Winning with Words: The Origins and Impact of Political Framing [1 ed.] 0415997933, 9780415997935

Citation preview

Winning with Words

Today’s politicians and political groups devote great attention and care to how their messages are conveyed. From policy debates in Congress to advertising on the campaign trail, they carefully choose which issues to emphasize and how to discuss them in the hope of affecting the opinions and evaluations of their target audience. This groundbreaking text brings together prominent scholars from political science, communication, and psychology in a tightly focused analysis of both the origins and the real-world impact of framing. Across the chapters, the authors discuss a broad range of contemporary issues, from taxes and health care to abortion, the death penalty, and the teaching of evolution. The chapters also illustrate the wide-ranging relevance of framing for many different contexts in American politics, including public opinion, the news media, election campaigns, parties, interest groups, Congress, the presidency, and the judiciary. Brian F. Schaffner is Associate Professor of Political Science at the University of Massachusetts, Amherst and editor of the journal, Congress & The Presidency. He is also co-author with John Bibby of Parties, Politics, and Elections in America. Patrick J. Sellers is Professor of Political Science at Davidson College, North Carolina. He is the author of Cycles of Spin: Strategic Communication in the U.S. Congress.

Winning with Words The Origins and Impact of Political Framing

Edited by Brian F. Schaffner and Patrick J. Sellers

First published 2010 by Routledge 270 Madison Ave, New York, NY 10016 Simultaneously published in the UK by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business

This edition published in the Taylor & Francis e-Library, 2009. To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk. © 2010 Routledge, Taylor and Francis All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging in Publication Data Winning with words : the origins and impact of political framing / [edited by] Brian Schaffner, Patrick J. Sellers. p. cm. [etc.] 1. Communication in politics–United States. 2. Rhetoric–Political aspects–United States. I. Schaffner, Brian F. II. Sellers, Patrick J. JA85.2.U6W56 2009 320.97301'4–dc22 2009011323 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

ISBN 0-203-88311-X Master e-book ISBN

ISBN 10: 0–415–99793–3 (hbk) ISBN 10: 0–415–99794–1 (pbk) ISBN 10: 0–203–88311–X (ebk) ISBN 13: 978–0–415–99793–5 (hbk) ISBN 13: 978–0–415–99794–2 (pbk) ISBN 13: 978–0–203–88311–2 (ebk)

Contents

List of Figures List of Tables Preface Acknowledgments

vii viii ix xi

1 Introduction BRIAN F. SCHAFFNER AND PATRICK J. SELLERS

1

PART I

Origins 2 Framing and Value Recruitment in the Debate Over Teaching Evolution

9

11

THOMAS E. NELSON, DANA E. WITTMER, AND ALLYSON F. SHORTLE

3 Partisan Framing in Legislative Debates

41

DOUGLAS B. HARRIS

4 Building a Framing Campaign: Interest Groups and the Debate on Partial-birth Abortion

60

JESSICA C. GERRITY

5 Mobilizing to Frame Election Campaigns

78

TAYLOR ANSLEY AND PATRICK J. SELLERS

PART II

Impact 6 Competing Frames in a Political Campaign JAMES N. DRUCKMAN

99 101

vi Contents

7 Taxing Death or Estates? When Frames Influence Citizens’ Issue Beliefs

121

BRIAN F. SCHAFFNER AND MARY LAYTON ATKINSON

8 Great Communicators? The Influence of Presidential and Congressional Issue Framing on Party Identification

136

MICHAEL W. WAGNER

9 The Decline of the Death Penalty: How Media Framing Changed Capital Punishment in America

159

FRANK R. BAUMGARTNER, SUZANNA LINN, AND AMBER E. BOYDSTUN

10 Framing Research: The Next Steps

185

SHANTO IYENGAR

Notes on Contributors Index

192 195

Figures

2.1 6.1 7.1 8.1 8.2 9.1 9.2 9.3 9.4 9.5 9.6

9.7 9.8 10.1

Frequencies of Values in Content Analysis Likelihood of Casino Support Use of “Estate Tax” and “Death Tax” Labels in Congressional Record, 1989–2004 The Variability of Presidential and Congressional Framing on Tax Issues, 1975–2000 The Variability of Presidential and Congressional Framing on Abortion, 1975–2000 The Number of Stories on Capital Punishment in the New York Times Index, 1960–2005 The Number of Stories on Each Topic, or Dimension, of Attention in the New York Times Index, 1960–2005 The Topic Determines the Tone Tone and Mention of Victim and Defendant The Number of Stories Mentioning the Victim and the Defendant, 1960–2005 The Number of Stories Mentioning Innocence, Evidence, Flaws in the System, or Defendant Characteristics, 1960–2005 The Net Tone of New York Times Index Coverage, 1960–2005 Capital Sentences, Executions, and the Death Row Population Example of Similarity Manipulation

