chernyshenko, O. S., & Drasgow, F. (2005). An IRT approach to constructing and scoring pairwise preference items involving stimuli on different dimensions: The multi-unidimensional pairwise preference model. Applied Psychological Measurement, 29, 184–201. Stark, S., Chernyshenko, O. S., & Drasgow, F., & Williams, B. A. (2006). Examining assumptions about item responding in personality assessment: Should ideal point methods be considered for scale development and scoring? Journal of Applied Psychology, 91, 25–39. Stark, S., & Drasgow, F. (2002). An EM approach to parameter estimation for the Zinnes and Griggs paired comparison IRT model. Applied Psychological Measurement, 26, 208–227. Stark, S., Drasgow, F., & Chernyshenko, O. S. (2008, October). Update on the Tailored Adaptive Personality Assessment System (TAPAS): The next generation of personality assessment systems to support personnel selection and classification decisions. Paper presented at the 50th annual conference of the International Military Testing Association, Amsterdam, the Netherlands. Stokes, G. S., & Searcy, C. A. (1999). Specification of scales in biodata form development: Rational vs. empirical and global vs. specific. International Journal of Selection and Assessment, 7, 72–85. Stokes, G. S., Toth, C. S., Searcy, C. A., Stroupe, J. P., & Carter, G. W. (1999). Construct/rational bio-data dimensions to predict salesperson performance: Report on the U.S. Department of Labor sales study. Human Resource Management Review, 9, 185–219. Tay, L., Drasgow, F., Rounds, J., & Williams, B. A. (2009). Fitting measurement models to vocational interest data: Are dominance models ideal? Journal of Applied Psychology, 94, 1287–1304. Tellegen, A. (1982). A brief manual for the Multidimensional Personality Questionnaire. Unpublished manuscript, University of Minnesota, Minneapolis. Tett, R. P., Jackson, D. N., & Rothstein, M. (1991). Personality measures as predictors of job performance: A meta-analytic review. Personnel Psychology, 44, 703–742. Thorndike, R. L. (1949). Personnel selection. New York: Wiley. Thurstone, L. L. (1927). A law of comparative judgment. Psychological Review, 34, 278–286. Thurstone, L. L. (1928). Attitudes can be measured. American Journal of Sociology, 33, 529–554.

150

Thurstone, L. L. (1929). Theory of attitude measurement. Psychological Review, 36, 222–241. Thurstone, L. L. (1938). Primary mental abilities. Psychometric Monographs, No. 1. Chicago: University of Chicago Press. Thurstone, L. L. (1947). Multiple factor analysis. Chicago: University of Chicago Press. Tinsley, H. E. (2000). The congruence myth: An analysis of the efficacy of the person–environment fit model. Journal of Vocational Behavior, 56, 147–179. Tippins, N. T., Beaty, J., Drasgow, F., Gibson, W. M., Pearlman, K., Segall, D., & Shepherd, W. (2006). Unproctored Internet testing in employment settings. Personnel Psychology, 59, 189–225. Tracey, T. J., & Rounds, J. (1993). Evaluating Holland’s and Gati’s vocational-interest models: A structural meta-analysis. Psychological Bulletin, 113, 229–246. Tracey, T. J. G., & Rounds, J. (1996). The spherical representation of vocational interests. Journal of Vocational Behavior, 48, 3–41. Tranberg, M., Slane, S., & Ekeberg, S. E. (1993). The relation between interest congruence and satisfaction: A meta-analysis. Journal of Vocational Behavior, 42, 253–264. Vasilopoulos, N. L., Cucina, J. M., Dyomina, N. V., Morewitz, C. L., & Reilly, R. R. (2006). Forcedchoice personality tests: A measure of personality and cognitive ability? Human Performance, 19, 175–199. Vernon, P. E. (1950). The structure of human abilities. New York: Wiley. Viswesvaran, C., Ones, D. S., & Schmidt, F. L. (1996). Comparative analysis of the reliability of job performance ratings. Journal of Applied Psychology, 81, 557–574. U.S. Department of Labor. (1991). Dictionary of occupational titles (rev. 4th ed.). Washington DC: U.S. Government Printing Office. Watson, G., & Glaser, E. M. (1980). Watson–Glaser Critical Thinking Appraisal: Forms A and B. San Antonio, TX: Psychological Corporation. Weiss, H. M., & Cropanzano, R. 1996. An affective events approach to job satisfaction. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 18, pp. 1–74). Greenwich, CT: JAI Press. White, L. A., & Young, M. C. (1998, May). Development and validation of the Assessment of Individual Motivation (AIM). Paper presented at the 106th Annual Convention of the American Psychological Association, San Francisco, CA.

Individual Differences

White, L. A., Young, M. C., Hunter, A. E., & Rumsey, M. G. (2008). Lessons learned in transitioning personality measures from research to operational settings. Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 291–295. Wonderlic Personnel Test. (1992). User’s manual for the Wonderlic Personnel Test and the Scholastic Level Exam. Libertyville, IL: Author.

Zickar, M. J., & Drasgow, F. (1996). Detecting faking on a personality instrument using appropriateness measurement. Applied Psychological Measurement, 20, 71–87. Zytowski, D. G. (1974). Predictive validity of the Kuder Preference Record, Form B, over a 25-year span. Measurement and Evaluation in Guidance, 7, 122–129.

Copyright American Psychological Association. Not for further distribution.

Yung, Y. F., Thissen, D., & McLeod, L. D. (1999). On the relationship between the higher-order factor model

and the hierarchical factor model. Psychometrika, 64, 113–128.

151

CHAPTER 5

PERSONALITY AND ITS ASSESSMENT IN ORGANIZATIONS: THEORETICAL AND EMPIRICAL DEVELOPMENTS Copyright American Psychological Association. Not for further distribution.

Frederick L. Oswald and Leaetta M. Hough

Whether a company is hiring customer sales representatives with an appropriate mix of friendliness and aggressiveness, or a space agency is training astronauts who require initiative, independent judgment, and team problem solving on long-duration space flights, individual personality traits are relevant to an organization’s successful achievement of its goals. Personality traits relevant to organizations are those psychological characteristics that predict consistent work-related thoughts, motivations, behavior, and other outcomes across situations and over time. Although laypeople take the influence of personality on individual and group outcomes for granted, the research evidence has not been as obvious. Indeed, the past 50 years have been tumultuous for researchers investigating the importance of personality variables to organizations and work outcomes. Critics in the mid-1960s challenged the power of personality traits to predict job performance. Today, the accumulated evidence has indicated that personality traits do predict variance in dependent variables of paramount importance to both organizations and their employees (e.g., career success, job performance, teamwork, job satisfaction, employee turnover). More important, this prediction is often incremental, meaning that personality measures often predict above and beyond the prediction afforded by cognitive ability measures, especially (and obviously) when the criteria rely more heavily on personality characteristics (e.g., police who must relate to members of the community to perform effectively). The magnitudes of the criterion-related validities for personality that we report weigh against classic critiques of the historical weakness of personality research

(Guion & Gottier, 1965; Mischel, 1968). Those critiques had concluded that average validities were weak or highly variable at best when really they may have been a function of the state of personality research at the time. Contemporary organizational researchers are taking on the challenge of continuing to refine the conceptualization, measurement, and modeling of personality and organizational outcomes. At the same time, they must grapple with the many practical realities that constrain any empirical research project that takes to the field. For instance, elaborate theories might specify more constructs than time allows for their appropriate measurement. In addition, although many personality measures designed for employment settings adhere to the standards required for appropriate psychological testing (e.g., American Educational Research Association, American Psychological Association, & National Council on Measurement in Education, 1999), they may lag behind the recommendations provided by current research. Finally, criterion-related validity may necessarily be based on administrative forms that reflect organizational goals such as promotion and compensation more than the outcome itself (e.g., job performance, turnover). Our chapter first provides a brief history of personality assessment, especially as it pertains to organizational research and variables that form the structure of personality. Included in this section is a review of useful theoretical frameworks, namely, the five-factor personality model (FFM), the HEXACO (six-factor) model, and a nomological-web clustering model, all of which have been used to organize available personality measures and validity evidence.

http://dx.doi.org/10.1037/12170-005 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

153

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

Using these models to examine validity evidence, we learn that analyses focused at the facet (subfactor) level can improve both theoretical understanding and prediction of dependent variables of critical interest to the field. Second, we summarize important findings concerning the criterion-related validity of personality measures, often relying on meta-analytic evidence to summarize quantitative effects found within a particular research domain. Third, we move from a discussion of the structure and validity of personality factors and facets to an examination of mean ethnic–culture, gender, and age differences on personality measures and their implications for efforts to reduce adverse impact in personnel selection. Fourth, we review and provide a perspective on test-score faking, a phenomenon that pervades the personality testing literature. We argue that research investigating various proactive attempts to prevent faking may be more productive than the vast body of research that explores attempts to induce it in the lab or to detect and correct for it statistically. Fifth, we discuss several broad integrated psychological models involving the prediction of performance outcomes, pointing out the integral role that personality plays, often at a level more refined than the Big Five or HEXACO factors. Sixth, we review several innovative approaches to personality assessment, many of which attempt to thwart faking, reduce subgroup mean differences, and either improve or broaden the prediction of organizational criteria. Seventh, although we provide directions for future research throughout the chapter, we close with a discussion of areas of inquiry that appear especially productive and timely to pursue. MODELS AND STRUCTURES OF PERSONALITY

Brief History and Background The meta-analysis by Barrick and Mount (1991) is widely cited as a turning point in the decades-long saga of academic disputes concerning the usefulness of self-report measures of personality constructs for predicting job performance. The critical feature of this meta-analysis was in summarizing criterionrelated validity coefficients by personality construct as well as by job or occupation. Specifically, they 154

used the FFM, or the Big Five factors of Emotional Stability, Neuroticism, Extraversion, Agreeableness, Conscientiousness, and Openness (see Table 5.1 for the factors and their facets), to organize personality measures, summarizing the criterion-related validities for each personality construct within and across job types. In particular, they reported that conscientiousness was a valid predictor across most jobs, and extraversion was a predictor for those interpersonal jobs requiring it (e.g., sales, law enforcement). There has been a virtual explosion of personality research in industrial and organizational (I/O) psychology subsequent to this seminal work. An important part of this history has been research examining the hierarchical nature of personality, in which facets are modeled as narrowly defined constructs whose covariances, in part, give rise to factors such as those of the Big Five (Costa & McCrae, 1995; Eysenck, 1947; J. Hogan & Roberts, 1996; Hough, 1992; John, Hampson, & Goldberg, 1991; Markon, Krueger, & Watson, 2005; Paunonen, 1998). Even the Big Five factors, broad as they are, have covariances that have been summarized by two even broader factors (labeled alpha and beta by Digman, 1997; agency and communion by Wiggins, 1991; or getting ahead and getting along by Hogan, 1983; although see Ashton, Lee, Goldberg, & de Vries, 2009, for a criticism). A discussion of the merits of factors versus facets requires an awareness of a tension between personality theory and the use of personality variables in employment settings to which organizational research is intended to generalize. More specifically, practitioners often need to know whether a specific personality measure will be useful within the assessment system of a particular company, whereas researchers seek a broader and more theoretical understanding of the underlying nature, dimensionality, and relationships between latent personality constructs. Both sides of the research–practice coin are clearly necessary because a theoretical understanding of personality constructs informs the development and practical usefulness of measures. The present-day practice of applying meta-analysis to a wide array of personality measures in organizational research has been beneficial for interpreting our results in terms of theoretical constructs. The theoretical question of whether we

Personality and Its Assessment

TABLE 5.1 Description and Alignment of Big Five (B5) and HEXACO (HX) Factors and Facets Model

Factor

HX-H

Honesty–Humility

Sincerity, fairness, greed avoidance, modesty

B5

Emotional Stability (reversed)/Emotionality

Anxiety, anger, depression, self-consciousness, immoderation, vulnerability Fearfulness, anxiety, dependence, sentimentality

Extraversion/Surgency

Friendliness, gregariousness, assertiveness, activity level, excitement seeking, cheerfulness Expressiveness, social boldness, sociability, liveliness

B5 HX-A

Agreeableness

Trust, morality, altruism, cooperation, modesty, sympathy Forgiveness, gentleness, flexibility, patience

B5

Conscientiousness

Self-efficacy, orderliness, dutifulness, achievement striving, self-discipline, cautiousness Organization, diligence, perfectionism, prudence

Openness to Experience

Imagination, artistic interests, emotionality, adventurousness, intellect, liberalism Aesthetic appreciation, inquisitiveness, creativity, unconventionality

HX-E B5 HX-X Copyright American Psychological Association. Not for further distribution.

Facets

HX-C B5 HX-O

Note. Facet descriptions are from the International Personality Item Pool (http://www.ipip.ori.org; Goldberg, 1999) for the Big Five and from K. Lee and Ashton (2004) for the HEXACO.

have identified—or ever will identify—fundamental constructs (taxons) that form the structure of human personality, however, may be less important than the extent to which we have consistent patterns of evidence for personality constructs predicting organizationally relevant constructs, thereby contributing meaningfully to practice and to our theoretical knowledge base. In the next section, we discuss in greater depth the FFM and the HEXACO model; we also discuss the usefulness of personality facets that, both theoretically and empirically, are one level more refined than the factors contained in these models (see Table 5.1). We then describe a nomological-web clustering approach that incorporates facets in a more flexible bottom-up manner that increases our understanding of the structure of personality and patterns of criterion-related validity.

FFM: Past, Present, and Future Regardless of whether the five personality constructs in the FFM are fundamental, it is obvious to anyone

in I/O psychology that personality research has relied heavily on it since the 1990s. The FFM has influenced our theory building and our practice. The FFM emerged from a factor analysis of ratings of personality-descriptive adjectives from the dictionary. The Big Five factors are thus natural-language constructs (Allport & Odbert, 1936; Goldberg, 1990, 1993; Tupes & Christal, 1992) and represent a shared or folk understanding of the structure of personality. In other words, the FFM’s roots are in the English lexicon, not in psychological theory (Hough & Ones, 2001; Hough & Schneider, 1996; Tellegen, 1993; see Hough & Schneider, 1996, for a history of the FFM). Strong proponents of the FFM have argued that it is a universal or nomothetic structure of personality variables that, like the very definition of a personality trait itself, is stable within individual across ethnic–cultural and gender groups and over time. Considerable research in the past 20 years has been devoted to examining the adequacy and universal nature of the FFM (e.g., Benet-Martínez & John, 1998; 155

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

Digman & Takemoto-Chock, 1981; Goldberg, 1990; Goldberg & Saucier, 1998; Katigbak, Church, & Akamine, 1996; McCrae & Costa, 1987, 1997; Norman, 1963; Ostendorf & Angleitner, 1994; Peabody & Goldberg, 1989; Somer & Goldberg, 1999). The FFM is often supported by the data, but there are trends showing that some parts of this model are more robust or replicable across studies than others. Extraversion and Emotional Stability are found in virtually all of the studies just cited, followed by Conscientiousness. Agreeableness is less robust; Openness to Experience tends to be the least replicable and thus the most controversial of the Big Five factors (Hough & Ones, 2001). Perhaps this latter construct is better understood at the level of its facets, which deal with openness to ideas and openness to art and culture. Openness to ideas correlates more strongly with intelligence measures, for instance (Zimprich, Allemand, & Dellenbach, 2009). Two streams of research have investigated the universality of the FFM across culture and language. One stream takes the lexical approach, an approach similar to the development of the FFM itself, based on collecting and analyzing responses to single adjectives relevant to personality in the language of a particular culture. The other stream is the persondescriptive approach, based on responses to statements as might be found on a typical Likert-scale personality test (e.g., “I like parties”). The two approaches tend to produce different results and conclusions for reasons that have not been well understood; it may be the result of the content, the format, or both. The lexical approach typically finds poor similarity of factors across culture and language (e.g., Di Blas & Forzi, 1999; De Raad, Perugini, & Szirmak, 1997), leading to the following conclusion: “A study of the various studies participating in the crusade for cross-lingual personality-descriptive universals makes it clear that researchers are unlikely to find one and only one canonical, cross-culturally valid trait structure” (De Raad, 1998, p. 122). By contrast, researchers using the person-descriptive sentences method have concluded that the FFM is a biologically based human universal (e.g., Eysenck & Eysenck, 1985; McCrae, Costa, Del Pilar, Rolland, & Parker, 156

1998), generalizing across culture, language, gender, type of assessment rating source, and type of factor extraction and rotation methods (e.g., McCrae & Costa, 1997). Continued research will surely qualify broad generalizations such as this one in important ways. For example, some factors appear to be more relevant in some cultures than in others (e.g., interpersonal relatedness in Chinese culture; F. M. Cheung et al., 2001), and measures that appear similar in content across cultures also need to have similar underlying psychometric properties and ensure that they are measuring the same constructs in a reliable manner (Church, 2000, 2001; van de Vijver & Leung, 2001). Exploratory factor analysis is the typical approach to analyzing personality data (Goldberg & Velicer, 2006), and although it often yields interpretable and consistent patterns of factor loadings for the Big Five, the claims regarding the similarity of exploratory factor structures across groups (e.g., race, culture, gender) are often based on statistical indices that tend to be deceptively high (congruence indices; see Bijnen, van der Net, & Poortinga, 1986). Moreover, repeated attempts to fit the Big Five model of personality to empirical data using confirmatory factor analysis models have failed in accordance with standard model-fit criteria (Hopwood & Donnellan, 2009), and adequate fit within groups should be a prerequisite for appropriate comparisons between groups. On the conceptual side of the coin, challenges to the FFM have resulted in a considerable body of research that questions the conclusion that the FFM is a comprehensive and universal taxonomy for predicting work behavior. Such a conclusion may very well be premature. Expanding personality constructs and content beyond the Big Five can lead to a more diverse set of personality constructs that might provide even greater insight into understanding and explaining work behavior, performance, and outcomes. Thus, although assuming a broad structure of personality such as the Big Five has the benefit of organizing a wide array of confusing and redundant constructs, treating broad factors as fundamental has the potential to stifle or misdirect the development of more sophisticated theory and research (Block, 1995).

Copyright American Psychological Association. Not for further distribution.

Personality and Its Assessment

In spite of continued controversies regarding the comprehensiveness and cross-cultural robustness of the FFM, I/O psychologists have generally embraced it as the model of choice for understanding direct and indirect relationships between personality and organizational outcomes, especially dimensions of job performance. Therefore, the FFM is the organizing structure for summarizing many of the criterionrelated validities that we discuss. Models that incorporate FFM personality variables tend to explain more of the variability in work behavior than do cognitive ability measures alone (Borman, White, Pulakos, & Oppler, 1991; Pulakos, Schmitt, & Chan, 1996; Schmidt & Hunter, 1992). Organizational researchers have begun to explore and refine the taxonomic structure of personality further. At the level of broad factors, the HEXACO is the major contender to the FFM in the organizational research literature.

HEXACO Model The HEXACO model extends the FFM to include a sixth factor. The six factors that have been recovered from personality data are described in Table 5.1 and are termed HEXACO, simultaneously an acronym and a reference to the six factors (Ashton & Lee, 2001; Ashton, Lee, Perugini, et al., 2004). HEXACO contains analogs of all the Big Five factors and adds an Honesty–Humility factor, which increases the representativeness of the model over the FFM. This factor also enhances the prediction of criteria, demonstrating a 10% to 15% increase over the FFM in the prediction of workplace delinquency across four cross-cultural samples, for instance (K. Lee, Ashton, & deVries, 2005). This sixth factor also appears to be in line with a sixth factor called the ideal-employee factor, found in an analysis of job applicant data but not student data (Schmit & Ryan, 1993). The HEXACO emerged from a critical reanalysis of a set of studies using the lexical approach across seven languages (Italian, French, German, Dutch, Hungarian, Polish, and Korean), yielding six major dimensions of personality, not just the Big Five (Ashton, Lee, Perugini, et al., 2004). One initial thought was that differences were based on the specificities of culture and language: Whereas five factors best described the data for research with English

adjectives that describe personality, other languages and cultures required six factors. However, contrary to many previous lexical studies in English that have found support for the Big Five, a lexical study of the structure of 1,710 English adjectives also found a set of six factors (Ashton, Lee, & Goldberg, 2004). The argument for six factors in English is that researchers were more inclusive of the lexicon and used hundreds more adjectives than in research done previously, allowing for richer descriptions of people (K. Lee & Ashton, 2008). For example, Tupes and Christal (1992), who were credited with the discovery of the FFM, gathered self-ratings on only 35 adjectives. Computational limitations at that time made the factor analysis of large data sets virtually or literally impossible. Even studies conducted in the 1980s and 1990s that could have made use of greater computational resources may have been burdened by this history because they tended to restrict the number of adjectives as well. The more expansive analysis is what has led to the addition of this sixth factor of Honesty–Humility.

Using Personality Facets Instead of Factors Although the HEXACO model might have greater conceptual breadth than the FFM, other data on criterion-related validity and mean score differences between ethnic–cultural and gender groups suggest that moving a level down the hierarchy of personality constructs to the facet level can yield even more useful information. Without a doubt, the FFM and HEXACO factors have been practically useful, but researchers and practitioners might incorrectly assume that they are in some sense comprehensive, fundamental, or the best that personality research can provide. We have long argued that FFM factors are too broad for advancing our understanding of personality traits and prediction of behavior in the workplace, and the HEXACO model, although adding a sixth factor, suffers from the same general problem. Research involving facets—more narrowly defined constructs that often define these superordinate factors—has provided a more substantive understanding of the relationships between personality and criterion variables of importance to organizations (Hough, 1989, 1992, 1997, 1998; 157

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

Hough & Oswald, 2000, 2005, 2008; Hough & Schneider, 1996; R. J. Schneider, Hough, & Dunnette, 1996). Research that has empirically derived facet-level taxonomies for each Big Five domain is in its relative infancy, but at least one study or meta-analysis for facets underlying each of the Big Five domains has been reported (Extraversion: S. E. Davies, Connelly, Ones, & Birkland, 2008; Emotional Stability: Birkland & Ones, 2006; Conscientiousness: Connelly & Ones, 2007; Roberts, Bogg, Walton, Chernyshenko, & Stark, 2004; Roberts, Chernyshenko, Stark, & Goldberg, 2005; Agreeableness: Connelly, Davies, Ones, & Birkland, 2008a; Openness to Experience: Connelly, Davies, Ones, & Birkland, 2008b). Somewhat different facets are identified when the lexical approach versus person-descriptive approach is used to investigate the structure of Conscientiousness (see Roberts et al., 2003, 2005). Nonetheless, results of these studies hold the promise of facets as components for improved theory, model building, and validity. Many other personality traits relevant to employee behavior appear to be either missing or not well represented in the FFM or HEXACO frameworks (Hough & Furnham, 2003, detailed 21 such variables). They include facets such as (a) rugged individualism (Hough, 1992) and masculinity–femininity (Costa, Zonderman, Williams, & McCrae, 1985); (b) social adroitness, social competence, and social insight (Ashton, Jackson, Helmes, & Paunonen, 1998; M. Davies, Stankov, & Roberts, 1998; Gough, 1968; Hogan, 1969; R. J. Schneider, Ackerman, & Kanfer, 1996); (c) self-regulation and ego resiliency (Baumeister, Gailliot, DeWall, & Oaten, 2006;

Conscientiousness

Block & Block, 2006); (d) villainy (De Raad & Hoskens, 1990); (e) fairness (Goldberg & Saucier, 1998); (f) tolerance for contradiction (Chan, 2004); (g) humorousness (Paunonen & Jackson, 2000); (h) prowess or heroism (Saucier, Georgiades, Tsaousis, & Goldberg, 2005); (i) social independence (Dancer & Woods, 2006); (j) work pace (Sanz, Gil, Garcia-Vera, & Barrasa, 2008); and (k) ease in decision making (Sanz et al., 2008) as well as needs contextualized for the work setting such as (a) need for rules and supervision and (b) need to be supportive (Sanz et al., 2008). There are more facets such as these that are clearly relevant to organizational behavior. They may be subsumed by broader factors, conceptually and empirically, but treating the facets as if they are interchangeable or the same as a broader factor may lead to ignoring or obscuring important patterns of validity. Facet-level constructs are promising building blocks for theory building. Figure 5.1 provides a simple illustration, breaking down the Conscientious construct into two facets, achievement and conformity, which are further broken down into six subfacets. At a conceptual level, one could see how some organizations may seek employees who are high on achievement and low on conformity (e.g., marketers for start-up companies) or, conversely, others may seek employees who are lower on achievement and high on conformity (e.g., security monitors). A conscientiousness composite that weights these two facets equally as part of a single compound construct will not be able to differentiate between these two different profiles of applicants and therefore would tend to select applicants who fit less well than if achieve-

Job Performance Achievement

Achievement-Striving Self-Efficacy Self-Discipline

Works Hard to Accomplish a Task Completes Tasks Successfully Avoids Distraction Conformity

Orderly Dutiful Cautious

Attends to Details Follows Rules Avoids Mistakes

FIGURE 5.1. Relating facets of conscientiousness to job performance behaviors. 158

Copyright American Psychological Association. Not for further distribution.

Personality and Its Assessment

ment and conformity were measured and used as separate facets in a linear regression. Of course, this scenario assumes that empirically there are differential relationships to be found, but research has in fact supported this notion. By now, there is a whole host of facet-level analyses that reveal different patterns of validity, and in many cases higher levels of validity, that were previously masked by analyses using broader factors. Here are several examples: In a newly hired sample of sales representatives, conformity was more predictive of obtaining new clients, whereas achievement was more predictive of the same criterion in later stages of employment (Stewart, 1999). Generally, achievement predicts performance more than conformity when it comes to overall performance, sales performance, and creativity (Hough, 1992). In a decision-making task, where performers had to discover that the rules were changing over time, those higher in conformity committed more errors, whereas achievement was unrelated to errors (LePine, Colquitt, Erez, 2000). For those who continue to believe that “broader is better,” we provide several more samples: Compared with Big Five scales, responsibility and risk taking demonstrated better prediction of self-reported delinquency such as unsafe workplace behavior and theft (Ashton, 1998); positive emotion and surgency, facets of Extraversion, predict citizenship behavior in opposite directions, obscuring any prediction by Extraversion itself (Moon, Hollenbeck, Marinova, & Humphrey, 2008); dependability, a facet of Conscientiousness, predicts counterproductive work behavior and job dedication better than does global conscientiousness (Dudley, Orvis, Lebiecki, & Cortina, 2006). Similarly, dependability and achievement, both facets of Conscientiousness, exhibit different patterns of criterion-related validity for job performance and law-abiding behaviors (Hough, 1992; Roberts et al., 2005), and dependability does not predict sales performance as strongly as achievement (Hough, 1992; Vinchur, Schippmann, Switzer, & Roth, 1998; Warr, Bartram, & Martin, 2005); harmony predicts interpersonal contextual behaviors but not personal contextual behaviors, and moral obligation and group loyalty predict personal contextual behaviors but not interpersonal ones (Kwong &

Cheung, 2003); facets such as ambition and adjustment were most predictive of their criterion counterparts (Hogan & Holland, 2003; Tett, Steele, & Beauregard, 2003). Finally, in four large-sample data sets, 16 personality factors accounted for roughly twice the variance in organizational criteria compared with 6 factors, even after statistical adjustment for greater capitalization on chance as a result of including more facets than factors in the prediction equation (Mershon & Gorsuch, 1988; see also Paunonen & Nicol, 2001). To be clear, however, we are not arguing for an infinite regress of increasingly narrow constructs, making distinctions that do not make a difference. We do not want to return to the “good old daze” (Hough, 1997) of our organizational research in personality, when each scale was considered a unique personality variable without any theoretical structure to organize it, making summaries of the literature unnecessarily difficult and subject to heavy criticism. Instead, we are suggesting that studies and metaanalyses of personality research in organizations should be conceptually and empirically examined at the facet level; they should use the FFM, HEXACO, or other broad factor models to organize facets, yet at the same time (a) acknowledge the existence of useful facets that do not fall conveniently under broad factors and (b) understand that facets may show criterion-related validities different from their parent factors. Regarding this latter point, when facets underlying a broader factor show similar correlations with criterion constructs, this indicates that the broader factor may account for more of the criterion variance—but only by researching facets can we know whether this is the case. In general, we are advocating more of a bottom-up approach, based on the accumulation and empirical examination of facet-level criterion-related validities, through metaanalysis, multilevel modeling, multigroups analysis, or other methods. Obviously, patterns of facet-level correlations that are revealed will depend on the criterion or criteria of interest (e.g., task performance, contextual performance, turnover).

Nomological-Web Clustering Approach Hough and colleagues have long argued that nomological-web clustering should be the basis for 159

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

forming clusters of homogeneous personality variables demonstrating high construct and criterion-related validity (e.g., Hough, 1992, 1997, 1998; Hough, Eaton, Dunnette, Kamp, & McCloy, 1990; Hough & Ones, 2001; Hough & Schneider, 1996; Schneider & Hough, 1995). More generally, this approach is aligned with the more stringent psychometric assumptions of convergent and discriminant validity in support of construct validity (Campbell & Fiske, 1959; Loevinger, 1957; McDonald, 1999). Nomological-web clustering is the general approach or philosophy that personality variables or facets that are grouped together (by factor analysis, expert-sorting methods, or some other method) should show similar patterns of correlation across an array of criterion variables. Its focus on both criterion-related validity and relationships between personality variables, and not either one in isolation, is what makes nomological-web clustering distinctive. It is an important way to test or extend the lexical hypothesis, which claims that analyzing personality content alone (without reference to criteria) yields a structure of personality that would be considered basic. An extensive and intensive nomological-web clustering of personality facets has been conducted through a rational sorting of facets based on criterion-related validity evidence, and further research should contribute to the taxonomy that resulted (for an extensive list of clusters, see the appendix of Hough & Ones, 2001). At least two different studies to date have been conducted to this end. One study (Dudley et al., 2006) meta-analyzed criterion-related validities of Conscientiousness facets, concluding that the facets (a) have only low to moderate correlations with each other and are best conceived of as distinct rather than as parts of a single global factor; (b) have different patterns of correlations with criteria; and (c) depending on the criterion and occupation, have higher criterion-related validities than a global Conscientiousness factor that ignores facets (e.g., 24% more variance in overall job performance predicted in skilled and semiskilled jobs). Another study (Foldes, Duehr, & Ones, 2008) meta-analyzed mean score differences between Whites and different cultural–ethnic groups according to a nomologicalweb clustering taxonomy, reporting that the pattern of mean scores differed for the subgroups on the 160

facets within the Big Five factors, even though subgroups scored similarly at the Big Five factor level.

Summary and Conclusion Although the FFM has enjoyed significant support as a foundation of personality in I/O psychology, and in psychology as a whole, its base may be cracking. Support for the FFM relies primarily on factor analysis of self-reported personality descriptions (either person-descriptive sentences or adjectives); other methods suggest different models of personality. The FFM also is not comprehensive—important variables are missing. The HEXACO model is an improvement in its representativeness, but both the FFM and the HEXACO model combine facets into broad heterogeneous factors that, we argue, may not be as useful as the facets themselves for prediction and understanding. We are undoubtedly biased in favor of the nomological-web clustering approach that organizes personality facets into groups or factors, based on similar patterns of criterion-related validity and similar patterns of relationships with other personality variables. More important than our bias, the validity evidence for facet-based criterion-related validities that we have reviewed is generally supportive of nomological-web clustering as opposed to broader factor models of personality. Regardless of the taxonomic approach, the weight of additional criterionrelated validity evidence at the personality facet level, based on more careful criterion measurement, should speak louder than anyone’s theory-driven agendas. As a general recommendation, we urge researchers to specify organizational criteria against which to pit different personality models against one another empirically, not just examining personality models on their own without respect to criteria. Creating a “fair fight” between competing models can be challenging. For example, measures of personality facets may contain fewer items than personality factors, and thus the predictive power of the former may be less than that of the latter. The comparison of non-nested models can pose an additional challenge because model fit can be confounded with model parsimony, although there are statistical methodologies that can contend with this issue (e.g., indices such as Bayes factors and root-mean-square error of approximation).

Personality and Its Assessment

Copyright American Psychological Association. Not for further distribution.

CRITERION-RELATED VALIDITY OF PERSONALITY TRAITS: BRIEF SUMMARY OF THE EVIDENCE Measures of different personality traits first need to support the psychometric prerequisites of reliability (e.g., test–retest and internal consistency reliability) and convergent and discriminant validity (e.g., a multifactor model fit between personality factors and related factors) before one can begin to say that personality traits are being measured and meaningfully predict organizational outcomes via criterion-related validity or, in a broader multilevel sense, that personality measurement is useful for those organizations that implement it in their selection, training, or employee development systems (see Vol. 2, chap. 13, this handbook). We point out clear bottom-line trends based on cumulative research evidence indicating that personality measures do predict important work outcomes. We do not provide an exhaustive review of the evidence in this chapter; rather, we seek to summarize important findings that should encourage organizational researchers and practitioners to continue their pursuits in the personality domain. Although we have long known that higher levels of conscientiousness relates to occupational success (Barrick & Mount, 1991), recent meta-analytic work has shown it also relates to successful occupational attainment over the life span (Roberts, Kuncel, Shiner, Caspi, & Goldberg, 2007), and greater emotional stability is associated with the positive mental health required to function effectively in the workplace (Lahey, 2009). Exhibit 5.1 highlights other broad findings regarding the criterion-related validity of personality measures: Personality predicts leadership and career success, various dimensions of job performance, contextual performance, counterproductive work behaviors, team performance, and job satisfaction (see Vol. 1, chaps. 7 and 19, this handbook; chaps. 3, 9, 10, and 19, this volume; and Vol. 3, chaps. 4 and 17, this handbook). Many important mediators partially account for these relationships, such as goal setting, self-regulation, positive emotions, and motivation (Elliot & Church, 1997), as we later discuss. In reporting criterion-related validity evidence, we tend to rely on meta-analytic correlations, each of which is a psychometrically corrected average validity

across a set of studies. Although we do not consider these “true” validity coefficients that generalize across all samples, settings, and measures, these averages provide useful descriptive (if not inferential) summaries of various research domains. Meta-analysis invokes sampling error variance as a more parsimonious explanation of the variability across reported effect sizes; however, this explanation is always provisional until more data are available that allow one to conduct moderator analyses with appropriate statistical power (Oswald & McCloy, 2003). More powerful statistical tests may determine that at least some of the variance in criterion-related validities may be moderated by the type of occupations, samples, and settings under study. The criterion-related validities we report are moderate in size, and—for a number of reasons—no apologies should be made for them not being larger: (a) Even moderate validities for personality prove to be highly valuable in practice, often incrementing the prediction afforded by ability measures and providing utility across an organizational workforce retained over time; (b) validities for personality in field settings are of similar magnitude to those found when examining the strength of main effects resulting from manipulating situations in social psychology research (r ≈ .2 or d ≈ 0.4 on average, see Richard, Bond, & Stokes-Zoota, 2003); (c) it is naive to think that all the variance of complex human behavior in the world of work can be fully predicted from a handful of personality scales and their bivariate relationships with criterion measures; (d) research findings often reflect the “criterion problem” of broad performance measures that were used for administrative purposes, and this cannot be corrected for psychometrically; and (e) the amount of variance predicted is not the same as the importance of variance predicted (e.g., no measure will predict much variance in low base-rate behaviors such as organizational theft, yet theft is a clearly important outcome). In addition, we are not unaware of the reality that validity coefficients are more trustworthy when the data are based on higher quality measurement, relevant and generalizable settings, and larger and more representative samples. No study achieves these ideals to the fullest. Thus, although meta-analytic validities corrected for psychometric artifacts may come closer to estimating 161

Oswald and Hough

Exhibit 5.1 Criterion-Related Validity of Personality Measures: General Conclusions Across Occupational Groups

■

■

Copyright American Psychological Association. Not for further distribution.

■

■

■

■

■

■

■

■

■ ■

■

■

■

■

162

Leadership and career success Personality predicts occupational attainment and advancement ([.20 to .30]; Judge & Hurst, 2007; meta-analysis: Roberts, Kuncel, Shiner, Caspi, & Goldberg, 2007). Personality predicts overall managerial effectiveness, promotion, and managerial level (achievement orientation [.15 to .20], dominance and energy level [.20 to .25]; meta-analysis: Hough, Ones, & Viswesvaran, 1998). Personality predicts leadership, transformational leadership, and leadership emergence (sociability, dominance, achievement, and dependability [.30 to .35], conscientiousness [.10 to .30], extraversion [.20 to .30], emotional stability [.15 to .25], openness [.10 to .25]; meta-analyses: Bartram, 2005; Bono & Judge, 2004; Judge, Bono, Ilies, & Gerhardt, 2002). Job performance outcomes Personality predicts overall job performance ratings, task performance, behavioral measures of performance, and career attainment (achievement and dependability [.15 to .25], conscientiousness [.15 to .30], agreeableness, emotional stability, extraversion [.10 to .20]; meta-analyses: Barrick, Mount, & Judge, 2001; Bartram, 2005; Dudley, Orvis, Lebiecki, & Cortina, 2006; Hurtz & Donovan, 2000). Personality predicts the performance of expatriates (cultural sensitivity and cultural flexibility [.25 to .30], conscientiousness and extraversion [.15 to .20], meta-analysis: Mol, Born, Willemsen, & Van Der Molen, 2005). Personality predicts goal setting, organizing, execution, and not procrastinating (self-efficacy [.45], depression [−.35], conscientiousness [.15 to .35], agreeableness [−.30], emotional stability [.30 to .35], extraversion [.15], openness [.20]; meta-analyses: Bartram, 2005; Judge & Ilies, 2002; Steel, 2007). Personality variables predict outcomes reflecting creativity and innovation (dominance and potency [.20], innovativeness [.30], conscientiousness [.35], openness [.15 to .40]; Grucza & Goldberg, 2007; Hough, 1992; Robertson & Kinder, 1993; meta-analyses: Bartram, 2005; Feist, 1998; Hough & Dilchert, 2007). Personality-based integrity tests predict overall job performance ([.30 to .45], meta-analysis: Ones, Viswesvaran, & Schmidt, 1993). Contextual performance Personality predicts contextual performance, encompassing constructs such as organizational citizenship, altruism, job dedication, interpersonal facilitation, and generalized compliance (empathy, helpfulness, positive affectivity [.15 to .20], altruism [.25], dependability [.10 to .20], affective commitment [.25], conscientiousness [.15 to .20], agreeableness [.10 to .15], emotional stability [.15]; meta-analyses: Borman, Penner, Allen, & Motowidlo, 2001; Dudley et al., 2006; Hurtz & Donovan, 2000; LePine, Erez, & Johnson, 2002; Organ & Ryan, 1995). Counterproductive work behaviors Personality variables predict counterproductive work behaviors (conscientiousness [−.20 to −.40], agreeableness [−.30 to −.45], emotional stability [−.25]; meta-analysis: Berry, Ones, & Sackett, 2007). Personality-based integrity tests predict counterproductive work behaviors ([.30]; meta-analysis: Ones et al., 1993). Personality-based integrity tests predict absenteeism ([.35]; meta-analysis: Ones, Viswesvaran, & Schmidt, 2003). Training outcomes Personality predicts learning and skill acquisition during training (achievement motivation [.15 to .35], conscientiousness [.15 to .30], extraversion [.15 to .30], agreeableness [.15], openness [.25 to .35]; meta-analyses: Barrick & Mount, 1991; Barrick et al., 2001; Colquitt, LePine, & Noe, 2000). Teamwork and team performance Personality variables predict team cohesion and teamwork (conscientiousness [.25], agreeableness [.35], emotional stability [.20], extraversion [.15], openness [.15]; meta-analyses: Barrick et al., 2001; Hogan & Holland, 2003). Personality variables predict team performance (conscientiousness [.25], agreeableness [.35], extraversion [.15], openness [.15]; meta-analysis: Peeters, Van Tuijl, Rutte, & Reymen, 2006). Job satisfaction Personality predicts both job and career satisfaction (proactivity and locus of control [.40], conscientiousness [.25], agreeableness [.15], emotional stability [.30 to .35], openness [.15]; meta-analyses: Judge, Heller, & Mount, 2002; Ng, Eby, Sorensen, & Feldman, 2005).

Personality and Its Assessment

Exhibit 5.1 (Continued) Criterion-Related Validity of Personality Measures: General Conclusions Across Occupational Groups ■

■

Copyright American Psychological Association. Not for further distribution.

■

■

■

Personality variables are positively correlated with measures of subjective well-being (conscientiousness [.25 to .30], agreeableness [.15 to .35], emotional stability [.35 to .50], extraversion [.35 to .60]; meta-analysis: Steel, Schmidt, & Shultz, 2008). Mediated models involving personality Conscientiousness predicts learning but is fully mediated by goal commitment (Klein & Lee, 2006). Relationships between personality variables and work-related criteria are influenced by motivational and self-regulatory mechanisms (Barrick, Mount, & Strauss, 1993; Erez & Judge, 2001; Kanfer & Heggestad, 1999; F. K. Lee, Sheldon, & Turban, 2003). Positive emotions help those high in psychological resilience recover from daily stress more effectively (Ong, Bergeman, Bisconti, &Wallace, 2006). Status striving mediates the relationship between extraversion and sales performance, such that those high in extraversion tend to be high-status strivers, who in turn are better performers (Barrick, Stewart, & Piotrowski, 2002).

Note. Brackets contain the approximate range of zero-order correlations, based on the most statistically stable estimates reported. Meta-analytic correlations were often corrected for statistical artifacts. Only practically significant correlations are reported (>.10).

both operational and latent relationships of interest, they should be taken with a grain of salt and a larger standard error (see discussions by Morgeson et al., 2007; Murphy & DeShon, 2000). Researchers may appreciate the validity coefficients we report in their correlational metric, but it is also important to translate these values into other metrics that employees, managers, and other stakeholders can understand and appreciate, such as dollars, odds ratios, and expected average increases in performance (Cascio & Boudreau, 2008; Kuncel, Cooper, & Owens, 2009). For example, the binomial effect size display (Rosenthal, 1991) converts a correlation into the percentage increase in success rate, so, for instance, a validity coefficient of .20, although seemingly small, translates roughly into a 10% increase in hiring success, a value that many managers view as meaningful, in a metric that is more understandable than a correlation coefficient. As a more specific metric, validity coefficients combined with personality test cutoffs used in selection, training, or promotion contexts can be translated into sensitivity and specificity indices that describe the usefulness of a personality test, with sensitivity referring to the percentage of those who passed the test and were identified as successful performers and specificity referring

to the converse, the percentage of those who failed the test and were identified as unsuccessful (Glaros & Kline, 1988). These percentages are themselves useful, but they can also be translated into the mean increase in performance of those selected, compared with either those not selected or with the total pool of test takers. Sensitivity and specificity metrics, however, assume that one has data- or model-based estimates for performance scores of those who fail the test. More generally, alternative metrics like these must either be corrected for the effects of direct and incidental range restriction, sampling error variance, and measurement unreliability or must be interpreted appropriately in the context of these known factors. Validity coefficients have the same problem; however, these alternative metrics may still provide managers, employees, and job applicants with a quicker appreciation for the usefulness of a valid personality measure than would a validity coefficient on its own. SUBGROUP MEAN DIFFERENCES IN PERSONALITY SCALES: IMPLICATIONS FOR ADVERSE IMPACT When personality assessment is used in personnel selection, the intent of personality theory and models 163

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

is to produce measures that show improved prediction and a fairer selection process overall. The basic wisdom of providing fair selection procedures to all job applicants, the diversity in the global marketplace and labor pools, and the legal prohibitions that expressly affect employment decision making, taken together, strongly encourage employers to attend to a variety of factors that affect protected groups of applicants during the selection process. In the United States, fundamental pieces of legislation, such as the Civil Rights Acts of 1964 and 1991, the Americans With Disabilities Act of 1990, and the Age Discrimination in Employment Act of 1967 (amended in 1978), protect classes such as people with disabilities, minorities, women, and people age 40 and older. Adverse impact against protected groups is often a critical factor when evaluating the fairness of a selection process. For instance, if the percentage of Black applicants who are hired is close to the percentage of White applicants who are hired (e.g., the former being at least 80% of the latter by the “four-fifths rule”), then any adverse impact that exists is unlikely to generate a lawsuit (see Morris & Lobsenz, 2000). Adverse impact in hiring is in part a function of the mean difference between applicant subgroups because larger mean differences will lead to disproportionate hiring of one group over another in topdown selection procedures. Therefore, it is important to be aware of subgroup mean differences for those personality constructs measured in personnel selection systems. Because we have argued for the usefulness of facet-level personality variables in future research and practice, this section not only provides mean score differences between various subgroups on the Big Five factors but also on its facets, as much as the data allow. We draw heavily on two meta-analyses in presenting summary results (Foldes et al., 2008; Hough, Oswald, & Ployhart, 2001), making use of conventional standards in the field for defining small, medium, and large standardized mean differences (i.e., ds = 0.2, 0.5, and 0.8, respectively; Cohen, 1988).

FFM: Mean Differences Ethnic–cultural subgroup comparisons. These two meta-analyses summarized mean-score differences for ethnic–cultural subgroups (i.e., Blacks, Hispanics, Asians, and American Indians) according 164

to the FFM of personality variables. At the broadly defined level of the Big Five, both meta-analyses found that Blacks, Hispanics, and Asians scored virtually the same when compared with Whites. One discrepancy was where in one meta-analysis (Hough et al., 2001), Blacks scored lower than Whites on average on Openness (d ≈ −0.2), but in the other (Foldes et al., 2008), that difference was negligible (d ≈ −0.1). Another discrepancy was for Asian–White mean differences on Openness to Experience; however, uniformly small sample sizes for Asians on this construct precluded accurate results. In both meta-analyses, mean scores for American Indians were also based on small sample sizes (even when accumulated across studies), making any mean differences too tenuous to report, in terms of both statistical power and appropriate representation of the subgroups. Compellingly large mean differences between Blacks and Asians were found in one of the meta-analyses (Foldes et al., 2008), with Blacks scoring higher on average on Emotional Stability (d ≈ 0.6), Extraversion (d ≈ 0.4), and Conscientiousness (d ≈ 0.4). Future personality research comparing diverse subgroups is clearly warranted, particularly given the major demographic shifts occurring in today’s global workforce. In short, both meta-analyses reached the same general conclusion about Black, Hispanic, and Asian mean scores, namely, that mean scores on the broadly defined Big Five personality variables tend to be very similar when they are compared with scores of Whites as the majority group. To the extent this is true, the Big Five–level personality variables at this aggregate level do not tend to contribute to adverse impact against these protected classes. Instead, they have the potential to reduce adverse impact when used in combination with cognitive ability measures—although not as much as one would hope, as we discuss in more detail. Gender comparisons. A meta-analysis averaged mean score differences between men and women across studies at the Big Five level (Hough et al., 2001), finding some noteworthy results: Women generally scored higher than men on Agreeableness (d ≈ 0.4) and lower than men on Emotional Stability

Personality and Its Assessment

Copyright American Psychological Association. Not for further distribution.

(d ≈ −0.2). On the other three Big Five factors— Extraversion, Conscientiousness, and Openness to Experience—this meta-analysis found essentially no differences between men and women, consistent with more recent work exploring gender differences in two of these factors (Extraversion and Conscientiousness; Duehr, Jackson, & Ones, 2003). Age group comparisons. A meta-analysis comparing older people (age 40 and older) and younger people (typically college age) on Big Five personality variables found similar scores across all Big Five factors (Hough et al., 2001). One exception was Agreeableness, on which older people scored somewhat higher on average than younger people (d ≈ 0.2). A large, nationally representative Dutch sample supported a similar factor structure (measurement invariance) across age cohorts, which allows for a stronger substantive interpretation of the mean increases in Agreeableness and Conscientiousness that were found with age (Allemand, Zimprich, & Hendriks, 2008). A meta-analysis of cross-sectional age-cohort data, largely based on U.S. samples, reported similar increases, with Conscientiousness showing a more steady linear increase over time and Agreeableness showing a sharp increase when people are in their 50s. Emotional Stability increased steadily until people reached age 40 and then remained stable; social dominance increased until the mid-30s and then remained relatively stable (Roberts, Walton, & Viechtbauer, 2006). Unlike this meta-analysis, two studies have reported curvilinear relationships between conscientiousness and age: a cross-sectional two-nation study finding peak levels of conscientiousness in the mid-40s (Donnellan & Lucas, 2008), and a longitudinal study in which conscientiousness found its peak in the mid-60s (Terracciano, McCrae, Brant, & Costa, 2005).

Facet-Level Personality Variables: Mean Differences Using facets identified through the aforementioned nomological-web clustering strategy (Hough & Ones, 2001), a meta-analysis summarized mean score differences between ethnic–cultural, gender, and age subgroups at the facet level of Big Five factors. Some interesting findings emerged that are not apparent at

the broader Big Five level (Hough et al., 2001), with similar findings replicated in a subsequent metaanalysis (Foldes et al., 2008). Ethnic–cultural subgroup comparisons. For achievement and dependability (two facets of Conscientiousness), Blacks scored about the same as Whites on achievement and only slightly lower than Whites on dependability (d ≈ −0.1; only Hough et al., 2001, is reported because of its much larger sample size). For dominance and sociability (two facets of Extraversion), Blacks scored about the same as Whites on dominance but lower (d ≈ −0.4) on sociability (Foldes et al., 2008; Hough et al., 2001). Hispanics scored on average higher (d ≈ 0.3) than Whites on self-esteem, a facet of Emotional Stability (Foldes et al., 2008). Other minor differences occurred between some of the comparisons for other ethnic–cultural subgroups, but sample sizes are too small (even for meta-analysis) to have confidence in those results. Gender comparisons. Small facet-level differences exist between men and women. For dependability (a facet of Conscientiousness), women tended to score somewhat higher than men (d ≈ 0.2; Hough et al., 2001). For dominance and activity (two facets of Extraversion), men scored somewhat higher than women on both (d ≈ −0.2; Duehr et al., 2003; Hough et al., 2001). Age group comparisons. Small facet-level differences exist between older (age 40 and older) people and younger people on some facets. For dependability and achievement (both facets of Conscientiousness), older people generally score higher (d ≈ 0.5) than younger people on dependability but somewhat lower than younger people on achievement (d ≈ −0.2; Hough et al., 2001). Overall conclusion. In general, differences between ethnic–cultural, gender, and age subgroups are small to moderate in size, but we have noted some sizable exceptions. Larger differences tend to be found at the level of narrowly defined facets that underlie Big Five–level personality constructs; in some cases, protected groups (e.g., Blacks or women) have higher average scores than traditional comparison groups (e.g., Whites or men). These 165

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

findings are especially important because although an organization may focus its measurement efforts on personality facets exhibiting the highest levels of validity, it is the mean differences and correlations between all subtests in the selection battery that jointly determine adverse impact against a protected group (without the need for any criterion data or validity information). The pattern of subgroup mean differences at the personality facet level between ethnic–cultural, gender, and age subgroups is complex and likely has meaningful moderator effects such as job complexity and requirements for engaging in teamwork on the job. We avoided a review of such effects because to our knowledge very little research evidence has spoken directly to this issue. However, both the attraction–selection–attrition model (Schneider, 1987) and theories of vocational interests (Holland, 1985) offer the general hypothesis that mean levels of personality traits should differ by both the type of jobs and the type of organizations in which employees are members (Schaubroeck, Ganster, & Jones, 1998); thus, it is reasonable to think that subgroup mean differences may vary as well, to the extent that occupations and organizations have varying personality requirements and members of different subgroups assort themselves differentially across occupations as a function of their personality (see chap. 3, this volume, and Vol. 3, chap. 1, this handbook). Future research may seek to explore this issue further to understand the extent to which mean differences in personality measures by particular subgroups (age, race, gender) lead to practically significant differences in hiring or promotion rates across different selection and assessment contexts. It is also an important practical issue to analyze large samples in terms of combinations of subgroup characteristics because examining race and gender differences independent of one another may not uncover more specific subgroup differences (e.g., Black women vs. White men). Regarding the selection context, we remind the reader that even though many personality measures have shown little evidence for subgroup mean differences, adding them to a selection battery that already demonstrates large subgroup mean differences will not reduce adverse impact as much as one might 166

anticipate. As a concrete example (see Sackett & Ellingson, 1997), when two group means differ on one measure by 1 standard deviation (e.g., the Black–White mean difference that has frequently been documented for cognitive ability measures), and there is no mean difference on a second uncorrelated measure (e.g., a measure of the trait of dominance), the resulting composite has a group mean difference of .71—not the difference of .50 that one might intuitively expect. As a result, minority rates of hiring will be improved, but the bad news is that adverse impact will not be reduced enough to satisfy the aforementioned four-fifths rule across a wide range of selection ratios and subgroup proportions in the applicant pool. This holds true even if it were legal to take a slight reduction in validity to reduce adverse impact (see the pareto-optimal tradeoff scenarios in DeCorte, Lievens, & Sackett, 2007). A more effective approach to reducing adverse impact would be to demonstrate that personality measures with low to no adverse impact predict noncognitive criteria that are highly relevant to the job (Sackett, Schmitt, Ellingson, & Kabin, 2001), although this approach is also not without its constraints, most notably because cognitive ability is a strong predictor of job performance criteria across almost every job and at the same time is accompanied by the persistence of large subgroup mean differences across racial–ethnic groups. PERSONALITY TEST-SCORE FAKING There has been an overwhelming concern in organizational research that people’s responses to personality tests do not reflect their true standings on underlying traits of interest, particularly the responses of job applicants in high-stakes selection settings. One can easily imagine applicants inflating their self-report of the positive job-relevant traits they lack—whether on a personality test, in a face-to-face interview, or on a resume—especially in cases in which there is little perceived negative consequence along with the potential reward of being hired into a highly desired (if not desirable) job. Applicants who feel a greater incentive to present themselves in a more positive light and inflate personality scores would be more likely to be hired, without necessarily showing higher levels of performance on the job (Mueller-Hanson,

Copyright American Psychological Association. Not for further distribution.

Personality and Its Assessment

Heggestad, & Thornton, 2003). This has implications for establishing personality test norms as well: If applicant means are substantially higher than incumbent means (as in Weekley, Ployhart, & Harold, 2004), then cutoff scores based on the scores of the incumbent sample would clearly be inappropriate to apply to the applicant sample (Bott, O’Connell, Ramakrishnan, & Doverspike, 2007). Research data have not unequivocally supported the commonsense notion of applicant score inflation. An early faking study found much higher means for people who were directly asked to fake a personality test like an ideal job applicant would versus those asked to respond honestly; however, the mean differences between applicant and incumbent scores in a military sample showed higher means for incumbents, meaning that effects of the entire selection process appear to have been stronger than any of the faking effects of applicants (Hough et al., 1990). Other evidence has also provided little reason to suspect score inflation in job applicants. Two recent large within-person studies found small differences in mean personality test scores when the first test was for selection purposes and the second was for developmental purposes or vice versa (Ellingson, Sackett, & Connelly, 2007), and they were near zero when job applicants took a personality test, were rejected on the basis of those scores, and then reapplied and were retested (J. Hogan, Barrett, & Hogan, 2007). However, mean differences are more evident in other studies. For instance, a meta-analysis comparing incumbents and nonincumbents on Big Five measures found moderate standardized mean differences for Emotional Stability (d = 0.45) and Conscientiousness (d = 0.45); those effects were even larger for studies with scales explicitly intended to measure the Big Five (Birkeland, Manson, Kisamore, Brannick, & Smith, 2006). In general, within-subject designs appear to have smaller effects than betweensubjects designs, but this conclusion must be qualified because between-subjects designs often compare applicants with incumbents, and within-subject designs do not compare these roles. Future research should continue to investigate differences between conditions of personality testing in a systematic manner, investigating lab versus field settings, role of the test taker, and strength of the motivational context.

Test-score faking can also be understood through Monte Carlo studies, in which the practical effects of known conditions can be investigated through simulated data. Two such studies have manipulated parameters relevant to personality test-score faking, such as the correlations between faking, personality, and performance outcomes; the selection ratio; and the prevalence and magnitude of faking. The simulations operationalized faking either as an inflated score on a personality measure (Komar, Brown, Komar, & Robie, 2008) or as a high score on a social desirability measure that serves as a flag to invalidate an applicant’s score (Schmitt & Oswald, 2006). Together, the results suggested that under a wide array of realistic applicant scenarios, faking does not tend to affect the criterion-related validity of personality tests or the mean levels of performance in those selected in any material way. Nonetheless, there are likely to be circumstances in which the effects of faking personality tests will alter the rank ordering of applicants in a top-down selection process. Although research has indicated that faking scales or lie scales do not reliably and sensitively determine who is prevaricating and who is not, a recent strategy using idiosyncratic patterns of item responses appears to show some promise (Kuncel & Borneman, 2007). Faking scores are developed from subsets of personality items that consistently show multimodal distributions under directed faking, thus creating large differences between faked and honest response distributions that are reflected in the scoring key. These scores show low correlations with cognitive ability; however, there is other evidence that cognitive ability may also improve people’s faking ability on some personality tests (Christiansen, Burns, Montgomery, 2005; Vasilopoulos, Cucina, Dyomina, Morewitz, & Reilly, 2006), and some people can avoid detection on lie scales and social desirability scales through coaching (Hurtz & Alliger, 2002). However, people with higher ability and who are coachable may tend to be hired anyway on the basis of those characteristics, even if they can be coached to appear a more desirable candidate on a personality test. Further understanding of the effects of faking and coaching will require more thinking on multiple fronts: theory development, measurement issues, and ethical 167

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

implications—and perhaps, above all, criterionrelated validity. Instead of attempting to detect faking after personality tests are administered, we recommend considering proactive approaches to preventing or reducing the likelihood of faking. Warnings may reduce the scores of those who would have inflated their scores—yet they also have the potential to reduce the scores of those who were responding honestly in the first place but were anxious about the warning. Perhaps the framing of a warning moderates this effect: a negative warning that dishonest responding will invalidate personality test scores, versus a positive warning providing encouragement that honest responding leads to usable scores, may have differential effects on mean scores and potentially on validity (Converse et al., 2008). Warning about response verification has also been shown to influence the measurement of conscientiousness such that it correlates more highly with cognitive ability (Vasilopoulos, Cucina, & McElreath, 2005), presumably because the thought required to respond carefully affects responses to the conscientiousness items. Warnings could also be provided during the test, so that test takers are alerted whenever their responses appear unusual (e.g., inconsistent with typical responders, randomly patterned, or too extreme). An integrated psychological model of faking that includes warnings has received empirical support; it identifies attitudes toward faking, perceived behavioral control, and group norms for faking as variables that predict the intent to fake, which in turn predicts faking outcomes, with warnings and the ability to fake as important moderators (McFarland & Ryan, 2000, 2006). This model of faking could be productively extended into a broader model of factors that induce or reduce personality test scores, influence their psychometric reliability and factor structure, and ultimately affect criterion-related validities (see Hough & Oswald, 2008). Such a broad model could incorporate variables that may affect the extent to which job applicants would tend to inflate their scores, such as test-taker understanding of how the personality test is used within a larger personnel selection process; the test taker’s own desire for person–job fit; the desirability of the job; the tightness or loose168

ness of the job sector’s labor market when applying; the test format (e.g., ipsative–normative, long form–short form, paper-and-pencil–Internet); testtaker individual differences (e.g., general and specific test experience [or test anxiety], reading ability, and impression management); and the strength of a perceived norm for faking. A recent model of faking incorporates many of the aforementioned individual differences and contextual factors that influence the ability and motivation to fake one’s response to a personality item (Goffin & Boyd, 2009). PROCESS MODELS: PUTTING IT ALL TOGETHER It is clear from our review up to this point that contemporary organizational scholars have made meaningful advances in understanding how personality relates to important workplace behaviors and the practical issues associated with personality assessment. A critical and perennial challenge of personality research in I/O psychology is addressing or reconciling the dual purposes of practical prediction (developing personality measures that correlate with organizational outcomes of importance) and theoretical explanation (identifying personality constructs and testing psychological models that account for those predictive relationships). To address this challenge, we review important theoretical models found in the literature that we call process models. Generally speaking, process models are conceptualized such that (a) personality traits predict performance outcomes directly; (b) this relationship is mediated (partially explained) by narrower and more proximal variables (e.g., motivation; Judge & Ilies, 2002; Kanfer, 1990) and is moderated by organizational characteristics (e.g., team structure, leadership style); (c) the model is multilevel, with the impact of personality occurring at the individual, team, unit, and organizational levels (Tett & Burnett, 2003); and (d) the process of influence occurs over time. Figure 5.2 is an example of a process model that integrates prediction and explanation. It incorporates goal setting, motivation (see Vol. 3, chap. 3, this handbook), goal orientation, episodic performance behaviors, and revisions of goals and behaviors over time. Relationships in this model may be enhanced or

Copyright American Psychological Association. Not for further distribution.

Personality and Its Assessment

FIGURE 5.2. Prediction and explanation: The “black box” of personality–performance validities.

diminished in the presence of key features of the work situation, whether those are concrete features, such as the physical arrangement of the workplace or the equipment used, or psychological features, such as organizational and team climate and perceived norms. We provide this example for two major reasons. First, it illustrates the reality that criterionrelated validities of personality variables, often relied on for personnel selection purposes, are necessarily removed from the “black box” of theory containing the more detailed psychological processes that employees undergo over time as they interact with supervisors, teams, and the organizational climate and culture. Data that describe these processes that are more proximal to performance outcomes simply are not available for an applicant at the point of selection. Instead, they often have to be inferred from measures of more distal constructs such as personality. Second, models such as this reinforce the reality that although organizational research on personality measures has largely focused on personnel selection settings with cross-sectional research

designs, additional research efforts could investigate the longitudinal process of how employee personality and motivation predict individual and organizational outcomes over time. We discuss this latter point shortly. There are many existing psychological models that fit our description of process models. One influential model reflects the influence of personality (achievement, anxiety, impulsivity) on motivation, which in turn affects cognitive task performance (Humphreys & Revelle, 1984). Another model of individual differences proposes that science–math, intellectual–cultural, clerical–conventional, and social “trait complexes” have their basis in personality, motivation, and interests and are a critical mediating or explanatory influence on the relationship between intellectual ability and occupational knowledge acquisition (process, personality, interests, and knowledge theory by Ackerman, 1996; see also investment theory by Cattell, 1971). A third integrated model proposes psychological relationships more specific to organizations, where personality, 169

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

ability, experience, and organizational context are a collection of distal characteristics influencing constructs that are more proximal to work performance, such as stress, autonomy, self-efficacy, and goal setting (Johnson & Hezlett, 2008). A final example is a model and theory of training motivation, where personality and situational variables are exogenous predictors of motivation and knowledge, which in turn predict training performance and ultimately job performance (Colquitt, LePine, & Noe, 2000). These models are all important variations on the broader model we provide in Figure 5.2. It may be impractical for empirical studies to test process models such as these in their entirety, or at least in as intensive or extensive a longitudinal manner as a researcher might find ideal, but research can provide empirical evidence (or lack thereof) for various parts of the model. A handful of examples is illustrative: One study investigated the conscientiousness–performance relationship in a sample of restaurant employees, finding that this relationship was dependent on high levels of work effort and a positive psychological climate; when either of these was low, the conscientiousness– performance relationship was not evident (Byrne, Stoner, Thompson, & Hochwarter, 2005). Another study found that personality traits influence performance through the types and levels of goals set, as well as through dedicated attention (called mental focus; F. K. Lee, Sheldon, & Turban, 2003). Meta-analyses can provide additional information about moderator effects that were not necessarily tested in the individual studies themselves because studies can be coded for the type of sample, the nature of the work setting, and other personal and situational factors. These relatively detailed models need to stand up against the principles of scientific parsimony. Empirically, simpler models generally are preferred to more complex ones when they fit the data just as well (e.g., given a nonsignificant chi-square difference test between two nested models). Furthermore, researchers should report alternative models that are identical in terms of their overall statistical fit (MacCallum, Wegener, Uchino, & Fabrigar, 1993), particularly when they have different theoretical implications. Conceptually, parsimony also dictates 170

that mediating variables should be psychologically distinct from the personality variables themselves if they are to be included as a meaningful component of a model. For instance, if a work motivation measure has the same content as a measure of conscientiousness except for the tag “at work” at the end of each item, then the finding that work motivation fully mediates the conscientiousness–performance relationship likely offers no meaningful contribution to our understanding of organizational settings. INNOVATIVE PERSONALITY ASSESSMENT METHODS

New Forced-Choice Measures Although forced-choice measures of personality have been in the literature for decades, they have resurfaced in the research literature and in real-world practice settings. The general notion prompting its use is that faking will tend to be reduced in a forced-choice format that requires test takers to endorse one of a pair of equally less desirable options (as suggested by Edwards, 1953). The Army’s Assessment of Individual Motivation was a recent large and resource-intensive effort that led to reexamining the use of forced-choice personality measures in highstakes selection testing as a predictor of attrition, with some promising validation results (Knapp, Heggestad, & Young, 2004). In a similar vein, the U.S. Navy has developed the Navy Computer Adaptive Personality System, a forced-choice computer adaptive measure (see Houston, Borman, Farmer, & Bearden, 2006). The Tailored Adaptive Personality Assessment System (Stark, Drasgow, & Chernyshenko, 2008) measures 23 narrow facets of personality that can be measured through unidimensional or multidimensional forced-choice formats (see chap. 4, this volume). An additional benefit of these three forced-choice measures for practice is their computerized adaptive nature. This means that tests can be shorter because items are only administered until the desired degree of trait precision is achieved (estimated using item response theory). This makes testing time more efficient, and it also reduces item exposure rates, which increases test security. These two advantages of adaptive testing are not limited to forced-choice measures, but they are evident in these new applications.

Copyright American Psychological Association. Not for further distribution.

Personality and Its Assessment

Forced-choice measures are psychometrically challenging because, regardless of the constructs being measured, they lead to negative intercorrelations because of the dependencies between item responses (Hicks, 1970). In other words, choosing one option in a forced-choice format necessarily means not choosing the other option. This dependency is reduced for larger numbers of scales, however (Baron, 1996); furthermore, modern statistical modeling and scoring tools are available to account for this interdependency, although large sample sizes (of at least 450 participants) are required for stable estimation (Stark & Drasgow, 2002; also see Cheung, 2004; Maydeu-Olivares & Böckenholt, 2005). In general, forced-choice personality measures may retain higher levels of criterion-related validity in those high-stakes situations that tend to reduce validity for their Likert-scale counterparts, presumably because the format reduces faking. However, the jury is still out. Studies have demonstrated that forced-choice measures yield validities that are at least no worse than Likert-scale measures (lab samples: Chernyshenko et al., 2009; Converse, Oswald, Imus, Hedricks, Roy, & Butera, 2008), with initial evidence that they may be better (lab samples: Christiansen, Burns, & Montgomery, 2005, and Jackson, Wroblewski, & Ashton, 2000; U.S. Navy sample: Houston et al., 2006, showing some validities in the .40s in predicting supervisory ratings).

Conditional Reasoning Measures Conditional reasoning measures (James, 1998) were developed as an innovative method for assessing personality traits in a manner that indirectly reveals motivational tendencies. A typical conditional reasoning item presents a scenario and then asks the test taker to select a response option that reflects a reasonable conclusion that would follow from it. The critical feature of conditional reasoning items lies in the response options, in that more than one response is logically correct, but the logically correct responses differ in terms of personality-relevant motives. For example, take the item “Why would a coworker comment on your interactions with a customer?” There may be two logical responses (among other distracters) that test takers usually endorse: “to provide developmental information to

improve my sales performance” and “to criticize me while making him or her look better.” If this were a conditional reasoning measure of aggression, then it is the latter response that gets scored, where test takers receive 1 point if they endorse it and 0 points otherwise. In fact, conditional reasoning measures of aggression have been developed on the basis of an in-depth theoretical model (James, McIntyre, Glisson, Bowler, & Mitchell, 2004; James et al., 2005). Both self-report and conditional reasoning measures of aggression interact to predict counterproductive and prosocial behavior criteria (Bing et al., 2007) as well as aggressive physical behaviors in basketball (Frost, Ko, & James, 2007). One general issue with conditional reasoning measures is that once the nature of the measure is revealed to test takers, research has shown that they can easily identify items that reflect the construct of interest (at least for aggression). In this case, even if test takers knew about the nature of the test yet remained committed to responding logically, they may still respond differently, and there could be greater potential for faking (e.g., d ≈ 0.40 in LeBreton, Barksdale, Robin, & James, 2007). In line with the conditional reasoning approach to personality measurement, response options to situational judgment tests can also be designed so that personality traits are indirectly measured across hypothetical situations presented to test takers (e.g., Westring et al., 2009; see chap. 8, this volume), but once again, if test takers can identify which traits are being measured in the response options, that may alter their responses and thus the psychometric properties of the measure. Future research in this area should continue to address testscore faking issues and develop measures of other personality constructs that might be better measured indirectly or where indirect measures increment validity over that of their explicit counterparts.

Third-Party Ratings of Personality Employees assume three simultaneous roles in their working lives, all of which are influenced by personality: actor (one who behaves), striver (one who plans his or her behavior), and narrator (one who self-reflects on his or her behavior and creates a self-identity; see McAdams & Olson, in press). 171

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

Personality-relevant information pertaining to all three roles can be gleaned from knowledgeable others (e.g., peers, supervisors, subordinates, customers), and some third-party ratings have been shown to yield higher validities in predicting work performance. A recent meta-analysis (Connelly, 2008) has shown that interrater reliability is generally high for the Big Five traits and highest for extraversion and conscientiousness, which is sensible given that these traits are often more visible than traits such as openness or emotional stability that may require observers to have more extensive contact or a deeper acquaintanceship to understand the person better. Perhaps the most practically important finding in this meta-analytic work is that “other” ratings of personality (e.g., ratings by coworkers, supervisors, and subordinates) also yielded significantly higher criterion-related validities than self-ratings for predicting job performance. We strongly encourage continued research on the validity of third-party ratings of personality traits, examining facet-level ratings, and ensuring that higher correlations are because of personality– performance relationships and not because of raters and supervisors sharing a mental model that is not related to actual performance (e.g., shared liking or disliking the ratee independent of performance; Lefkowitz, 2000). Certainly, distorting effects of third-party ratings may arise just as much as they do in the self-report context, although perhaps for different reasons. But fundamentally, to have any hope of being accurate and valid, third-party ratings require that personality-relevant information about a person be available to the rater, that it be perceived, and that perception be accurate and translate into an accurate judgment (Funder, 1995). Given the promising findings for higher criterionrelated validities than are typically found in selfreport measures of personality, future research should work toward understanding third-party cognitive, affective, and relationship biases when judging personality traits, determining whether those biases are helpful or harmful to criterion-related validity, and investigating what situation manipulations or incentives will provide more veridical or valid personality ratings. 172

PROMISING RESEARCH DIRECTIONS IN PERSONALITY ASSESSMENT

Within-Person Structures of Personality One avenue of promise in our understanding of personality constructs will come from research comparing intraindividual trends and covariances. Decades of organizational research on personality and criterion-related validity has focused primarily on correlations between typical characteristics of people, where scores reflect people’s trait level (e.g., conscientiousness), both implicit and averaged, on a criterion (e.g., performance) over time. In addition to the average, people also show reliable and important levels of variability associated with their overall trait level (Fleeson, 2001); we reviewed, for example, longitudinal studies indicating mean age differences in personality over time. People also appear to show reliable differences in patterns of covariances as well. Repeated-measures personality data have demonstrated that intraindividual factor structures of personality can be very different from the factor structure found across individuals (Molenaar & Campbell, 2009). For example, the FFM or HEXACO factors may be represented in a set of personality items across people, and at the same time these items may support more factors for one person, yet fewer factors for another. Only under rigorous mathematical conditions that one generally would not expect from the data (measurement invariance across people and over time) would the within-person and betweenpersons structures be equivalent. Therefore, organizational research and practice can benefit by extending traditional between-person analyses to within-person analyses that investigate which traits are more (or less) malleable over time given specific combinations of work situations and employees (Tett & Burnett, 2003). When they are reliable and take sampling error variance into account, within-person analyses can be a rich bottom-up approach to understanding people and situations more fully than assuming similar variable interrelationships and personality–situation interactions in the population as a whole (Cervone, 2005). Distinctions between dimensions of personality and personality types may emerge from such analyses (Ruscio & Ruscio, 2008), where types may be tied to advances in the biological underpinnings of per-

Copyright American Psychological Association. Not for further distribution.

Personality and Its Assessment

sonality and temperament (Cloninger, 2008). Historically, longitudinal research examining individual lives over time has yielded a wealth of useful within-person information in the general domains of human development (Block, 1971) and managerial work (Bray, Campbell, & Grant, 1974). Likewise, we hope that future researchers will invest their resources into well-planned intensive longitudinal projects that investigate the effects of personality on task and contextual performance over the spans of years, jobs, and careers. Few within-person analyses of this sort have appeared in the organizational literature to date. However, several multilevel models have examined the within-person question of whether personality traits predict individuals’ job performance level and rate of change over time (Ployhart, Weekley, & Baughman, 2006; Stewart & Nandkeolyar, 2006; Thoresen, Bradley, Bliese, & Thoresen, 2004; Wallace & Chen, 2006). Similar analyses have been conducted in the domain of skill acquisition and learning performance (Chen, Gully, Whiteman, & Kilcullen, 2000; Yeo & Neal, 2004) and could serve as models for examining whether personality predicts trajectories of learning outcomes in relatively autonomous settings such as self-paced Web-based training and active learning environments (Bell & Kozlowski, 2008; Ford & Oswald, 2003; Gully, Payne, Koles, & Whiteman, 2002). Because longitudinal designs such as these operate on the traditional assumption that trait measures are enduring characteristics of people, it is important to determine the long-term reliability of trait measures. Reliance on Cronbach’s alpha as an indicator of reliability has long been known to be generally problematic (see Sijtsma, 2009), but it is particularly problematic for trait measurement because alpha is only a rough indicator of how well scale items covary with one another; alpha does not indicate the stability of the scale score over time. Appropriate psychometric models of trait reliability distinguish between actual changes in the trait over time (“true change”) and changes resulting from random error over time (“transient error”). Appropriate reliability estimates (i.e., coefficients of reliability and stability) tend to indicate meaningful amounts of variance resulting from transient error, and therefore Cronbach’s alpha

tends to be an overestimate of the reliability of personality traits (e.g., by 5%–13% for conscientiousness in F. L. Schmidt, Le, & Ilies, 2003; by 25% across traits in Chmielewski & Watson, 2009, with wide variation by trait).

Understanding the Situation In addition to understanding the person through within-person factor analyses and multilevel models, we also need to have a greater understanding of the situations in which personality is more or less predictive of important work outcomes. This understanding has been only roughly described as strong and weak situational cues for behavior, with the notion that stronger cues (e.g., a micromanaging supervisor, peer pressure from teammates) will tend to dictate an employee’s behavior, making personality less potent of a predictor than when cues are weak (e.g., an autonomous work environment; see Ackerman, Kanfer, & Goff, 1995; Beaty, Cleveland, & Murphy, 2001; Klehe & Anderson, 2007; Ployhart, Lim, & Chan, 2001). Strong situational cues may be real or perceived; when they are perceived, then maximally motivated performance may emerge, whereas when cues are perceived as weaker, typical performance results. A typical transition from maximal to typical performance is when new hires are in the initial “honeymoon” period of a job, motivated to perform at their maximum, but then transition into more typical performance, when personality is more likely to emerge as an influence on behavior (Helmreich, Sawin, & Carsrud, 1986). Although evidence for the situational-strength hypothesis could benefit from more carefully conducted research (Cooper & Withey, 2009), situations might never be classified much further than strong versus weak because they are best defined by examining the complex goals (and constraints) within them. Employees in the same organization or on the same team likely have shared goals, but some of those goals are more clearly communicated, perceived, and shared than others; furthermore, there are personal goals held by each employee (Bagozzi, Bergami, & Leone, 2003), and still other goals are negotiated between organizations and teams or employees. Additionally, personal and organizational goals are dynamic; employees evaluate and revise the multiple goals they 173

Copyright American Psychological Association. Not for further distribution.

Oswald and Hough

hold over the course of time, such as when feedback is obtained or a deadline is fast approaching (A. M. Schmidt, Dolis, & Tolli, 2009). Despite this real-world complexity, the hope is that variance in work behavior can be partially understood by examining how individual differences in personality interact with situational characteristics at the individual, team, and organizational levels that facilitate or constrain individual behavior and team performance (Pervin, 1989; Stewart, Fulmer, Barrick, 2005; Tett & Burnett, 2003). Studying important situations is likely to be more productive than developing an abstract theory of situations (Funder, 2009). Personality traits likely have an influence on the selection, evaluation, revision, and pursuit of goals. For instance, leaders who define organizational goals clearly (a strong situation) may help conscientious workers direct their energies toward those goals (Colbert & Witt, 2009). Over a broader time scale, individuals may self-select their goals, in which case conscientiousness and emotional stability have been found to predict the goals of intrinsic career success (e.g., satisfaction) and extrinsic career success (e.g., job position, income), above and beyond general cognitive ability (Judge, Higgins, Thoresen, & Barrick, 1999). These latter findings may be more relevant in individualistic cultures such as that of the United States, where career success is highly valued. There are also personal goals (self-esteem) and interpersonal goals (affiliation; DeShon & Gillespie, 2005) that may likely be more important in other cultures. In general, cultural and multicultural factors need to have an increasing influence in models of personality and work behavior in today’s global economy (Gelfand, Leslie, & Fehr, 2008). In addition to national culture, the culture of organizations may also create relatively strong or weak situations, and it may emphasize or facilitate some personality characteristics more than others (Schneider, 1987). One recent study identified how employees higher in extraversion are more influential in a team-driven organization and, by contrast, how conscientious individuals benefitted most in organizations in which the work was largely done alone (Anderson, Spataro, & Flynn, 2008). 174

CONCLUSION Personality assessment has accumulated a noteworthy history in I/O psychology, yet researchers and practitioners alike continue to challenge and explore the core assumptions and boundary conditions related to the nature of personality and its measurement. Both personality theory and research that advances organizational practice tend to address time-honored questions such as the following: What is the most useful level of refinement for measuring personality from the perspectives of reliability and validity (how broad is too broad, and how narrow is too narrow)? Just how stable are personality traits over an employee’s life span, and at what critical points in time are personality traits most (and least) predictive of employee behaviors? Can personality predict inter- and intraindividual changes in behavior over time? How do subgroups such as age, race, and culture and team composition moderate validities? These questions will never have definitive answers. Instead, it is the neverending pursuit of those answers that contributes to the progress and promise of personality assessment in organizations.

References Ackerman, P. L. (1996). A theory of adult intellectual development: Process, personality, interests and knowledge. Intelligence, 22, 227–257. doi:10.1016/ S0160-2896(96)90016-1 Ackerman, P. L., Kanfer, R., & Goff, M. (1995). Cognitive and noncognitive determinants and consequences of complex skill acquisition. Journal of Experimental Psychology: Applied, 1, 270–304. doi:10.1037/1076898X.1.4.270 Age Discrimination in Employment Act of 1967, Pub. L. 90-202, 81 Stat. 602 (29 U.S.C. 621 et seq.). Allemand, M., Zimprich, D., & Hendriks, A. A. J. (2008). Age differences in five personality domains across the life span. Developmental Psychology, 44, 758–770.doi:10.1037/0012-1649.44.3.758 Allport, G. W., & Odbert, H. S. (1936). Trait names: A psycholexical study. Psychological Monographs, 47, 1–171. American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (1999). Standards for educational and psychological testing. Washington, DC: American Educational Research Association.

Personality and Its Assessment

Americans With Disabilities Act of 1990, Pub. L. 101–336, 42 U.S.C. § 12101. Anderson, C., Spataro, S. E., & Flynn, F. J. (2008). Personality and organizational culture as determinants of influence. Journal of Applied Psychology, 93, 702–710.

Copyright American Psychological Association. Not for further distribution.