18 113 124 143 145 163 164 166 167 168

169 169 173 189

Tables

2.1 2.2 2.3 5.1 5.2 5.3 6.1 6.2 6.3 6.4 6.5 6.A1 6.B1 7.1 7.2 8.1 8.2 8.3 8.4 8.A1 8.B1 9.1

Tactics from Content Analysis Experimental Treatments Experimental Results Campaign Website Tools Models of Campaign Tools Mentions of Issues in the 2006 Ohio Senate Race Casino Frames Pre-test Results Experimental Casino Frames Demographic and Political Profile of Sample Experimental Conditions Experimental Condition Means Experimental Framing Effects on Support for Casino Proposal Effect of “Death Tax” Frame on Support for Inheritance Tax Effect of “Death Tax” Frame on Beliefs About Who Pays Tax Distribution of Frame Sources for Taxes, 1975–2000 Distribution of Frame Sources for Abortion, 1975–2000 Effects of Presidential and Congressional Framing of Taxes on Party Identification Effects of Presidential and Congressional Framing of Abortion on Party Identification Tax Frames and Changes in Party Identification, 1975–2000 Abortion Frames and Changes in Party Identification, 1975–2000 Explaining the Number of Annual Death Sentences

24 31 33 87 90 92 106 107 108 110 111 115 116 129 130 142 144 149 151 153 154 176

Preface

The idea for this book originated during a series of discussions (and sometimes arguments) that we had with each other over several years. During this time, Sellers was working on his book, Cycles of Spin, examining how parties in Congress attempt to frame policy debates and win news coverage of their frames. Schaffner was conducting research on public opinion, particularly examining how different messages influenced the public in different ways. When discussing these research projects, our conversations often focused on framing and our frustration with the concept’s amorphous nature, despite its ubiquity in political science research. Framing plays a central role in studies of Congress, interest groups, the presidency, the news media, campaigns, and public opinion; yet, little research integrates framing across these different contexts. In addition, the concept is often defined vaguely or in vastly different ways from study to study. To address these frustrations, we held a conference where we brought in scholars studying framing in a variety of contexts and from widely divergent approaches. We hoped to foster a dialogue among scholars who were all studying framing, but in such different ways that they might not normally talk with each other. Our two main goals were to (1) increase the linkages between elite-based and mass-based research on framing, and (2) bring some clarity to the way the field defines the concept of framing. Ultimately, we believe that this conference was quite successful in achieving the first goal, and this volume is a testament to that success. However, there was much less agreement on how to define framing.

What is Framing? During a particularly lively debate at the conference, participants diverged widely in defining the nature and scope of framing. Robert Entman, who has written widely about framing from a mass communication perspective, argued that framing included any effort to influence public opinion through the formulation of messages. Shanto Iyengar, whose work on framing is just as expansive and influential as Entman’s, offered a narrower definition, limiting framing to instances of presenting the same information in different ways. This definition is reflected in Iyengar’s conclusion to this volume, where he argues for “a return

x Preface to equivalency framing.” Nevertheless, Iyengar acknowledges in the conclusion that the field has largely accepted the broader view of framing: the concept of framing is applied quite differently across psychology, sociology, political science, and communications. In operational terms, however, there is less diversity; researchers have converged on a relatively loose definition of framing as information that conveys differing perspectives on some event or issue. The studies in this volume mostly fall under this more inclusive definition of framing. However, the question of how best to define framing remains an open one, a point that receives close attention in the introduction and conclusion to this volume. We encourage readers to consider this debate as they assess the different studies in this volume.