Ashton, M. C. (1998). Personality and job performance: The importance of narrow traits. Journal of Organizational Behavior, 19, 289–303. doi:10.1002/(SICI) 1099-1379(199805)19:33.0.CO; 2-C Ashton, M. C., Jackson, D. N., Helmes, E., & Paunonen, S. V. (1998). Joint factor analysis of the Personality Research Form and the Jackson Personality Inventory: Comparisons with the Big Five. Journal of Research in Personality, 32, 243–250. doi:10.1006/jrpe.1998.2214 Ashton, M. C., & Lee, K. (2001). A theoretical basis for the major dimensions of personality. European Journal of Personality, 15, 327–353. doi:10.1002/per.417 Ashton, M. C., Lee, K., & Goldberg, L. R. (2004). A hierarchical analysis of 1,710 English personalitydescriptive adjectives. Journal of Personality and Social Psychology, 87, 707–721.doi:10.1037/00223514.87.5.707 Ashton, M. C., Lee, K., Goldberg, L. R., & de Vries, R. E. (2009). Higher order factors of personality: Do they exist? Personality and Social Psychology Review, 13, 79–91. doi:10.1177/1088868309338467 Ashton, M. C., Lee, K., Perugini, M., Szarota, P., De Vries, R. E., Di Blas, L., et al. (2004). A six-factor structure of personality-descriptive adjectives: Solutions from psycholexical studies in seven languages. Journal of Personality and Social Psychology, 86, 356–366. doi:10.1037/0022-3514.86.2.356 Ashton, M. C., Lee, K., & Son, C. (2000). Honesty as the sixth factor of personality: Correlations with Machiavellianism, psychopathy, and social adroitness. European Journal of Personality, 14, 359–368. doi:10.1002/1099-0984(200007/08)14:43.0.CO;2-Y Bagozzi, R. P., Bergami, M., & Leone, L. (2003). Hierarchical representation of motives in goal setting. Journal of Applied Psychology, 88, 915–943. doi:10.1037/0021-9010.88.5.915 Baron, H. (1996). Strengths and limitations of ipsative measurement. Journal of Occupational and Organizational Psychology, 69, 49–56. Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A metaanalysis. Personnel Psychology, 44, 1–26. doi:10.1111/ j.1744-6570.1991.tb00688.x Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and performance at the beginning of the new millennium: What do we know and where do

we go next? International Journal of Selection and Assessment, 9, 9–30. doi:10.1111/1468-2389.00160 Barrick, M. R., Mount, M. K., & Strauss, J. P. (1993). Conscientiousness and performance of sales representatives: Test of the mediating effects of goal setting. Journal of Applied Psychology, 78, 715–722. doi:10.1037/0021-9010.78.5.715 Barrick, M. R., Stewart, G. L., & Piotrowski, M. (2002). Personality and job performance: Test of the mediating effects of motivation among sales representatives. Journal of Applied Psychology, 87, 43–51. doi:10.1037/ 0021-9010.87.1.43 Bartram, D. (2005). The Great Eight competencies: A criterion-centric approach to validation. Journal of Applied Psychology, 90, 1185–1203. doi:10.1037/0021-9010. 90.6.1185 Baumeister, R. F., Gailliot, M., DeWall, C., & Oaten, M. (2006). Self-regulation and personality: How interventions increase regulatory success, and how depletion moderates the effects of traits on behavior. Journal of Personality, 74, 1773–1802. doi:10.1111/j.14676494.2006.00428.x Beaty, J. C., Cleveland, J. N., & Murphy, K. R. (2001). The relationship between personality and contextual performance in “strong” versus “weak” situations. Human Performance, 14, 125–148. doi:10.1207/ S15327043HUP1402_01 Bell, B. S., & Kozlowski, S. W. J. (2008). Active learning: Effects of core training design elements on selfregulatory processes, learning, and adaptability. Journal of Applied Psychology, 93, 296–316. doi:10.1037/0021-9010.93.2.296 Benet-Martínez, V., & John, O. P. (1998). Los Cinco Grandes across cultures and ethnic groups: Multitrait multimethod analyses of the big five in Spanish and English. Journal of Personality and Social Psychology, 75, 729–750. doi:10.1037/0022-3514.75.3.729 Berry, C. M., Ones, D. S., & Sackett, P. R. (2007). Interpersonal deviance, organizational deviance, and their common correlates: A review and meta-analysis. Journal of Applied Psychology, 92, 410–424. doi:10. 1037/0021-9010.92.2.410 Bijnen, E. J., van der Net, T. Z., & Poortinga, Y. H. (1986). On cross-cultural comparative studies with the Eysenck Personality Questionnaire. Journal of CrossCultural Psychology, 17, 3–16. doi:10.1177/ 0022002186017001001 Bing, M. N., Stewart, S. M., Davison, H. K., Green, P. D., McIntyre, M. D., & James, L. R. (2007). Integrative typology of personality assessment for aggression: Implications for predicting counterproductive workplace behavior. Journal of Applied Psychology, 92, 722–744. doi:10.1037/0021-9010.92.3.722 Birkeland, S. A., Manson, T. M., Kisamore, J. L., Brannick, M. T., & Smith, M. A. (2006). A meta-analytic 175

Oswald and Hough

investigation of job applicant faking on personality measures. International Journal of Selection and Assessment, 14, 317–335. Birkland, A. S., & Ones, D. S. (2006, July). The structure of emotional stability: A meta-analytic investigation. Paper presented at the International Congress of Applied Psychology, Athens, Greece. Block, J. (1971). Lives through time. Berkeley, CA: Bancroft Books.

Copyright American Psychological Association. Not for further distribution.

Block, J. (1995). Going beyond the five factors given: Rejoinder to Costa and McCrae (1995) and Goldberg and Saucier (1995). Psychological Bulletin, 117, 226–229. doi:10.1037/0033-2909.117.2.226 Block, J., & Block, J. H. (2006). Venturing a 30-year longitudinal study. American Psychologist, 61, 315–327. doi:10.1037/0003-066X.61.4.315 Bono, J. E., & Judge, T. A. (2004). Personality and transformational and transactional leadership: A metaanalysis. Journal of Applied Psychology, 89, 901–910. doi:10.1037/0021-9010.89.5.901 Borman, W. C., Penner, L. A., Allen, T. D., & Motowidlo, S. J. (2001). Personality predictors of citizenship performance. International Journal of Selection and Assessment, 9, 52–69. doi:10.1111/1468-2389.00163 Borman, W. C., White, L. A., Pulakos, E. D., & Oppler, S. H. (1991). Models of supervisory job performance ratings. Journal of Applied Psychology, 76, 863–872. doi:10.1037/0021-9010.76.6.863 Bott, J. P., O’Connell, M. S., Ramakrishnan, M., & Doverspike, D. (2007). Practical limitations in making decisions regarding the distribution of applicant personality test scores based on incumbent data. Journal of Business and Psychology, 22, 123–134. doi:10.1007/ s10869-007-9053-x

Psychology, 56, 423–452. doi:10.1146/annurev. psych.56.091103.070133 Chan, D. (2004). Individual differences in tolerance for contradiction. Human Performance, 17, 297–324. doi:10.1207/s15327043hup1703_3 Chen, G., Gully, S. M., Whiteman, J. A., & Kilcullen, R. N. (2000). Examination of relationships among trait-like individual differences, state-like individual differences, and learning performance. Journal of Applied Psychology, 85, 835–847. doi:10.1037/0021-9010.85. 6.835 Chernyshenko, O. S., Stark, S., Prewett, M. S., Gray, A. A., Stilson, F. R., & Tuttle, M. D. (2009). Normative scoring of multidimensional pairwise preference personality scales using IRT: Empirical comparison with other formats. Human Performance, 22, 105–127. doi:10.1080/08959280902743303 Cheung, F. M., Leung, K., Zhang, J.-X., Sun, H.-F., Gan, Y.-Q., Song, W.-Z., & Zie, D. (2001). Indigenous Chinese personality constructs: Is the five-factor model complete? Journal of Cross-Cultural Psychology, 32, 407–433. doi:10.1177/0022022101032004003 Cheung, M. W. L. (2004). A direct estimation method on analyzing ipsative data with Chan and Bentler’s (1993) method. Structural Equation Modeling, 11, 217–243. doi:10.1207/s15328007sem1102_5 Chmielewski, M., & Watson, D. (2009). What is being assessed and why it matters: The impact of transient error on trait research. Journal of Personality and Social Psychology, 97, 186–202. doi:10.1037/a0015618 Christiansen, N. D., Burns, G. N., & Montgomery, G. E. (2005). Reconsidering forced-choice item formats for applicant personality assessment. Human Performance, 18, 267–307. doi:10.1207/s15327043hup1803_4

Bray, D. W., Campbell, R. J., & Grant, D. L. (1974). Formative years in business: A long-term AT&T study of managerial lives. Malabar, FL: Krieger.

Church, A. T. (2000). Culture and personality: Toward an integrated cultural trait psychology. Journal of Personality, 68, 651–703. Medline doi:10.1111/14676494.00112

Byrne, Z. S., Stoner, J., Thompson, K. R., & Hochwarter, W. (2005). The interactive effects of conscientiousness, work effort, and psychological climate on work performance. Journal of Vocational Behavior, 66, 326–338. doi:10.1016/j.jvb.2004.08.005

Church, A. T. (2001). Personality measurement in a cross-cultural perspective. Journal of Personality, 69, 979–1006. Medline doi:10.1111/1467-6494.696172

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait–multimethod matrix. Psychological Bulletin, 56, 81–105. doi:10.1037/ h0046016

Civil Rights Act of 1991, Pub. L. 102-166, Nov. 21, 1991, 105 Stat. 1071.

Civil Rights Act of 1964, Pub. L. 88–352, 42 U.S.C., Ch. 21 § 1981-2000.

Cascio, W., & Boudreau, J. (2008). Investing in people: Financial impact of human resource initiatives. Upper Saddle River, NJ: FT Press.

Cloninger, C. R. (2008). The psychobiological theory of temperament and character: Comment on Farmer and Goldberg (2008). Psychological Assessment, 20, 292–299. doi:10.1037/a0012933

Cattell, R. B. (1971). Abilities: Their structure, growth and action. Amsterdam: North Holland.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Cervone, D. (2005). Personality architecture: Withinperson structures and processes. Annual Review of

Colbert, A. E., & Witt, L. A. (2009). The role of goalfocused leadership in enabling the expression of

176

Personality and Its Assessment

Copyright American Psychological Association. Not for further distribution.

conscientiousness. Journal of Applied Psychology, 94, 790–796. doi:10.1037/a0014187

struct. Journal of Personality and Social Psychology, 75, 989–1015. doi:10.1037/0022-3514.75.4.989

Colquitt, J. A., LePine, J. A., & Noe, R. A. (2000). Toward an integrative theory of training motivation: A metaanalytic path analysis of 20 years of research. Journal of Applied Psychology, 85, 678–707. doi:10.1037/00219010.85.5.678

Davies, S. E., Connelly, B. S., Ones, D. S., & Birkland, A. S. (2008, April). Enhancing the role of extraversion for work related behaviors. Poster session presented at the annual conference of the Society for Industrial and Organizational Psychology, San Francisco.

Connelly, B. S. (2008). The reliability, convergence, and predictive validity of personality ratings: An other perspective. Unpublished doctoral dissertation, University of Minnesota.

De Corte, W., Lievens, P., & Sackett, P. R. (2007). Combining predictors to achieve optimal trade-offs between selection quality and adverse impact. Journal of Applied Psychology, 92, 1380–1393.doi:10.1037/ 0021-9010.92.5.1380

Connelly, B. S., & Ones, D. S. (2007, April). Combining conscientiousness scales: Can’t get enough of the trait, baby. Poster session presented at the annual conference of the Society for Industrial and Organizational Psychology, New York. Connelly, B. S., Davies, S. E., Ones, D. S., & Birkland, A. S. (2008a, April). Agreeableness: A meta-analytic review of structure, convergence, and predictive validity. In J. P. Thomas & C. Viswesvaran (Chairs), Personality in the workplace: Advances in measurement and assessment. Symposium conducted at the annual conference of the Society for Industrial and Organizational Psychology, San Francisco. Connelly, B. S., Davies, S. E., Ones, D. S., & Birkland, A. S. (2008b, January). Opening up openness: A metaanalytic review of measures of the personality construct. Poster session presented at the annual conference of the Society for Personality and Social Psychology, Albuquerque, NM. Converse, P. D., Oswald, F. L., Imus, A., Hedricks, C., Roy, R., & Butera, H. (2008). Comparing personality test formats and warnings: Effects on criterionrelated validity and test-taker reactions. International Journal of Selection and Assessment, 16, 155–169. doi:10.1111/j.1468-2389.2008.00420.x Cooper, W. H., & Withey, M. J. (2009). The strong situation hypothesis. Personality and Social Psychology Review, 13, 62–72. doi:10.1177/1088868308329378 Costa, P. T., & McCrae, R. R. (1995). Domains and facets: Hierarchical personality assessment using the Revised NEO Personality Inventory. Journal of Personality Assessment, 64, 21–50. doi:10.1207/s15327752 jpa6401_2 Costa, P. T., Zonderman, A. B., Williams, R. B., & McCrae, R. R. (1985). Content and comprehensiveness in the MMPI: An item factor analysis in a normal adult sample. Journal of Personality and Social Psychology, 48, 925–933. Dancer, L. J., & Woods, S. A. (2006). Higher-order factor structures and intercorrelations of the 16PF5 and FIRO-B. International Journal of Selection and Assessment, 14, 385–391. Davies, M., Stankov, L., & Roberts, R. D. (1998). Emotional intelligence: In search of an elusive con-

De Raad, B. (1998). Five big, Big Five issues: Rationale, content, structure, status and cross-cultural assessment. European Psychologist, 3, 113–124. De Raad, B., & Hoskens, M. (1990). Personalitydescriptive nouns. European Journal of Personality, 4, 131–146. doi:10.1002/per.2410040206 De Raad, B., Perugini, M., & Szirmak, Z. (1997). In pursuit of a cross-lingual reference structure of personality traits: Comparisons among five languages. European Journal of Personality, 11, 167–185. doi:10.1002/ (SICI)1099-0984(199709)11:3 3.0.CO;2-B DeShon, R. P., & Gillespie, J. Z. (2005). A motivated action theory account of goal orientation. Journal of Applied Psychology, 90, 1096–1127. doi:10.1037/00219010.90.6.1096 Di Blas, L., & Forzi, M. (1999). Refining a descriptive structure of personality attributes in the Italian language. Journal of Personality and Social Psychology, 76, 451–481. doi:10.1037/0022-3514.76.3.451 Digman, J. M., & Takemoto-Chock, M. K. (1981). Factors in the natural language of personality: Re-analysis and comparison of six major studies. Multivariate Behavioral Research, 16, 149–170. doi:10.1207/s15327906mbr1602_2 Digman, J. M. (1997). Higher-order factors of the Big Five. Journal of Personality and Social Psychology, 73, 1246–1256. doi:10.1037/0022-3514.73.6.1246 Donnellan, M. B., & Lucas, R. E. (2008). Age differences in the Big Five across the life span: Evidence from two national samples. Psychology and Aging, 23, 558–566. Dudley, N. M., Orvis, K. A., Lebiecki, J. E., & Cortina, J. M. (2006). A meta-analytic investigation of Conscientiousness in the prediction of job performance: Examining the intercorrelations and the incremental validity of narrow traits. Journal of Applied Psychology, 91, 40–57. doi:10.1037/00219010.91.1.40 Duehr, E. E., Jackson, H. L., & Ones, D. S. (2003, April). Gender differences in conscientiousness and extraversion: A multiple-culture, multiple-instrument meta-analysis. 177

Oswald and Hough

Copyright American Psychological Association. Not for further distribution.

Poster presented at the annual conference of the Society for Industrial and Organizational Psychology, Orlando, FL.

Funder, D. C. (2009). Naïve and obvious questions. Perspectives on Psychological Science, 4, 340–344. doi:10.1111/j.1745-6924.2009.01135.x

Edwards, A. L. (1953). The relationship between the judged desirability of a trait and the probability that it will be endorsed. Journal of Applied Psychology, 37, 90–93. doi:10.1037/h0058073

Gelfand, M. J., Leslie, L. M., & Fehr, R. (2008). To prosper, organizational psychology should . . . adopt a global perspective. Journal of Organizational Behavior, 29, 493–517. doi:10.1002/job.530

Ellingson, J. E., Sackett, P. R., & Connelly, B. S. (2007). Personality assessment across selection and development contexts: Insights into response distortion. Journal of Applied Psychology, 92, 386–395. doi:10. 1037/0021-9010.92.2.386

Glaros, A. G., & Kline, R. B. (1988). Understanding the accuracy of tests with cutting scores: The sensitivity, specificity, and predictive value model. Journal of Clinical Psychology, 44, 1013–1023.doi:10.1002/10974679(198811)44:63.0.CO;2-Z

Elliot, A. J., & Church, M. A. (1997). A hierarchical model of approach and avoidance achievement motivation. Journal of Personality and Social Psychology, 72, 218–232. doi:10.1037/0022-3514. 72.1.218 Erez, A., & Judge, T. A. (2001). Relationship of core selfevaluations to goal setting, motivation, and performance. Journal of Applied Psychology, 86, 1270–1279. doi:10.1037/0021-9010.86.6.1270 Eysenck, H. J. (1947). Dimensions of personality. London: Routledge & Kegan Paul. Eysenck, H. J., & Eysenck, M. W. (1985). Personality and individual differences: A natural science approach. New York: Plenum Press. Feist, G. J. (1998). A meta-analysis of personality in scientific and artistic creativity. Personality and Social Psychology Review, 2, 290–309. doi:10.1207/ s15327957pspr0204_5 Fleeson, W. (2001). Toward a structure- and processintegrated view of personality: Traits as density distribution states. Journal of Personality and Social Psychology, 80, 1011–1027. doi:10.1037/00223514.80.6.1011 Foldes, H. J., Duehr, E. E., & Ones, D. S. (2008). Group differences in personality: Meta-analyses comparing five U.S. racial groups. Personnel Psychology, 61, 579–616. doi:10.1111/j.1744-6570.2008.00123.x Ford, J. K., & Oswald, F. L. (2003). Understanding the dynamic learner: Linking personality traits, learning situations, and individual behavior. In M. R. Barrick & A. M. Ryan (Eds.), Personality and work: Reconsidering the role of personality in organizations (pp. 229–261). San Francisco: Jossey-Bass. Frost, B. C., Ko, C.-H. E., & James, L. R. (2007). Implicit and explicit personality: A test of a channeling hypothesis for aggressive behavior. Journal of Applied Psychology, 92, 1299–1319.doi:10.1037/00219010.92.5.1299 Funder, D. C. (1995). On the accuracy of personality judgment: A realistic approach. Psychological Review, 102, 652–670.doi:10.1037/0033-295X.102. 4.652 178

Goffin, R. D., & Boyd, A. C. (2009). Faking and personality assessment in personnel selection: Advancing models of faking. Canadian Psychology, 50, 151–160. Goldberg, L. R. (1990). An alternative “description of personality:” The Big Five factor structure. Journal of Personality and Social Psychology, 59, 1216–1229. doi:10.1037/0022-3514.59.6.1216 Goldberg, L. R. (1993). The structure of phenotypic personality traits. American Psychologist, 48, 26–34. doi:10.1037/0003-066X.48.1.26 Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe (Vol. 7, pp. 7–28). Tilburg, the Netherlands: Tilburg University Press. Goldberg, L. R., & Saucier, G. (1998). What is beyond the Big Five? Journal of Personality, 66, 495–524. doi:10.1111/1467-6494.00022 Goldberg, L. R., & Velicer, W. F. (2006). Principles of exploratory factor analysis. In S. Strack (Ed.), Differentiating normal and abnormal personality (2nd ed., pp. 209–237). New York: Springer Publishing Company. Gough, H. G. (1968). The Chapin Social Insight Test manual. Palo Alto, CA: Consulting Psychologists Press. Grucza, R. A., & Goldberg, L. R. (2007). The comparative validity of 11 modern personality inventories: Prediction of behavioral acts, informant reports, and clinical indicators. Journal of Personality Assessment, 89, 167–187. doi:10.1002/pmh.3 Guion, R. M., & Gottier, R. F. (1965). Validity of personality measures in personnel selection. Personnel Psychology, 18, 135–164. doi:10.1111/j.1744-6570. 1965.tb00273.x Gully, S. M., Payne, S. C., Koles, K. L. K., & Whiteman, J. A. K. (2002). The impact of error training and individual differences on training outcomes: An attribute–treatment interaction perspective. Journal of Applied Psychology, 87, 143–155. doi:10.1037/ 0021-9010.87.1.143

Personality and Its Assessment

Helmreich, R. L., Sawin, L. L., & Carsrud, A. L. (1986). The honeymoon effect in job performance: Temporal increases in the predictive power of achievement motivation. Journal of Applied Psychology, 71, 185–188. doi:10.1037/0021-9010.71.2.185 Hicks, L. E. (1970). Some properties of ipsative, normative, and forced-choice normative measures. Psychological Bulletin, 74, 167–184.

Copyright American Psychological Association. Not for further distribution.

Hogan, J., Barrett, P., & Hogan, R. (2007). Personality measurement, faking, and employment selection. Journal of Applied Psychology, 92, 1270–1285. doi:10.1037/0021-9010.92.5.1270 Hogan, J., & Holland, B. (2003). Using theory to evaluate personality and job-performance relations: A socioanalytic perspective. Journal of Applied Psychology, 88, 100–112. doi:10.1037/0021-9010.88.1.100 Hogan, J., & Roberts, B. W. (1996). Issues and non-issues in the fidelity-bandwidth trade-off. Journal of Organizational Behavior, 17, 627–637. Hogan, R. (1969). Development of an empathy scale. Journal of Consulting and Clinical Psychology, 33, 307–316. doi:10.1037/h0027580 Hogan, R. (1983). A socioanalytic theory of personality. In M. M. Page (Ed.), Nebraska Symposium on Motivation: Vol. 30. Personality—Current theory and research (pp. 336–355). Lincoln: University of Nebraska Press. Holland, J. L. (1985). Making vocational choices: A theory of careers. Englewood Cliffs, NJ: Prentice-Hall. Hopwood, C. J., & Donnellan, M. B. (2009). Assessing the internal structure of personality inventories. Unpublished manuscript. Hough, L. M. (1989). Development of personality measures to supplement selection decisions. In B. J. Fallon, H. P. Pfister, & J. Brebner (Eds.), Advances in industrial organizational psychology (pp. 365–375). Holland: Elsevier Science. Hough, L. M. (1992). The “Big Five” personality variables—Construct confusion: Description versus prediction. Human Performance, 5, 139–155. doi:10.1207/s15327043hup0501&2_8 Hough, L. M. (1997). The millennium for personality psychology: New horizons or good old daze. Applied Psychology, 47, 233–261.

Hough, L. M., Eaton, N. K., Dunnette, M. D., Kamp, J. D., & McCloy, R. A. (1990). Criterion-related validities of personality constructs and the effects of response distortion on those validities. Journal of Applied Psychology, 75, 581–595. doi:10.1037/0021-9010. 75.5.581 Hough, L. M., & Furnham, A. (2003). Importance and use of personality variables in work settings. In I. B. Weiner (Ed. in Chief) & W. Borman, D. Illgen, & R. Klimoski (Vol. Eds.), Comprehensive handbook of psychology. Vol. 12: Industrial & organizational psychology (pp. 131–169). New York: Wiley. Hough, L. M., & Ones, D. S. (2001). The structure, measurement, validity, and use of personality variables in industrial, work, and organizational psychology. In N. Anderson, D. S. Ones, H. K. Sinangil, & C. Viswesvaran (Eds.), International handbook of work and organizational psychology (pp. 233–377). London: Sage. Hough, L. M., Ones, D. S., & Viswesvaran, C. (1998, April). Personality correlates of managerial performance constructs. In R. Page (Chair), Personality determinants of managerial potential, performance, progression and ascendancy. Symposium conducted at the 13th Annual Conference of the Society for Industrial and Organizational Psychology, Dallas, TX. Hough, L. M., & Oswald, F. L. (2000). Personnel selection: Looking toward the future—Remembering the past. Annual Review of Psychology, 51, 631–664. doi:10.1146/annurev.psych.51.1.631 Hough, L. M., & Oswald, F. L. (2005). They’re right, well . . . mostly right: Research evidence and an agenda to rescue personality testing from 1960s insights. Human Performance, 18, 373–387. doi:10.1207/s15327043hup1804_4 Hough, L. M., & Oswald, F. L. (2008). Personality testing and industrial–organizational psychology: Reflections, progress, and prospects. Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 272–290. doi:10.1111/j.1754-9434.2008.00048.x Hough, L. M., Oswald, F. L., & Ployhart, R. E. (2001). Determinants, detection, and amelioration of adverse impact in personnel selection procedures: Issues, evidence, and lessons learned. International Journal of Selection and Assessment, 9, 152–194. doi:10.1111/ 1468-2389.00171

Hough, L. M. (1998). Personality at work: Issues and evidence. In M. Hakel (Ed.), Beyond multiple choice: Evaluating alternatives to traditional testing for selection (pp. 131–166). Hillsdale, NJ: Erlbaum.

Hough, L. M., & Schneider, R. J. (1996). Personality traits, taxonomies, and applications in organizations. In K. Murphy (Ed.), Individual differences and behavior in organizations (pp. 31–88). San Francisco: Jossey-Bass.

Hough, L., & Dilchert, S. (2007). Inventors, innovators, and their leaders: Selecting for conscientiousness will keep you “inside the box” (SIOP Leading Edge Consortium). Kansas City, MO: Society for Industrial and Organizational Psychology.

Houston, J. S., Borman, W. C., Farmer, W. F., & Bearden, R. M. (2006). Development of the Navy Computer Adaptive Personality Scales (NCAPS) (NPRST-TR06-2). Millington, TN: Navy Personnel, Research, Studies, & Technology. 179

Oswald and Hough

Humphreys, M. S., & Revelle, W. (1984). Personality, motivation, and performance: A theory of the relationship between individual differences and information processing. Psychological Review, 91, 153–184. doi:10.1037/0033-295X.91.2.153 Hurtz, G. M., & Alliger, G. M. (2002). Influence of coaching on integrity test performance and unlikely virtues scales. Human Performance, 15, 255–273. doi:10.1207/S15327043HUP1503_02

Copyright American Psychological Association. Not for further distribution.

Hurtz, G. M., & Donovan, J. J. (2000). Personality and job performance: The Big Five revisited. Journal of Applied Psychology, 85, 869–879. doi:10.1037/00219010.85.6.869 Jackson, D. N., Wroblewski, V. R., & Ashton, M. C. (2000). The impact of faking on employment tests: Does forced-choice offer a solution? Human Performance, 13, 371–388. doi:10.1207/S15327043HUP1304_3 James, L. R. (1998). Measurement of personality via conditional reasoning. Organizational Research Methods, 1, 131–163. doi:10.1177/109442819812001

Judge, T. A., & Hurst, C. (2007). Capitalizing on one’s advantages: Role of core self-evaluations. Journal of Applied Psychology, 92, 1212–1227. doi:10.1037/00219010.92.5.1212 Judge, T. A., & Ilies, R. (2002). Relationship of personality to performance motivation: A meta-analytic review. Journal of Applied Psychology, 87, 797–807. doi:10.1037/0021-9010.87.4.797 Kanfer, R. (1990). Motivation theory and industrial/ organizational psychology. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of industrial and organizational psychology. Volume 1: Theory in industrial and organizational psychology (pp. 75–170). Palo Alto, CA: Consulting Psychologists Press. Kanfer, R., & Heggestad, E. D. (1999). Individual differences in motivation: Traits and self-regulatory skills. In P. L. Ackerman & P. C. Kyllonen (Eds.), Learning and individual differences: Process, trait, and content determinants (pp. 293–313). Washington, DC: American Psychological Association.

James, L. R., McIntyre, M. D., Glisson, C. A., Bowler, J. L., & Mitchell, T. R. (2004). The Conditional Reasoning Measurement System for aggression: An overview. Human Performance, 17, 271–295. doi:10.1207/s15327043hup1703_2

Katigbak, M. S., Church, A. T., & Akamine, T. X. (1996). Cross-cultural generalizability of personality dimensions: Relating indigenous and imported dimensions in two cultures. Journal of Personality and Social Psychology, 70, 99–114. doi:10.1037/0022-3514.70. 1.99

James, L. R., McIntyre, M. D., Glisson, C. A., Green, P. D., Patton, T. W., LeBreton, J. M., et al. (2005). A conditional reasoning measure for aggression. Organizational Research Methods, 8, 69–99. doi:10.1177/1094428104272182

Klehe, U.-C., & Anderson, N. (2007). Working hard and working smart: Motivation and ability during typical and maximum performance. Journal of Applied Psychology, 92, 978–992. doi:10.1037/0021-9010.92. 4.978

John, O. P., Hampson, S. E., & Goldberg, L. R. (1991). The basic level in personality-trait hierarchies: Studies of trait use and accessibility in different contexts. Journal of Personality and Social Psychology, 60, 348–361. doi:10.1037/0022-3514.60.3.348

Klein, H. J., & Lee, S. (2006). The effects of personality on learning: The mediating role of goal setting. Human Performance, 19, 43–66. doi:10.1207/ s15327043hup1901_3

Johnson, J. W., & Hezlett, S. A. (2008). Modeling the influence of personality on individuals at work: A review and research agenda. In S. Cartwright & C. L. Cooper (Eds.), Oxford handbook of personnel psychology (pp. 59–92). Oxford, England: Oxford University Press. Judge, T. A., Bono, J. E., Ilies, R., & Gerhardt, M. W. (2002). Personality and leadership: A qualitative and quantitative review. Journal of Applied Psychology, 87, 765–780. doi:10.1037/0021-9010.87.4.765 Judge, T. A., Heller, D., & Mount, M. K. (2002). Five-factor model of personality and job satisfaction: A metaanalysis. Journal of Applied Psychology, 87, 530–541. Medline doi:10.1037/0021-9010.87.3.530 Judge, T. A., Higgins, C. A., Thoresen, C. J., & Barrick, M. R. (1999). The Big Five personality traits, general mental ability, and career success across the life span. Personnel Psychology, 52, 621–652. doi:10.1111/j. 1744-6570.1999.tb00174.x 180

Knapp, D. J., Heggestad, E. D., & Young, M. C. (Eds.). (2004). Understanding and improving the Assessment of Individual Motivation (AIM) in the Army’s GED Plus program (ARI Study Note 2004–03). Alexandria, VA: U.S. Army Research Institute for the Behavioral and Social Sciences. Komar, S., Brown, D. J., Komar, J. A., & Robie, C. (2008). Faking and the validity of conscientiousness: A Monte Carlo investigation. Journal of Applied Psychology, 93, 140–154. doi:10.1037/0021-9010. 93.1.140 Kuncel, N. R., & Borneman, M. J. (2007). Toward a new method of detecting deliberately faked personality tests: The use of idiosyncratic item responses. International Journal of Selection and Assessment, 15, 220–231. doi:10.1111/j.1468-2389.2007.00383.x Kuncel, N. R., Cooper, S. R., & Owens, K. (2009). Communicating the predictive power of selection and admissions measures: Three alternatives to correlations. Unpublished manuscript.

Personality and Its Assessment

Kwong, J. Y. Y., & Cheung, F. M. (2003). Prediction of performance facets using specific personality traits in the Chinese context. Journal of Vocational Behavior, 63, 99–110. doi:10.1016/S0001-8791(02)00021-0 Lahey, B. B. (2009). Public health significance of neuroticism. American Psychologist, 64, 241–256. doi:10. 1037/a0015309

Copyright American Psychological Association. Not for further distribution.

LeBreton, J. M., Barksdale, C. D., Robin, J., & James, L. R. (2007). Measurement issues associated with conditional reasoning tests: Indirect measurement and test faking. Journal of Applied Psychology, 92, 1–16. doi:10.1037/0021-9010.92.1.1 Lee, F. K., Sheldon, K. M., & Turban, D. B. (2003). Personality and the goal-striving process: The influence of achievement goal patterns, goal level, and mental focus on performance and enjoyment. Journal of Applied Psychology, 88, 256–265. doi:10.1037/ 0021-9010.88.2.256 Lee, K., & Ashton, M. C. (2004). Psychometric properties of the HEXACO personality inventory. Multivariate Behavioral Research, 39, 329–358. doi:10.1207/ s15327906mbr3902_8 Lee, K., & Ashton, M. C. (2008). The HEXACO personality factors in the indigenous personality lexicons of English and 11 other languages. Journal of Personality, 76, 1001–1053. Lee, K., Ashton, M. C., & deVries, R. E. (2005). Predicting workplace delinquency and integrity with the HEXACO and five-factor models of personality structure. Human Performance, 18, 179–197. doi:10.1207/ s15327043hup1802_4 Lefkowitz, J. (2000). The role of interpersonal affective regard in supervisory performance ratings: A literature review and proposed causal model. Journal of Occupational and Organizational Psychology, 73, 67–85. doi:10.1348/096317900166886 LePine, J. A., Colquitt, J. A., & Erez, A. (2000). Adaptability to changing task contexts: Effects of general cognitive ability, conscientiousness, and openness to experience. Personnel Psychology, 53, 563–593. doi:10.1111/j.1744-6570.2000.tb00214.x

Markon, K. E., Krueger, R. F., & Watson, D. (2005). Delineating the structure of normal and abnormal personality: An integrative hierarchical approach. Journal of Personality and Social Psychology, 88, 139–157. doi:10.1037/0022-3514.88.1.139 Maydeu-Olivares, A., & Böckenholt, U. (2005). Structural equation modeling of paired comparison and ranking data. Psychological Methods, 10, 285–304. doi:10.1037/ 1082-989X.10.3.285 McAdams, D. P., & Olson, B. D. (in press). Personality development: Continuity and change over the life course. Annual Review of Psychology. McCrae, R. R., & Costa, P. T., Jr. (1987). Validation of the five-factor model of personality across instruments and observers. Journal of Personality and Social Psychology, 52, 81–90. doi:10.1037/0022-3514.52.1.81 McCrae, R. R., & Costa, P. T., Jr. (1997). Personality trait structure as a human universal. American Psychologist, 52, 509–516. doi:10.1037/0003-066X.52.5.509 McCrae, R., Costa, P., Del Pilar, G., Rolland, J. P., & Parker, W. (1998). Cross-cultural assessment of the five-factor model: The revised NEO Personality Inventory. Journal of Cross-Cultural Psychology, 29, 171–188. doi:10.1177/0022022198291009 McDonald, R. P. (1999). Test theory: A unified treatment. Mahwah, NJ: Erlbaum. McFarland, L. A., & Ryan, A. M. (2000). Variance in faking across noncognitive measures. Journal of Applied Psychology, 85, 812–821. doi:10.1037/0021-9010. 85.5.812 McFarland, L. A., & Ryan, A. M. (2006). Toward an integrated model of applicant faking behavior. Journal of Applied Social Psychology, 36, 979–1016. Mershon, B., & Gorsuch, R. L. (1988). Number of factors in personality sphere: Does increase in factors increase predictability of real life criteria? Journal of Personality and Social Psychology, 55, 675–680. doi:10.1037/0022-3514.55.4.675 Mischel, W. (1968). Personality and assessment. New York: Wiley.

LePine, J. A., Erez, A., & Johnson, D. E. (2002). The nature and dimensionality of organizational citizenship behavior: A critical review and meta-analysis. Journal of Applied Psychology, 87, 52–65. doi:10.1037/0021-9010.87.1.52

Mol, S. T., Born, M. P. H., Willemsen, M. E., & Van Der Molen, H. T. (2005). Predicting expatriate job performance for selection purposes: A quantitative review. Journal of Cross-Cultural Psychology, 36, 590–620. doi:10.1177/0022022105278544

Loevinger, J. (1957). Objective tests as instruments of psychological theory. Psychological Reports, 3, 635–694. doi:10.2466/PR0.3.7.635-694

Molenaar, P. C. M., & Campbell, C. G. (2009). The new person-specific paradigm in psychology. Current Directions in Psychological Science, 18, 112–117. doi:10.1111/j.1467-8721.2009.01619.x

MacCallum, R. C., Wegener, D. T., Uchino, B. N., & Fabrigar, L. R. (1993). The problem of equivalent models in applications of covariance structure analysis. Psychological Bulletin, 114, 185–199. doi:10.1037/ 0033-2909.114.1.185

Moon, H., Hollenbeck, J. R., Marinova, S., & Humphrey, S. E. (2008). Beneath the surface: Uncovering the relationship between extraversion and organizational citizenship behavior through a facet approach. 181

Oswald and Hough

International Journal of Selection and Assessment, 16, 143–154. doi:10.1111/j.1468-2389.2008.00419.x Morgeson, F. P., Campion, M. A., Dipboye, R. L., Hollenbeck, J. R., Murphy, K., & Schmitt, N. (2007). Reconsidering the use of personality tests in personnel selection contexts. Personnel Psychology, 60, 683–729.

Copyright American Psychological Association. Not for further distribution.

Morris, S. B., & Lobsenz, R. (2000). Significance tests and confidence intervals for the adverse impact ratio. Personnel Psychology, 53, 89–111. doi:10.1111/ j.1744-6570.2000.tb00195.x Mueller-Hanson, R., Heggestad, E. D., & Thornton, G. C., III. (2003). Faking and selection: Considering the use of personality from select-in and select-out perspectives. Journal of Applied Psychology, 88, 348–355. doi:10.1037/0021-9010.88.2.348 Murphy, K. R., & DeShon, R. D. (2000). Progress in psychometrics: Can industrial and organizational psychology catch up? Personnel Psychology, 53, 913–924. doi:10.1111/j.1744-6570.2000.tb02423.x Ng, T. W. H., Eby, L. T., Sorensen, K. L., & Feldman, D. C. (2005). Predictors of objective and subjective career success. A meta-analysis. Personnel Psychology, 58, 367–408. doi:10.1111/j.1744-6570.2005.00515.x Norman, W. T. (1963). Toward an adequate taxonomy of personality attributes: Replicated factor structure in peer nomination personality ratings. Journal of Abnormal and Social Psychology, 66, 574–583. doi:10.1037/h0040291

Paunonen, S. V. (1998). Hierarchical organization of personality and prediction of behavior. Journal of Personality and Social Psychology, 74, 538–556. doi:10.1037/0022-3514.74.2.538 Paunonen, S. V., & Jackson, D. N. (2000). What is beyond the Big Five? Plenty! Journal of Personality, 68, 821–835. Paunonen, S. V., & Nicol, A. A. A. M. (2001). The personality hierarchy and the prediction of work behaviors. In B. W. Roberts & R. T. Hogan (Eds.), Personality psychology in the workplace (pp. 161–191). Washington, DC: American Psychological Association. Peabody, D., & Goldberg, L. R. (1989). Some determinants of factor structures from personality-trait descriptors. Journal of Personality and Social Psychology, 57, 552–567. doi:10.1037/0022-3514.57.3.552 Peeters, M. A. G., Van Tuijl, H. F. J. M., Rutte, C. G., & Reymen, I. M. M. J. (2006). Personality and team performance: A meta-analysis. European Journal of Personality, 20, 377–396. doi:10.1002/per.588 Pervin, L. A. (1989). Persons, situations, interactions: The history of a controversy and a discussion of theoretical models. Academy of Management Review, 14, 350–360. doi:10.2307/258172 Ployhart, R. E., Lim, B.-C., & Chan, K.-Y. (2001). Exploring relations between typical and maximum performance ratings and the five factor model of personality. Personnel Psychology, 54, 809–843. doi:10.1111/j.1744-6570.2001.tb00233.x

Ones, D. S., Viswesvaran, C., & Schmidt, F. L. (1993). Comprehensive meta-analysis of integrity test validities: Findings and implications for personnel selection and theories of job performance. Journal of Applied Psychology, 78, 679–703. doi:10.1037/00219010.78.4.679

Ployhart, R. E., Weekley, J. A., & Baughman, K. (2006). The structure and function of human capital emergence: A multilevel examination of the attraction– selection–attrition model. Academy of Management Journal, 49, 661–667.

Ones, D. S., Viswesvaran, C., & Schmidt, F. L. (2003). Personality and absenteeism: A meta-analysis of integrity tests. European Journal of Personality, 17(Suppl.), S19–S38. doi:10.1002/per.487

Pulakos, E. D., Schmitt, N., & Chan, D. (1996). Models of job performance ratings: An examination of ratee race, ratee gender, and rater level effects. Human Performance, 9, 103–119. doi:10.1207/s15327043 hup0902_1

Ong, A. D., Bergeman, C. S., Bisconti, T. L., & Wallace, K. A. (2006). Psychological resilience, positive emotions, and successful adaptation to stress in later life. Journal of Personality and Social Psychology, 91, 730–749. doi:10.1037/0022-3514.91.4.730

Richard, F. D., Bond, C. F., Jr., & Stokes-Zoota, J. J. (2003). One hundred years of social psychology quantitatively described. Review of General Psychology, 7, 331–363. doi:10.1037/1089-2680.7.4.331

Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional predictors of organizational citizenship behavior. Personnel Psychology, 48, 775–802. doi:10.1111/j.1744-6570.1995.tb01781.x

Roberts, B. W., Bogg, T., Walton, K., Chernyshenko, O. S., & Stark, S. (2004). A lexical investigation of the lower-order structure of conscientiousness. Journal of Research in Personality, 38, 164–178.

Ostendorf, F., & Angleitner, A. (1994). The five-factor taxonomy: Robust dimensions of personality description. Psychologica Belgica, 34, 175–194.

Roberts, B. W., Chernyshenko, O. S., Stark, S., & Goldberg, L. R. (2005). The structure of conscientiousness: An empirical investigation based on seven major personality questionnaires. Personnel Psychology, 58, 103–139. doi:10.1111/j.1744-6570.2005.00301.x

Oswald, F. L., & McCloy, R. A. (2003). Meta-analysis and the art of the average. In K. R. Murphy (Ed.), Validity generalization: A critical review (pp. 311–338). Mahwah, NJ: Erlbaum. 182

Roberts, B. W., Kuncel, N. R., Shiner, R., Caspi, A., & Goldberg, L. R. (2007). The power of personality:

Personality and Its Assessment

Copyright American Psychological Association. Not for further distribution.

The comparative validity of personality traits, socioeconomic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science, 2, 313–345. doi:10.1111/j.1745-6916.2007. 00047.x

Schmidt, F. L., & Hunter, J. E. (1992). Development of a causal model of processes determining job performance. Current Directions in Psychological Science, 1, 89–92. doi:10.1111/1467-8721.ep10768758

Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132, 1–25. doi:10. 1037/0033-2909.132.1.1

Schmidt, F. L., Le, H., & Ilies, R. (2003). Beyond alpha: An empirical examination of the effects of different sources of measurement error on reliability estimates for measures of individual differences constructs. Psychological Methods, 8, 206–224. doi:10.1037/1082989X.8.2.206

Robertson, I. T., & Kinder, A. (1993). Personality and job competencies: An examination of the criterion related validity of some personality variables. Journal of Occupational and Organizational Psychology, 66, 225–244.

Schmit, M. J., & Ryan, A. M. (1993). The Big Five in personnel selection: Factor structure in applicant and non-applicant populations. Journal of Applied Psychology, 78, 966–974. doi:10.1037/0021-9010. 78.6.966

Rosenthal, R. (1991). Effect sizes: Pearson’s correlation, its display via the BESD, and alternative indices. American Psychologist, 46, 1086–1087. doi:10. 1037/0003-066X.46.10.1086

Schmitt, N., & Oswald, F. L. (2006). The impact of corrections for faking on the validity of noncognitive measures in selection settings. Journal of Applied Psychology, 91, 613–621. doi:10.1037/0021-9010. 91.3.613

Ruscio, J., & Ruscio, A. M. (2008). Categories and dimensions: Advancing psychological science through the study of latent structure. Current Directions in Psychological Science, 17, 203–207. doi:10.1111/j.1467-8721.2008.00575.x

Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437–453. doi:10.1111/j.1744-6570. 1987.tb00609.x

Sackett, P. R., & Ellingson, J. E. (1997). The effects of forming multi-predictor composites on group differences and adverse impact. Personnel Psychology, 50, 707–721. doi:10.1111/j.1744-6570.1997.tb00711.x

Schneider, R. J., Ackerman, P. L., & Kanfer, R. (1996). To “act wisely in human relations”: Exploring the dimensions of social competence. Personality and Individual Differences, 21, 469–481. doi:10.1016/0191-8869 (96)00084-0

Sackett, P. R., Schmitt, N., Ellingson, J. E., & Kabin, M. B. (2001). High-stakes testing in employment, credentialing, and higher education: Prospects in a post-affirmative-action world. American Psychologist, 56, 302–318. doi:10.1037/0003-066X.56.4.302

Schneider, R. J., & Hough, L. M. (1995). Personality and industrial/organizational psychology. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (pp. 75–129). Chichester, England: Wiley.

Sanz, J., Gil, F., Garcia-Vera, M. P., & Barrasa, A. (2008). Needs and cognition/behavior patterns at work and the Big Five: An assessment of the Personality and Preference Inventory-Normative (PAPI-N) from the perspective of the five-factor model. International Journal of Selection and Assessment, 16, 46–58. doi:10.1111/j.1468-2389.2008.00408.x

Schneider, R. J., Hough, L. M., & Dunnette, M. D. (1996). Broadsided by broad traits: How to sink science in five dimensions or less. Journal of Organizational Behavior, 17, 639–655. doi:10.1002/(SICI)10991379(199611)17:63.0.CO;2-9

Saucier, G., Georgiades, S., Tsaousis, I., & Goldberg, L. R. (2005). The factor structure of Greek personality adjectives. Journal of Personality and Social Psychology, 88, 856–875. doi:10.1037/0022-3514. 88.5.856 Schaubroeck, J., Ganster, D. C., & Jones, J. R. (1998). Organization and occupation influences in the attraction–selection–attrition process. Journal of Applied Psychology, 83, 869–891. doi:10.1037/0021-9010. 83.6.869 Schmidt, A. M., Dolis, C. M., & Tolli, A. P. (2009). A matter of time: Individual differences, contextual dynamics, and goal progress effects on multiple-goal self-regulation. Journal of Applied Psychology, 94, 692–709. doi:10.1037/a0015012

Sijtsma, K. (2009). On the use, misuse, and the very limited usefulness of Cronbach’s alpha. Psychometrika, 74, 107–120. doi:10.1007/s11336-008-9101-0 Somer, O., & Goldberg, L. R. (1999). The structure of Turkish trait-descriptive adjectives. Journal of Personality and Social Psychology, 76, 431–450. Stark, S., & Drasgow, F. (2002). An EM approach to parameter estimation for the Zinnes and Griggs paired comparison IRT model. Applied Psychological Measurement, 26, 208–227. doi:10.1177/01421602026002007 Stark, S., Drasgow, F., & Chernyshenko, O. S. (2008, October). Update on the Tailored Adaptive Personality Assessment System (TAPAS): The next generation of personality assessment systems to support personnel selection and classification decisions. Paper presented at the 50th annual conference of the International Military Testing Association, Amsterdam. 183

Oswald and Hough

Steel, P. (2007). The nature of procrastination: A metaanalytic and theoretical review of quintessential selfregulatory failure. Psychological Bulletin, 133, 65–94. doi:10.1037/0033-2909.133.1.65

van de Vijver, F. J. R., & Leung, K. (2001). Personality in cultural context: Methodological issues. Journal of Personality, 69, 1007–1031. doi:10.1111/1467-6494. 696173

Steel, P., Schmidt, J., & Shultz, J. (2008). Refining the relationship between personality and subjective wellbeing. Psychological Bulletin, 134, 138–161. doi:10.1037/0033-2909.134.1.138

Vasilopoulos, N. L., Cucina, J. M., Dyomina, N. V., Morewitz, C. L., & Reilly, R. R. (2006). Forced-choice personality tests: A measure of personality and cognitive ability? Human Performance, 19, 175–199. doi:10.1207/s15327043hup1903_1

Copyright American Psychological Association. Not for further distribution.

Stewart, G. (1999). Trait bandwidth and stages of job performance: Assessing differential effects for conscientiousness and its subtraits. Journal of Applied Psychology, 84, 959–968. doi:10.1037/0021-9010. 84.6.959 Stewart, G., Fulmer, I. S., & Barrick, M. R. (2005). An exploration of member roles as a multilevel linking mechanism for individual traits and team outcomes. Personnel Psychology, 58, 343–365. doi:10.1111/j. 1744-6570.2005.00480.x Stewart, G., & Nandkeolyar, A. K. (2006). Adaptation and intraindividual variation in sales outcomes: Exploring the interactive effects of personality and environmental opportunity. Personnel Psychology, 59, 307–322. Tellegen, A. (1993). Folk concepts and psychological concepts of personality and personality disorder. Psychological Inquiry, 4, 122–130. doi:10.1207/ s15327965pli0402_12 Terracciano, A., McCrae, R. R., Brant, L. J., & Costa, P., Jr. (2005). Hierarchical linear modeling analyses of the NEO-PI-R scales in the Baltimore Longitudinal Study of Aging. Psychology and Aging, 20, 493–506. doi:10.1037/0882-7974.20.3.493 Tett, R. P., & Burnett, D. D. (2003). A personality traitbased interactionist model of job performance. Journal of Applied Psychology, 88, 500–517. doi:10.1037/0021-9010.88.3.500 Tett, R. P., Steele, J. R., & Beauregard, R. S. (2003). Broad and narrow measures on both sides of the personalityjob performance relationship. Journal of Organizational Behavior, 24, 335–356. doi:10.1002/ job.191 Thoresen, C. J., Bradley, J. C., Bliese, P. D., & Thoresen, J. D. (2004). The Big Five personality traits and individual job performance growth trajectories in maintenance and transitional job stages. Journal of Applied Psychology, 89, 835–853. doi:10.1037/0021-9010. 89.5.835 Tupes, E. C., & Christal, R. E. (1961). Recurrent personality factors based on trait ratings (USAF Tech. Rep. No. 61-97). Lackland Air Force Base, TX: U.S. Air Force. Tupes, E. C., & Christal, R. E. (1992). Recurrent personality factors based on trait ratings. Journal of Personality, 60, 225–251. doi:10.1111/j.1467-6494. 1992.tb00973.x 184

Vasilopoulos, N. L., Cucina, J. M., & McElreath, J. M. (2005). Do warnings of response verification moderate the relationship between conscientiousness and cognitive ability? Journal of Applied Psychology, 90, 306–322. doi:10.1037/0021-9010.90.2.306 Vinchur, A. J., Schippmann, J. S., Switzer, F. S., & Roth, P. L. (1998). A meta-analytic review of predictors of job performance for salespeople. Journal of Applied Psychology, 83, 586–597. doi:10.1037/0021-9010. 83.4.586 Wallace, C., & Chen, G. (2006). A multilevel integration of personality, climate, self-regulation, and performance. Personnel Psychology, 59, 529–557. doi:10. 1111/j.1744-6570.2006.00046.x Warr, P., Bartram, D., & Martin, T. (2005). Personality and sales performance: Situational variation and interactions between traits. International Journal of Selection and Assessment, 13, 87–91. doi:10.1111/ j.0965-075X.2005.00302.x Weekley, J. A., Ployhart, R. E., & Harold, C. M. (2004). Personality and situational judgment tests across applicant and incumbent settings: An examination of validity, measurement, and subgroup differences. Human Performance, 17, 433–461. doi:10.1207/ s15327043hup1704_5 Westring, A. J. F., Oswald, F. L., Schmitt, N., Drzakowski, S., Imus, A., Kim, B., & Shivpuri, S. (2009). Estimating trait and situational variance in a situational judgment test. Human Performance, 22, 44–63. doi:10.1080/08959280802540999 Wiggins, J. S. (1991). Agency and communion as conceptual co-ordinates for the understanding and measurement of interpersonal behavior. In W. M. Grove & D. Cicchetti (Eds.), Thinking clearly about psychology: Essays in honor of Paul E. Meehl (Vol. 2, pp. 89–113). Minneapolis: University of Minnesota Press. Yeo, G. B., & Neal, A. (2004). A multilevel analysis of effort, practice, and performance: Effects of ability, conscientiousness and goal orientation. Journal of Applied Psychology, 89, 231–247. doi:10.1037/00219010.89.2.231 Zimprich, D., Allemand, M., & Dellenbach, M. (2009). Openness to experience, fluid intelligence, and crystallized intelligence in middle-aged and old adults. Journal of Research in Personality, 43, 444–454. doi:10.1016/j.jrp.2009.01.018

CHAPTER 6

INTERVIEWS

Copyright American Psychological Association. Not for further distribution.

Allen I. Huffcutt and Satoris S. Culbertson

The employment interview is something of an enigma. We know it predicts job performance, at least when properly designed (Campion, Palmer, & Campion, 1998; Huffcutt & Arthur, 1994; McDaniel, Whetzel, Schmidt, & Maurer, 1994), but we are not completely sure why. For instance, the degree to which it captures specific job-related knowledge, skills, abilities, and other characteristics (KSAOs; e.g., ability to soothe irate customers) versus general constructs such as mental ability or conscientiousness (Cortina, Goldstein, Payne, Davison, & Gilliland, 2000; Hunter & Hirsch, 1987) versus other things like general impressions (Sackett, 1982) is not perfectly clear. What adds further complexity is that the interview can be influenced by additional variables such as the applicant’s impression management skills (Ellis, West, Ryan, & DeShon, 2002; Levashina & Campion, 2007), self-monitoring (Dipboye, 1992), or interview-specific self-efficacy (Tay, Ang, & Van Dyne, 2006). Arthur, Woehr, and Maldegen (2000) regarded the combination of assessment centers with criterion-related validity without clear construct validity as a validity paradox, a term that appears to apply equally well to the interview. Despite the uncertainty, the interview continues to be used almost universally. It is rare, even unthinkable, for someone to be hired without some type of interview. When someone is hired without one, it is often because of extenuating circumstances, such as that the hiring organization is desperately short-handed (e.g., nursing). The interview is used far more frequently than any other selection technique (e.g., psychological testing, work samples, assessment centers; Nyfield &

Baron, 2000; Sharf & Jones, 2000). The only other comparable component in the selection process (at least in terms of frequency of use) is the application form. Of interest for several reasons is the question of why interviews are so popular despite uncertainty regarding what they measure and, for certain types (e.g., unstructured; Wiesner & Cronshaw, 1988), their reliability and validity. As a practical matter, the interview is not needed because critical KSAOs usually can be assessed by other means and often done so more accurately (Hunter & Hunter, 1984; Schmidt & Hunter, 1998). It would appear that there is a basic human need to want personal contact with others before placing them in a position of importance even if they have a proven track record, a tendency from which personnel managers and others involved in organizational selection do not appear to be exempt. It is almost as if a part of the human makeup does not trust objective information completely, even if it is accurate; mere facts do not supersede an underlying desire for personal verification. The purpose of this chapter is to provide a comprehensive overview of the employment interview. We begin by tracing the history of its research, which provides a background of what issues have been addressed and where future research is headed. Then we go back and take a closer look at structure of the interview, arguably the single most important element of interview process and outcomes. Throughout this chapter, we discuss the substantive issues of why, when, and how interviews are the most reliable and valid. Finally, we close with a discussion of future research needs and directions.

http://dx.doi.org/10.1037/12170-006 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

185

Huffcutt and Culbertson

Copyright American Psychological Association. Not for further distribution.

HISTORY OF INTERVIEW RESEARCH Thanks to a series of major narrative reviews (in chronological order: Wagner, 1949; Mayfield, 1964; Ulrich & Trumbo, 1965; Wright, 1969; Schmidt, 1976; Arvey & Campion, 1982; Harris, 1989; Posthuma, Morgeson, & Campion, 2002), it is possible to trace the history of interview research in some detail. When doing so, it becomes apparent that interview research has gone through a series of distinct phases or periods, five of which are outlined in this section. Two disclaimers are warranted. First, once a phase and its timeline are put forth, there is some tendency to assume that no other type of research was being conducted. That is clearly not the case, as research of many types tends to be scattered throughout all periods. Second, identifying the number of phases and the nature of these phases, choosing labels for them, and delineating their dates is an extremely subjective process, and there could be some noticeable differences if other researchers were to do the same thing (cf. Eder, Kacmar, & Ferris, 1989).

Introductory Period (Circa 1915–1950) A key theme of this phase is the opening of the interview process to scientific inquiry and analysis. Rather than treating the interview as a given or as something that was not a proper topic for science, a host of issues were raised at once, tantamount to opening Pandora’s box. The earliest reference to appear in modern journals describing a formal study was Scott (1915). In his study, six personnel managers individually interviewed and rank ordered 36 sales candidates as to their suitability for the position. Wide disagreement resulted. For instance, for 28 of the candidates the six managers could not agree whether they should be placed in the upper half or lower half of the group. Similar studies were conducted by Scott, Bingham, and Whipple (1916) and Hollingworth (1922), with similar results. Collectively, these studies called into question the reliability and general efficacy of the interview and established an early pessimism that influenced thinking for a number of decades. Wagner (1949), the first major narrative review to appear in the literature, provided an excellent 186

summary of the issues raised during this phase and the empirical studies that were conducted to address them. Concerns over basic reliability and validity not surprisingly were at the forefront, and the empirical data were too mixed to provide any real answers. There was considerable debate over the scope of the interview, including whether the interview should be used to evaluate a broad range of characteristics or used just to assess overall suitability. Those favoring the former raised the additional issues of what traits should be assessed (e.g., general versus job-specific) and whether the interviewers should integrate the ratings themselves or combine them mechanically. Yet others advocated a focus on a single or very select number of specific traits. For instance, Rundquist (1947) felt that social interaction skill was the only characteristic that should be assessed in the interview and that all other important job characteristics should be left to standardized testing. Another issue was whether interviewers should be allowed access to ancillary information (e.g., test scores), as that information appeared to increase the accuracy of the interview ratings. Lastly, a smattering of researchers were proponents of a standardized approach. Wonderlic (1942), for example, noted that without such an approach the interview “generally amounts to a disorganized conversation resulting in a series of impressions based upon impulsive reactions” (cited on p. 33, Wagner, 1949). In summary of this phase, there were numerous questions raised and few answers developed. A number of these issues remain unresolved even today. For instance, many companies still use unstructured interviews that are not overtly designed around job requirements and which vary in scope from assessing specific job skills to overall suitability. In fairness, use of a standardized format has become much more common, so the trend is in the right direction. Per Rundquist (1947), social skill (discussed later in this chapter) appears to be getting more formal attention lately and could hold considerable promise. The method of combining ratings is also still an issue, as some interviewers continue to integrate the ratings and arrive at an overall evaluation subjectively.

Interviews

Copyright American Psychological Association. Not for further distribution.

Early Psychometrical Period (Circa 1950–1980) The momentum created in the first phase that resulted from placing the interview under the lens of scientific scrutiny continued and expanded during this second phase, increasing not only in quantity but also in scope and sophistication. Three major narrative reviews emerged during this time frame: Mayfield (1964), Ulrich and Trumbo (1965), and Wright (1969). All three reviewers summarized reliability and validity data and again found mixed results, although they did note that the data appeared stronger for standardized interviews. The advent of meta-analysis was still years away, and researchers of this time did not have a full understanding that sampling error and other artifacts were a major source of variability among the findings. These reviewers raised additional issues, which, while not being addressed strongly during this phase, planted some seeds that grew in subsequent ones or at least are starting to grow in current research. Mayfield (1964) questioned why the interview works, what is the effect of varying degrees of structure, and how we deal with the general lack of comparability across interviews (e.g., length, content, rating characteristics). Furthermore, he raised awareness of the importance of individual differences among interviewers, driven in part by data suggesting acceptable reliability across candidates by the same interviewer (intrarater) but poor consistency across interviews by different interviewers (interrater), and called for research into their decision-making processes. Ulrich and Trumbo (1965) called for analysis of how much variance in the ratings is attributable to ancillary data like test scores and how much is due to the interview itself and, in addition, raised the question of how accurate the information is that is obtained during an interview. Wright (1969) noted that the culture (e.g., ethnic group) of a candidate could affect both semantics and the rapport developed with the interviewer and the interviewer’s ratings and/or decision and that interview research should expand and incorporate theories and concepts from other literatures such as communication. He also echoed Mayfield’s concern that the influence of the interviewer on the applicant and the process had been overlooked far too much.

A line of research emerged during this phase that is unparalleled in terms of overall prominence and distinguishability: the McGill University studies led by E. C. Webster. In terms of timing, Mayfield (1964) noted that these studies had been conducted over the previous 10 years (i.e., since the early 1950s), whereas Wright (1969) noted that Webster’s seminal 1964 report of work by him and his colleagues (e.g., dissertations by Springbett, 1958, and Anderson, 1960) covered 8 years of research. In the words of Wright, “It would be difficult to over-estimate the importance of the work done by Webster and his colleagues” (p. 394). The defining nature of this line of research was its microanalytic nature, that is, that the interview was broken up into small segments and processes that were studied in isolation. Wright (1969) identified seven major principles from this work, as listed below. 1. Interviewers have a stereotype of a good candidate and match interviewees to that. 2. Interviewers establish biases early in the interview, which affect their subsequent decision. 3. Unfavorable information is more influential than favorable information. 4. Interviewers seek/focus on information that confirms their impressions. 5. Development and display of empathy varies by interviewer. 6. The decision interviewers make is different when given information is processed piece by piece rather than simultaneously in its totality. 7. Experienced interviewers tend to rank applicants in the same order although they differ in the number they would accept. Schmitt (1976) and Arvey and Campion (1982), the next major reviews, highlighted the increase in research complexity. For instance, additional variables were being looked at (e.g., visual cues) and conceptual models of the interview process were being put forth (e.g., both of the narrative reviews cited contained such a conceptual model). The reviewers further noted that research designs appeared to be more realistic and generalizable (e.g., not as many “paper people” formats). Reliability and validity were still being questioned, although both continued to 187

Copyright American Psychological Association. Not for further distribution.

Huffcutt and Culbertson

note stronger results for standardized formats. Arvey and Campion noted the influence of artifacts (from the early beginnings of meta-analysis) and that other literatures (e.g., person perception) were still not being utilized enough. They planted another seed by noting the importance of interviewees’ preexisting motives and expectations. One interesting but limited line of decisionmaking research was noted in the next major review by Harris (1989): use of the Brunswik lens model (Brunswik, 1955, 1956). This model proposed that interviewer perceptions and attributions of applicants are based in part on aspects of the physical or social environment, including qualities of the applicant, as well as on cues, such as applicant nonverbal behaviors. Several studies employed this technique, which included analysis of how interviewers arrive at an overall rating by looking at the correlation between it and the ratings for the individual dimensions (e.g., Dougherty, Ebert, & Callender, 1986; Gifford, Ng, & Wilkinson, 1985; Zedeck, Tziner, & Middlestadt, 1983). These studies showed considerable individual differences among interviewers, including with utilization of information, reliability, and validity. Harris also noted greater use of theories from other literatures, including attribution theory (Herriot, 1981), decision theory (Rowe, 1989), and confirmatory bias (Sackett, 1982), greater analysis of applicant characteristics (e.g., race), more focus on interviewer training, and more consideration of external validity (e.g., fewer paper-people studies).

Modern Structured Interviewing (Circa 1980–Present) Few things have changed the field as much as the introduction of two specific structuring techniques: the situational interview (Latham, Saari, Purcell, & Campion, 1980) and the behavior description interview (Janz, 1982). Although there was a plethora of approaches and suggestions for how to structure an interview prior to this time, these two pioneering techniques provided an anchor for both researchers and practitioners. These techniques were not innovative in their standardization of questions, as a number of prior interviews had standardized questions. Moreover, use of these types of questions was not new, as we see earlier references to researchers posing 188

questions regarding how candidates think they would react to a given problem situation (O’Rourke, 1932) and inquiring about past situations the candidates had that were relevant to the position (McMurry, 1947). Rather, what differentiated these two techniques from prior techniques was the exclusive use of a single type of question for the entire interview (hypothetical scenarios in the former, past behavior in the latter) coupled with a formal rating system, the combination of which made these techniques unique and powerful. The situational interview had two further distinctions in that the questions were completely standardized across candidates and responses were rated individually by question using a customized scale developed specifically for that question. These latter two practices helped move the interview closer to psychological tests that have standardized content and are scored at the item level (e.g., those for mental ability and personality). In its original creation, the behavior description interview was introduced by Janz (1982) as the “patterned behavior description interview” because banks of questions were created for each job dimension (e.g., problem-solving, initiative) and interviewers could choose freely from among them. Then the interviewer made ratings by dimension afterward based on synthesis of whatever questions were asked. There is some advantage to banks of questions in that interviewers can adapt the interview to some degree to the unique background of each applicant, a format consistent with the modern idea of “adaptive testing” (see Belov & Armstrong, 2008; Lee, Ip, & Fuh, 2008). However, doing so loses consistency in “procedural variability” (Huffcutt & Arthur, 1994) across candidates. In more recent times, it has become common practice with behavior description interviews to standardize questions and rate responses individually by question using a benchmarked rating scale as is done in the situational interview (e.g., Campion, Campion, & Hudson, 1994; Pulakos & Schmitt, 1995). Of course, this is not to say that question scoring is better than dimensional scoring for the interview, as which one is “better” likely depends on the purpose of the interview (e.g., selection vs. recruitment) and the specific dimensions captured by the interview. Nevertheless, what makes question-level scor-

Copyright American Psychological Association. Not for further distribution.

Interviews

ing attractive is the large body of research in the cognitive area regarding the limitations of human information processing. Simply put, it appears that most people are limited in how much information they can process at one time, as exemplified by short-term memory being limited in capacity and easily interfered with (Baddely, 1986; Miller, 1956; Nairne, Neath, & Serra, 1997) and by the tendency to utilize mental shortcuts (heuristics) in the processing of that information (e.g., Kahneman & Tversky, 1973, 2000; Tversky & Kahneman, 1973, 1982). Furthermore, this research suggests that recall of information from long-term memory tends to be based on reconstruction of stored fragments (Loftus, 2007; Loftus & Cahill, 2007). Rating the response to each question individually using a carefully devised benchmarked rating scale should reduce the amount of information that must be processed at any one time considerably, help to prevent shortcuts in the processing of that information, and keep processing “immediate” rather than involving extensive memory recall. Conceptually, both techniques are grounded in theory. Latham (1989) noted that the situational interview is grounded in goal setting, that is, that intentions are a precursor to actions. Janz (1982, 1989) noted that the behavior description interview is grounded in behavioral consistency, specifically, that past behavior is the best predictor of future behavior. While both premises are very reasonable, it might be possible to expand the theoretical basis behind these techniques further, which could generate some meaningful future research. To illustrate, the theory of planned behavior (Ajzen, 1991) from the social psychological literature lists three antecedents to behavioral intentions and behavior: attitudes, subjective norms (i.e., the perception that important others will approve of that behavior), and perceived behavioral control (i.e., the perception of how difficult it would be to perform that behavior). Each of these three components may in turn have its own antecedents, such as attitudes having both a learned (Bandura, 1986; Krosnick, Betz, Jussim, & Lynn, 1992) and a genetic (Bouchard, 2004; Olson, Vernon, Harris, & Jang, 2001) component. One avenue for future research is to examine these components and their antecedents specifically in relation to aspects of the interview process, such as the likelihood of appli-

cants’ faking answers (see Levashina & Campion, 2007). For example, what are their attitudes regarding deception, do they think others would approve of their deceptions, and how difficult would it be to falsify information in a convincing manner? These two structured formats have remained at the forefront of interview research and practice, a span of almost 30 years. Moreover, they have remained largely unchanged, with the noted exception of complete standardization and question-level scoring with the behavior description interview. It would not be surprising to see the introduction of new structuring techniques in coming years, either modifications of these two techniques or entirely new approaches. For instance, rather than simply asking what an applicant would do in a specific situation, the underlying motivation for why they would that could be explored. Alternatively, there may be new types of questions that could be introduced.

Synthesis Era (Circa 1982–1996) Four events stand out in the history of interview research as having changed the field in prominent, significant, and lasting ways. Two have already been mentioned: the microanalytic research program by Webster and his colleagues (1964) and the introduction of the situational and behavior description interview formats. The third is Title VII of the 1964 Civil Rights Act, which established legal considerations as a permanent fixture in the interview arena. Legal considerations are worthy of review as well but are beyond the scope of this chapter. The fourth is what we refer to as the synthesis era. It involves the introduction of meta-analysis, a technique that allows researchers to summarize data that have been collected over time, in different venues, and by different researchers. This era began formally with publication of the pioneering book by Hunter, Schmidt, and Jackson (1982; see Hunter & Schmidt, 2004, for the latest version). Meta-analytic techniques have aided in the synthesis of interview information by providing much more precise estimates of criterion-related validity, overall and for specific levels of variables that moderate interview validity, such as structure. Prior to meta-analysis there was a tendency to take the results of a given study at face value, which helps 189

Copyright American Psychological Association. Not for further distribution.

Huffcutt and Culbertson

explain among other things why most of the narrative reviewers noted a general lack of consistency across studies. The capability of meta-analysis to take into account the influence of artifacts such as sampling error, range restriction, and measurement error has led to a better understanding of the accuracy of interviews as a predictor of job performance and variables that moderate that prediction. There have been three large-scale meta-analyses of the criterion-related validity of the interview: Wiesner and Cronshaw (1988), McDaniel et al. (1994), and Huffcutt and Arthur (1994). All three meta-analyses found greater validity for structured interviews (vs. unstructured ones), although the difference was smaller in McDaniel et al., most likely because of the way in which studies were coded (e.g., their unstructured category appeared to include studies with at least some degree of structure). There is also a largescale meta-analysis of the reliability of the interview by Conway, Jako, and Goodman (1995), which found that the interview could achieve reasonable reliability, upward of .75 under the right design conditions (i.e., use of structure and a panel format). The impact of the three validity meta-analyses cannot be understated. There was widespread pessimism about the interview prior to this time, resulting largely from the combination of psychometric effects that were not well understood (e.g., range restriction reducing magnitudes, sampling error inducing inconsistency) and failure to take the degree of structure into full consideration. After correction for artifacts, the results indicated that when properly designed (mainly structured and based on a job analysis), the employment interview appears to reach a level of criterion-related validity that is highly comparable with mental ability tests, job knowledge tests, work samples/simulations, and other top predictors (see Hunter & Hunter, 1984; Schmidt & Hunter, 1998). In short, meta-analytic techniques have not only helped to summarize and synthesize existing information but have also established the interview as a major validity component in the selection process, revitalized interest in it, and spurred new lines of research for it (which, as described later, continue today). However, some caution in the interpretation of the result of interview meta-analyses is warranted. 190

Like the assessment center, the interview is a method. As such, the constructs measured can vary from situation to situation and can impact resulting validity. As noted in the Principles for the Validation and Use of Personnel Selection Procedures (4th ed.; Society for Industrial and Organizational Psychology, 2003), Because methods such as the interview can be designed to assess widely varying constructs (from job knowledge to integrity), generalizing from cumulative findings is only possible if the features of the method that result in positive method-criterion relationships are clearly understood, if the content of the procedures and meaning of the scores are relevant for the intended purpose, and if generalization is limited to other applications of the method that include those features. (p. 30) What this limitation emphasizes is the need for greater understanding of what constructs interviews measure and the factors that influence that measurement. Large-scale validity analysis of the interview appears to have reached a saturation point during the mid 1990s. The meta-analyses cited earlier used hundreds of studies and tested most if not all of the major moderators (e.g., structure, job complexity). Metaanalytic research of varying types has continued to be conducted and published, of course, but its frequency clearly has dropped. Although there certainly is room for differing opinions, one can argue that its decline coincided to some degree with the emergence of the next phase. Although the issue of what interviews measure has always been of interest, a more concentrated effort to understand their constructs appears to be under way.

Interview Construct Research (Circa 1996–Present) Throughout the history of interview research there has always been interest, speculation, and primarystudy research (mainly correlating interview ratings to scores on one or more psychological tests) regarding what interviews measure. In 1996, Huffcutt, Roth, and McDaniel used meta-analysis to do a large-scale

Copyright American Psychological Association. Not for further distribution.

Interviews

summary of the relationship between interview ratings and mental ability, which provided meaningful construct information and sharpened the focus on constructs. Understanding what constructs interviews measure is a critical issue to its continued advancement as a formal selection technique. At a scientific level, even though an interview may provide comparable validity, the interview as a selection method cannot reach the same standing as cognitive ability tests unless a basic understanding of what it measures is obtained. At a more pragmatic level, issues such as incremental validity (Dipboye & Gaugler, 1993; Hakel, 1989; Harris, 1989) and ethnic group differences (Huffcutt & Roth, 1998) come into play. (Fortunately, the latter study suggests fairly low differences overall, certainly much lower than with ability tests.) Unfortunately, identifying the constructs that interviews capture is difficult, first, because the dimensions they tend to assess are often complex and multifaceted (Roth, Van Iddekinge, Huffcutt, Eidson, & Schmit, 2005; Schmidt & Hunter, 1998) and, second, because of the variability in design, content, interviewees, and interviewers across settings. To illustrate the former point, a question about handling a difficult customer could include aspects of mental ability (e.g., thinking through options), conscientiousness (e.g., doing the right thing), and emotional stability (maintaining a calm composure). To illustrate the latter point, several studies involved careful matching of situational and behavior description questions to assess the same job dimensions, and they did not find a strong correspondence between the two parallel sets of questions. For instance, Huffcutt, Weekley, Weisner, DeGroot, and Jones (2001) found a mean correlation of .09 between matching situational and behavior description questions in their first study and a mean correlation of .05 in their second study. There appear to be at least six types or lines of interview construct research. The first line consists of correlations between interview ratings and psychological measures and is illustrated by Huffcutt, Roth, and McDaniel’s (1996) summary of the interview– mental ability association. They found a mean corrected correlation of .40 between interview ratings

and mental ability, suggesting modest overall saturation and reasonable potential for incremental validity. Berry, Sackett, and Landers (2007) reanalyzed the relationship using a more modern approach for the range restriction correction (Le & Schmidt, 2006; Schmidt, Oh, & Le, 2006) and found a somewhat lower value. No large-scale meta-analyses have yet to be done for other psychological characteristics such as personality, mainly because those data tend to be scarce. The second line revolves around applicant fit. Posthuma et al. (2002), the only major review of the employment interview literature since Harris (1989), noted the emergence of research on the fit between applicants and either the job or the company with which they are interviewing. This research has the potential to provide a better understanding of the dynamics of the interviewee–interviewer exchange, particularly for low- to medium-structure interviews. Research cited by Posthuma et al. suggested that fit is associated with characteristics such as interpersonal skills, goal orientation, values, attractiveness, and liking of the candidate (Adkins, Russell, & Werbel, 1994; Cable & Judge, 1997; Rynes & Gerhart, 1990). The third line is analysis of the content of the dimensions rated in interviews, as illustrated by Huffcutt, Conway, Roth, and Stone’s (2001) summary of the frequency at which various characteristics were rated across a sample of 47 interviews. Just over 60% of the dimensions rated in those studies pertained to personality and/or applied social skills, suggesting more effort to measure these types of characteristics than perhaps anticipated. A limitation of their work is that they were not able to verify the degree to which these characteristics were actually assessed, only the degree to which they were chosen as selection criteria. The fourth line, which is particularly promising, pertains specifically to the dynamics of the interview process. Previous research tended to look at surface features such as structure, length, amount of applicant speaking time, and demographics (e.g., race, gender, age). A more concentrated effort has emerged that deals directly with the dynamics of the interviewee–interviewer exchange. The strongest venue appears to be with impression management tactics such as self-promotion, ingratiation, and 191

Copyright American Psychological Association. Not for further distribution.

Huffcutt and Culbertson

even outright lying (Ellis et al., 2002; Higgins & Judge, 2004; Higgins, Judge, & Ferris, 2003; Levashina & Campion, 2007; McFarland, Ryan, & Kriska, 2003; Stevens & Kristof, 1995). Perhaps the most surprising find from this line is that candidates appear to engage in these behaviors frequently, even in tightly structured interviews (Ellis et al. 2002; Levashina & Campion, 2007). The related concept of social skills (Ferris, Witt, & Hochwarter, 2001) has been studied to a lesser extent (e.g., Schuler & Funke, 1989; see Salgado & Moscoso, 2002, for a meta-analysis of six correlations) but also holds considerable promise for understanding what interviews measure. Also less studied but potentially of importance are self-monitoring (e.g., Dipboye, 1992; Higgins & Judge, 2004) and relational control (e.g., Bateson, 1958; Engler-Parish & Millar, 1989; Tuller, 1989). There are other aspects of the dynamics of the interview process that, although researched during and since the microanalytic era, continue to be assessed. Some of these aspects include attractiveness, verbal quality, and nonverbal behavior. Despite the abundant existing research in these particular areas, there are some differences in the more modern line regarding how they are being examined. For instance, more factors such as attractiveness (Burnett & Motowidlo, 1998) are being looked at in relation to specific types of interviews (e.g., structured interviews) rather than interviews in general. Utilizing concepts from the communication area, DeGroot and Motowidlo (1999, second study) analyzed voice quality by putting segments of videotaped structured interviews through a voice analyzer and found a moderate correlation (r = .32) between what they termed a vocal cue index and interview judgments. The fifth line is barely beginning, but it answers the call by Arvey and Campion (1982) to look more at the applicants’ state of mind coming into the interview. For example, applicants may have preexisting motives and expectations that they bring with them to the interview that could potentially influence their interview performance as well as their own perceptions of it. To further test these assertions, researchers have started to look more at factors such as interviewspecific self-efficacy (Tay et al., 2006) and interviewspecific motivation (Maurer, Solamon, Andrews, & 192

Troxtel, 2001). As researchers continue to examine such issues, the role of applicants in the interview process will become better understood. The last line is the most recent and is just beginning; it is model-based construct research. Previous models of the employment interview dealt with the process and/or the outcome (e.g., Arvey & Campion, 1982; Dipboye, 2005; Dipboye & Macan, 1988; Raza & Carpenter, 1987; Schmitt, 1976). A new model by Huffcutt, Van Iddekinge, and Roth (2007) focuses directly on the constructs captured by employment interviews. Their model, shown in Figure 6.1, has three main components. First, there are constructs related to job performance, which are proposed to carry over into the interview since their purpose is to assess potential to perform. Second, the model contains interviewee performance constructs, which are related to the notion that the interview is an interpersonal performance by the interviewees (e.g., use of impression management tactics), which can raise or lower ratings relative to actual qualifications. Finally, the model includes influences from interviewer rating tendencies (e.g., idiosyncrasies). Whereby research in the area of interview constructs appears to be in the very early stages of development, much like interview validity research a decade or two ago, this and other models yet to come should help to provide a solid foundation for the advancement of construct research. For instance, in an empirical summary of an updated (2009) version of their model, Huffcutt, Van Iddekinge, and Roth (unpublished manuscript) found that constructs relating to interviewee performance have a mean correlation with interview ratings that is twice as large as the mean correlation for constructs pertaining to job-related interview content. What have we learned by viewing the history of interview research? One thing that is apparent is that interview research is very dynamic. It has evolved from its infancy with simplistic assessment of reliability into a major area of organizational study replete with rich and complex veins of research, and it should continue to evolve for the indefinite future, as there are a number of issues yet to be resolved (or for some issues, yet to be addressed). Table 6.1 summarizes what we see as 10 of the most major issues raised from this review of interview history (collectively across all five eras) and their current status.

Interviews

JOB-RELATED INTERVIEW CONTENT INTERVIEWEE PERFORMANCE

General Human Attributes • Mental ability • Personality • Interests • Goals & values

Social Interaction Skills • Social influence behaviors • Interpersonal presentation Core Job Elements • Declarative knowledge • Procedural skills & abilities • Motivation

INTERVIEW RATINGS

Copyright American Psychological Association. Not for further distribution.

Experiential Factors

Personal / Contextual Influences

• Experience • Education • Training

• Interview training & experience • Interview self-efficacy • Interview motivation

INTERVIEWER RATING TENDENCIES

• Limited use of information • Idiosyncratic rater tendencies • Susceptibility to nonrelevant information

FIGURE 6.1. The Huffcutt, Van Iddekinge, and Roth (2007) model of the constructs that employment interviews capture. The curved arrows indicate that basic human attributes and experiential factors can influence both core job elements (declarative knowledge, procedural skills & abilities, motivation) and interviewee performance.

TABLE 6.1 Summary of Issues From the Historical Review Issue 1. Reliability validity/structure 2. Scope of the interview 3. Ancillary information 4. Combining ratings 5. Constructs captured 6. Interviewer influences 7. Use of other literatures 8. Applicant individual differences 9. Interviewee–interviewer dynamics 10. Interview to interview variability

Status Largely resolved because of large-scale meta-analyses such as Wiesner and Cronshaw’s (1988). Very much unresolved. To reach resolution, we need a much better understanding of the constructs interviews capture and the factors that influence their measurement. Resolved scientifically by Dipboye (1982, 1989) and others. Although extensive review of preinterview information is not advised, it is still done commonly in practice. Largely unresolved, although method of combining ratings is not as much of an issue since ratings are generally summed in modern structured interviews. Progress being made; perhaps the new frontier of interview research. Unresolved, with a long history of being overlooked in interview research. Slightly better but still vastly underutilized. Other than for mental ability and to a lesser extent personality, this aspect of interview research has for the most part been overlooked. Definite progress being made; perhaps another new frontier of interview research. Other than for structure and to a lesser extent purpose (selection versus recruitment), identification of interview variability factors and their influence on the interview process and the constructs assessed has been vastly overlooked and may be a particularly ripe area for future research.

193

Copyright American Psychological Association. Not for further distribution.

Huffcutt and Culbertson

We wish to highlight two issues in particular because they were raised decades ago but are still ongoing and relate to the fundamental nature or purpose of the interview. One is the scope. By default, interviews in our modern era seem to have gravitated toward being a broad measure of job-related characteristics, yet the potential remains for the interview to be more effective with a narrower scope. Resolving this issue will require obtaining a better understanding of constructs captured by interview ratings, one of the exciting directions of current research. It is entirely possible that the interview will be found to routinely assess at least some characteristics that could be better assessed through psychological testing, in which case the interview could be focused more selectively on characteristics not as easily assessed through psychological testing (e.g., social skills, per Rundquist, 1947). The other is the idea that there is no one “interview,” but rather that the interview is a method that can vary widely from situation to situation depending on features such as the nature of the job, the type of questions, the interpersonal skills and personality of the interviewer, and the purpose of the interview (per Mayfield, 1964). In this line of thinking, what is measured by interview ratings is dependent on the specific combination of features for a given interview situation. Other than for level of structure and possibly interview purpose (selection vs. recruitment), this aspect of the interview has been largely overlooked. THE PARAMOUNT ROLE OF STRUCTURE If a sample of researchers and practitioners in the field were asked what has been the single most important influence on the interview process and its outcome, a majority would no doubt answer that it is structure. Structure has literally changed the way interviews are conducted, from being essentially a loosely constructed conversation to a uniform and carefully defined process, one that can approach standardized psychological tests psychometrically.

The Meaning of Structure There are several operationalizations of the term structure as applied to the employment interview. In a classic definition, Huffcutt and Arthur (1994) 194

defined structure in terms of reducing procedural variability across candidates. Such a definition has a somewhat psychometric flavor in that it positions the interview in potentially the same context as mental ability and personality tests. Put in terms of stimulus and response, interviews can be structured to such a degree that all candidates are exposed to the same stimuli (i.e., exactly worded questions with no probing) as is done in most of specifically psychometric tests (adaptive testing would be the exception), and responses are evaluated individually according to carefully defined and benchmarked rating scales. However, evaluation of responses in the interview can never match that of mental ability tests because the latter use absolute scoring (generally right or wrong), while rating of interview responses will always entail at least some degree of subjectivity. Still, interviews can be structured to the point where they come close to standardized tests, a point that is verified empirically by interview interrater reliabilities that can meet or even exceed test-retest reliabilities of some mental ability and personality tests (cf. Conway et al., 1995). More practically, Huffcutt and Arthur (1994) identified four levels of question standardization and three levels of response standardization, which they combined into four overall levels of structure. In the first level, the lowest level, there are no formal constraints on the questions asked and a global evaluation of responses is used. The second level includes limited constraints on the questions and some degree of structure in how the responses are evaluated. Level 3 has some prespecified questions and responses that are evaluated along a specified set of job dimensions after the interview. Finally, Level 4, the highest level of structure, has completely prespecified questions, the responses to each of which are rated on a customized scale with benchmark answers. The corrected validities of these four levels (corrected for range restriction in the interview and unreliability in the performance criteria) are shown in Table 6.2 along with the mean corrected validity of other predictors. As is evident, the validity of Level 3 and Level 4 structure compares favorably with the corrected mean validity of .51 for mental ability tests, .54 for work sample tests, and .48 for job knowledge tests (Schmidt & Hunter, 1998). Not

Interviews

TABLE 6.2 Corrected Validities for the Employment Interview and Other Selection Predictors

Copyright American Psychological Association. Not for further distribution.