This Volume in the Classroom We believe that this volume not only makes an important contribution to the research on framing but also provides a useful tool for bringing this research into the classroom. The entire text should be clearly applicable in advanced courses on political communication, public opinion, and media and politics (among others). At the same time, this text is ideal for lower-division undergraduate courses such as introductory classes on American Government or Public Policy. The book includes eight chapters of original research on different aspects of the framing process. Each is relevant to different parts of the American political system and tackles different substantive policy issues. In an American Government course, one might assign the chapters in the following way: • • • • • • • •

State and local government: Chapter 2 Congress: Chapters 3 and/or 4 Presidency: Chapter 8 Judiciary: Chapter 9 Interest groups: Chapter 4 Parties: Chapters 3 and/or 8 Campaigns and elections: Chapters 5 and/or 6 Public opinion: Chapters 6 and/or 7

For a course on American Public Policy, the chapters may be broken down as follows: • • • • • •

Education: Chapter 2 Health care: Chapter 3 Abortion: Chapters 4 and 8 Casino regulation: Chapter 6 Tax policy: Chapters 7 and 8 Crime: Chapter 9

Acknowledgments

This volume grew out of a conference that we hosted at American University in June, 2007. Without that conference, the book would never have happened. Therefore, we owe our first and most significant debt of gratitude to Jim Thurber. As Founder and Director of the Center for Congressional and Presidential Studies, Jim has been an institution at American University and in the Washington political sphere for decades. It is hard to imagine anyone giving more to his colleagues while asking less, and Jim’s sponsorship of the conference on framing research was yet another example of this spirit. He provided us with the resources needed to sponsor the conference and the advice necessary to carry it out successfully, while also allowing us the freedom to assemble a program that reflected our interests. We can never repay Jim for his assistance with this project, but we hope that this book will serve as a worthy example of the contributions he has made to the discipline over the years. This edited volume is our first, and we were fortunate to have a wonderful group of contributors who provided excellent material and responded to our suggestions and emails in a timely and conscientious way. While all deserve our gratitude, we especially thank Jamie Druckman. We came to him first with our idea for a conference on framing, and he provided incredible support and advice for moving the project forward. Of course, anyone who knows Jamie has a similar story to tell—he is a wonderful contributor to the discipline who is always generous with his time and advice. We also thank others who participated in the conference, including Bob Entman, Candy Nelson, Tim Groeling, Kim Gross, Christine Degregorio, Joumana Moukarim, and Adam Simon. We are also grateful to the entire editorial team at Routledge, but particularly Michael Kerns, who is one of the sharpest and most thoughtful editors in this business. Finally, we would like to thank our families for their patience and support.

1

Introduction Brian F. Schaffner and Patrick J. Sellers

On his fourth day in office, President Barack Obama delivered his first radio address to the nation. His remarks focused on the American Recovery and Reinvestment Act, legislation that Obama hoped Congress would pass quickly in order to stem the economic downturn. Despite the fact that the news media and many politicians commonly referred to the legislation as a “stimulus package,” Obama did not use the word stimulus once in his 806-word address. Democratic pollsters had indicated that the public was more willing to support an economic recovery package than an economic stimulus. Their research also indicated that the public was more supportive of the legislation when told about how the money would be spent than about the total amount of spending (Brown 2009). Not surprisingly, Obama’s speech framed the legislation in a favorable way by focusing almost entirely on the programs that the spending would support, without ever mentioning the total size of the bill. This story illustrates how today’s politicians devote great attention and care to framing their messages to the public and each other. From policy debates in Washington to advertising on the campaign trail, our political leaders carefully choose which issues to emphasize, and how to discuss those issues. These crafted messages can significantly affect the opinions and evaluations of target audiences (Iyengar 1991; Nelson et al. 1997; Jacobs and Shapiro 2000; Entman 2004; Druckman 2001). Yet, important questions remain about politicians’ attempts to win political battles with carefully worded messages. Exactly what strategies guide politicians’ creation of frames? Do these strategies differ across issues and contexts? What makes a message more persuasive in the real world? Are some voters more susceptible to persuasion than others? Does this impact extend beyond the public’s short-term attitudes? These questions are fundamental for American politics. Public discussions about policy issues inherently involve framing, as politicians intentionally and necessarily emphasize certain aspects of those issues and deemphasize others. If politicians can effectively use framing to shape public opinion, it becomes unclear whether that opinion even exists absent framing. In such a scenario, voters may possess weak or even non-existent attitudes about policy issues and their elected leaders (Asher 2007). Politicians could conceivably define and manipulate public opinion and the political process to further their own interests (Lippman 1922;