Predictor

Mean validity

All interviews Level 1 Structure Level 2 Structure Level 3 Structure Level 4 Structure

.37 .20 .35 .56 .57

Mental Ability Tests Work Sample Tests Integrity Tests Job Knowledge Tests Reference Checks Biographical Data Measures Assessment Centers Graphology

.51 .54 .41 .48 .26 .35 .37 .02

Note. The validities for all interviews and for the four levels of interview structure are from Huffcutt and Arthur (1994) and are corrected for range restriction in the interview and unreliability in the performance criteria. All remaining validities, which are from Schmidt and Hunter (1998), are from various other meta-analyses and corrected for range restriction in the predictor and downward bias due to measurement error in job performance ratings.

surprisingly, increased structure has improved the utility of interview ratings as well, from very low to a level that is comparable to mental ability tests and other top predictors (Hunter & Hunter, 1984; Schmidt & Hunter, 1998). An interesting aspect of the Huffcutt and Arthur (1994) definition of structure is that it focuses largely on operational consistency. In their scheme, structure is enhanced by reducing variability in how candidates are interviewed. It is conceptually possible to have an interview developed at the highest level of structure (Level 4 questions, Level 3 ratings) that contains totally irrelevant content. Fortunately, structure and job-relatedness tend to covary because professionals who take the time to create questions and relevant rating scales usually do so directly in relation to the requirements of the position. Another seminal article that has provided an operationalization of interview structure is Campion, Palmer, and Campion (1997). They identified

15 ways in which an interview can be structured, separated into two categories: those that influence interview content and those that influence the evaluation process. The components of an interview that relate to content include (a) basing the questions on a job analysis, (b) asking the same questions of candidates, (c) limiting the extent to which the interviewer prompts the respondent, (d) using better types of questions (e.g., situational and past behavior questions rather than general background questions), (e) using longer interviews or a larger number of questions, (f) controlling ancillary information, and (g) not allowing questions from the candidate until after the interview. The interview components that relate to the evaluation process include (a) rating each answer or using multiple scales, (b) using detailed anchored rating scales, (c) taking detailed notes, (d) using multiple interviewers, (e) using the same interviewer for all candidates, (f ) refraining from discussing candidates and their answers between interviews, (g) providing extensive interviewing training, and (h) using statistical rather than clinical prediction when combining data. Their operationalization of interview structure is quite helpful, particularly from a practical standpoint. As these authors noted, the reliability, validity, and user reactions of the interview can be easily enhanced by incorporating at least some of these 15 components. Although there is no magic number of components for making an ideal interview, and it is unclear which, if any, of the components could be considered more critical than others, it is presumed that the more components that can be taken into account, the better. Building on Campion et al.’s (1997) conceptualization, Chapman and Zweig (2005) sought to identify the factor structure of interview structuring practices in order to develop a measure of interview structure. Using a field survey of 812 interviewees and 592 interviewers, they found four dimensions of interview structure: evaluation standardization, question consistency, question sophistication, and rapport building. In general, they noted that more structure is afforded to interviews in which evaluation is more standardized, questions are more consistent and sophisticated, and rapport building within the interview is lower. 195

Huffcutt and Culbertson

Copyright American Psychological Association. Not for further distribution.

The Importance of Structure Why is structure so important? Several major metaanalyses have confirmed that structured interviews fare considerably better than unstructured interviews psychometrically. For instance, Wiesner and Cronshaw (1988) reported mean criterion-related validity coefficients of .31 for unstructured interviews versus .62 for structured interviews. Conway et al. (1995) reported mean interrater reliability coefficients of .37 for Level 1 interviews versus .66 for Level 4 interviews (when conducted individually). But what is it about structured interviews that makes them superior? One obvious reason is greater consistency in content and process, which influences reliability and is important because reliability provides an upper limit for validity (Conway et al., 1995). Schmidt and Zimmerman (2004) found some evidence that the superiority of structured interviews can be traced largely to greater reliability. Put in a different way, unstructured interviews, by their very nature, provide more opportunity for random error in responses, more random error in what the responses represent, and more random error in evaluation. This random error throughout the interview process likely leads to lower reliability and validity for unstructured interviews. Another reason is that the standardization inherent in structured interviews serves to make the content more job-related. It is not surprising, therefore, that the first component on which to base interviews that Campion et al. (1997) identified for structuring interviews was a quality job analysis. Without this basis, unstructured interviews run the risk of tapping constructs that are unrelated to the job, and hence result in low validity. Finally, the standardization and consistency found in structured interviews helps to reduce such troubling information-processing tendencies as contrast and similarity effects that often plague unstructured interviews (Latham, Wexley, & Pursell, 1975; Wexley, Sanders, & Yukl, 1973). For example, because the questions and scoring for structured interviews are the same across participants, there is less opportunity to compare applicants to each other rather than to the constructs and dimensions of interest. Along these lines, the cognitive complexity of response processing can be reduced through the increased structure (Tsai, Chen, & Chiu, 2005). 196

Similarly, there is likely a reduction in interviewer susceptibility to influence tactics, which would thereby help keep the interview job-related and the ratings consequently more valid.

Creating Structure in Interviews Given the clear advantages of increased structure in interviews, a logical transition is to discuss how structured interviews are developed. There are two general approaches for making a job-related structured interview. One approach is to create questions directly from the KSAOs. For example, Campion, Pursell, and Brown (1988) created a list of duties and requirements and wrote questions based on the KSAOs needed to perform them. The other approach, which is used more frequently, involves collecting critical incidents and turning them into questions. For example, Motowidlo et al. (1992) conducted critical incident workshops in which participants wrote critical incidents describing management performance and the researchers identified and defined various interpersonal and problem-solving dimensions based on them. Both approaches involve a fair degree of literary license, which could lead to differences in content and validity across interview developers. Additionally, there is some evidence that formal interviewer training contributes to greater standardization, consistency, and formalization in the evaluation processes of interviews, as does having more of a selection (vs. recruitment) focus for the interview (Chapman & Zweig, 2005). Certainly, this is not altogether surprising, as Campion et al. (1997) noted that training is one of the components that influence the evaluation process. Nevertheless, if one is interested in creating more structure in interviews, proper interviewer training is clearly of importance. Selecting the questions to use involves a variety of options. The two types of questions described earlier, situational and behavior description, have become very popular in recent decades. Several meta-analyses have demonstrated that both of these formats yield exceptional predictability overall, with mean corrected criterion-validity estimates ranging from .47 to .63 (e.g., Latham & Sue-Chan, 1999; Taylor & Small, 2002). The validity of these interviews may vary, however, depending on the type of job for which the interview is used. For example, several studies have shown

Copyright American Psychological Association. Not for further distribution.

Interviews

that the validity of the situational interview is not as great as the behavioral description for higher-level jobs or those higher in complexity, such as managerial jobs (e.g., Huffcutt, Conway, Roth, & Klehe, 2004; Huffcutt, Weekley, et al., 2001; Krajewski, Goffin, McCarthy, Rothstein, & Johnston, 2006; Pulakos & Schmitt, 1995). Additionally, the specific question type used may actually influence the way in which candidates behave during an interview. For example, Ellis et al. (2002) found that situational questions led to applicants using more ingratiation techniques, while behavior description questions resulted in candidates using more self-promotion tactics. There are also structured formats that utilized multiple types of questions. For example, Campion, Pursell, and Brown (1988) developed a comprehensive structured interview for entry-level employees at a pulp-and-paper mill that contained job knowledge questions assessing mechanical comprehension, simulation questions gauging reading ability, and worker characteristic or willingness questions determining fear of heights. Schuler (1989, 1992) devised a multimodal employment interview that included selfpresentation, vocational, biographical, and situational questions. Developers of interviews may also consider making the questions somewhat transparent, or clear as to what the question is trying to assess. A recent study conducted by Klehe, Konig, Richter, Kleinmann, and Melchers (2008) found that interviewees tended to perform better when interviews were transparent. Perhaps more important, although both nontransparent and transparent interviews were not significantly different in terms of their criterion-related validity, transparent interviews yielded greater construct validity than nontransparent interviews. Thus, although not directly related to the structure of interviews, it is definitely a consideration worthy of noting. CONCLUDING SUMMARY AND DIRECTIONS FOR FUTURE RESEARCH In this chapter, we have provided an overview of the history of employment interview research, tracing progress in this area from the early 1900s through to the present day. This overview makes it apparent that

some key issues have largely been resolved (e.g., reliability, validity, structure), some have not been addressed in meaningful ways yet (e.g., scope), and some are being explored in new and exciting ways in current research (e.g., constructs, interviewee– interviewer dynamics). Given its paramount role in the interview process, we also took a closer look at structure, including how it has been conceptualized and operationalized, what makes it so critical to enhancing validity and reliability, and how to increase it. Of all the advances over almost a century of interview research, arguably the most clear and practically important finding is the superiority of structured interviews over unstructured interviews. We now conclude with a brief discussion of future research needs and directions. The issues summarized in Table 6.1 that have not been fully resolved provide an excellent starting point, including scope, individual differences in interviewees and interviewers, and interviewee–interviewer dynamics. To illustrate, there may be differences in the effectiveness of interviews in different countries and across cultures. As organizations become more global and outsourcing becomes more prevalent, it is important to consider the effects of any selection technique in the context of the organization’s business territory as well as the applicant’s national and cultural background. Although studies have been conducted using samples from a variety of countries, most of the research in this area has studied samples from the United States. It is unclear, at this point, whether differences exist in interviews across countries, yet there is reason to believe they might (Moy, 2006). Similarly, the recent shift to a greater focus on constructs highlights a plethora of potential avenues. For example, there has been limited research to date that has examined the roles of social skills, self-monitoring, and relational control as key dynamics of the interview process. Further research could certainly aid in a better understanding not only of how these different activities influence the interview process and outcomes but also of how they might change what constructs the interview captures. Returning to the use of meta-analysis to help summarize what we know about the validity and reliability of interviews while accounting for various artifacts, future researchers should revisit some 197

Copyright American Psychological Association. Not for further distribution.

Huffcutt and Culbertson

lingering questions. As noted earlier, there have been recent and substantial advances in the ways to deal with the range restriction correction (Hunter, Schmidt, & Le, 2006; Le & Schmidt, 2006; Schmidt et al., 2006). Although researchers have already begun to reanalyze some relationships (e.g., Berry et al., 2007), more work is certainly called for in this area. Another area worthy of research is the role of intentional response distortion within the employment interview. Although a plethora of research exists that examines the role of faking for such selection methods as personality testing, relatively little research has examined such issues for the employment interview (see Fletcher, 1990; Levashina & Campion, 2007, for exceptions). Although the research in this area has provided useful theoretical and practical insight (e.g., it appears questions that assess past behavior are more resistant to faking than are situational questions, and follow-up questions increase faking; Levashina & Campion, 2007), it remains unknown what the effect of faking (or outright lying) is on the reliability and validity of the employment interview. This is particularly important considering Levashina and Campion’s finding that more than 90% of undergraduate job candidates engaged in faking during employment interviews. Researchers have only begun to examine the impact of coaching as it relates to the employment interview, with the majority of work in this area coming from Maurer and colleagues (Maurer et al., 2001; Maurer, Solamon, & Troxtel, 1998; Maurer, Solamon, & Lippstreu, 2008). In general the research from this area has demonstrated that interviewees tend to perform better when they are coached, and the predictive validity and reliability of the interview is higher for those applicants who are coached than for those who are not coached. More work in the area of coaching and its effects on the various types of interviews, in terms of interviewee performance as well as the psychometric properties of the interviews, is still warranted. Posthuma et al. (2002) raised the question of how the cognitive demands of the interview influence the process and outcomes, including in relation to cognitive capability of the interviewees and interviewer. For instance, are the correlations that have 198

typically been found between interview ratings and mental ability (e.g., Berry et al., 2007; Huffcutt, Roth, et al., 1996) more a function of the cognitive demands of the interview or the cognitive capabilities of the candidates? An examination of Huffcutt, Van Iddekinge, and Roth’s (2007, 2009) model of interview constructs (the 2007 version is shown in Figure 6.1 and described earlier in this chapter) is also likely to provide ideas for years of research going forward. One could quite easily take any component of the model and find a multitude of questions awaiting research. For example, regarding job-related content, more research is needed to clarify the role of personality factors in the interview, as well as an examination of how different interview dimensions might relate to different kinds of mental ability (e.g., verbal ability vs. memory). Similarly, it is not clear to what extent other general human attributes (e.g., applicant goals, interests, and values), core job elements (e.g., KSAOs, motivation), or experiential factors (e.g., experience, training, education) may be assessed in both structured and unstructured interviews. These points, of course, are only for the job-related content portion of Huffcutt et al.’s (2007) model. Certainly, more research is justified for the interviewee performance and interviewer rating tendencies sections of their model. Future research on employment interviews may also be influenced by technological advances. For example, as organizations have become more global, alternate mediums (other than face-to-face) have become more common. Researchers have examined the impact of such alternatives as telephone and videoconference interviews on applicant ratings and reactions (e.g., Chapman & Rowe, 2001; Chapman, Uggerslev, & Webster, 2003; Chapman & Webster, 2001; Silvester, Anderson, Haddleton, CunninghamSnell, & Gibb, 2000; Straus, Miles, & Levesque, 2001). Certainly, as more and more technological advances occur and business adapts to capture these advances, it will be important to determine whether changes in the reliability, validity, and practicality of the interview will be hindered or enhanced. Finally, there is the potential for incorporating some research from neuroscience into the research domain. For example, behavioral description inter-

Copyright American Psychological Association. Not for further distribution.

Interviews

views clearly require higher mental functions, such as memory searching and analyzing. Along these lines, there is evidence that these higher mental functions require large amounts of brain sugars (glucose, glycogen) to function optimally and that they tend to use up these sugars rather quickly (Gailliot, 2008; Gailliot & Baumeister, 2007). Furthermore, these sugars appear to be used up even more rapidly in situations involving high self-regulation (Masicampo & Baumeister, 2008), which again is characteristic of most interviews. When brain sugar levels drop, there is a tendency for individuals to rely more on mental shortcuts (heuristics) rather than doing full cognitive processing. This is just one example of how findings from neuroscience may shed light on interview processes. As such, this appears to be an area into which future researchers may want to delve. In conclusion, it is clear that the interview has a long and rich history of research, almost a century at this point. Researchers have learned much over this time, including how to structure an interview and the psychometric benefits that result, and they have raised numerous questions that are being addressed currently and can continue to be addressed in future research. We look forward to many more exciting years of interview research, which should serve to both strengthen our understanding of this unique selection device and increase its efficacy in the organizational arena.

References Adkins, C. L., Russell, C. J., & Werbel, J. D. (1994). Judgments of fit in the selection process: The role of work value congruence. Personnel Psychology, 47, 605–623. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–204. Anderson, C. W. (1960). The relation between speaking times and decision in the employment interview. Journal of Applied Psychology, 44, 267–268. Arthur, W., Jr., Woehr, D. J., & Maldegen, R. (2000). Convergent and discriminant validity of assessment center dimensions: A conceptual and empirical reexamination of the assessment center construct-related validity paradox. Journal of Management, 26, 813–815. Arvey, R. D., & Campion, J. E. (1982). The employment interview: A summary and review of recent research. Personnel Psychology, 35, 281–322.

Baddely (1986). Working memory. New York: Oxford University Press. Bandura, A. (1986). Psychological mechanisms of aggression. In R. G. Green & E. I. Donnerstein (Eds.), Aggression: Theoretical and empirical reviews (Vol. 1, pp. 1–40). New York: Academic Press. Bateson, G. (1958). Naven. Stanford, CA: Stanford University Press. Belov, D. I., & Armstrong, R. D. (2008). A Monte Carlo approach to the design, assembly, and evaluation of multistage adaptive tests. Applied Psychological Measurement, 32, 119–137. Berry, C. M., Sackett, P. R., & Landers, R. N. (2007). Revisiting interview–cognitive ability relationships: Attending to specific range restriction mechanisms in meta-analysis. Personnel Psychology, 60, 837–874. Bouchard, T. J., Jr. (2004). Genetic influences on human psychological traits. Current Directions in Psychological Science, 13, 148–151. Brunswik, E. (1955). Representative design and probabilistic theory. Psychological Review, 62, 236–242. Brunswik, E. (1956). Perception and the representative design of psychological experiments. Berkeley: University of California Press. Burnett, J. R., & Motowidlo, S. J. (1998). Relations between different sources of information in the structured selection interview. Personnel Psychology, 51, 963–983. Cable, D. M., & Judge, T. A. (1997). Interviewers’ perceptions of person–organization fit and organizational selection decisions. Journal of Applied Psychology, 82, 546–561. Campion, M. A., Campion, J. E., & Hudson, J. P. (1994). Structured interviewing: A note on incremental validity and alternative question types. Journal of Applied Psychology, 79, 998–1102. Campion, M. A., Palmer, D. K., & Campion, J. E. (1997). A review of structure in the selection interview. Personnel Psychology, 50, 655–702. Campion, M. A., Palmer, D. K., & Campion, J. E. (1998). Structuring employment interviews to improve reliability, validity and users’ reactions. Current Directions in Psychological Science, 7, 77–82. Campion, M. A. Pursell, E. D., & Brown, B. K. (1988). Structured interviewing: Raising the psychometric properties of the employment interview. Personnel Psychology, 41, 25–42. Chapman, D. S., & Rowe, P. M. (2001). The impact of videoconference technology, interview structure, and interviewer gender on interviewer evaluations in the employment interview: A field experiment. Journal of Occupational and Organizational Psychology, 74, 279–298. 199

Huffcutt and Culbertson

Chapman, D. S., Uggerslev, K. L., & Webster, J. (2003). Applicant reactions to face-to-face and technologymediated interviews: A field investigation. Journal of Applied Psychology, 88, 944–953. Chapman, D. S., & Webster, J. (2001). Rater correction processes in applicant selection using videoconference technology: The role of attributions. Journal of Applied Social Psychology, 31, 2518–2537.

Copyright American Psychological Association. Not for further distribution.

Chapman, D. S., & Zweig, D. I. (2005). Developing a nomological network for interview structure: Antecedents and consequences of the structured selection interview. Personnel Psychology, 58, 673–702. Conway, J. M., Jako, R. A., & Goodman, D. F. (1995). A meta-analysis of interrater and internal consistency reliability of selection interviews. Journal of Applied Psychology, 80, 565–579. Cortina, J. M., Goldstein, N. B., Payne, S. C., Davison, H. K., & Gilliland, S. W. (2000). The incremental validity of interview scores over and above g and conscientiousness scores. Personnel Psychology, 53, 325–351. DeGroot, T., & Motowidlo, S. J. (1999). Why visual and vocal interview cues can affect interviewers’ judgments and predict job performance. Journal of Applied Psychology, 84, 986–993. Dipboye, R. L. (1982). Self-fulfilling prophecies in the selection/recruitment interview. Academy of Management Review, 7, 579–587. Dipboye, R. L. (1989). Threats to the incremental validity of interviewer judgments In R. W. Eder & G. R. Ferris (Eds.), The employment interview: Theory, research, and practice (pp. 45–60). Newbury Park, CA: Sage. Dipboye, R. L. (1992). Selection interviews: Process perspectives. Cincinnati, OH: South-Western. Dipboye, R. L. (2005). The selection/recruitment interview: Core processes and contexts. In A. Evers, N. R. Anderson, & O. F. Smit-Voskuijl (Eds.), The Blackwell handbook of personnel selection (pp. 121–142). Malden, MA: Blackwell. Dipboye, R. L., & Gaugler, B. B. (1993). Cognitive and behavioral processes in the selection interview. In N. Schmitt & W. Borman (Eds.), Personnel selection in organizations (pp. 135–170). San Francisco: Jossey-Bass.

sis. In R. W. Eder & G. R. Ferris (Eds.), The employment interview: Theory, research, and practice (pp. 17–31). Newbury Park, CA: Sage. Ellis, A. P., West, B. J., Ryan, A. M., & DeShon, R. P. (2002). The use of impression management tactics in structured interviews: A function of question type? Journal of Applied Psychology, 87, 1200–1208. Engler-Parish, P. G., & Millar, F. E. (1989). An exploratory relational control analysis of the employment screening interview. Western Journal of Speech Communication, 53, 30–51. Ferris, G. R., Witt, L. A., & Hochwarter, W. A. (2001). Interaction of social skill and general mental ability on job performance and salary. Journal of Applied Psychology, 86, 1075–1082. Fletcher, C. (1990). The relationships between candidate personality, self-presentation strategies, and interview assessments in selection interviews: An empirical study. Human Relations, 43, 739–749. Gailliot, M. T. (2008). Unlocking the energy dynamics of executive functioning. Perspectives on Psychological Science, 3, 245–263. Gailliot, M. T., & Baumeister, R. F. (2007). The physiology of willpower: Linking blood glucose to self-control. Personality and Social Psychology Review, 11, 303–327. Gifford, R., Ng, C. F., & Wilkinson, M. (1985). Nonverbal cues in the employment interview: Links between applicant qualities and interviewer judgments. Journal of Applied Psychology, 70, 729–736. Hakel, M. D. (1989). The state of employment interview theory and research. In R. W. Eder & G. R. Ferris (Eds.), The employment interview: Theory, research, and practice (pp. 285–293). Newbury Park, CA: Sage. Harris, M. M. (1989). Reconsidering the employment interview: A review of recent literature and suggestions for future research. Personnel Psychology, 42, 691–726. Herriot, P. (1981). Towards an attributional theory of the selection interview. Journal of Occupational Psychology, 54, 165–173. Higgins, C. A., & Judge, T. A. (2004). The effect of applicant influence tactics on recruiter perceptions of fit and hiring recommendations: A field study. Journal of Applied Psychology, 89, 622–632.

Dipboye, R. L., & Macan, T. M. (1988). A process view of the selection-recruitment interview. In R. Schuler, V. Huber, & S. Youngblood (Eds.), Readings in personnel and human resource management (pp. 217–232). New York: West.

Higgins, C. A., Judge, T. A., & Ferris, G. R. (2003). Influence tactics and work outcomes: A meta-analysis. Journal of Organizational Behavior, 24, 89–106.

Dougherty, T. W., Ebert, R. J., & Callender, J. C. (1986). Policy capturing in the employment interview. Journal of Applied Psychology, 71, 9–15.

Huffcutt, A. I., & Arthur, W., Jr. (1994). Hunter & Hunter (1984) revisited: Interview validity for entrylevel jobs. Journal of Applied Psychology, 79, 184–190.

Eder, R. W., Kacmar, K. M., & Ferris, G. R. (1989). Employment interview research: History and synthe-

Huffcutt, A. I., Conway, J. M., Roth, P. L., & Klehe, U.-C. (2004). The impact of job complexity and study

200

Hollingworth, H. L. (1922). Judging human character. New York: Appleton.

Interviews

design on situational and behavior description interview validity. International Journal of Selection and Assessment, 12, 262–273. Huffcutt, A. I., Conway, J. M., Roth, P. L., & Stone, N. J. (2001). Identification and meta-analytic assessment of psychological constructs measured in employment interviews. Journal of Applied Psychology, 86, 897–913.

Copyright American Psychological Association. Not for further distribution.

Huffcutt, A. I., Roth, P. L., & McDaniel, M. A. (1996). A meta-analytic investigation of cognitive ability in employment interview evaluations: Moderating characteristics and implications for incremental validity. Journal of Applied Psychology, 81, 459–473. Huffcutt, A. I., Van Iddekinge, C. H., & Roth, P. L. (2007, April). A comprehensive review of the constructs captured by employment interview ratings. Symposium held at the 22nd Annual Conference of the Society for Industrial and Organizational Psychology, New York.

Klehe, U.-C., Konig, C. J., Richter, G. M., Kleinmann, M., & Melchers, K. G. (2008). Transparency in structured interviews: Consequences for construct and criterionrelated validity. Human Performance, 21, 107–137. Krajewski, H. T., Goffin, R. D., McCarthy, J. M., Rothstein, M. G., & Johnston, N. (2006). Comparing the validity of structured interviews for manageriallevel employees: Should we look to the past or focus on the future? Journal of Occupational and Organizational Psychology, 79, 411–432. Krosnick, J. A., Betz, A. L., Jussim, L. J., & Lynn, A. R. (1992). Subliminal conditioning of attitudes. Personality and Social Psychology Bulletin, 18, 152–152. Latham, G. P. (1989), The reliability, validity and practicality of the situational interview. In R. W. Eder & G. R. Ferris (Eds.), The employment interview (pp. 162–182). Newbury Park, CA: Sage.

Huffcutt, A. I., Van Iddekinge, C. H., & Roth, P. L. (2009). An empirical review of the employment interview construct literature. Manuscript submitted for publication.

Latham, G. P., Saari, L. M., Pursell, E. D., & Campion, M. A. (1980). The situational interview. Journal of Applied Psychology, 65, 422–427.

Huffcutt, A. I., Weekley, J., Weisner, W. H., DeGroot, T., & Jones, C. (2001). Comparison of situational and behaviour description interview questions for higherlevel positions. Personnel Psychology, 54, 619–644.

Latham, G. P., & Sue-Chan, C. (1999). A meta-analysis of the situational interview: An enumerative review of the reasons for its validity. Canadian Psychology, 40, 56–67.

Hunter, J. E., & Hirsch, H. R. (1987). Applications of metaanalysis. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (Vol. 2, pp. 321–357). London: Wiley.

Latham, G. P., Wexley, K. N., & Pursell, E. D. (1975). Training managers to minimize rating errors in the observation of behavior. Journal of Applied Psychology, 60, 550–555.

Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternate predictors of job performance. Psychological Bulletin, 96, 72–98.

Lee, Y.-H., Ip, E. H., & Fuh, C. D. (2008). A strategy for controlling item exposure in multidimensional computerized adaptive testing. Educational and Psychological Measurement, 68, 215–232.

Hunter, J. E., & Schmidt, F. L. (2004). Meta-analysis: Correcting errors and bias in research findings (2nd ed.). Thousand Oaks, CA: Sage. Hunter, J. E., Schmidt, F. L., & Jackson, G. B. (1982). Meta-analysis: Cumulating research findings across studies. Beverly Hills, CA: Sage. Hunter, J. E., Schmidt, F. L., & Le, H. (2006). Implications of direct and indirect range restriction for meta-analysis methods and findings. Journal of Applied Psychology, 91, 594–612. Janz, T. (1982). Initial comparison of patterned behavior description interviews versus unstructured interviews. Journal of Applied Psychology, 67, 577–580. Janz, T. (1989). The patterned behavior description interview: The best prophet of the future is the past. In R. W. Eder & G. R. Ferris (Eds.), The employment interview: Theory, research, and practice (pp. 158–168). Thousand Oaks, CA: Sage.

Le, H., & Schmidt, F. L. (2006). Correcting for indirect range restriction in meta-analysis: Testing a new meta-analytic procedure. Psychological Methods, 11, 416–438. Levashina, J., & Campion, M. A. (2007). Measuring faking in the employment interview: Development and validation of an interview faking behavior scale. Journal of Applied Psychology, 92, 1638–1656. Loftus, E. (2007). Memory distortions: Problems solved and unsolved. In M. Garry & H. Hayne (Eds.), Do justice and let the sky fall (pp. 1–14). Mahwah, NJ: Erlbaum. Loftus, E. & Cahill, L. (2007). Memory distortion: From misinformation to rich false memory. In J. S. Nairne (Ed.), The foundation of remembering: Essays in honor of Henry L. Roediger, III (pp. 413–425). New York: Psychology Press.

Kahneman, D. & Tversky, A. (1973). On the psychology of prediction. Psychological Review, 80, 237–251.

Masicampo, E. J., & Baumeister, R. F. (2008). Toward a physiology of dual-process reasoning and judgment: Lemonade, willpower, and expensive rule-based analysis. Psychological Science, 19, 255–260.

Kahneman, D. & Tversky, A. (2000). Choices, values, and frames. New York: Cambridge University Press.

Maurer, T. J., Solamon, J. M., Andrews, K. D., & Troxtel, D. D. (2001). Interviewee coaching, preparation 201

Huffcutt and Culbertson

strategies, and response strategies in relation to performance in situational employment interviews: An extension of Maurer, Solamon, and Troxtel (1998). Journal of Applied Psychology, 86, 709–717. Maurer, T. J., Solamon, J. M., & Lippstreu, M. (2008). How does coaching interviewees affect the validity of a structured interview? Journal of Organizational Behavior, 29, 355–371.

Copyright American Psychological Association. Not for further distribution.

Maurer, T. J., Solamon, J. M., & Troxtel, D. D. (1998). Relationship of coaching with performance in situational employment interviews. Journal of Applied Psychology, 83, 128–136. Mayfield, E. C. (1964). The selection interview: A reevaluation of published research. Personnel Psychology, 17, 239–260. McDaniel, M. A., Whetzel, D. L., Schmidt, F. L., & Maurer, S. (1994). The validity of employment interviews: A comprehensive review and meta-analysis. Journal of Applied Psychology, 79, 599–617. McFarland, L. A., Ryan, A. M., & Kriska, S. D. (2003). Impression management use and effectiveness across measurement methods. Journal of Management, 29, 641–661. McMurry, R. N. (1947). Validating the patterned interview. Personnel, 23, 263–272. Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81–97. Motowidlo, S. J., Carter, G. W., Dunnette, M. D., Tippins, N., Werner, S., Burnett, J. R., & Vaughan, M. J. (1992). Studies of the structured behavioral interview. Journal of Applied Psychology, 77, 571–587. Moy, J. M. (2006). Are employers assessing the right traits in hiring? Evidence from Hong Kong companies. International Journal of Human Resource Management, 17, 734–754. Nairne, J. S., Neath, I., & Serra, M. (1997). Positional distinctiveness and ratio rule in free recall. Journal of Memory and Language, 37, 155–166. Nyfield, G., & Baron, H. (2000). Cultural context in adapting selection practices across borders. In J. F. Kehoe (Ed.), Managing selection in changing organizations (pp. 242–270). San Francisco: Jossey-Bass. Olson, J. M., Vernon, P. A., Harris, J. A., & Jang, K. L. (2001). The heritability of attitudes: A study of twins. Journal of Personality and Social Psychology, 80, 845–860. O’Rourke, L. J. (1932). Measuring judgment and resourcefulness. Personnel Journal, 7, 427–440. Posthuma, R. A., Morgeson, F. P., & Campion, M. A. (2002). Beyond employment interview validity: A comprehensive narrative of review of recent research and trends over time. Personnel Psychology, 55, 1–81. 202

Pulakos, E. D., & Schmitt, N. (1995). Experience-based and situational interview questions: Studies of validity. Personnel Psychology, 48, 289–308. Raza, S. M., & Carpenter, B. N. (1987). A model of hiring decisions in real employment interviews. Journal of Applied Psychology, 72, 596–603. Roth, P. L., Van Iddekinge, C. H., Huffcutt, A. I., Eidson, C. E. Jr., & Schmit, M. J. (2005). Personality saturation in structured interviews. International Journal of Selection and Assessment, 13, 261–273. Rowe, P. M. (1989). Unfavorable information and interview decisions. In R. W. Eder & G. R. Ferris (Eds.), The employment interview: Theory, research, and practice. (pp. 77–89). Beverly Hills: Sage. Rundquist, E. A. (1947). Development of an interview for selection purposes. In G. A. Kelly (Ed.), New methods in applied psychology (pp. 85–95). College Park: University of Maryland. Rynes, S. L., & Gerhart, B. (1990). Interviewer assessments of applicant “fit”: An exploratory investigation. Personnel Psychology, 43, 13–35. Sackett, P. R. (1982). The interviewer as hypothesis tester: The effects of impressions of an applicant on interviewer questioning strategy. Personnel Psychology, 35, 789–803. Salgado, J. F., & Moscoso, S. (2002). Comprehensive metaanalysis of the construct validity of the employment interview. European Journal of Work and Organizational Psychology, 11, 299–324. Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124, 262–274. Schmidt, F. L., Oh, I.-S., & Le, H. (2006). Increasing accuracy of corrections for range restriction: Implications for selection procedure validities and other research results. Personnel Psychology, 59, 281–305. Schmidt, F. L., & Zimmerman, R. D. (2004). A counterintuitive hypothesis about employment interview validity and some supporting evidence. Journal of Applied Psychology, 89, 553–561. Schmitt, N. (1976). Social and situational determinants of interview decisions: Implications for the employment interview. Personnel Psychology, 29, 79–101. Schuler, H. (1989). Construct validity of a multi-model employment interview. In B. J. Fallon, H. P. Pfister, & J. Brebner (Eds.), Advances in industrial and organizational psychology. New York: North Holland. Schuler, H. (1992). Das multimodale einstellungsinterview [The multimodal employment interview]. Diagnostica, 38, 281–300. Schuler, H., & Funke, U. (1989). The interview as a multimodal procedure. In R. W. Eder & G. R. Ferris (Eds.),

Interviews

The employment interview: Theory, research, and practice (pp. 183–192). Thousand Oaks, CA: Sage. Scott, W. D. (1915, October). Scientific selection of salesmen. Advertising and Selling Magazine, pp. 5–6, 94–96. Scott, W. D., Bingham, W. V., & Whipple, G. M. (1916). Scientific selection of salesmen. Salesmanship, 4, 106–108.

Copyright American Psychological Association. Not for further distribution.

Sharf, J. C., & Jones, D. P. (2000). Employment risk management. In J. F. Kehoe (Ed.), Managing selection in changing organizations (pp. 271–318). San Francisco: Jossey-Bass. Silvester, J., Anderson, N., Haddleton, E., CunninghamSnell, N., & Gibb, A. (2000). A cross-modal comparison of telephone and face-to-face selection interviews in graduate recruitment. International Journal of Selection and Assessment, 8, 16–21. Society for Industrial and Organizational Psychology, Inc. (2003). Principles for the validation and use of personnel selection procedures (4th ed.). College Park, MD: Author. Springbett, B. M. (1958). Factors affecting the final decision in the employment interview. Canadian Journal of Psychology, 12, 13–22. Stevens, C. K., & Kristof, A. L. (1995). Making the right impression: A field study of applicant impression management during job interviews. Journal of Applied Psychology, 80, 587–606. Straus, S. G., Miles, J. A., & Levesque, L. L. (2001). The effects of videoconference, telephone, and face-toface media on interviewer and applicant judgments in employment interviews. Journal of Management, 27, 363–381. Tay, C., Ang, S., & Van Dyne, L. (2006). Personality, biographical characteristics, and job interview success: A longitudinal study of the mediating effects of interviewing self-efficacy and the moderating effects of internal locus of causality. Journal of Applied Psychology, 91, 446–454. Taylor, P. J., & Small, B. (2002). Asking applicants what they would do versus what they did do: A meta-analytic comparison of situational and past

behaviour employment interview questions, Journal of Occupational and Organizational Psychology, 75, 277–294. Tsai, W.-C., Chen, C.-C., & Chiu, S.-F. (2005). Exploring boundaries of the effects of applicant impression management tactics in job interviews. Journal of Management, 31, 108–125. Tuller, W. L. (1989). Relational control in the employment interview. Journal of Applied Psychology, 74, 971–977. Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5, 207–232. Tversky, A. & Kahneman, D. (1982). Judgment under uncertainty: Heuristics and biases. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 251–258). New York: Cambridge University Press. Ulrich, L., & Trumbo, D. (1965). The selection interview since 1949. Psychological Bulletin, 63, 110–116. Wagner, R. (1949). The employment interview: A critical summary. Personnel Psychology, 2, 17–46. Webster, E. C. (1964). Decision-making in the employment interview. Montreal: Industrial Relations Center, McGill University. Wexley, K. N., Sanders, R. E., & Yukl, G. A. (1973). Training interviewers to eliminate contrast effects in employment interviews. Journal of Applied Psychology, 57, 233–236. Wiesner, W. H., & Cronshaw, S. F. (1988). The moderating impact of interview format & degree of structure on interview validity. Journal of Occupational Psychology, 61, 275–290. Wonderlic, E. F. (1942). Improving interview technique. Personnel, 18, 232–238. Wright, O. R., Jr. (1969). Summary of research on the selection interview since 1964. Personnel Psychology, 22, 391–413. Zedeck, S., Tziner, A., & Middlestadt, S. E. (1983). Interviewer validity and reliability: An individual analysis approach. Personnel Psychology, 36, 355–370.

203

CHAPTER 7

ASSESSMENT CENTERS

Copyright American Psychological Association. Not for further distribution.

Winfred Arthur Jr. and Eric Anthony Day

The objective of this chapter is to provide the reader with an overview of the science and practice of assessment centers (ACs). To accomplish this objective, we present a broad summary overview of the pertinent issues in the extant AC literature. Thus, the chapter starts with a definition and description of ACs along with their historical background. This is followed with a description of the design, development, administration, and scoring of ACs. The next section pertains to the psychometric properties— namely, the reliability and validity of AC ratings— which is then followed by a discussion of costs and practical issues associated with the use of ACs. In the next two sections, we discuss the issue of subgroup differences and the potential for adverse impact as well as the international scope of AC research and practice. Finally, we conclude with a brief discussion of directions for future research and practice. ASSESSMENT CENTERS: DEFINITION AND DESCRIPTION In keeping with the predictor construct–predictor method distinction, Arthur and Villado (2008) described a predictor as a specific behavioral domain, information about which is sampled via a specific method. Thus, depending on one’s focus, predictors can be represented in terms of what they measure [i.e., the constructs or content of the predictor] and how they measure what they are designed to

measure [i.e., the method of assessment or measurement]. (p. 435) AC science and practice specifies the behavioral domain of ACs in terms of dimensions (although in some of the recent practitioner literature the term competencies is being used; e.g., see International Task Force on Assessment Center Guidelines, 2000). Hence, as predictors, ACs are best conceptualized as a method by which information concerning multiple behavioral dimensions is collected. The traditional use of ACs has been in managerial contexts for administrative purposes (e.g., selection and promotion), although the flexibility inherent in the AC methodology has led to its use in a wide variety of human resource settings and purposes. A general overview of the AC process is illustrated in Figure 7.1. In the AC method, participants work through a series of behavioral exercises (e.g., situational exercises and job simulations such as leaderless group discussions, in-baskets, and role plays). Multiple assessors observe and document participants’ behavior in the exercises and then sort and organize their observations in terms of the focal dimensions of interest. After they have observed and documented participants’ performance in all the exercises, the team of assessors meets and arrives at final dimension ratings for each participant. Finally, if warranted, such as in a developmental AC, assessors then meet with participants in one-on-one feedback interviews to review and discuss the participant’s performance on the assessed behavioral dimensions and the strategies and activities for improvement. Participants receive a written feedback report at a later date.

http://dx.doi.org/10.1037/12170-007 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

205

Arthur and Day

1. Participants work through exercises.

Copyright American Psychological Association. Not for further distribution.

2. Assessors observe and record behaviors, focusing on the participants to whom they have been assigned (see Figure 7.6). Assessors are typically assigned to observe two participants.

3. Assessors sort and organize observations and may generate initial dimension-level scores for the exercise.

4. After all the exercises are completed, assessors meet (typically for hours) to generate participants’ dimension-level scores. See the Scoring and Rating Approach section of this chapter for a detailed description of this process. 5. Depending on the purpose of the assessment center, assessors may meet with and provide oneon-one feedback to each participant. Participants may also later receive a formal, written feedback report.

FIGURE 7.1. A general overview of the assessment center process.

In summary, an AC is a comprehensive, standardized procedure that uses multiple techniques (exercises) and assessors to assess multiple behavioral dimensions of interest (International Task Force on Assessment Center Guidelines, 2000). A noteworthy distinguishing feature embedded in this definition is the use of multiple exercises (i.e., methods) to obtain multiple dimension scores. Figure 7.2 and Figure 7.3 present descriptions of some commonly used AC dimensions and exercises, respectively. Hence the standard design and use of ACs is to cross dimensions and exercises, and then collapse performance across exercises to obtain dimension scores from a “single” method; this feature is illustrated in Figure 7.4. It is important to note that dimension scores are also collapsed across multiple raters or assessors. So, the triangulation of dimensions through 206

the use of multiple methods and multiple assessors is, in fact, the major defining characteristic of ACs. Therefore, because the set of specified multiple methods represents an AC, ACs are best conceptualized as a method instead of multiple methods. HISTORICAL BACKGROUND The origins of the AC can be traced to military and industrial efforts surrounding World War II, primarily first with the Germans, then the British War Office Selection Boards, and then the United States Office of Strategic Services (OSS; Thornton & Byham, 1982). With the massive human resources efforts needed to meet the demands of such a large-scale operation, military psychologists were faced with the critical task of developing procedures for identifying and

Assessment Centers

COMMUNICATION: The extent to which an individual conveys oral and written information and responds to questions and challenges. CONSIDERATION/AWARENESS OF OTHERS: The extent to which an individual’s actions reflect a consideration for the feelings and needs of others as well as an awareness of the impact and implications of decisions relevant to other components both inside and outside the organization.

Copyright American Psychological Association. Not for further distribution.