2 Introduction Bartels 2003). This worrisome possibility makes it important to understand politicians’ framing strategies and the limits to their effectiveness, that is, the conditions under which the public holds opinions in resistance to framing. In this volume, prominent scholars from political science, communication, and psychology come together to explore the dynamics of framing. Across the chapters, the authors discuss political messages about a range of contemporary issues, from taxes to abortion and the death penalty. The chapters also illustrate the wide-ranging relevance of framing for many different contexts in American politics, including the news media, election campaigns, interest groups, Congress, and public opinion. To explore framing across these contexts, the authors employ diverse methodologies, from surveys and laboratory experimentation to elite interviews and content and archival analysis. When combined, the diverse issues, contexts, and methods go beyond the existing studies of framing and paint a broad picture of many ways in which framing influences American politics. Such a unique picture can help students of contemporary politics better understand how competing politicians carefully choose their words in order to win. This introductory chapter places the chapters in the broader context of the expansive literature concerning framing. We discuss two dimensions of this literature and explain how subsequent chapters in this volume fit into these dimensions. We conclude the chapter by emphasizing the importance of considering both framing successes and failures.

Framing Debates The term “framing” is one of the most widely used concepts in political science, communication studies, psychology, and sociology.1 Two dimensions of the literature figure prominently in this volume. The first dimension concerns competing conceptions of framing. The broadest conception, what we might call “emphasis framing,” describes a process in which competing frames emphasize different messages and arguments in a policy debate (Nelson et al. 1997; Druckman 2001). Communication scholar Robert Entman uses these types of frames to explain how political and media elites interact to shape US foreign policy and public opinion (Entman 2004). In a similar manner, cognitive linguist George Lakoff describes how Democrats and Republicans employ very different arguments and evidence in contemporary policy debates (Lakoff 2004). Republicans often use a “strict father” model to frame issues, with emphasis on protecting the family in a dangerous, difficult world and teaching children right from wrong. In contrast, the “nurturant parent” model, employed by Democrats, emphasizes making the world a better place in order to nurture the needs of children and others. On the issue of poverty, for example, the models lead Republicans to emphasize self-reliance and personal responsibility, and the Democrats to focus on society’s responsibility to help those in need. Many of this book’s chapters also employ similar types of “emphasis framing.” The Nelson et al. chapter, for example, examines the policy debate over the

Introduction 3 teaching of intelligent design. In this debate, supporters of intelligent design framed the scientific process as based on a “marketplace of ideas.” Their opponents, in contrast, labeled intelligent design as a “Trojan Horse” designed to force religious beliefs on school children. Other chapters focus on emphasis framing as well. Gerrity describes how interest groups attempted to emphasize different aspects of the partial-birth abortion issue; Druckman demonstrates how focusing on different arguments for or against a casino influenced support for the proposal; and Baumgartner et al. argue that a new emphasis on the “innocence frame” led to a significant decline in the application of the death penalty. An alternative conception of framing focuses less broadly on competing arguments, and more narrowly on the ways in which a single argument is presented. Early and prominent examples of this “equivalence framing” (Druckman 2001) emerged from the work of Tversky and Kahneman (1974), who demonstrated how presenting the same policy proposal as a gain or loss shaped individuals’ evaluations of that proposal. In this volume, the chapter by Schaffner and Atkinson uncovers a similar effect: the authors illustrate how the competing labels of “death tax” and “estate tax” shape citizens’ perceptions and evaluations of inheritance tax reform. In this book’s conclusion, Iyengar argues for a focus on this type of equivalence framing, suggesting that it is more precisely defined and offers greater potential for distinguishing among competing influences on public opinion. These competing conceptions of framing matter because they suggest different ways that political elites may influence citizens’ perceptions and evaluations. This influence forms the basis of a second dimension of framing that emerges from this volume: the distinction between the origin and impact of frames.