DRIVE: The extent to which an individual originates and maintains a high activity level, sets performance standards and persists in his or her achievement, and expresses the desire to advance to higher job levels. INFLUENCING OTHERS: The extent to which an individual persuades others to do something or adopt a point of view in order to produce desired results and takes action in which the dominant influence is his or her own convictions rather than the influence of others’ opinions. ORGANIZING AND PLANNING: The extent to which an individual systematically arranges his or her own work and resources as well as that of others for efficient task accomplishment, and the extent to which an individual anticipates and prepares for the future. PROBLEM SOLVING: The extent to which an individual gathers information; understands relevant technical and professional information; effectively analyzes data and information; generates viable options, ideas, and solutions; selects supportable courses of action for problems and situations; uses available resources in new ways; and generates and recognizes imaginative solutions.

FIGURE 7.2. Some commonly used assessment center dimensions (see Arthur et al., 2003).

selecting military personnel, primarily officers and intelligence agents in the case of the OSS. On the basis of a behavioral consistency philosophy that “the best predictor of future performance is past performance” (Wernimont & Campbell, 1968, p. 372), psychologists were particularly interested in developing performance tests that presented officer candidates with complex stimuli requiring complex behavioral responses that translated well to actual performance situations. In other words, psychologists were interested in simulating the demands of real-world situations in the assessment of potential. Moreover, because no single simulation or test could fully represent all of the demands associated with serving as a military officer, psychologists placed great emphasis on the need for multiple simulations, both individual- and group-based, in conjunction with batteries of paper-and-pencil tests, projective tests, biographical inventories, and interviews. It was also thought that such complex simulations could not be feasibly scored in an objective manner; rather, scoring could be better accomplished by

integrating the judgments of multiple psychologists who observe the behavior of the officer candidates during the performance simulations. Thus, the core distinguishing aspects of contemporary ACs— simulation and triangulation via methods and assessors—were born. Commonalities and conspicuous differences among the German, the British War Office Selection Boards, and OSS efforts are evident when reading early accounts of their backgrounds, principles, and specific procedures (e.g., Ansbacher, 1941; Assessment Staff, 1948; Farago, 1942; Harris, 1949; Jennings, 1949; Morris, 1949). The overriding theoretical conceptualization of potential, and human nature in general, was very similar. Drawing on Freudian, neo-Freudian, and gestalt theories as well as Lewin’s burgeoning group dynamics perspective, assessment had a holistic focus, involving a complex interplay between general intellectual ability, specific aptitudes, interests, personality, needs, and motives. This holistic approach emphasized the dynamic interplay among human characteristics (Bray, 1982), 207

Arthur and Day

IN-BASKET: The in-basket presents a high-volume of information to participants including letters, memos, informal correspondence, reports, and announcements that have accumulated in an in-box. The materials are designed to represent the full scope of contextual, procedural, and financial challenges of managerial work. Participants are given background information about the organization and key personnel as well as a calendar that may already show previously scheduled meetings and events. Participants are asked to take action on the materials in the in-box. These actions may take the form of drafting letters and memos, writing instructions, delegating responsibilities, and scheduling meetings. Responses may be handwritten or computer-mediated. Stimulus materials typically vary in urgency, job relevance, interrelatedness, complexity, and significance. The in-basket typically involves substantial time pressure. In many assessment centers (ACs), an assessor interviews the participant shortly after the in-basket is completed to provide better insight into the participant’s priorities and rationale underlying the actions taken.

Copyright American Psychological Association. Not for further distribution.

ORAL PRESENTATION: The oral presentation requires participants to formally present information and new ideas to one or more persons. Examples include meeting with a board of directors, managers from various organizational divisions, or members of the press. Participants may be given little or no background information on a topic, or they may be given some time to review more extensive background information and prepare for the presentation. The presentation is usually made to an assessor or group of assessors, and in many ACs the assessors also serve as role-players who have standardized scripted behaviors and questions intended to challenge the participant and simulate a more dynamic interpersonal interaction. WRITTEN CASE ANALYSIS: The case analysis requires participants to review a set of background materials that describe a specific organizational problem or issue and then individually prepare a written report with a specific action plan. The problem or issue at hand may require a new set of policies or procedures, specific financial decisions, or a more systemwide strategy. In some ACs, an assessor interviews the participant afterward on the use of the information provided and the reasons for the recommendations made. LEADERLESS GROUP DISCUSSION: The leaderless group discussion is designed to simulate the dynamics associated with decision making in small groups. Participants are given time to review background materials regarding a specific organizational problem or issue, and then they meet with three to seven other participants to discuss and resolve the matter. No one is designated as the leader or chairperson. In a competitive leaderless group discussion, various roles of equal status are assigned to the participants, and the participants are presented with a problem that involves the distribution of limited resources. Each participant is instructed to develop a solution that maximizes the payoff to them but also benefits the entire group. All participants in the group may be required to sign off on the final solution. In a cooperative leaderless group discussion, no specific roles are assigned and the discussion is similar to ad hoc committees formed to examine a specific organization problem. Participants are simply instructed to generate and integrate their ideas into a single course of action. ONE-ON-ONE ROLE PLAY: The one-on-one role play requires participants to meet with another person to address a specific problem or issue. Examples include interviewing a job applicant, meeting with an employee who has performance problems, meeting with a potential new client or business partner, seeking information from a knowledgeable colleague, or dealing with a dissatisfied customer. The participant must talk with the role-player to gather new information and generate a solution or course of action. Role-players answer questions and they might also ask questions, make suggestions, or display specific emotions. Standardization of role-players is critical. Role-players typically receive general background information about their role and the general scenario along with a specific script for how to respond to potential participant behaviors and questions. In some ACs, the role-players also serve as assessors. Participants are typically given a brief period to review background materials prior to the role play. Afterward, participants may be asked to generate a brief report or plan of action.

FIGURE 7.3. Descriptions of some commonly used AC exercises.

208

Assessment Centers

ASSESSOR RATINGS

EXERCISES Competitive LGD (Resource Allocation)

In-Basket Exercise

In-Basket Interview

Cooperative LGD (Management Problem)

Written Analysis

Initial Rating

Final Rating

D

Consideration/ Awareness of I Others

M Influencing

Copyright American Psychological Association. Not for further distribution.

E

Others Oral

N Communication S Organizing and Planning

I Problem

O Solving N Written

Communication

S

FIGURE 7.4. Dimension × Exercise matrix. LGD = leaderless group discussion. Shaded areas represent dimensions that are not observable in the specified exercise. The scores in the Dimension × Exercise cells will be filled in by the assigned assessor. The initial rating is generated independently by the assessor after incorporating all dimension-level information that has been shared by the assigned assessors, and so ratings will vary across assessors. The final rating is generated either mechanically or via consensus, and so these ratings will be the same across all assessors.

which required equally dynamic assessments in which individuals must face challenging and stressful circumstances. It was important for the simulations to focus on how individuals adjusted to frustrating social circumstances and failure. This early period in the history of ACs can be described as a time of experimentation and inspiration. Because of their unfamiliarity with behavioral assessment, military psychologists had to experiment with new simulation techniques, much of the time by trial and error. Changing the assessment protocols and simulations midstream was common. Indeed, many of the chief scientists and authors of the early reports referred to these assessment efforts as experiments. The term assessment center was by no means in vogue at the time. The German assessment efforts inspired the British, and the British efforts during the war in turn inspired the first nonmilitary industrial

application of this assessment approach used by the British Civil Service Selection Boards (Anstey, 1977). Although connections between the British War Office Selection Boards and OSS efforts are evident, Henry Murray and his associates at the Harvard Psychological Clinic (Murray, 1938), who used a similar assessment approach in a nonindustrial study of the normal human personality, also inspired the OSS effort. In turn, the OSS endeavor was the key source of inspiration for the AC methodology used in American Telephone and Telegraph’s (AT&T’s) seminal Management Progress Study (MPS). The MPS began in 1956. It was originally conceived for purely research purposes as a longitudinal investigation of adult development in the context of managerial work (Howard & Bray, 1988). Douglas Bray introduced and directed the AC methodology applied to the MPS. The use of the AC methodology 209

Copyright American Psychological Association. Not for further distribution.

Arthur and Day

as a means of selecting and promoting managerial talent was a by-product of the extensive validation efforts showing how AC scores were predictive of salary and promotion progress (e.g., Bray & Grant, 1966). It is important to note that great care was taken to keep the AC data from AT&T executives to prevent later salary and promotion decisions from being contaminated by knowledge of the managers’ AC performance. The MPS established the use of multiple behavioral simulations with pooled assessor judgments as the hallmark of the AC methodology. Managerial potential was articulated via 26 dimensions derived from an extensive review of the management literature and interviews with behavioral scientists and personnel executives. The 26 dimensions reflected administrative and interpersonal skills, intellectual ability, and work, career, and social motives. Like its predecessors, the MPS approach combined behavioral simulations with a large battery of paper-and-pencil tests, projective tests, and unstructured interviews. However, in contrast to its predecessors, the explicit purpose of frustrating participants and examining adjustment in the face of failure was not part of the assessment process. Individual- and group-based simulations simply presented participants with the responsibility of addressing a wide range of managerial issues and problems. The AT&T MPS quickly sparked the widespread development and adoption of the AC methodology throughout the public and private sectors in the United States (Bray, 1982; Dunnette, 1971; Mayes, 1997). By the end of the 1960s, organizations such as Standard Oil, IBM, Caterpillar Tractor, General Electric, Sears & Roebuck, J. C. Penney, the Peace Corps, and the U.S. Internal Revenue Service were using the AC model developed by AT&T. This period in AC history also witnessed the birth of consulting firms offering AC services to private and public organizations. The 1970s to the present day can be characterized as a period of both proliferation and examination in AC history. Not only has the AC method remained popular as a means of selecting and promoting managerial talent, but it has also been extended to a wide variety of nonmanagerial employees, such as salespeople, teachers and principals, engineers, 210

rehabilitation counselors, pilots, police officers, and firefighters. Advancements in video and computer technologies have opened doors to more streamlined and technologically based assessment and scoring procedures. Moreover, the use of ACs has been expanded to a variety of human resources purposes other than selection and promotion, such as training, development, recruitment, performance appraisal, human resource planning, layoffs, and organizational development (Joiner, 2002; Spychalski, Quiñones, Gaugler, & Pohley, 1997; Thornton & Rupp, 2006). Using ACs for developing talent rather than selecting and promoting talent is rather popular in today’s business environment. In these instances, the AC may be labeled as a development center (Tillema, 1998). The scholarly literature is steadily becoming more populated with research and recommendations regarding the use of ACs for developmental purposes (e.g., Chen & Naquin, 2006; Jackson, Stillman, Burke, & Englert, 2007; Melancon & Williams, 2006; Rupp et al., 2006). ACs have also been used to assist high school and college students with career planning and the prediction of early career success (Arthur & Benjamin, 1999; Rowe & Mauer, 1991; Waldman & Korbar, 2004) and, even more interesting, to select political candidates (Silvester & Dykes, 2007). The number of companies in the United States using ACs is estimated in the thousands, with probably just as high a number outside the United States. It has also been noted that the vast majority of Fortune 500 companies use the AC methodology in some capacity (Mayes, 1997). In spite of this growing popularity, the explanatory mechanisms underlying why ACs work have been greatly scrutinized over the past 3 decades. Despite ample evidence of criterion-related validity and assurances of content-related validity, scholars have hotly debated the extent to which AC scores actually represent the behavioral dimensions purported to be measured. The major impetus for this debate is that construct-related validity studies (e.g., factor analyses, multitrait–multimethod correlations) typically show that scores derived from ACs reflect exercise variance more so than dimension variance. Researchers (e.g., Klimoski & Brickner, 1987) have proposed multiple alternative explanations for why AC scores are predictive of performance-related job criteria

Copyright American Psychological Association. Not for further distribution.

Assessment Centers

such as ratings of performance, ratings of potential, performance in training, and career advancement (Gaugler, Rosenthal, Thornton, & Bentson, 1987). Scholars also continue to debate the construct-related validity evidence yielded by ACs. This debate is reviewed in more detail in the Reliability and Validity section of this chapter. In short, this debate can be summarized by two statements. On the one hand, scholars tend to agree that the overall assessment score derived from the entire AC process is a valid reflection of overall potential and thus, a valid predictor of future performance. On the other hand, scholars tend not to agree on the extent to which AC scores actually reflect the dimensions purported to be measured. Although this second statement may not be as critical a concern in the context of selection and promotion decisions, it does suggest that the use of ACs for developmental purposes may not be tenable. If the AC dimension scores are not valid reflections of the specific constructs purported to be assessed, then feedback concerning one’s relative strengths and weaknesses vis-à-vis the dimension scores is unfounded and subsequent recommendations regarding developmental action steps are misguided. Regardless of this debate, the AC continues to be a popular tool for achieving a variety of human resource objectives.

Job analysis—Collect job-related information

Determine major work behaviors

DESIGN, DEVELOPMENT, ADMINISTRATION, AND SCORING The steps involved in the design and development of ACs are, to a large extent, not any different from those involved in the development of other predictors in industrial and organizational (I/O) psychology. Figure 7.5 presents a general overview of the broad sequence of steps that one may follow in designing and developing an AC from a “best practices” perspective. However, because there is no such thing as the AC or a single AC, we acknowledge that there may be variations in the steps outlined in Figure 7.5 as a function of the specific situation. That is, there is no one way to structure an AC; the specific design, content, and administration will be a function of the objectives of the AC and the specific target group that is being assessed. Nevertheless, as illustrated in the Figure 7.5, the typical first step would be the implementation of a detailed job analysis using a variety of job analysis techniques. (See also chap. 1, this volume.) As the second step, this information would be used to generate a draft job description that will then be reviewed by incumbents and supervisors. On the basis of incumbent and supervisor feedback, the job description would be revised and finalized. Third, on the basis of the job description, critical knowledge, skills, abilities, and other characteristics

Identify KSAOs or constructs underlying major work behaviors

Identify behavioral dimensions related to the KSAOs

Select or develop exercises to measure dimensions

Implement assessment center

Refine as warranted

Pilot test assessment center

Train assessors and administrators

FIGURE 7.5. Assessment center design, development, and implementation sequence. KSAO = knowledge, skills, abilities, and other characteristics. 211

Copyright American Psychological Association. Not for further distribution.

Arthur and Day

(KSAOs) necessary for successful job performance would be identified. The fourth step calls for the identification of behavioral dimensions related to the identified KSAOs. For purposes of selection and promotion, dimensions should reflect abilities and traits that underlie potential, whereas knowledge- and skillbased dimensions are more applicable for training and development purposes. The implementation of this step (i.e., the identification, labeling, and definition of dimensions) is a particularly critical one in the development of the AC because it has been identified as a weak link in the common practice of AC design and development because of the proclivity to rely on espoused constructs (dimensions; Arthur, Day, & Woehr, 2008). In fact, the AC literature appears to be a domain in which assertions about the dimensions being measured are rarely, if ever, subjected to the psychometric standards that characterize test development in other domains (Arthur et al., 2008; Brannick, 2008). Merely labeling data as reflecting a particular dimension (espoused dimension or construct) does not mean that is the dimension being assessed (actual dimension or construct). Consequently, the overreliance on espoused dimensions or constructs has been highlighted as a threat to the construct-related validity of ACs. As noted by Arthur et al. (2008), the formulation and explication of construct definitions should not be approached casually, and support for the construct representation of the AC’s content should be pellucidly demonstrated using multistage data collection and refinement efforts before the AC is put into operational or research use. Thus, it should be noted that contrary to common practice, efforts to establish the constructrelated validity of the AC dimension ratings should be undertaken before the AC is put into operational or research use. The fifth step entails the selection or development of exercises to measure the behavioral dimensions that were identified. Following the preceding developmental process is intended to ensure a high level of content-related validity. Finally, one would train the assessors and AC administrators, pilot test and refine the AC as warranted, and then, of course, operationally implement the AC. In contrast to construct-related validity evidence, if one were 212

interested in demonstrating the criterion-related validity of the AC ratings, this validation effort would be undertaken after the implementation of the AC because the collection of predictor and criterion data is necessary for any criterion-related validation endeavor. Still building on our illustrative example and assuming that the implementation of the first five steps resulted in the six dimensions and five exercises presented in Figure 7.4 (see also Figures 7.2 and 7.3), in the next section we present a description of a “typical” assessor training and an operational AC administration set of activities.

A Best-Practices Assessor Training and Assessment Center Administration Illustrative Example The specific practices presented here were arrived at on the basis of a summation of the pertinent extant literature. Supporting research evidence for specified AC design and methodological choices summarized later is reviewed in the Important Methodological and Design-Related Characteristics and Features section of this chapter. With regard to the selection or choice of assessors, the research evidence indicates that it should be tilted in favor of I/O psychologists and similarly trained human resources management (HRM) professionals, and less so in favor of managers and supervisors (Gaugler et al., 1987; Lievens & Conway, 2001; Woehr & Arthur, 2003). After their selection, the next step is the training of the assessors, which will typically take the form of a full- or multiday frame-of-reference training program (Jackson, Atkins, Fletcher, & Stillman, 2005; Lievens, 2001; Schleicher, Day, Mayes, & Riggio, 2002; Woehr & Huffcutt, 1994) comprising the following steps: 1. Initial explanation of the AC method and the purpose of conducting the AC. 2. Definitions of the dimensions and descriptions of the exercises. Assessors are given examples of behaviors representing each dimension, and each set of behaviors (i.e., by dimension) is accompanied by an anchored rating scale (e.g., a 1- to 5-point scale) indicating the effectiveness levels of the specified dimension. 3. Once assessors are familiar with the dimensions and exercises, they then practice observing

Assessment Centers

Copyright American Psychological Association. Not for further distribution.

4. 5.

6.

7.

8. 9.

10.

behaviors elicited by the exercises. In our example, this is accomplished by having them view videotapes of past participant groups participating in the specified exercises, such as the resource allocation exercise (see Figure 7.4), and recording the behaviors of the participants. Prior to the practice observations, the trainers would emphasize the observing and recording of participant behaviors without trying to categorize the behaviors into dimensions while observing the participants. Assessors categorize the behaviors into dimensions and provide ratings on each dimension. Assessors discuss their observations and ratings, using this as an opportunity to build a common frame of reference. Assessors practice another exercise, such as the in-basket. In this example, they are each given a completed in-basket and directed to record any information provided by the participant. Assessors practice the in-basket interview. This is accomplished via a role-played in-basket interview during which assessors have the opportunity to clarify responses and seek explanations from the participant (e.g., why the participant took specific action[s]). Assessors categorize the in-basket behaviors into dimensions and provide dimension ratings. Again, assessors discuss behaviors and dimension ratings so that a common frame of reference can be developed. Assessors train on and practice the remaining exercises using the general approach described earlier.

The next training segment has the objective of familiarizing assessors with the rating process. As discussed later, there are various approaches to generating assessment ratings and scores—specifically, the within- versus across-exercise (i.e., withindimension) approach. The approach used in this example is the across-exercise approach. Hence assessors are trained to work across exercises and finish rating a dimension before moving on to the next one. If required, the final segment of training focuses on how to prepare feedback reports and conduct feedback interviews. Finally, as part of their training, new and inexperienced assessors shadow

a more experienced assessor for at least one AC administration before they serve as a full assessor whose ratings are used operationally. When the AC is administered, participants will participate in all the exercises and will be evaluated on the six dimensions (see Figure 7.4). Groups of three to four assessors—who serve as assessor groups—are assigned to observe and rate the materials of a group of four to six participants (a design feature that results in a 2:1 participant-to-assessor ratio [see Figure 7.6]; Woehr & Arthur, 2003). For the group exercises, assessors sit in the back of the room (to minimize their obtrusiveness) and record behaviors displayed by the participants (see Figure 7.1). It should be noted that behavioral checklists may be used as an aid in the recording of behaviors. Because of the 2:1 participant-to-assessor ratio, for the group exercises (e.g., the leaderless group discussions), each assessor observes and records the behavior of one or two AC participants. However, all assessors in the group are present during each exercise, and each assessor observes different participants in each exercise using a rotation schedule that ensures that each assessor has the responsibility of being the primary observer for each participant in at least one exercise. Figure 7.6 presents an illustration of the rotation sequence. On the completion of each exercise, assessors categorize their recorded observations into dimensions using materials that describe each dimension in detail, along with a list of some representative behaviors. For the in-basket exercise (including the follow-up in-basket interview) and written exercise, assessors review and rate the materials for each participant to whom they are assigned. Following the completion of all exercises, the assessors in each assessor group meet to rate the participants’ performance. Each assessor provides independent ratings of each participant (in their participant group) on each dimension. For exercises in which an assessor was not assigned to record behaviors of a participant, ratings are based on a verbatim listing of observed behaviors recorded by the assigned assessor as well as the assessor’s own observations. Consistent with the across-exercise (within-dimension) approach, the rating process proceeds as follows. Selecting a participant, the 213

Arthur and Day

Participant A

Participant B

Assessor 1

Competitive LGD (Resource Allocation)

Copyright American Psychological Association. Not for further distribution.

Participant A

Participant B

In-Basket

Participant A

Participant C

Assessor 2

Participant C

Assessor 1

Participant B

Cooperative LGD (Management Problem)

Participant D

Participant C

Assessor 1

Participant E

Participant F

Participant E

Participant F

Assessor 3

Participant D

Assessor 2

Participant F

Assessor 3

Participant D

Assessor 2

Participant E

Assessor 3

FIGURE 7.6. Assignments and rotation of assessors observing and recording participant behavior in a three-exercise assessment center. LGD = leaderless group discussion.

assessors start with a specified dimension (e.g., oral communication). A verbatim listing of observed behaviors for the first relevant exercise (e.g., resource allocation) for that dimension is presented by the assigned assessor. After the presentation of observed behaviors for the exercise, an independent rating is made on the dimension in question at the exercise level; this process is repeated for all the exercises (e.g., in-basket, management problem) relevant to the dimension in question. It is important to note that no group discussion or consensus process occurs prior to each assessor’s ratings. Ratings are typically made on a Likert scale with descriptive anchors. After the assessors have completed their dimension rating for each exercise (i.e., cells in Figure 7.4), they then again make an independent initial dimension rating (see Figure 7.4). The members of the assessor 214

group then share their initial ratings and discuss any differences before assigning a final dimension score (Roch, 2006; Schmitt, 1977). In this final step, assessors may be required to reach consensus; if that is the case, then they will have identical final dimension ratings. On the other hand, the AC designers may have opted to generate the final scores mechanically (e.g., average the assessors’ ratings); if so, then assessors would obviously not be instructed to reach consensus. Regardless of whether they are instructed to reach consensus or not, the preceding sequence of steps would be repeated for the remaining (five) dimensions. The entire process would then be repeated for each participant assigned to the assessor group. It is also worth noting that it is common for AC designers to be interested in an overall assessment center rating (OAR) that is intended to

Copyright American Psychological Association. Not for further distribution.

Assessment Centers

represent the participant’s overall future potential or overall performance in the AC collapsed across all the AC dimensions (cf. Arthur, Day, McNelly, & Edens, 2003). The AC process does not end with the final determination of ratings. Once ratings have been finalized, feedback needs to be delivered to relevant constituents in a manner that is consistent with the purpose of the AC. Recognizing that federal and state privacy laws may also be applicable, ethical guidelines (American Educational Research Association, American Psychological Association, and National Council on Measurement in Education, 1999; American Psychological Association, 1992; International Task Force on Assessment Center Guidelines, 2000; Society for Industrial and Organizational Psychology, Inc., 2003) state that there should be policies for providing feedback to participants as well as for restricting access to AC ratings and reports to specific relevant persons within the organization. The nature of the feedback given to participants will likely vary depending on the purpose of the AC. General feedback regarding overall performance and resulting decisions (i.e., pass–fail, yes–no) is relevant in cases of selection and promotion, and more specific feedback regarding strengths and weaknesses is required in the case of training and development. With regard to developmental ACs, the assessment itself is just the beginning of what should be a larger program that involves the delivery of accurate feedback from a credible source and the development and implementation of an action plan designed to remediate areas in need of improvement. At present there is very little empirical research examining the feedback and follow-up phases in developmental ACs (Thornton & Rupp, 2006). However, the limited research there is shows a positive relationship between AC ratings and feedback acceptance and feedback-seeking behavior. For instance, Bell and Arthur (2008) reported a correlation of .30 (N = 141) between OARs and feedback acceptance in a developmental AC used by a state auditing office and master’s graduate program in public administration. In a case study of 20 call center workers, Dhanju (2007) reported that a majority of the participants who were promoted to center team leader as a result

of their AC performance accepted their feedback whereas a majority of the participants who were not promoted did not accept their feedback. In a sample of 189 newly hired supervisors, Abraham, Morrison, and Burnett (2006) showed a standardized difference of 0.39 in OARs between participants who sought feedback after participating in a developmental AC versus those who did not seek feedback. Although other individual differences (e.g., agreeableness; Bell & Arthur, 2008) may also play roles in feedback acceptance and seeking behavior, this state of affairs is unfortunate given that those individuals most in need of development are the ones least likely to accept feedback and address their development needs. Without participation in follow-up developmental activities, it is unlikely that participation in the AC will benefit participants ( Jones & Whitmore, 1995).

Important Methodological and Design-Related Characteristics and Features As previously noted, because they are methods, there is not a single AC or the AC. Consequently, as with other methods such as the interview, ACs are only as good as their design and administration. However, there is a rich and reasonably large body of research that indicates what the desirable features of ACs are or should be (e.g., see International Task Force on Assessment Center Guidelines, 2000; Lievens, 1998; Woehr & Arthur, 2003) and subsequently provides some guidance on how to design and implement “good” ACs. This literature served as the basis for the practices described in the preceding section on best practices. It is obvious that poor enactment and implementation of any of the steps in Figure 7.5—such as poor planning; inadequate job analysis; weakly defined, ambiguous, and poorly explicated dimensions; poor exercise selection or design; absence of pilot testing and evaluation; use of unqualified or inadequately trained assessors; and poor behavior documentation and scoring—is likely to have adverse effects on the effectiveness and validity of the resultant AC (Caldwell, Thornton, & Gruys, 2003; Chen, 2006). Furthermore, drawing from a relatively large body of research, one can identify specific AC methodological factors and design 215

Arthur and Day

Copyright American Psychological Association. Not for further distribution.

characteristics that have discernable hypothesized positive and negative effects on the effectiveness and validity of ACs as measurement tools (Lievens, 1998; Woehr & Arthur, 2003). These include (a) the number of dimensions assessors are asked to observe, record, and subsequently rate; (b) the conceptual distinctiveness of dimensions; (c) the transparency of dimensions; (d) the choice and design of exercises, and the type and number of exercises; (e) the participant-to-assessor ratio; (e) the scoring and rating approach; (f ) the type of assessor used; and (g) assessor training.

Number of Dimensions Primary studies (e.g., Bycio, Alvares, & Hahn, 1987; Gaugler & Thornton, 1989; Russell, 1985; Sackett & Hakel, 1979; Schmitt, 1977), qualitative reviews (e.g., Lievens, 1998), and meta-analytic studies (e.g., Arthur et al., 2003; Bowler & Woehr, 2006; Lievens & Conway, 2001; Meriac, Hoffman, Woehr, & Fleisher, 2008; Woehr & Arthur, 2003) have converged on the summary conclusion that all things being equal, a smaller number of dimensions is preferable to a larger number. This is primarily because of the high cognitive demands placed on assessors when they are asked to process a large number of dimensions. Thus, for instance, in a meta-analysis of 48 AC studies, Woehr and Arthur (2003) obtained partial support for the effect of the number of dimensions on the construct-related validity of ACs; fewer dimensions were associated with higher levels of convergent validity (r = .37) and more dimensions with a lower level (r = .27). In this meta-analysis, the mean number of dimensions was 10.60 (SD = 5.11, min = 3, max = 25)— a number that is a little higher than that recommended by the extant literature (Arthur et al., 2003; Gaugler & Thornton, 1989; Lievens, 1998; Russell, 1985; Sackett & Hakel, 1979; Schmitt, 1977).

Conceptual Distinctiveness of Dimensions An issue related to the number of dimensions is the conceptual distinctiveness of the dimensions. One result of the reliance on espoused constructs is a lengthy, extensive list of dimensions purported to be measured by ACs such that recent meta-analyses (see Arthur et al., 2003; Bowler & Woehr, 2006; 216

Woehr & Arthur, 2003) have extracted anywhere from 79 to 168 different dimension labels. Such large numbers of dimensions and the reliance on espoused constructs has perpetuated a lack of dimension distinctiveness within primary studies. Consequently, recent meta-analyses have generally obtained more favorable and supportive results for the construct validity of AC ratings after collapsing the myriad of dimensions in the extant literature into a smaller, conceptually distinct set of dimensions. Using Arthur et al.’s (2003) six-dimension taxonomy (communication, consideration and awareness of others, drive, influencing others, organizing and planning, and problem solving; see Figure 7.2), these meta-analyses provide evidence for the criterionrelated validity of AC dimensions (Arthur et al., 2003) and the impact of dimension factors (Bowler & Woehr, 2006; Connelly, Ones, Ramesh, & Goff, 2008). For instance, using a regression-based composite consisting of four out of their six dimensions, Arthur et al. (2003) explained more variance in performance (20%) than Gaugler et al. (1987) were able to explain using the OAR (14%). In addition, Arthur et al.’s (2003) taxonomy has also been used to effectively demonstrate the differentiation from and incremental criterion-related validity of AC dimensions over cognitive ability and personality (Meriac et al., 2008).

Transparency of Dimensions Lievens (1998) strongly recommended informing participants in developmental ACs of the dimensions to be assessed and suggested that this practice should be considered for ACs for selection purposes as well. This recommendation is based primarily on Kleinmann (1993) and Kleinmann, Kuptsch, and Köller (1996), who found that disclosing the dimensions to participants and informing them of which behaviors represented specified dimensions resulted in better validity evidence. Similar findings were obtained by Kolk, Born, and Van der Flier (2003), who found an increase in convergent validities from a nontransparent (r = .22) to a transparent condition (r = .31) with relatively similar discriminant validities (r = .52 and .50, respectively). The conceptual basis for this effect is that informed participants display higher levels of behavioral consistency which in turn

Assessment Centers

enables assessors to differentiate the dimensions and also rate participants consistently across the AC exercises.

Copyright American Psychological Association. Not for further distribution.

Choice and Design of Exercises and the Number and Type of Exercises The choice and design of exercises are critical to the effectiveness of ACs, not just in terms of achieving face validity via replicating the job environment but, more important, also in terms of how well behaviors associated with the identified performance dimensions are elicited. All too often there is an overemphasis on replicating the job environment at the expense of the coupling between dimensions and exercises. For instance, over the past 2 decades practitioners have increasingly incorporated exercises in which stimulus materials and responses are computer-mediated (e.g., e-mails) in real time rather than the traditional approach of using printed and handwritten memos, letters, and reports. Although these higher fidelity formats may indeed reflect the nature of work today, the importance of dimensions in the assessment process should not be deemphasized. We were unable to locate any published research comparing the validity of such higher fidelity ACs to the more traditional paper-and-pencil approach. It is crucial that the development of AC exercises includes an explicit blueprint for how the dimensions are represented in the exercises (Lievens, 1998; Thornton & Rupp, 2006). The use of standardized role-players can aid in the elicitation of dimensionrelevant behaviors (Lievens, 1998). In light of the early history of ACs, it is important that a variety of exercises be used to better triangulate the assessment of dimensions. Indeed, Gaugler et al.’s (1987) meta-analysis showed that stronger predictive validities are associated with the use of a larger number of different types of exercises (ρ = .25–.95 depending on the criterion). Figure 7.3 provides a brief overview of commonly used AC exercises. For an instructive and detailed account of how to build AC exercises, the interested reader is referred to Thornton and Mueller-Hanson (2004). With regard to the use of AC scores or ratings, if the AC is part of a larger assessment system that includes interviews, résumés, aptitude tests, and other forms of assessment, and one is interested in

obtaining or focusing exclusively on performance on the AC, then it is recommended that information from the other tests be kept from the assessors during the administration of the AC to keep the AC ratings free from contamination. However, if the focus is on performance on the larger assessment system of which the AC is a component, then of course information from the other tests would be made available to the assessors who would in turn use them to arrive at the participants’ scores. However, it should be emphasized that the performance scores in this latter instance reflect performance on the larger assessment system and not just the AC and indeed cannot be accurately described as being the AC score. To facilitate clear communication and discussions of transportability of assessment, one must distinguish ACs from ostensibly similar types of assessment, including both low- and other high-fidelity simulations (Thornton & Rupp, 2006). Low-fidelity simulations such as situational judgment tests (video and paper-and-pencil administered), performance walkthroughs, and situational interviews should not be confused with AC exercises as they elicit behavioral intentions and not actual job-relevant behaviors. Likewise, there are meaningful distinctions that can be drawn between work samples (i.e., performance tests) and AC exercises. Work samples can be used as stand-alone tests or incorporated into ACs as one of several exercises. However, although work samples, like other AC exercises, also elicit job-relevant behaviors, in contrast to their use as stand-alone tests, when used as AC exercises, there is a deemphasis on current knowledge and skill in a specific domain and a strong focus on the assessment of future potential via multiple underlying performance dimensions. In fact, it is important to design AC exercises that do not closely resemble the specific demands of the job to prevent job knowledge and expertise from contaminating the assessment of broader underlying performance dimensions. In other words, work samples as stand-alone tests are designed to simulate actual job tasks, whereas AC exercises are designed to represent the general contexts surrounding the demands of the job. In the case of managerial work, such contexts include multitasking (in-basket exercises); interacting one-on-one with subordinates, clients, and other 217

Copyright American Psychological Association. Not for further distribution.

Arthur and Day

organizationally relevant constituents (role-play, interview, and fact-finding exercises); conducting meetings and working jointly with others to address organizational issues and problems (cooperative and competitive leaderless group discussion exercises and business games); developing action plans and proposals (case analysis exercise); and giving presentations (oral presentation exercise). Furthermore, work samples are frequently scored in terms of accuracy in accomplishing a given task, whereas ACs are scored in terms of the manner by which tasks are addressed. As methods, work samples and ACs are both adaptable to a variety of purposes. Therefore, there are exceptions to the descriptions and distinctions made earlier. For example, it is thought that exercises that more closely resemble the actual job are appropriate for training and developmental ACs (Thornton & Rupp, 2006). In general, a single simulation is not an AC, and a system of work samples that separately assess different performance dimensions is also not an AC. Simulations in and of themselves do not define ACs; rather, ACs are defined by the use of simulations in an attempt to triangulate underlying performance components. For a comprehensive review of work samples, the interested reader is referred to Truxillo, Donahue, and Kuang (2004) and Roth, Bobko, and McFarland (2005) and the situational judgment test chapter by Ployhart and MacKenzie (see chap. 8, this volume).

Participant-to-Assessor Ratio The arguments for using a small number of dimensions are similar to those for using a low participantto-assessor ratio. As the number of participants that a given assessor is required to observe and evaluate in any given exercise increases, the cognitive demands placed on the assessor may make it more difficult for the assessor to process information at a dimension level. In addition, AC ratings are more likely to be susceptible to bias and information-processing errors under conditions of high cognitive demand. So for instance, depending on the criterion, Gaugler et al. (1987) reported rhos ranging from −.15 to −.29 for the relationship between AC ratings and specified criteria, indicating that higher ratios are associated with lower criterion-related validities. In Woehr and 218

Arthur’s (2003) meta-analysis, of the 26 studies that reported information on the participant-to-assessor ratio, the ratio ranged from 1 participant to 4 assessors to 4 participants for each assessor with a mean ratio of 1.71 (mode = 2) participants per assessor. The extant literature would recommend a participantto-assessor ratio that does not exceed 2:1.

Scoring and Rating Approach As previously noted, there are two primary rating approaches in the AC literature (Robie, Adams, Osburn, Morris, & Etchegaray, 2000; Sackett & Dreher, 1982; Woehr & Arthur, 2003): the withinexercise and across-exercise (i.e., within-dimension) approaches. In the within-exercise approach, participants are rated on each dimension after the completion of each exercise. Two variations of the within-exercise approach have been described: In one variation, the same assessors observe all exercises and provide dimension ratings after observing each exercise, and in the second variation, different sets of assessors observe each exercise and provide ratings for each dimension. In the across-exercise approach, evaluation occurs after all the exercises have been completed, and dimension ratings are based on performance in all of the exercises. Two variations of the across-exercise approach have also been described: In the first, assessors provide an overall rating for each dimension reflecting performance across all exercises, and in the second, assessors provide dimension ratings for each exercise, but only after all exercises are observed. In addition, as noted by Arthur et al. (2008), common to both of these rating approaches is the use of postconsensus dimension ratings to assess dimensions. Postconsensus dimension ratings represent the combination (either judgmentally or mechanically) of postexercise dimension ratings into a summary rating representing dimension-level performance information across multiple exercises and raters. From a traditional psychometric perspective, postexercise dimension ratings may be viewed as item-level information, whereas postconsensus dimension ratings represent scale-level information. With regard to which is more appropriate and therefore the preferred practice, research examining the effect of the rating approach on the construct-

Copyright American Psychological Association. Not for further distribution.

Assessment Centers

related validity of AC ratings suggests that the across-exercise (within-dimension) approach is likely to result in higher levels of construct-related validity (Woehr & Arthur, 2003). Woehr and Arthur (2003) obtained a higher convergent validity for the acrossexercise (r = .43) compared with the within-exercise approach (r = .29) and lower discriminant validity for the across-exercise (r = .48) compared with the within-exercise approach (r = .58). Thus, it would seem that the across-exercise approach results in dimension factors, whereas the within-exercise approach results in exercise factors. Another aspect of the evaluation and scoring process is the use of behavioral checklists to record behaviors and even rate dimension performance. Behavioral checklists can reduce the assessors’ load in terms of the amount of time spent taking notes and the number of judgments needed to be made, but assessors still need to match observed behaviors to listed behaviors, and they still have to make judgments about observed behaviors that do not appear on the checklists (Thornton & Rupp, 2006). An overreliance on behavioral checklists can also confine assessors to limited definitions of performance. An inspection of primary studies (e.g., Donahue, Truxillo, Cornwell, & Gerrity, 1997; Reilly, Henry, & Smither, 1990) reveals variability in the manner of use and efficacy of behavioral checklists. Overall, the influence that behavioral checklists have on validity appears to be modest (Lievens & Conway, 2001). In the end, information from the multiple exercises and assessors must be integrated to derive final dimension ratings (as well as an OAR if one is desired). Information can be integrated either judgmentally (sometimes referred to as the clinical or consensus approach) or mechanically (sometimes referred to as the statistical or actuarial approach), or through some hybrid of the two. The judgmental approach is the traditionally advocated approach, and as previously described it involves assessors meeting to discuss their independent ratings (and observations) for each participant with the goal of arriving at a consensus about each participant for each dimension and then for the OAR. The mechanical approach removes discussion of participants’ ratings from the integration process. At the dimension level, a mechanical approach would typically involve

averaging assessor judgments. With respect to the OAR, dimension ratings can be unit-weighted, differentially weighted by importance to the job, or differentially weighted based on a cross-validated regression analysis. (See Bobko, Roth, & Buster, 2007, for a review of the use of unit weights in creating composite scores.) A mechanical approach can streamline the rating process and dramatically reduce the cost of an AC. Consensus meetings can account for 25% to 33% of the AC time devoted to administering the AC (Gilbert, 1981). However, with a holistic perspective on understanding human nature, it is thought that a judgmental approach can provide deeper insights into each participant’s potential by drawing on his or her unique combination of strengths and weaknesses (Thornton & Rupp, 2006). What does the research say? Numerous investigations from a wide variety of ACs have compared the two approaches. As a whole, these investigations show that mechanically derived ratings yield criterion-related validities that are as strong if not stronger than those derived through consensus (e.g., Feltham, 1988a; Pynes & Bernardin, 1992; Tziner, Meir, Dahan, & Birati, 1994; Wollowick & McNamara, 1969). Despite this body of research, it has been estimated that as many as 84% of operational ACs use the consensus approach (Spychalski et al., 1997). However, it must be noted that comparisons between the two approaches have always involved operational ACs using a within-subjects design. Recent research has shown that the anticipation of a discussion about ratings has a positive influence on the accuracy of assessors’ observations, and that requiring assessors to reach consensus further improves accuracy (Roch, 2006, 2007). Thus, decisions to eliminate the consensus meeting from the AC to reduce costs should be made with caution.

Type of Assessor The type of assessor, specifically the use of (I/O) psychologists (and similarly trained HRM professionals) versus managers and supervisors, has also been demonstrated to influence the validity of ACs. In an explanation of their meta-analytic results in which they obtained higher criterion-related validities for ACs that used psychologists as assessors compared 219

Copyright American Psychological Association. Not for further distribution.

Arthur and Day

with those that used managers and supervisors (ρ = .26–.39 depending on the criterion), Gaugler et al. (1987) posited that psychologists make better assessors because, as a result of their education and training, they are better equipped to observe, record, and rate behavior. Both Lievens and Conway (2001) and Woehr and Arthur (2003) provided additional support for this proposition by obtaining stronger construct-related validity evidence for ACs that used psychologists as assessors compared with those that used managers and supervisors. For instance, Woehr and Arthur obtained a higher convergent validity for psychologists (r = .45) compared with managers and supervisors (r = .38) and lower discriminate validity for the former (r = .40) compared with the latter (r = .64).