Origins Political elites may create frames for different reasons and in different contexts. Legislative leaders may use arguments or amendments to frame a policy debate and therefore shape legislators’ votes on a policy proposal (Riker 1986). Policy making more broadly has grown increasingly reliant on public communication in recent decades (Kernell 1997; Jacobs and Shapiro 2000; Sellers 2010), encouraging political elites to promote favorable frames to the news media and the public, again in hopes of shaping opinions and policy outcomes.2 The chapters in the first half of this volume help unpack the decision making process behind elites’ choice of frames for political communication. Nelson and his co-authors examine how elites use values as tools of persuasion in political debates. “Value recruitment” describes how communicators adopt social and political values to make persuasive policy arguments and influence public opinion. The recurring debate over the teaching of evolution in public schools provides a vivid and instructive case study of value recruitment in action, as captured through in-depth interviews and content analysis of news coverage about the debate.

4 Introduction In the next chapter, Harris traces the recent history of how congressional parties have framed policy debates in the US House of Representatives. Drawing extensively from House leadership archives (from Speakers Tip O’Neill to Newt Gingrich), Harris demonstrates how congressional leaders consciously use framing strategies to create party messages. These strategies have evolved across congresses, parties, and leaders. In Chapter 4, Gerrity explores how interest groups have developed frames on the issue of abortion. Building on in-depth analysis of interest groups’ public documents and interviews with the group leaders, Gerrity outlines how the groups’ goals varied in the abortion debate and how the groups created different frames to realize their particular goals. The chapter by Ansley and Sellers shifts the focus to recent election campaigns and a new way to create frames. These authors argue that candidates can more effectively frame electoral and policy choices by changing from the traditional top-down management structure to a more decentralized organization. This alternative approach empowers volunteers to contribute to the campaign’s message and help promote that message in new and different ways. Those volunteers and the public subsequently find that message more persuasive and credible. Ansley and Sellers support these arguments with extensive content analysis of campaign websites, advertisements, and news coverage of the 2006 Senate elections.

Impact After creating frames in these various ways, the political elites expect these frames to shape the perceptions and evaluations of target audiences. The second half of this volume examines this impact, that is, how frames influence individual citizens, overall public opinion, and ultimately policy making. The context and substantive content of this impact vary widely across the chapters. In Chapter 6 Druckman asks how competition among frames affects elites’ efforts to shape public opinion. He first defines the range of competitive contexts that might surround a policy debate, which political scientists have largely ignored. He then explains how audiences, messages, and competitive environments interact to shape the impact of framing. As noted above, Schaffner and Atkinson in Chapter 7 focus on a specific issue common to real-world policy debates: reforming the federal inheritance tax. Conservatives and Republicans labor mightily to frame this debate over the “death tax.” Schaffner and Atkinson illustrate how the parties’ actual frames in this debate shaped not only citizens’ preferred positions but also their knowledge of the tax and how widely it applied. Wagner’s chapter considers how framing may have a more lasting impact on the public’s attitudes. He captures how US presidents talked about two issues, taxes and abortion, over a 25-year period (1975–2000). Wagner then shows how the presidents’ varying frames on these issues helped shape the public’s party identification over the same period.

Introduction 5 In the final empirical chapter, Baumgartner et al. illustrate how framing may go beyond public opinion to influence policy making directly. In the decadeslong debate over the death penalty, the authors document how the rise of the powerful new “innocence” frame ended the long-term expansion of the death penalty in the United States. Since the rise of the “innocence” frame, the number of death sentences has actually decreased by over half.

The News Media When discussing the origins and impact of framing, it is important to note the news media’s role in mediating the relationship between political elites and the public. Journalists may attempt to provide coverage that mirrors actual events, including politicians’ frames (Gans 2004; Graber 2006). But, the public may not receive competing arguments from all sides, particularly if the news media merely “index” the degree of disagreement on an issue among political elites (Bennett 1990). In addition, market pressures may push the news media toward presenting their coverage in certain ways. For example, competitive concern over ratings often pushes television outlets to personalize their coverage of public policy debates, using more episodic than thematic framing (Iyengar 1991). While not a central focus of this volume, the news media’s mediating role emerges in several chapters. Political elites often construct frames in anticipation of how the news media will cover them. The Harris chapter explains how this anticipation shaped the public relations decisions of congressional leaders. The Gerrity chapter does the same for interest groups involved in the partial-birth abortion debate. Baumgartner et al. track coverage of the death penalty in the New York Times since 1960 and find that, as the Times paid more attention to the “innocence frame,” their coverage came to present an increasingly negative view of the death penalty.