Assessor Training The generation of AC ratings is inherently judgmental and subjective in nature. Hence, the training of assessors to observe, record, and categorize behaviors by dimension and subsequently rate performance is a very important element in the design, development, and administration of ACs. For instance, Woehr and Arthur (2003) obtained higher convergent validity when training was indicated in the description of the AC (r = .36) in contrast to when none was indicated (r = .29), and lower discriminant validity for the former (r = .51) compared with the latter (r = .63). In addition, the type and quality of training is also an important feature of ACs. For instance, there is consensus in the literature that the frame-of-reference approach, which imposes a common theory of performance on assessors enabling better judgment of behavior, is a highly effective approach to rater training (Lievens, 2001; Noonan & Sulsky, 2001; Schleicher & Day, 1998; Woehr & Huffcutt, 1994). Consistent with this, a conclusion that can be drawn from the extant AC literature is that the use of frame-of-reference training and also the extensiveness and quality of assessor training play an important role in establishing and enhancing the validity of AC ratings (Jackson et al., 2005; Lievens, 1998; Schleicher et al., 2002; Woehr & Arthur, 2003). In summary, although empirical studies have demonstrated the role of each design factor (on its own) on the validity of AC ratings, we suspect that 220

synergistic gains to validity will result by aligning all of these optimal features in one AC (Arthur, Woehr, & Maldegen, 2000). Consequently, in line with the results of multiple meta-analyses and the conceptual basis for their efficacy, we posit that a well-designed AC would be characterized by the following features: (a) the AC should be based on sound planning and an adequate job analysis; (b) the number of dimensions assessors are asked to observe, record, and subsequently rate is limited to a reasonable and manageable number (somewhere in the region of six to eight dimensions); (c) the dimensions are conceptually distinct, clearly defined, and unambiguously explicated; (d) the dimensions may be made transparent to the participants; (e) the choice and design of exercises are sound, multiple and different exercises are used, and each exercise elicits behaviors across multiple dimensions; (f ) the marriage between dimensions and exercises should be strong; (g) the AC should be pilot tested and the exercises evaluated and refined accordingly; (h) the participantto-assessor ratio is limited to about a 2:1 ratio; (i) an across-exercise (i.e., within-dimension) rating approach is used and, resources permitting, consensus meetings are incorporated into the generation of participants’ scores; (j) the rating process generates postconsensus dimension-level scores (as well as an OAR if warranted); (k) I/O psychologists or other similarly trained HRM professionals are used as assessors or as part of assessor teams; and (l) assessors are extensively trained using methodologies such as frame-of-reference training. RELIABILITY AND VALIDITY OF ASSESSMENT CENTER RATINGS As with just about any test, the two psychometric properties of most interest in a discussion of AC ratings are the reliability and validity of said ratings. In reviewing the psychometric properties of ACs, it is important to keep in mind that because an AC is a method, there is a great deal of variability in operational ACs. Even if the purpose, exercises, dimensions, and assessor training are held constant, differences in the administration of an AC can dramatically influence validity conclusions (Schmitt, Schneider, & Cohen, 1990). So although we try to

Assessment Centers

present an overall summary of the psychometric standing of the AC, one should keep in mind that the differences in psychometric quality from one AC to another can be substantial.

Copyright American Psychological Association. Not for further distribution.

Reliability of AC Ratings Within the context of ACs, reliability pertains to the consistency of assessor ratings and the extent to which they are free of measurement error. To this end, metrics of interrater reliability—the extent to which multiple raters assign the same dimension scores to the same participant—would seem to be the facets of most interest in the context of AC ratings. Although assessments of interrater reliability of dimension scores within exercises are not uncommon, a detailed search and reading of the extant literature would suggest that the reliability of dimension scores across exercises has rarely been examined. A similar observation was made by Schmitt (1977) and more recently by Brannick (2008). The paucity of interrater reliability data in the extant literature may be a result of the nature of AC design. That is, the use of postconsensus dimension ratings as the basis for dimension scores obviates the need to assess interrater reliability. However, what limited research there is suggests that assessors rate participants similarly. For instance, Schmitt (1977) found that postdiscussion interrater correlations were all between .74 and .91. In contrast, the prediscussion correlations were in the .60s and .70s. To obtain additional summary information on the reliability of AC ratings, we turned to AC criterion-related validity meta-analyses to see how many had corrected for unreliability in the predictor scores and what the estimated levels of reliability were. We identified eight meta-analyses that had investigated the criterion-related validity of AC ratings. Some of these meta-analyses (i.e., Arthur et al., 2003; Meriac et al., 2008) corrected for predictor unreliability, but most did not (i.e., Aamodt, 2004; Gaugler et al., 1987; Hardison & Sackett, 2007; Hermelin, Lievens, & Robertson, 2007; Hunter & Hunter, 1984; Schmitt, Gooding, Noe, & Kirsch, 1984). Consistent with the previous point about the lack of data, the common reason for not reporting this information is best summarized by Gaugler et al. (1987), who stated that “the validities were

not corrected for predictor unreliability because we were unable to obtain a reasonable estimate of the distribution of reliabilities for the overall assessment center ratings” (p. 495). Of the two meta-analyses that provided information on the reliability of AC ratings, both used an artifact distribution correction approach. Hence, Arthur et al. (2003) reported a mean interrater reliability of .86 that was based on 37 data points obtained from six articles. In contrast, although they also used an artifact distribution approach, Meriac et al. (2008) used a different approach to estimate the reliability of the AC ratings: AC dimensions were corrected based on the mean dimension intercorrelations. For example, if two primary study dimensions (e.g., analysis and judgment) were both coded into one of Arthur et al.’s (2003) dimensions (e.g., problem solving), and the correlation between these two dimensions was provided, the square root of this value was used as input to form an artifact distribution for the AC dimensions. In essence, this represents a type of alternate form reliability in that alternate measures of the same construct (i.e., within each of the [6] AC dimensions) were used to assess reliability. (p. 1046) Consequently, Meriac et al. reported the following dimension-level reliability estimates: consideration and awareness of others = .80, communication = .86, drive = .86, influencing others = .87, organizing and planning = .87, and problem solving = .91. So these data, in their totality, would suggest that AC ratings are fairly reliable. Finally, if the AC is an ongoing one, with assessors participating in multiple temporally spaced ACs, then an important question is the extent to which the ratings made by individual assessors change over time. To address this question, Sackett and Hakel (1979) directly compared the correlation matrices generated from the ratings (of 719 participants) made by four assessor teams over 6 months. Their results indicated fairly high levels of consistency in assessors and assessor teams’ rating patterns over time. 221

Arthur and Day

Copyright American Psychological Association. Not for further distribution.

Validity of AC Ratings In broad terms, construct validity pertains to an assessment of the extent to which a test is measuring what it purports to measure, how well it does so, and the appropriateness of inferences that are drawn from the test’s scores (American Educational Research Association, American Psychological Association, and National Council on Measurement in Education, 1999; Society for Industrial and Organizational Psychology, 2003). Furthermore, there are several evidential bases for demonstrating the construct validity of a test or measure, some of which include but are not limited to the use of content-related, criterion-related, and constructrelated validity evidence (Binning & Barrett, 1989; Messick, 1995; see also chap. 13, this volume). In fact, it is often noted that the continued popularity of ACs is due in large part to their strong criterionrelated validity evidence and high content-related validity (and fidelity). In contrast, the adequacy of their construct-related validity evidence has been strongly debated and questioned. We review and discuss the research literature and evidence on these issues in the following sections.

selves are articulated and explicated (before linkages to exercises are made); job analysis by itself per se does not speak to the explication and articulation of dimensions. Furthermore, although there are examples of comprehensive approaches taken to demonstrate all three linkages (e.g., I. L. Goldstein et al., 1993; Haymaker & Grant, 1982; Schmitt & Noe, 1983), the vast majority of published studies on ACs provide very little information that speaks to all of these linkages and none appear to speak to the dimension explication issues (Arthur et al., 2008; Sackett, 1987). At present, we believe that there are low expectations in the published literature for reporting how AC dimensions are explicated. It is common practice for authors to simply list the dimensions assessed and, in some cases, provide a single-sentence definition of each dimension. Without more complete descriptions of how dimensions were developed and what procedures were taken to ensure dimension representativeness, it is difficult to make strong conclusions regarding the content-related validity of ACs, even in the presence of job analysis (Arthur et al., 2008).

Content-Related Validity Evidence

Criterion-Related Validity Evidence

The content-related validity of ACs is commonly established by evidence pertaining to the links between (a) the underlying performance dimensions and job activities, (b) job activities and exercise content, and (c) exercise content and performance dimensions (Thornton & Mueller-Hanson, 2004; see also I. L. Goldstein, Zedeck, & Schneider, 1993). However, a careful reading of widely accepted definitions of content validation—demonstrating how and the extent to which the content of a test adequately represents the domain of interest (Anastasi & Urbina, 1997)—indicates that these three linkages do not fully address content-related validity. So in spite of the long-standing notion that job analysis is the primary mechanism by which the content-related validity of ACs is established, it is erroneous to assume that this is sufficient to establish or ensure the content-related validity of ACs (Sackett, 1987). The domain to be assessed in an AC is the set of dimensions. As such, demonstrations of representativeness pertain to how well the dimensions them-

In contrast to the paucity of published papers investigating the content-related validity of AC ratings, as previously noted, we identified eight meta-analyses investigating the criterion-related validity of ACs. However, of these, six (i.e., Aamodt, 2004; Gaugler et al., 1987; Hardison & Sackett, 2007; Hermelin et al., 2007; Hunter & Hunter, 1984; Schmitt et al., 1984) were based on the OAR, and two (i.e., Arthur et al., 2003; Meriac et al., 2008) focused on dimensionlevel ratings. Although using the OAR as the level of analysis is a fairly common practice, Arthur et al. (2003) and Arthur and Villado (2008) have noted problems with it. For example, collapsing dimension ratings to generate an OAR results in a loss of construct-level information. As a result, it is conceptually unclear exactly what an OAR represents, making it difficult if not impossible to interpret comparative validity statements because OARs represent method-level data. Contrary to this common practice, Principles for the Validation and Use of Personnel Selection Procedures (Society for Industrial

222

Copyright American Psychological Association. Not for further distribution.

Assessment Centers

and Organizational Psychology, Inc., 2003) clearly and explicitly states the need to consider constructs when conducting a meta-analysis of predictors. Furthermore, Principles also notes that “when studies are cumulated on the basis of common methods (e.g., interviews, biodata) instead of constructs, a different set of interpretational difficulties arise” (p. 30)—namely, to what extent do the predictor methods measure the same predictor construct? Because OARs are aggregates of dimensions, comparing ACs at the level of OARs is almost guaranteed to result in divergent findings because one is, after all, comparing amalgamations of different dimensions. In other words, no two OARs are necessarily going to be the same without holding the dimensions (that comprise the OARs) constant. So from this perspective, it is not surprising that meta-analytic estimates of the criterion-related validity of ACs, operationalized as OARs, have obtained a wide range of estimates, ranging from .43 (Hunter & Hunter, 1984) to .41 (Schmitt et al., 1984), to .37 (Gaugler et al., 1987), to .28 (Hardison & Sackett, 2007; Hermelin et al., 2007), to .22 (Aamodt, 2004). Although these differences could be due to differences in meta-analytic methodology, inclusion criteria, and historical changes in the quality of ACs, we submit that a major potential reason for the range and divergence in AC findings may be the pervasive focus on OARs. Arthur et al. (2003), on the other hand, collapsed 168 AC dimension labels into an overriding set of six dimensions. Their meta-analysis of these six dimensions resulted in criterion-related validities that ranged from .25 (drive) to .39 (problem solving). In addition, as previously noted, a regression-based composite consisting of four out of the six dimensions accounted for the criterion-related validity of AC ratings and explained more variance in performance (20%) than Gaugler et al. (1987) were able to explain using the OAR (14%). Another question pertaining to the criterionrelated validity of ACs ratings (either dimension-level or OAR) has been the extent to which they contribute incremental validity over other predictor constructs and methods (e.g., see Collins et al., 2003). To this end, using Arthur et al.’s (2003) summary dimensions, Meriac et al. (2008) examined the degree of overlap

between the six AC dimensions and cognitive ability and the five-factor model (FFM) personality dimensions. They also investigated the extent to which the AC dimensions explained incremental variance in job performance beyond these individual difference variables. Their results indicated that the six AC dimensions—communication, consideration and awareness of others, drive, influencing others, organizing and planning, and problem solving— are distinguishable from cognitive ability and the FFM personality dimensions. In addition, they also explained a sizable proportion of variance in job performance beyond these individual difference variables. Primary studies (e.g., Dayan, Kasten, & Fox, 2002; Goffin, Rothstein, & Johnston, 1996) that have compared the criterion-related validities of OARs and cognitive ability and conscientiousness have obtained similar results.

Construct-Related Validity Evidence In sharp contrast to the consensus on the adequacy of AC content-related and criterion-related validity evidence, the status of the construct-related validity of AC ratings is controversial and has been strongly and actively debated. This debate originates from the fact that although ACs are designed to evaluate individuals on specific dimensions across multiple exercises, correlations among ostensibly different dimensions within exercises are typically larger than correlations among the same dimensions across exercises—a finding that is interpreted as a lack of construct-related validity. Furthermore, because ACs display relatively satisfactory levels of contentrelated and criterion-related validity evidence but allegedly not construct-related validity, this has been referred to as the AC construct-related validity paradox (Arthur et al., 2000; Woehr & Arthur, 2003). It is called a paradox because within the unitarian framework of validity, at a theoretical level, if a measurement tool demonstrates criterion-related and content-related validity evidence, as is widely accepted with ACs, then it should also be expected to demonstrate construct-related validity evidence (Binning & Barrett, 1989). Indeed, concerns with this issue have resulted in recent calls for the “redesign of ACs toward task- or role-based ACs and away from traditional dimension-based ACs” 223

Copyright American Psychological Association. Not for further distribution.

Arthur and Day

(Lance, 2008a, p. 84) because AC participant behaviors are “inherently cross-situationally (i.e., cross-exercise) specific, not cross-situationally consistent as was once thought” (Lance, 2008a, p. 84). Although there are contrarian opinions about their adequacy as explanations for the weak constructrelated validity evidence observed for AC ratings (e.g., Lance, 2008a, 2008b), several explanations have been advanced to account for the weak evidence. These include (a) the aforementioned methodological and design-related factors (Arthur et al., 2000; Woehr & Arthur, 2003); (b) the use of espoused versus actual constructs (Arthur et al., 2008) and the misspecification of the target constructs; (c) issues with analytical approaches specifically, the use of postexercise dimension ratings to represent AC scores, and the misapplication of the multitrait–multimethod design (Arthur et al., 2008; Bowler & Woehr, 2006; Howard, 2008; Lance, Woehr, & Meade, 2007); and (d) differential activation of traits depending on the demands of a particular exercise (Lievens, Chasteen, Day, & Christiansen, 2006). Of course, a full treatment and resolution of construct-related validity debate is beyond the scope of the present chapter. The reader is, however, strongly encouraged to read the “Why Assessment Centers Do Not Work the Way They Are Supposed To” focal article and commentaries in Industrial and Organizational Psychology: Perspectives on Science and Practice (2008). Our interpretation of the gist and tone of the papers in this issue is that the critique of AC construct-related validity and the subsequent call to abandon a dimension-based approach to ACs for a task- or role-based one was not found to be particularly compelling by most of the commentary authors and that for theoretical, conceptual, and empirical reasons, a dimension-based approach will continue to be the primary focus in AC research and practice. Consistent with this, a process analysis of ACs might be an informative approach to investigating and understanding what ACs measure and the boundary conditions under which they may or may not do so effectively—a point that was made over 20 years ago by Zedeck (1986). Furthermore, a focus on dimensions has applied usefulness. As noted by Howard (1997), 224

There are a number of practical reasons why assessment center users prefer dimensions based on human attributes rather than tasks. Lists of tasks can be long and generalize to fewer situations. Tasks are an unnatural way to describe people and are less meaningful than attributes for developmental feedback. . . . Task descriptions have little explanatory power . . . psychologists should be studying human qualities, not tasks. (p. 28) Moreover, organizing ACs around exercises and evaluating participants against exercises rather than dimensions is challenging in applied practice (Mayes, 1997).

Response Distortion and Validity Because participants in ACs know they are being assessed and that positive evaluations will result in favorable outcomes, issues pertaining to response distortion (i.e., faking) are germane. It is often thought that the realism and length of assessment involved in ACs make it difficult for participants to sustain disingenuous behavior. However, there is little empirical evidence to support or refute such a claim. To our knowledge, AC studies involving explicit instructions to “fake good” in comparison with control instructions have not been conducted. The few studies examining relationships between self-monitoring and AC performance have yielded mostly weak and statistically nonsignificant correlations. For example, using data from five administrations of an operational assessment center (n = 92), Arthur and Tubré (1999) found that self-monitoring was related to 360° performance data collected after the implementation of the assessment center (r = −.20 [average of 360° dimension scores] and −.29 [single overall effectiveness item]) but was not related to assessment center scores (r = .05). Anderson and Thacker (1985) showed a statistically nonsignificant correlation of −.14 between self-monitoring and AC performance in a sample of 45 men applying for computer sales positions while also showing a statistically significant correlation of .45 in a sample of 15 women applying for the same positions. Heon (1989) showed a statistically nonsignificant cor-

Copyright American Psychological Association. Not for further distribution.

Assessment Centers

relation of .14 between self-monitoring and overall AC performance in a sample of 75 state police officers in a promotional AC, with dimension correlations ranging from .05 to a statistically significant .24 for oral communication (the only dimension yielding a statistically significant correlation) across eight dimensions. In a sample of 191 managers from a developmental AC in a large telecommunications organization, Warech, Smither, Reilly, Millsap, and Reilly (1998) showed statistically nonsignificant correlations between self-monitoring and dimensions related to business competence (−.07) and interpersonal skills (−.01) and between self-monitoring and dimensions related to business competence (−.08) but a statistically significant correlation between self-monitoring and dimensions related to interpersonal skills (.14). Rubin, Bartels, and Bommer (2002) reported a statistically significant correlation of .12 between self-monitoring and candidate peer nominations of leadership emergence in a budget meeting exercise (leaderless group exercise) with a sample of 346 undergraduate business students participating in a developmental AC. In a similar vein, McFarland, Yun, Harold, Viera, and Moore (2005) showed predominantly weak relationships between impression management tactics and AC ratings in two samples of 30 firefighters participating in promotional ACs (Study 1 average r across 27 correlations = .07, SD = .15; Study 2 average r across 52 correlations = .11, SD = .24). However, their results also suggested relationships might depend on a complex interaction between type of tactic, AC exercise, and AC dimension. In contrast to the predominantly weak relationships for self-monitoring and impression management, König, Melchers, Kleinmann, Richter, and Klehe (2007) showed how the ability to identify the criteria for effective performance in AC exercises is positively related to AC performance (r = .39, N = 95). However, it is likely that such an extraneous influence of ability to identify criteria could be eliminated by making dimensions transparent to all participants (Kleinmann, 1993; Kleinmann et al., 1996). In general, the extent to which AC ratings are unduly influenced by intentional response distortion is an underresearched topic that warrants empirical attention.

COSTS AND PRACTICAL CONSIDERATIONS The total costs associated with the design, development, and administration of ACs can be quite high. Although specific costs are likely to vary substantially, it is not uncommon for ACs to cost over $2,000 per participant (Spychalski et al., 1997), making them one of the most expensive human resources assessment tools. These costs result from expenses associated with conducting job analyses, identifying and defining dimensions, designing exercises, and training assessors, as well as the expenses associated with compensating the time for multiple assessors to observe, document, categorize, discuss, and rate the performance of participants (as well as generating written reports and conducting feedback sessions in the case of training or developmental ACs). However, these relatively high costs must be weighed against the benefits accrued from using ACs. Consequently, multiple utility analysis studies (e.g., Burke & Frederick, 1986; Cascio & Ramos, 1986; Feltham, 1988b; Goldsmith, 1990; Hoffman & Thornton, 1997; Hogan & Zenke, 1986; Thornton, Murphy, Everest, & Hoffman, 2000; Tziner et al., 1994) have consistently reported favorable utility gains from the use of ACs for a variety of purposes in a variety of domains. For instance, within the specified boundary conditions that they investigated, Burke and Frederick’s (1986) utility estimates per selectee ranged from $998 to $5,306 per year (in 1986 dollars). Similarly, Cascio and Ramos’s (1986) utility gains over the interview were $2,676 per selectee per year. Finally, Hogan and Zenke (1986) reported that using a pool of 115 applicants for seven school principal positions, the AC had utility gains in excess of $58,000 over the traditional interview. There are additional nonfiscal benefits to using ACs that may also contribute to justifying their costs. For instance, applicants tend to view ACs as more face valid than (paper-and-pencil) cognitive ability tests and consequently have more favorable reactions to them and the selection process (Macan, Avedon, Paese, & Smith, 1994). Finally, in light of the relative cost disparities between ACs and other predictors, the practical value of ACs is further enhanced to the extent that they demonstrate incremental validity over these 225

Arthur and Day

Copyright American Psychological Association. Not for further distribution.

predictors. To this end, although they have typically been at the level of OARs, the results of primary studies (e.g., Dayan et al., 2002; Goffin et al., 1996; H. W. Goldstein, Yusko, Braverman, Smith, & Chung, 1998; Krause, Kersting, Haggestad, & Thornton, 2006) generally indicate that ACs demonstrate incremental validity over commonly used predictor constructs such as cognitive ability and personality (cf. Arthur & Villado, 2008), an effect that was also replicated in Meriac et al.’s (2008) dimensionlevel meta-analysis. SUBGROUP DIFFERENCES AND POTENTIAL FOR ADVERSE IMPACT Efforts to reduce subgroup differences and adverse impact continue to be of interest to I/O psychologists and organizations. A test or other employment-related decision-making tool is said to display adverse impact when decisions made on the basis of the test’s scores violate the 80% rule (i.e., the selection ratio in a subgroup of a protected class is less than 80% of the selection ratio of the other subgroup) or the differences between the scores of specified subgroups of the protected class (as per the Civil Rights Act of 1991) are statistically significant (Equal Employment Opportunity Commission, Civil Service Commission, Department of Labor, and Department of Justice, 1978). This definition characterizes adverse impact as a legal and administrative term or concept that arises from the psychological phenomena of subgroup differences on measures of psychological constructs or behavioral variables. Thus, at some specified selection ratio, adverse impact is primarily a function of the magnitude of the observed subgroup differences. Consequently, it is possible to have subgroup differences without adverse impact but less likely to have adverse impact without subgroup differences unless it is due to chance selection outcomes resulting from small sample sizes. Therefore, a focus on reducing subgroup differences would seem to be one direct approach to minimizing or eliminating adverse impact. In the context of such a focus, one suggested approach to reducing subgroup differences (see Arthur and Doverspike, 2005, and also Ployhart and Holtz, 2008, for extensive reviews of various 226

techniques and strategies) is the method-change approach. This approach to reducing subgroup differences focuses on changing the test method with the intention of altering test perceptions and attitudes and also reducing non-job-related reading demands. Thus, this approach posits that observed subgroup differences may partially arise from the mode of testing. Consequently, alternatives to traditional paper-and-pencil multiple-choice tests have included the use of ACs (H. W. Goldstein et al., 1998), performance tests (Chan & Schmitt, 1997), and situational judgment tests (Nguyen, McDaniel, & Whetzel, 2005). This stream of research has resulted in widely professed conclusions to the effect that adverse impact is less of a problem with ACs and work samples than it is for cognitive ability (Cascio & Aguinis, 2005, p. 372; Hoffman & Thornton, 1997; Thornton & Rupp, 2006; cf. Bobko, Roth, & Buster, 2005). However, the results of Dean, Roth, and Bobko’s (2008) meta-analysis question this received doctrine. They found an overall Black–White d of −0.52 and a Hispanic–White d of −0.28. These results suggest that subgroup differences involving race, especially Black–White differences, may be larger than previously thought and, subsequently, that ACs may be associated with more adverse impact against Blacks than has typically been characterized in the extant literature. Dean et al. (2008) also showed a small difference in favor of women over men (d = 0.19). Research has also shown negative relationships between age and AC scores (e.g., Burroughs, Rollins, & Hopkins, 1973), but it is thought such observed relationships have been primarily a function of sampling issues such as preassessment selection (Thornton & Byham, 1982). In addition, recent research also suggests that lower AC scores as a function of age do not occur for older workers who are high on certain personality variables such as exhibition and dominance (Krajewski, Goffin, Rothstein, & Johnston, 2007). It is worth noting a pervasive problem with the method-change approach in general, and especially as it applies to ACs in particular. Because this approach, as practiced, fails to disaggregate predictor constructs and predictor methods, it is unclear whether the observed reductions in subgroup differences or lack thereof are due to changes in the

Copyright American Psychological Association. Not for further distribution.

Assessment Centers

constructs being measured or the methods of assessment (Anderson, Lievens, van Dam, & Born, 2006; H. W. Goldstein, Yusko, & Nicolopoulos, 2001). So, for example, in the case of ACs, because the methods and constructs are confounded, it is unclear whether the posited reduction in subgroup differences is due to the assessment method of ACs or the constructs measured by ACs (Arthur et al., 2003). With this in mind, research has shown that more cognitively loaded dimensions yield larger Black–White differences (H. W. Goldstein et al., 2001) and that interpersonally loaded dimensions yield larger female–male differences (Anderson et al., 2006). In short, any method of assessment can display high or low levels of subgroup differences; it depends on the construct(s) being measured. From a practical perspective, to clearly delineate which predictor will result in a lower level of subgroup differences, we need to know both the method of assessment (e.g., ACs) and the constructs (e.g., communication, consideration and awareness of others, drive, influencing others, organizing and planning, problem solving) assessed by said methods. Thus, although the results of a meta-analysis like that of Dean et al. (2008) are of some informational value, they are limited by the fact that they are conducted at the level of the OAR and are not disaggregated by dimensions. It is not unreasonable to posit that the presence and magnitude of subgroup differences is likely to be a function of the constructs assessed (e.g., see Foldes, Duehr, & Ones, 2008; Hough, Oswald, & Ployhart, 2001). However, we were unable to locate any empirical AC studies that investigated this issue. Any future studies investigating these issues will need to disaggregate the constructs (dimensions) assessed from the method of assessment to permit clear conclusions as to whether observed subgroup differences are dimension or method effects (Arthur & Villado, 2008). In summary, in spite of (or maybe because of ) its history and growing volume of case law (see Thornton and Rupp, 2006, for a list and review of cases), it would seem that there is no a priori favorable or unfavorable legal predisposition toward ACs, and rightfully so. The ability of an AC to withstand legal challenge is primarily a function of the demonstrable job-relatedness of that specific AC. This fact

highlights the importance of recognizing the scientific versus legal demands and guidelines: that whereas they may on occasion be related and overlap, this is not always the case. Hence, whereas scientific endeavors stress validity generalization, legal challenges focus on situational specificity. INTERNATIONAL SCOPE From the historical account provided in the Historical Background section, it is evident that the science and practice of ACs is international in scope. The rapid growth of ACs in the public and private sectors in the United States sparked by the AT&T MPS was quickly followed by growth in AC use outside the United States. The first International Congress on Assessment Center Methods met in 1973. The Congress facilitates the sharing of knowledge on both research and practice through annual meetings, and it provides the guidelines and ethical considerations for AC operations emulated throughout the world. A clear sense for the international scope of ACs can be seen in the authorship of empirical journal articles appearing in the scholarly literature. As shown in Table 7.1, 75% of empirical journal articles on ACs over the past 10 years (identified via a PsycINFO

TABLE 7.1 Frequency Count of Empirical Journal Articles on Assessment Centers by Country of Lead Author Identified via a PsycINFO ( January 1998–August 2008) Search Country United States Germany Netherlands Belgium New Zealand United Kingdom China Switzerland Canada Ireland Israel Australia France

No. of articles

% of articles

20 14 11 11 5 5 3 3 2 2 2 1 1

25 18 14 14 6 6 4 4 3 3 3 1 1

227

Copyright American Psychological Association. Not for further distribution.

Arthur and Day

search) are from scholars outside the United States. For the most part, the vast majority of this empirical research focuses on issues pertaining to the criterionrelated and construct-related validity of AC ratings. Despite the international scope of AC practice, there is little published empirical research regarding the role cross-cultural factors play in AC use and practice. For instance, recent research has shown that compared with North Americans, Western Europeans are more likely to use presentation, group discussion, and case analysis exercises, and they are more likely to assess dimensions pertaining to drive and consideration of others (Krause & Thornton, 2009). So not only is there limited published research on differences in AC practice around the world, the empirical literature also has not adequately addressed the cross-cultural generalizability of AC scores. This is an especially important issue confronted by multinational companies that must balance applying universal standards for managing human resources and building a unifying organizational culture with sensitivity to local cultural differences. In light of the array of design elements to the AC methodology, although questions regarding cross-cultural generalizability can be targeted at a large number of specific issues, questions regarding the cross-cultural generalizability of exercises and dimensions are perhaps the most fundamental. For instance, it is conceivable that an exercise developed in one culture that effectively evokes behaviors for a specific set of dimensions may not evoke such behaviors in another culture. For example, competitive leaderless group discussions are frequently designed to spark contention among a group of peers; they are zero-sum exercises. Such an exercise may not be as effective in collectivistic cultures that value conformity and consensus seeking. With regard to dimensions, again it is conceivable that they may be viewed differently across cultures. For example, the expectation in one culture could be that coaching (or leadership in general) involves a large degree of subordinate participation and empowerment, whereas a more autocratic and direct style may be expected in another culture. This is just a snippet of the challenging issues facing many AC practitioners. The notion that a given AC would be generalizable across varying cultures can easily be 228

dismissed, but the manner by which ACs are to be designed, implemented, and evaluated vis-à-vis cross-cultural differences presents a great challenge that should not be ignored. As such, empirical research addressing international and cross-cultural issues surrounding ACs is sorely needed (Briscoe, 1997; Thornton & Rupp, 2006). Along these lines, empirical research presented at the 2008 International Congress on Assessment Center Methods showed significant cultural differences in AC performance across cultures with Anglo-Germanics generally achieving higher scores than Asians (Bernthal & Lanik, 2008). As with issues regarding examinations of adverse impact discussed earlier in this chapter, it is difficult to tease out the extent to which these culture-based performance differences reflect true differences in dimension performance or cultural bias in the exercises. With this in mind, empirical research examining the cross-cultural relevance of specific dimensions (i.e., constructs) is a strong starting point for a stream of research on crosscultural differences in AC scores (e.g., Gibbons, Rupp, Kim, & Woo, 2006). This could then subsequently be followed by a focus on the cross-cultural generalizability of specific exercise formats, instructions, and the scenarios depicted in exercises. CONCLUSIONS AND FUTURE DIRECTIONS The AC has a rich history in the science and practice of identifying managerial talent. The simulation of job performance through behavioral exercises and the triangulation of performance dimensions through multiple exercises and multiple assessors distinguish the AC from other forms of assessment. Substantial meta-analytic criterion-related validity evidence supports the use of ACs for selection and promotion purposes. Utility analyses and evidence of favorable participant reactions also support the use of ACs. However, the AC approach has been criticized for weaknesses in how content-related validity is established and reported. In addition, the evidence surrounding construct-related validity has been strongly and actively debated. The flexibility of the AC as a method is evident in its use by a wide range of public and private organizations around the world and application

Copyright American Psychological Association. Not for further distribution.

Assessment Centers

to a wide variety of jobs for a number of different purposes. Overall, it appears that the science and practice of ACs is quite vibrant. We hope to see more research in the near future concerning the use of ACs for purposes other than selection and promotion, especially training and development. More research is also warranted regarding the use of technology to improve AC fidelity as well as to streamline the scoring process. In light of the international scope of AC practice, future research needs to address the cross-cultural issues associated with the design, implementation, and evaluation of ACs. We also hope that the critiques concerning the evidence of content-related validity and the recent debates over the construct-related validity spark more research better targeted at understanding the constructs underlying work performance. Such efforts should include a better explication of dimensions and external validation approaches that examine the nomological network surrounding AC dimensions. Finally, consonant with the centrality of constructs in the study and understanding of behavior, we call for a stronger focus and emphasis on dimension-level data and results. Thus, we recommend that OAR-only research should be discouraged and instead, whenever OAR results and data are reported, dimension-level data and results should be reported as well. This call is motivated by the fact that a focus on OARs makes it difficult if not impossible to meaningfully compare ACs with other predictor constructs such as cognitive ability and conscientiousness because the resultant comparison is one between a method and constructs (Arthur & Villado, 2008). Furthermore, with such comparisons, one is in effect comparing the validity of a single construct (e.g., cognitive ability or conscientiousness) to that of an aggregate of constructs (i.e., multiple AC dimensions). Finally, the use of OARs results in a loss of construct- (dimension-) level information and obscures the fact that some dimensions may be more predictive of performance than others (Arthur et al., 2003). We recognize and acknowledge that for selection and other administrative purposes, a composite score is required to aid in decision making. However, because AC dimensions represent different constructs, the procedures used to generate composite scores can be the same as those typically used to combine scores from multiple

predictor batteries such as the use of multiple regression (e.g., see Arthur et al., 2003, Meriac et al., 2008) or other means of assigning weights to different dimensions (e.g., see Arthur, Doverspike, & Barrett, 1996; Bobko et al., 2007).

References Aamodt, M. G. (2004). Research in law enforcement selection. Boca Raton, FL: Brown Walker. Abraham, J. D., Morrison, J. D., Jr., & Burnett, D. D. (2006). Feedback seeking among developmental assessment center participants. Journal of Business and Psychology, 20, 383–394. doi:10.1007/s10869-005-9008-z American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (1999). Standards for educational and psychological testing. Washington, DC: American Educational Research Association. American Psychological Association. (1992). Ethical principles of psychologists and code of ethics. American Psychologist, 47, 1597–1611. doi:10.1037/ 0003-066X.47.12.1597 Anastasi, A., & Urbina, S. (1997). Psychological testing (7th ed.). Upper Saddle River, NY: Prentice-Hall. Anderson, N., Lievens, F., van Dam, K., & Born, M. (2006). A construct-driven investigation of gender differences in a leadership-role assessment center. Journal of Applied Psychology, 91, 555–566. doi:10.1037/ 0021-9010.91.3.555 Anderson, L. R., & Thacker, J. (1985). Self-monitoring and sex as related to assessment center ratings and job performance. Basic and Applied Social Psychology, 6, 345–361. doi:10.1207/s15324834basp0604_5 Ansbacher, H. L. (1941). German military psychology. Psychological Bulletin, 38, 370–392. doi:10.1037/ h0056263 Anstey, E. (1977). A 30-year follow-up of the CSSB procedure, with lessons for the future. Journal of Occupational Psychology, 50, 149–159. Arthur, W., Jr., & Benjamin, L. T., Jr. (1999). Psychology applied to business. In A. M. Stec & D. A. Bernstein (Eds.), Psychology: Fields application (pp. 98–115). Boston: Houghton Mifflin. Arthur, W., Jr., Day, E. A., McNelly, T. L., & Edens, P. S. (2003). Meta-analysis of the criterion-related validity of assessment center dimensions. Personnel Psychology, 56, 125–154. doi:10.1111/j.1744-6570.2003.tb00146.x Arthur, W., Jr., Day, E. A., & Woehr, D. J. (2008). Mend it, don’t end it: An alternate view of assessment center construct-related validity evidence. Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 105–111. doi:10.1111/j.1754-9434. 2007.00019.x 229

Arthur and Day

Arthur, W., Jr., & Doverspike, D. (2005). Achieving diversity and reducing discrimination in the workplace through human resource management practices: Implications of research and theory for staffing, training, and rewarding performance. In R. L. Dipboye & A. Colella (Eds.), Discrimination at work: The psychological and organizational bases (pp. 305–327). San Francisco: Jossey-Bass.

Copyright American Psychological Association. Not for further distribution.

Arthur, W., Jr., Doverspike, D., & Barrett, G. V. (1996). Development of a job analysis-based procedure for weighting and combining content-related tests into a single test battery score. Personnel Psychology, 49, 971–985. doi:10.1111/j.1744-6570.1996.tb02457.x Arthur, W., Jr., & Tubré, T. C. (1999, April). The assessment center construct-related validity paradox: A case of construct misspecification? In M. A. Quiñones (Chair), Assessment centers, 21st century: New issues and new answers to old problems. Symposium presented at the 14th Annual Conference of the Society for Industrial and Organizational Psychology, Atlanta, GA. Arthur, W., Jr., & Villado, A. J. (2008). The importance of distinguishing between constructs and methods when comparing predictors in personnel selection research and practice. Journal of Applied Psychology, 93, 435–442. doi:10.1037/0021-9010.93.2.435 Arthur, W., Jr., Woehr, D. J., & Maldegen, R. (2000). Convergent and discriminant validity of assessment center dimensions: An empirical re-examination of the assessment center construct-related validity paradox. Journal of Management, 26, 813–835. doi:10.1016/ S0149-2063(00)00057-X Assessment Staff, OSS. (1948). Assessment of men. New York: Rinehart. Bell, S. T., & Arthur, W., Jr. (2008). Feedback acceptance in developmental assessment centers: The role of feedback message, participant personality, and affective response to the feedback session. Journal of Organizational Behavior, 29, 681–703. doi:10.1002/ job.525

and meta-analysis. Organizational Research Methods, 10, 689–709. doi:10.1177/1094428106294734 Bowler, M. C., & Woehr, D. J. (2006). A meta-analytic evaluation of the impact of dimension and exercise factors on assessment center ratings. Journal of Applied Psychology, 91, 1114–1124. doi:10.1037/ 0021-9010.91.5.1114 Brannick, M. T. (2008). Back to basics of test construction and scoring. Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 131–133. doi:10.1111/j.1754-9434.2007.00025.x Bray, D. W. (1982). The assessment center and the study of lives. American Psychologist, 37, 180–189. doi:10.1037/0003-066X.37.2.180 Bray, D. W., & Grant, D. L. (1966). The assessment center in the measurement of potential for business management. Psychological Monographs, 80 (17, Whole No. 625), 1–27. Briscoe, D. (1997). Assessment centers: Cross-cultural and cross-national issues. Journal of Social Behavior and Personality, 12, 261–270. Burke, M. J., & Frederick, J. T. (1986). A comparison of economic utility estimates for alternative SDy estimation procedures. Journal of Applied Psychology, 71, 334–339. doi:10.1037/0021-9010.71.2.334 Burroughs, W. A., Rollins, J. B., & Hopkins, J. J. (1973). The effect of age, departmental experience, and prior rater experience on performance in assessment center exercises. Academy of Management Journal, 16, 335–339. doi:10.2307/255335 Bycio, P., Alvares, K. M., & Hahn, J. (1987). Situation specificity in assessment center ratings: A confirmatory analysis. Journal of Applied Psychology, 72, 463–474. doi:10.1037/0021-9010.72.3.463 Caldwell, C., Thornton, G. C., III, & Gruys, M. L. (2003). Ten classic assessment center errors: Challenges to selection validity. Public Personnel Management, 23, 73–88.

Bernthal, P., & Lanik, M. (2008, September). Crosscultural comparisons of assessment center performance. Presentation at the 34th International Congress on the Assessment Center Method, Washington, DC.

Cascio, W. F., & Aguinis, H. (2005). Applied psychology in human resource management (6th ed.). Upper Saddle River, NJ: Prentice Hall.

Binning, J. F., & Barrett, G. V. (1989). Validity of personnel decisions: A conceptual analysis of the inferential and evidential bases. Journal of Applied Psychology, 74, 478–494. doi:10.1037/0021-9010.74.3.478

Cascio, W. F., & Ramos, R. A. (1986). Development and application of a new method for assessing job performance in behavioral/economic terms. Journal of Applied Psychology, 71, 20–28. doi:10.1037/0021-9010.71.1.20

Bobko, P., Roth, P. L., & Buster, M. A. (2005). Work sample tests and expected reduction in adverse impact: A cautionary note. International Journal of Selection and Assessment, 13, 1–10. doi:10.1111/ j.0965-075X.2005.00295.x

Chan, D., & Schmitt, N. (1997). Video-based versus paper-and-pencil method of assessment in situational judgment tests: Subgroup differences in test performance and face validity perceptions. Journal of Applied Psychology, 82, 143–159. doi:10.1037/ 0021-9010.82.1.143

Bobko, P., Roth, P. L., & Buster, M. A. (2007). The usefulness of unit weights in creating composite scores: A literature review, application of content validity, 230

Chen, H.-C. (2006). Assessment center: A critical mechanism for assessing HRD effectiveness and account-

Assessment Centers

ability. Advances in Developing Human Resources, 8, 247–264. doi:10.1177/1523422305286155 Chen, H.-C., & Naquin, S. S. (2006). An integrative model of competency development, training design, assessment center, and multi-rater assessment. Advances in Developing Human Resources, 8, 265–282. doi:10.1177/1523422305286156

Copyright American Psychological Association. Not for further distribution.

Civil Rights Act of 1991, Pub. L. No. 102-166, 105 Stat. 1071 (1991). Collins, J. M., Schmidt, F. L., Sanchez-Ku, M., Thomas, L., McDaniel, M. A., & Le, H. (2003). Can basic individual differences shed light on the construct meaning of assessment center evaluations? International Journal of Selection and Assessment, 11, 17–29. doi:10.1111/ 1468-2389.00223 Connelly, B. S., Ones, D. S., Ramesh, A., & Goff, M. (2008). A pragmatic view of assessment center exercises and dimensions. Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 121–124. doi:10.1111/j.1754-9434.2007.00022.x Dayan, K., Kasten, R., & Fox, S. (2002). Entry-level police candidate assessment center: An efficient tool or a hammer to kill a fly? Personnel Psychology, 55, 827–849. doi:10.1111/j.1744-6570.2002.tb00131.x Dean, M. A., Roth, P. L., & Bobko, P. (2008). Ethnic and gender subgroup differences in assessment center ratings: A meta-analysis. Journal of Applied Psychology, 93, 685–691. doi:10.1037/0021-9010.93.3.685 Dhanju, N. (2007). The perceived effectiveness of assessment centers for successful and unsuccessful participants for the selection of team leaders in an Indian call center: A case study. Unpublished doctoral dissertation, University of Minnesota, Minneapolis. Donahue, L. M., Truxillo, D. M., Cornwell, J. M., & Gerrity, M. J. (1997). Assessment center construct validity and behavioral checklists: Some additional findings. Journal of Social Behavior and Personality, 12, 85–108.

Foldes, H., Duehr, E. E., & Ones, D. S. (2008). Group differences in personality: Meta-analyses comparing five U.S. racial groups. Personnel Psychology, 61, 579–616. doi:10.1111/j.1744-6570.2008.00123.x Gaugler, B. B., Rosenthal, D. B., Thornton, G. C., III, & Bentson, B. (1987). Meta-analysis of assessment center validity. Journal of Applied Psychology, 72, 493–511. doi:10.1037/0021-9010.72.3.493 Gaugler, B. B., & Thornton, G. C., III. (1989). Number of assessment center dimensions as a determinant of assessor generalizability of the assessment center ratings. Journal of Applied Psychology, 74, 611–618. doi:10.1037/0021-9010.74.4.611 Gibbons, A. M., Rupp, D. E., Kim, M., & Woo, S. E. (2006). Perceptions of managerial performance dimensions in Korea. The Psychologist Manager Journal, 9, 125–143. doi:10.1207/s15503461tpmj0902_5 Gilbert, P. J. (1981). An investigation of clinical and mechanical combination of assessment center data. Journal of Assessment Center Technology, 4, 1–10. Goffin, R. D., Rothstein, M. G., & Johnston, N. G. (1996). Personality testing and the assessment center: Incremental validity for managerial selection. Journal of Applied Psychology, 81, 746–756. doi:10.1037/ 0021-9010.81.6.746 Goldsmith, R. F. (1990). Utility analysis and its application to the study of cost effectiveness of the assessment center method. In K. R. Murphy & F. E. Saal (Eds.), Psychology in organizations: Integrating science and practice (pp. 92–110). Hillsdale, NJ: LEA. Goldstein, H. W., Yusko, K. P., Braverman, E. P., Smith, D. B., & Chung, B. (1998). The role of cognitive ability in the subgroup differences and incremental validity of assessment center exercises. Personnel Psychology, 51, 357–374. doi:10.1111/j.1744-6570. 1998.tb00729.x

Dunnette, M. D. (1971). The assessment of managerial talent. In P. McReynolds (Ed.), Advances in psychological assessment (Vol. 2, pp. 79–108). Palo Alto, CA: Science and Behavior Books.

Goldstein, H. W., Yusko, K. P., & Nicolopoulos, V. (2001). Exploring Black-White subgroup differences of managerial competencies. Personnel Psychology, 54, 783–807. doi:10.1111/j.17446570.2001.tb00232.x

Equal Employment Opportunity Commission, Civil Service Commission, Department of Labor, and Department of Justice. (1978). Uniform guidelines on employee selection procedures. Federal Register, 43, 38295–38309.

Goldstein, I. L., Zedeck, S., & Schneider, B. (1993). An exploration of the job analysis-content validity process. In N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations (pp. 3–34). San Francisco: Jossey-Bass.

Farago, L. (1942). German psychological warfare. New York: G. P. Putnam’s Sons.

Hardison, C. M., & Sackett, P. R. (2007). Kriterienbezogene Validität des Assessment Centers: Lebendig und wohlauf? [Assessment center criterion-related validity: Alive and well?]. In H. Schuler (Ed.), Assessment Center zur Potenzialanalyse (pp. 192–202). Göttingen, Germany: Hogrefe.

Feltham, R. (1988a). Assessment centre decision making: Judgmental vs. mechanical. Journal of Occupational Psychology, 61, 237–241. Feltham, R. (1988b). Validity of a police assessment centre: A 1-19-year follow-up. Journal of Occupational Psychology, 61, 129–144.

Harris, H. (1949). The group approach to leadership testing. London: Routledge and Paul. 231

Arthur and Day

Haymaker, J. C., & Grant, D. L. (1982). Development of a model for content validation of assessment centers. Journal of Assessment Center Technology, 5, 1–7. Heon, S. M. (1989). Self-monitoring and cognitive complexity as related to assessment center performance. Unpublished doctoral dissertation, George Washington University, Washington, DC.

Copyright American Psychological Association. Not for further distribution.

Hermelin, E., Lievens, F., & Robertson, I. T. (2007). The validity of assessment centres for the prediction of supervisory performance ratings: A meta-analysis. International Journal of Selection and Assessment, 15, 405–411. doi:10.1111/j.1468-2389.2007.00399.x Hoffman, C. C., & Thornton, G. C., III. (1997). Examining selection utility where competing predictors differ in adverse impact. Personnel Psychology, 50, 455–470. doi:10.1111/j.1744-6570.1997.tb00916.x Hogan, J., & Zenke, L. L. (1986). Dollar-value utility of alternative procedures for selecting school principals. Educational and Psychological Measurement, 46, 935–945. doi:10.1177/001316448604600413 Hough, L. M., Oswald, F. L., & Ployhart, R. E. (2001). Determinants, detection and amelioration of adverse impact in personnel selection procedures: Issues, evidence and lessons learned. International Journal of Selection and Assessment, 9, 152–194. doi:10.1111/ 1468-2389.00171 Howard, A. (1997). A reassessment of assessment centers: Challenges for the 21st century. Journal of Social Behavior and Personality, 12, 13–52. Howard, A. (2008). Making assessment centers work the way they are supposed to. Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 98–104. doi:10.1111/j.1754-9434.2007.00018.x Howard, A., & Bray, D. W. (1988). Managerial lives in transition: Advancing age and changing times. New York: Guilford Press. Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternative predictors of job performance. Psychological Bulletin, 96, 72–98. doi:10.1037/0033-2909.96.1.72 Industrial and Organizational Psychology: Perspectives on Science and Practice. (2008). 1(1). International Task Force on Assessment Center Guidelines. (2000). Guidelines and ethical considerations for assessment center operations. Public Personnel Management, 29, 315–331. Jackson, D. J. R., Atkins, S. G., Fletcher, R. B., & Stillman, J. A. (2005). Frame of reference training for assessment centers: Effects on interrater reliability when rating behaviors and ability traits. Public Personnel Management, 34, 17–30. Jackson, D. J. R., Stillman, J. A., Burke, S., & Englert, P. (2007). Self versus assessor ratings and their classification in assessment centres: Profiling the self-rater. New Zealand Journal of Psychology, 36, 93–99. 232

Jennings, H. H. (1949). Military use of sociometric and situation tests in Great Britain, France, Germany, and the United States. Sociometry, 12, 191–201. doi:10.2307/2785386 Joiner, D. A. (2002). Assessment centers: What’s new? Public Personnel Management, 31, 179–185. Jones, R. G., & Whitmore, M. D. (1995). Evaluating developmental assessment centers as interventions. Personnel Psychology, 48, 377–388. doi:10.1111/ j.1744-6570.1995.tb01762.x Kleinmann, M. (1993). Are rating dimensions in assessment centers transparent to participants? Consequences for criterion and construct validity. Journal of Applied Psychology, 78, 988–993. doi:10.1037/0021-9010.78.6.988 Kleinmann, M., Kuptsch, C., & Köller, O. (1996). Transparency: A necessary requirement for the construct validity of assessment centres. Applied Psychology: An International Review, 45, 67–84. doi:10.1111/j.1464-0597.1996.tb00849.x Klimoski, R., & Brickner, M. (1987). Why do assessment centers work? The puzzle of assessment center validity. Personnel Psychology, 40, 243–260. doi:10.1111/j.1744-6570.1987.tb00603.x Kolk, N. J., Born, M. P., & Van der Flier, H. (2003). The transparent assessment center: The effect of revealing dimensions to applicants. Applied Psychology: An International Review, 52, 648–668. doi:10.1111/ 1464-0597.00156 König, C. J., Melchers, K. G., Kleinmann, M., Richter, G. M., & Klehe, U.-C. (2007). Candidates’ ability to identify criteria in nontransparent selection procedures: Evidence from an assessment center and a structured interview. International Journal of Selection and Assessment, 15, 283–292. doi:10.1111/j.1468-2389. 2007.00388.x Krajewski, H. T., Goffin, R. D., Rothstein, M. G., & Johnston, N. G. (2007). Is personality related to assessment center performance? That depends on how old you are. Journal of Business and Psychology, 22, 21–33. doi:10.1007/s10869-007-9043-z Krause, D. E., Kersting, M., Haggestad, E. D., & Thornton, G. C., III. (2006). Incremental validity of assessment center ratings over cognitive ability tests: A study at the executive management level. International Journal of Selection and Assessment, 14, 360–371. Krause, D. E., & Thornton, G. C., III. (2009). A crosscultural look at assessment center practices: Survey results from Western Europe and North America. Applied Psychology: An International Review, 58, 557–585. Lance, C. E. (2008a). Why assessment centers do not work the way they are supposed to. Industrial and Organiza-

Assessment Centers

Copyright American Psychological Association. Not for further distribution.

tional Psychology: Perspectives on Science and Practice, 1, 84–97. doi:10.1111/j.1754-9434.2007.00017.x

of Applied Psychology, 93, 1042–1052. doi:10.1037/ 0021-9010.93.5.1042

Lance, C. E. (2008b). Where have we been, how did we get there, and where shall we go? Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 140–146. doi:10.1111/j.1754-9434. 2007.00028.x

Messick, S. J. (1995). The validity of psychological assessment: Validation of inferences from persons’ responses and performances as scientific inquiry into score meaning. American Psychologist, 50, 741–749. doi:10.1037/0003-066X.50.9.741

Lance, C. E., Woehr, D. J., & Meade, A. W. (2007). Case study: A Monte Carlo investigation of assessment center construct validity models. Organizational Research Methods, 10, 430–448. doi:10.1177/ 1094428106289395

Morris, B. S. (1949). Officer selection in the British Army, 1942-1945. Occupational Psychology, 23, 219–234.

Lievens, F. (1998). Factors which improve the construct validity of assessment centers: A review. International Journal of Selection and Assessment, 6, 141–152. doi:10.1111/1468-2389.00085 Lievens, F. (2001). Assessor training strategies and their effects on accuracy, interrater reliability, and discriminant validity. Journal of Applied Psychology, 86, 255–264. doi:10.1037/0021-9010.86.2.255 Lievens, F., Chasteen, C. S., Day, E. A., & Christiansen, N. D. (2006). Large-scale investigation of the role of trait activation theory for understanding assessment center convergent and discriminant validity. Journal of Applied Psychology, 91, 247–258. doi:10.1037/ 0021-9010.91.2.247 Lievens, F., & Conway, J. M. (2001). Dimension and exercise variance in assessment center scores: A large-scale evaluation of multitrait-multimethod studies. Journal of Applied Psychology, 86, 1202–1222. doi:10.1037/0021-9010.86.6.1202 Macan, T. H., Avedon, M. J., Paese, M., & Smith, D. E. (1994). The effects of applicants’ reactions to cognitive ability tests and an assessment center. Personnel Psychology, 47, 715–738. doi:10.1111/j.1744-6570. 1994.tb01573.x Mayes, B. T. (1997). Insights into the history and future of assessment centers: An interview with Dr. Douglas W. Bray and Dr. William Byham. Journal of Social Behavior and Personality, 12, 3–12. McFarland, L. A., Yun, G., Harold, C. M., Viera, L., Jr., & Moore, L. G. (2005). An examination of impression management use and effectiveness across assessment center exercises: The role of competency demands. Personnel Psychology, 58, 949–980. doi:10.1111/ j.1744-6570.2005.00374.x Melancon, S. C., & Williams, M. S. (2006). Competencybased assessment center design: A case study. Advances in Developing Human Resources, 8, 283–314. doi:10.1177/1523422305286157 Meriac, J. P., Hoffman, B. J., Woehr, D. J., & Fleisher, M. S. (2008). Further evidence for the validity of assessment center dimensions: A meta-analysis of the incremental criterion-related validity of dimension ratings. Journal

Murray, H. A. (1938). Explorations in personality. New York: Oxford University Press. Nguyen, N. T., McDaniel, M. A., & Whetzel, D. (2005, April). Subgroup differences in situational judgment test performance: A meta-analysis. Paper presented at the 20th Annual Conference of the Society for Industrial and Organizational Psychology, Los Angeles, CA. Noonan, L. E., & Sulsky, L. M. (2001). Impact of frameof-reference and behavioral observation training on alternative training effectiveness criteria in a Canadian military sample. Human Performance, 14, 3–26. doi:10.1207/S15327043HUP1401_02 Ployhart, R. E., & Holtz, B. C. (2008). The diversityvalidity dilemma: Strategies for reducing racioethnic and sex subgroup differences and adverse impact in selection. Personnel Psychology, 61, 153–172. doi:10.1111/j.1744-6570.2008.00109.x Pynes, J. E., & Bernardin, J. H. (1992). Mechanical versus consensus-derived assessment center ratings: A comparison of job performance validities. Public Personnel Management, 21, 17–28. Reilly, R. R., Henry, S., & Smither, J. W. (1990). An examination of the effects of using behavior checklists on the construct validity of assessment center dimensions. Personnel Psychology, 43, 71–84. doi:10.1111/j.1744-6570.1990.tb02006.x Robie, C., Adams, K. A., Osburn, H. G., Morris, M. A., & Etchegaray, J. M. (2000). Effects of the rating process on the construct validity of assessment center dimension evaluations. Human Performance, 13, 355–370. doi:10.1207/S15327043HUP1304_2 Roch, S. G. (2006). Discussion and consensus in rater groups: Implications for behavioral and rating accuracy. Human Performance, 19, 91–115. doi:10.1207/ s15327043hup1902_1 Roch, S. G. (2007). Why convene rater teams: An investigation of the benefits of anticipated discussion, consensus, and rater motivation. Organizational Behavior and Human Decision Processes, 104, 14–29. doi:10.1016/j.obhdp.2006.08.003 Roth, P. L., Bobko, P., & McFarland, L. A. (2005). A meta-analysis of work sample test validity: Updating and integrating some classic literature. Personnel Psychology, 58, 1009–1037. doi:10.1111/j.1744-6570. 2005.00714.x 233

Arthur and Day

Rowe, F. A., & Mauer, K. A. (1991). Career guidance, career assessment, and consultancy. Journal of Career Development, 17, 223–233. Rubin, R. S., Bartels, L. K., & Bommer, W. H. (2002). Are leaders smarter or do they just seem that way? Exploring perceived intellectual competence and leadership emergence. Social Behavior and Personality, 30, 105–118. doi:10.2224/sbp.2002.30.2.105

Copyright American Psychological Association. Not for further distribution.

Rupp, D. E., Gibbons, A. M., Baldwin, A. M., Snyder, L. A., Spain, S. M., Woo, S. E., et al. (2006). An initial validation of developmental assessment centers as accurate assessments and effective training interventions. The Psychologist Manager Journal, 9, 171–200. doi:10.1207/s15503461tpmj0902_7

assessment center. Personnel Psychology, 43, 1–12. doi:10.1111/j.1744-6570.1990.tb02003.x Silvester, J., & Dykes, C. (2007). Selecting political candidates: A longitudinal study of assessment centre performance and political success in the 2005 UK General Election. Journal of Occupational and Organizational Psychology, 80, 11–25. doi:10.1348/ 096317906X156287 Society for Industrial and Organizational Psychology, Inc. (2003). Principles for the validation and use of personnel selection procedures (4th ed.). Bowling Green, OH: SIOP.

Russell, C. (1985). Individual decision processes in an assessment center. Journal of Applied Psychology, 70, 737–746. doi:10.1037/0021-9010.70.4.737

Spychalski, A. C., Quiñones, M. A., Gaugler, B. B., & Pohley, K. (1997). A survey of assessment center practices in organizations in the United States. Personnel Psychology, 50, 71–90. doi:10.1111/ j.1744-6570.1997.tb00901.x

Sackett, P. R. (1987). Assessment centers and content validity: Some neglected issues. Personnel Psychology, 40, 13–25. doi:10.1111/j.1744-6570.1987.tb02374.x

Thornton, G. C., III, & Byham, W. C. (1982). Assessment centers and managerial performance. San Diego, CA: Academic Press.

Sackett, P. R., & Dreher, G. F. (1982). Constructs and assessment center dimensions: Some troubling empirical findings. Journal of Applied Psychology, 67, 401–410. doi:10.1037/0021-9010.67.4.401

Thornton, G. C., III, & Mueller-Hanson, R. A. (2004). Developing organizational simulations: A guide for practitioners and students. Mahwah, NJ: Erlbaum.

Sackett, P. R., & Hakel, M. D. (1979). Temporal stability and individual differences in using assessment center information from overall ratings. Organizational Behavior and Human Performance, 23, 120–137. doi:10.1016/0030-5073(79)90051-5 Schleicher, D. J., & Day, D. V. (1998). A cognitive evaluation of frame-of-reference rater training: Content and process issues. Organizational Behavior and Human Decision Processes, 73, 76–101. doi:10.1006/ obhd.1998.2751 Schleicher, D. J., Day, C. V., Mayes, B. T., & Riggio, R. E. (2002). A new frame for frame-of-reference training: Enhancing the construct validity of assessment centers. Journal of Applied Psychology, 87, 735–746. doi:10.1037/0021-9010.87.4.735 Schmitt, N. (1977). Interrater agreement in dimensionality and combination of assessment center judgments. Journal of Applied Psychology, 62, 171–176. doi:10.1037/0021-9010.62.2.171 Schmitt, N., Gooding, R. Z., Noe, R. A., & Kirsch, M. (1984). Meta-analysis of validity studies published between 1964 and 1982 and the investigation of study characteristics. Personnel Psychology, 37, 407–422. doi:10.1111/j.1744-6570.1984.tb00519.x Schmitt, N., & Noe, R. A. (1983). Demonstration of content validity: Assessment center example. Journal of Assessment Center Technology, 6, 5–11. Schmitt, N., Schneider, J. R., & Cohen, S. A. (1990). Factors affecting validity of a regionally administered 234

Thornton, G. C., III, Murphy, K. R., Everest, T. M., & Hoffman, C. C. (2000). Higher cost, lower validity, and higher utility: Comparing utilities of two tests that differ in validity, costs, and selectivity. International Journal of Assessment and Selection, 8, 61–75. doi:10.1111/1468-2389.00134 Thornton, G. C., III, & Rupp, D. E. (2006). Assessment centers in human resource management: Strategies for prediction, diagnosis, and development. Mahwah, NJ: LEA. Tillema, H. H. (1998). Assessment of potential from assessment centers to development centers. International Journal of Selection and Assessment, 6, 185–191. doi:10.1111/1468-2389.00088 Truxillo, L. M., Donahue, L. M., & Kuang, D. (2004). Work samples, performance tests, and competency testing. In J. C. Thomas & M. Hersen (Eds.), Comprehensive handbook of psychological assessment: Vol. 4. Industrial and organizational assessment (pp. 345–370). Hoboken, NJ: Wiley. Tziner, A., Meir, E. I., Dahan, M., & Birati, A. (1994). An investigation of the predictive validity and economic utility of the assessment center for the high-management level. Canadian Journal of Behavioural Science, 26, 228–245. doi:10.1037/ 0008-400X.26.2.228 Waldman, D. A., & Korbar, T. (2004). Student assessment center performance in the prediction of early career success. Academy of Management Learning & Education, 3, 151–167.

Assessment Centers

Warech, M. A., Smither, J. W., Reilly, R. R., Millsap, R. E., & Reilly, S. P. (1998). Self-monitoring and 360-degree ratings. The Leadership Quarterly, 9, 449–473. doi:10.1016/S1048-9843(98)90011-X

Woehr, D. J., & Huffcutt, A. I. (1994). Rater training for performance appraisal: A meta-analytic review. Journal of Occupational and Organizational Psychology, 67, 189–205.

Wernimont, P. F., & Campbell, J. P. (1968). Signs, samples, and criteria. Journal of Applied Psychology, 52, 372–376. doi:10.1037/h0026244

Wollowick, H. B., & McNamara, W. J. (1969). Relationship of the components of an assessment center to management success. Journal of Applied Psychology, 53, 348–352. doi:10.1037/h0028102 Zedeck, S. (1986). A process analysis of the assessment center method. In B. Staw & L. Cummings (Eds.), Research in organizational behavior (Vol. 8, pp. 259–296). Greenwich, CT: JAI Press.

Copyright American Psychological Association. Not for further distribution.

Woehr, D. J., & Arthur, W., Jr. (2003). The constructrelated validity of assessment center ratings: A review and meta-analysis of the role of methodological factors. Journal of Management, 29, 231–258. doi:10.1177/ 014920630302900206

235

CHAPTER 8

SITUATIONAL JUDGMENT TESTS: A CRITICAL REVIEW AND AGENDA FOR THE FUTURE

Copyright American Psychological Association. Not for further distribution.

Robert E. Ployhart and William I. MacKenzie Jr.

Situational judgment tests (SJTs) are measurement methods that present respondents with work-related situations and ask them how they would or should handle the situations. SJTs have become very popular among industrial and organizational (I/O) psychologists in the last 20 years. Much of their popularity has been attributed to the fact that SJTs exhibit moderately strong criterion-related validities but are more face valid and tend to exhibit smaller racial and sex subgroup differences. It is important to recognize that at least to date, they are generally considered to be measurement methods, which means they do not measure a homogeneous construct but rather can be developed to measure several different types of constructs (often simultaneously). In this sense, SJTs are very much multidimensional measurement methods. The purpose of this chapter is to critically review and evaluate the theoretical basis of SJTs and their use in practice. We briefly review the historical antecedents of SJTs, offer a selective review of key scholarly research, and discuss practical issues. One important element of our review is an examination of SJT validity in terms of criterionrelated, content, and construct validity. We conclude by emphasizing that SJT scholarship must move beyond criterion-related issues and metaanalytic methods if we are to understand this very useful, if not enigmatic, predictor.

HISTORICAL OVERVIEW AND SUMMARY OF SITUATIONAL JUDGMENT TESTS SJTs have increased in popularity and use over the past 2 decades. According to McDaniel, Morgeson, Finnegan, Campion, and Braverman (2001), the origin of “modern” SJTs may be traced back to the 1920s, but tests designed to assess situational judgment have been around for quite some time. Early examples of United States Civil Service exams included items designed to assess situational judgment and date as far back as 1873 (DuBois, 1970). These early exams included open-ended questions that presented applicants with a situation and asked them to respond with how they would react in the described situation. The George Washington Social Intelligence Test is considered one of the earliest known examples of a widely used SJT that provided subjects with closed-ended (multiple choice) response options (McDaniel et al., 2001; Moss, 1926). Subjects were given a situation and asked to choose a solution from a list of several possible alternatives. The use of SJTs continued through World War II, when Army psychologists administered SJTs to assess the judgment of soldiers (Northrop, 1989). The 1940s saw a surge in the use of SJTs to assess individual levels of supervisory potential (Cardall, 1942; File, 1945). SJTs were used in the late 1950s and through the 1960s to predict managerial potential

We thank Sheldon Zedeck for his helpful comments and suggestions on this chapter.

http://dx.doi.org/10.1037/12170-008 APA Handbook of Industrial and Organizational Psychology, Vol 2: Selecting and Developing Members for the Organization, edited by S. Zedeck Copyright © 2011 American Psychological Association. All rights reserved.

237

Copyright American Psychological Association. Not for further distribution.

Ployhart and MacKenzie

of employees for corporations (Bruce, 1965; Bruce & Learner, 1958; Greenberg, 1963; Kirkpatrick & Planty, 1960). Most of these early SJTs were criticized because of their high correlation with cognitive ability (e.g., File & Remmers, 1971; Millard, 1952; Taylor, 1949a; Taylor, 1949b; Thorndike, 1941; Thorndike & Stein, 1937), and they were viewed as being factorially complex (e.g., Northrop, 1989). The 1990s saw a reemergence of research on SJTs after Motowidlo, Dunnette, and Carter (1990) “reintroduced” SJTs as a low fidelity simulation. Research increased geometrically after publication of the Motowidlo et al. (1990) article, with most research focusing on two core issues. The first was establishment of the criterion-related validity of SJTs and their incremental validity over other commonly used predictors (e.g., cognitive ability; Clevenger, Pereira, Wiechmann, Schmitt, & Harvey, 2001). The second was examination of how differences in response instructions could influence SJT criterionrelated and construct validity (McDaniel & Nguyen, 2001). Only recently have there been theoretical attempts to understand what SJTs measure and why they may be related to job performance (Motowidlo, Hooper, & Jackson, 2006b). There have been several excellent review articles published on SJTs (Chan & Schmitt, 2005; Lievens, Peeters, & Schollaert, 2008; McDaniel & Nguyen, 2001; Schmitt & Chan, 2006), and an edited book tried to promote greater theoretical understanding (Weekley & Ployhart, 2006). SJTs are generally considered to be multidimensional measurement methods (Chan & Schmitt, 1997). In some ways, it is difficult to define or operationalize SJTs because they share many key elements with other popular predictor methods such as interviews, certain assessment center exercises (e.g., role plays), and simulations (Ployhart, Schneider, & Schmitt, 2006). SJTs may be administered in a variety of formats that include paper, video, orally, and the Internet/computer. Therefore, we first describe the typical SJT structure, then compare/contrast this structure with other predictors, and conclude with identifying the unique features of SJTs. The defining feature of an SJT is that it presents respondents with realistic work situations and then requires these respondents to pick the best/worst (or 238

most/least likely) behavioral response to address the situation. Thus, SJTs are by definition context-bound, although these contexts may be described fairly generically to apply to different firms. The following is a sample SJT item that could be used to hire entry level managers: Imagine you are in charge of a project team with five members from different functional areas. The team members do not get along and several do not see the value in this project. The bickering and fighting are starting to jeopardize the quality of the project. a. Inform the team members that if they don’t stop fighting, they will be kicked off the team. b. Try to instill greater commitment to the project by emphasizing the team’s goals. c. Go to each of the team member’s supervisors to inform him/her of the poor behavior. d. Wait just a bit longer and see if the fighting stops. Which is the most effective option? _____ Which is the least effective option? _____ Notice that in this example there are multiple goals and competing goals that must be balanced. Several potentially plausible response options are provided, and these options reflect behaviors that may be effective or ineffective. The respondent must exercise his or her judgment and choose what he or she believes is the appropriate response. At this point it may be apparent that SJTs have many similarities to other commonly used predictor methods. In particular, SJTs are very similar to situational interviews, with the main difference being that SJTs provide a more structured response format (i.e., multiple choice). SJTs can be similar to computer-based simulations. For example, the sample item shown here could be presented in a videobased format so that respondents watch a video of the team’s interactions and then must indicate which options are most/least effective. Therefore,

Copyright American Psychological Association. Not for further distribution.

Situational Judgment Tests

the SJT would be less interactive than the typical computer-based simulation. If the instructions ask about what one should do in a situation, the SJT functions very much like a maximum-performance test because it focuses on ability and knowledge; if the instructions ask about what one would do in a situation, the SJT functions more like a typical performance test (McDaniel, Hartman, Whetzel, & Grubb, 2007). Thus, because the boundaries or defining features of SJTs are fairly amorphous, it has been somewhat difficult to define exactly what makes a predictor an SJT. This in turn has sometimes created confusion over fundamental issues with SJTs and how one goes about understanding them. For example, because SJTs are measurement methods, different SJT instructions or “structure” may affect SJT criterion-related validity. Likewise, SJTs with similar structure may be designed to measure different constructs, which may again affect SJT criterion-related validity. Therefore, when combining (e.g., meta-analysis) or contrasting SJTs, it is necessary to ensure that they are similar in terms of structure and constructs assessed (see Arthur & Villado, 2008, for a clear treatment of this issue). These are more than purely academic issues because being able to define an SJT has consequences for how validity should be established and how the SJT should be used in practice. SJT DEVELOPMENT AND SCORING Most of the variations around SJT measurement and scoring options are reviewed in Weekley, Ployhart, and Holtz (2006). Here we review the most common way to develop an SJT that involves a three-part process of situation generation, response option generation, and scoring.

Step 1: Situation Generation The first step uses job incumbents (or supervisors) to identify critical incidents at work that are challenging, difficult, important, but reasonably common. A long list of potential situations is created to sample from the work environment domain. SJT item stems are most commonly generated using one of two sources: critical incidents or understanding of the job. The use of critical incidents is the most widely used of the two

sources, even though it can be fairly expensive and time consuming. In general, job experts are asked to think of the context of a work situation, the behaviors that occurred in it, and the antecedents that lead up to it (e.g., Weekley & Jones, 1999). For example, a retail employee may discuss from personal experience a situation in which a customer wanted to return merchandise without a receipt (the antecedent), which created a situation (context) in which the employee was forced to choose between promoting customer satisfaction and following store policy stating no returns without receipts (behaviors). In addition to job experts, critical incidents may be generated using archival data sources. The alternative approach to generating item stems is to use a theoretical basis. The first step to this approach is identifying theories related to the job of interest. Once the pertinent job-related theories have been identified, the researcher then generates items that capture the required underlying job-related constructs. For example, one might review the extensive psychological and marketing literatures on customer service provision to identify common types of service encounters (e.g., relational, long term, brief encounters) and challenges (e.g., balancing customer, peer, and supervisor demands) to comprise the basis of SJT stems. Lievens and Sackett (2007) discussed some variations on this approach. Although the various approaches for situation generation have certainly been useful, future research needs to identify how to best write or structure SJT items.

Step 2: Response Option Generation The second step in SJT development involves another group of job experts identifying behavioral response options for each situation. Note that both effective and ineffective options are generated, and they are behavioral in nature. Use of both effective and ineffective options is done to increase variability in the item responses and help identify “good judgment.” Response options, like item stems, may be generated by subject matter experts (SMEs) or SJT developers. As noted by Weekley et al. (2006), there has been little research exploring the differences between SJT response options generated by SMEs and those generated by test developers. However, experience has suggested that response options generated by job experts 239

Ployhart and MacKenzie

are likely to exceed those generated by developers in terms of both quality and quantity. A long list of possible response options for each situation will be developed and then reduced to approximately four to six options for each situation.

Copyright American Psychological Association. Not for further distribution.

Step 3: Scoring Finally, a third group of SMEs (usually supervisors) will indicate which options are effective/ineffective and hence create the scoring key. There are several different methods used to score SJTs, but in practice most methods follow one of two strategies: having respondents make a forced choice among response options or having respondents rate the effectiveness of each response option along a continuum. Common applications of the forced-choice method involve having respondents choose the one best response (Hunter, 2003; O’Connell, Doverspike, Norris-Watts, & Hattrup, 2001; Sinar, Paquet, & Scott, 2002) or identifying both the best and worst (most/least effective) response to the situation (Motowidlo et al., 1990). Several methods are used to score both of these approaches. When respondents are asked to choose only the one best response, they may receive 1 point for each situation in which they identify the one best response and 0 points for incorrect responses. Because not all incorrect response options are equally wrong, another option is to rate the effectiveness of specific response options, which in essence allows for partial credit when respondents select a response option that falls between the best and worst option. The partial credit scoring method described by Motowidlo et al. (1990) has been particularly influential. Their approach asks respondents to identify both the best and worst options, so each item receives a score ranging from −2 to +2. Correctly identifying the best and worst option would each merit receiving 1 point, resulting in a score on the SJT item of +2. Identifying the best solution as the worst option, or vice versa, would result in losing 1 point, and incorrectly identifying both the best as the worst and the worst as the best would result in an SJT item score of −2. Incorrectly identifying any other response options (distracters) as the best or worst option would result in 0 points. This leads to respondents receiving a score of 1 when they correctly identify either the best or worst response option but 240

incorrectly choosing a distracter as the correct response option for the other. A final method for forced-choice scoring is to have respondents rank order all response options and compare their rank with those of SMEs using Spearman’s rank-order correlations. In a study conducted by Weekley, Harding, Creglow, and Ployhart (2004), respondents were asked to rank order all response options from best to worst. This method resulted in a small but significant improvement in validity over pick the best and pick the best/worst SJTs in two of the four samples. The second strategy for scoring SJTs is to have respondents rate the effectiveness of each response option along a continuum using a Likert-type scale. For example, the options from the sample SJT item presented previously would appear as shown in Exhibit 8.1. In this example, respondents would evaluate the effectiveness of each option, so the number of scores would be the number of options times the number of items. To score rating responses, job expert ratings are compared with respondent ratings. Common methods of calculating respondent scores for this strategy include assigning values depending on the percentage of agreement (Chan & Schmitt, 2002), taking the inverse of the absolute value of the difference between SME and respondent ratings (Sacco, Scheu, Ryan, & Schmitt, 2000; Sacco, Schmidt, & Rogg, 2000), and summing the squared deviations of respondent and SME ratings (Wagner, 1987). When compared with forced-choice strategies, there are some distinct advantages to having respondents rate each response option (McDaniel & Nguyen, 2001). Most notably, there is the potential for validity and reliability improvements because rating each response option provides more available items to score than with a forced-choice method. When it comes to scoring SJTs, correct/incorrect responses are occasionally identified using empirical methods (e.g., correlating each response with the criterion) but more frequently on the basis of the consensus by a group of job experts—frequently supervisors. For those items for which consensus cannot be reached, the item is usually eliminated and will not be included in operational use. Consequently, scoring keys will often be organization specific. Clearly the organization’s norms, cli-

Situational Judgment Tests

Exhibit 8.1 Effectiveness Scale for Response Items in Sample Situational Judgment Test

Copyright American Psychological Association. Not for further distribution.

Ineffective

Effective

Response option

1

2

3

4

5

a. Inform the team members that if they don’t stop fighting, they will be kicked off the team. b. Try to instill greater commitment to the project by emphasizing the team’s goals. c. Go to each of the team member’s supervisors to inform him/her of the poor behavior. d. Wait just a bit longer and see if the fighting stops.

1 1 1 1

2 2 2 2

3 3 3 3

4 4 4 4

5 5 5 5

mate, HR practices, and related contextual factors will affect the nature of the scoring key. Thus, it is quite common for a set of SJT items to be used in multiple organizations but the scoring of those items to be done separately within each firm. Finally, please realize that at each stage, editing of the situations will be necessary, and the items should be written so that the language and jargon are not so specific that the targeted population will be unable to understand them. Additionally, item stems should take into account the content of the job, whereas response options should consider the desired construct to be measured.

Other SJT Development Issues When developing the situations and response options, it is critical to consider their complexity and fidelity. Stem complexity refers to the level of detail that may be used in the item stem and response options. Simple stems are short and use simple language, for example, “Imagine you are trying to satisfy a difficult customer.” Complex stems are longer, use more complex language, and will frequently require greater attentional resources. For example, A customer wishes to return some expensive clothing but does not have a receipt. Store policy is to never accept returns without a receipt. You state this to the customer, who quickly becomes so irate that he or she starts to make other customers leave the store. Unfortunately, the store manager is out to lunch and you have to do something fast.

Highly complex item stems may sometimes include subsituations (e.g., branching). The key feature of a branching SJT is that the choices you make “early” in the SJT influence the situations you are presented with as you proceed through the test. SJTs with branching questions, therefore, have multiple situations, but respondents will only complete a subset of the situations. As an extreme example of branching, the first author developed an SJT to assess leadership skills in U.S. Army captains. The SJT first presented all captains with a situation in which a town that should have been empty was in fact inhabited by civilians and insurgents. If captains charged into the town, they found themselves outnumbered and the situation becoming critical. On the other hand, if captains maneuvered around the town, they found themselves unable to achieve their objectives. Thus, the choices that captains made early in the assessment influenced the types of situations they subsequently entered and needed to respond to. The point is that just as in life, the judgments and decisions people make have implications for their future circumstances. Research has produced conflicting results on the ideal level of stem complexity (McDaniel et al., 2001; Reynolds, Winter, & Scott, 1999). However, it has been shown that complex item stems that require a higher reading level to comprehend tend to lead to racial subgroup differences in SJT performance (Sacco, Scheu, et al., 2000; Sacco, Schmidt, & Rogg, 2000). The level of realism in the item, or its fidelity, must also be considered when developing item stems and response options. Whereas Motowidlo and colleagues (1990) referred to the SJT as a low-fidelity 241

Copyright American Psychological Association. Not for further distribution.

Ployhart and MacKenzie

simulation, several more recent studies have used alternative formats of SJTs to increase their fidelity. Using richer formats, such as video, retains the basic attributes of an SJT (i.e., providing the subject with a situation and alternative response options) but provides a more vivid account of the situation (particularly nonverbal information). These multimedia situational judgment tests (MMSJTs) often use video clips as a means of presenting the SJT item stems. After watching each video clip (i.e., situation), respondents are then presented with multiple choice options as previously described. By administering the situations in a video-based format, the physical and potentially psychological fidelity of the SJT is enhanced (see Ployhart et al., 2006). Furthermore, MMSJTs typically exhibit higher levels of face validity (Chan & Schmitt, 1997; Richman-Hirsch, OlsonBuchanan, & Drasgow, 2000) and slightly reduced subgroup differences (Chan & Schmitt, 1997; OlsonBuchanan et al., 1998). It is important for test developers to also give consideration to SJT response instructions. Response instructions have been shown to have a significant influence on SJT construct validity (McDaniel, Whetzel, Hartman, Nguyen, & Grubb, 2006; McDaniel et al., 2007). SJTs using “would do” instructions tend to correlate with different constructs than do SJTs using “should do” instructions. Specifically, SJTs that ask respondents which option they “would do” or “would most likely perform” tend to correlate more strongly with personality constructs, whereas SJTs that ask respondents what they “should do” or is the “best choice” tend to correlate more strongly with cognitive ability and knowledge constructs. As such, “would do” versions of SJTs are considered to be measures of behavioral tendencies or typical performance, and “should do” versions of SJTs are considered to be measures of cognitive ability, knowledge, or maximum performance (McDaniel et al., 2007). In addition to moderating construct validity, prior research has examined the effects of response instructions on the criterion-related validity of SJTs (e.g., Ployhart & Ehrhart, 2003). However, a recent meta-analysis by McDaniel et al. (2007) suggested criterion-related validity may be unaffected by SJT response instructions. Using meta-analytic tech242

niques to assess SJT construct and criterion-related validity poses certain challenges, chief among them is the variation of SJT content. In the few studies that held SJT content constant for both behavioral tendency and knowledge instructions, McDaniel et al. found knowledge SJT instructions exhibited higher levels of criterion-related validity than did behavioral tendency SJT instructions. In part this difference might be explained by differences in criterion construct, but unfortunately the meta-analysis did not code for or examine this question. Given the literature included in their analysis, our estimate is that the criteria were primarily ratings by supervisors or other individuals. As noted by McDaniel and colleagues, future research is needed to truly understand the moderating effect of response instructions on SJT criterion-related validity. Although much research has been done on creating response options that target specific constructs, there has been only limited success (Motowidlo, Diesch, & Jackson, 2003; Ployhart & Ryan, 2000; Trippe & Foti, 2003). For example, Ployhart and Ryan (2000) tried to create a construct-oriented SJT that targeted traits most likely to be important in customer service (i.e., neuroticism, agreeableness, and conscientiousness). Further, and in contrast to the usual approach for developing SJT items, they tried to ensure that the options for each item represented a continuum of that trait’s expression in that situation (the scaling of these options was derived from extensive pilot testing). Shown below is a sample item for agreeableness: You’re working at the cash register on a very busy shift. There is a long line of customers waiting to check out, and you can see several individuals are getting impatient. What do you do? (a) Tell the waiting customers that you appreciate their patience. (b) Regularly look at those waiting so that they know you are aware of them. (c) Once in a while look at those waiting so that they know you are aware of them. (d) Focus on your job as a cashier and don’t look at those waiting.

Copyright American Psychological Association. Not for further distribution.

Situational Judgment Tests

Despite the effort and care that went into developing this SJT, it still failed to reach high levels of convergent validity with the three personality traits, although it did correlate more strongly with the intended traits than a more “traditional” SJT. On the other hand, Motowidlo, Hooper, and Jackson (2006a) found that by building response options that expressed varying levels of specific traits, having individuals rate the effectiveness of each response option, and estimating the magnitude of a trait’s influence on the effectiveness rating, they were able to measure the implicit trait policy for the traits of agreeableness and extraversion. Additionally, they were able to pr