Framing Failures If political elites can create frames that effectively shape public opinion, the normative implications are indeed worrisome, as outlined at the start of this chapter. The elites could conceivably shape public opinion to further their own interests. But, this volume’s chapters and authors offer a hopeful response to this concern: framing does not always work. In the first half of this volume, the authors document how political elites in different debates and contexts considered a wide range of frames. The elites did not actually use all these frames, and not all of the frames used actually proved effective. For example, Gerrity describes how NARAL lost the debate over partial-birth abortion because it was unable to find a frame that worked for its side before opinion turned against it. Ultimately, the analyses in the book’s first half accurately portray the difficulty of finding the best message, as well as politicians’ actual ability (or inability!) to choose that message. The chapters in

6 Introduction this section demonstrate that politicians are not always the confident manipulators of public opinion that we often perceive them to be. In the volume’s second half, the chapter authors document the limitations of framing by demonstrating the contingent nature of framing effects. Druckman’s chapter shows that not all frames are inherently convincing and that even strong frames become less convincing when they are forced to compete with other strong frames. Several chapters also demonstrate that the power of a frame is limited by the individual or side who is promoting it. The overall impact of a frame may further depend on how frequently it is repeated and which medium is used to present it. These dynamics make it difficult to conclude that all frames are effective. The failures and qualifications of framing significantly limit the negative effects of framing on our democratic system. Framing is not as universally effective as suggested by the critics mentioned at the start of this chapter. Political elites can find it difficult to create and distribute the frames necessary for extended attitude change among the public. The contingencies of framing make it harder for manipulation to occur. Thus, when reading the chapters that follow, it is important to pay heed not only to the influence of framing but also to its limitations.

Notes 1. For more detailed reviews of framing, see Entman (1993), Pan and Kosicki (2001), Iyengar and McGrady (2007), and the Iyengar chapter in this volume. 2. These promotional efforts are often categorized as agenda setting (Kingdon 2003; Cobb and Elder 1983; Schattschneider 1960), but the actual process closely resembles the emphasis framing described above. For example, participants in an experiment about government spending may receive competing frames emphasizing either specific government programs or overall levels of spending. By presenting (i.e., emphasizing) one frame or the other, the experiment intends to assess whether the competing frames shape participants’ opinions about the desirability of government spending. The content of this hypothetical experiment finds many parallels in national partisan debates about the same topic. Democrats and Republicans each attempt to set the public agenda by focusing (respectively) on specific government programs or overall levels of spending. Each side believes that if it can focus the agenda on their preferred frames, it can shape the public’s opinions about the desirability of government spending (Sellers 2010). This ambiguity illustrates Iyengar’s concern in the concluding chapter about how emphasis framing may too easily overlap with traditional aspects of persuasion such as agenda setting.

Works Cited Asher, Herbert. 2007. Polling and the Public. Washington, DC: CQ Press. Bartels, Larry. 2003. “Democracy with Attitudes.” In Michael MacKuen and George Rabinowitz, editors, Electoral Democracy. Ann Arbor: University of Michigan Press. Bennett, W. Lance. 1990. “Toward a Theory of Press–State Relations in the United States.” Journal of Communication 40: 103–124. Brown, C. Budoff. 2009. “Who Will Be First to Define Stimulus?” Politico (January 26).

Introduction 7 Cobb, Roger and Charles Elder. 1983. Participation in American Politics: The Dynamics of Agenda-Building. Baltimore, MD: Johns Hopkins University Press. Druckman, James. 2001. “The Implications of Framing Effects for Citizen Competence.” Political Behavior 23(3): 225–256. Entman, Robert. 1993. “Framing: Toward Clarification of a Fractured Paradigm.” Journal of Communication 43(4): 51–58. Entman, Robert. 2004. Projections of Power: Framing News, Public Opinion, and U.S. Foreign Policy. Chicago, IL: The University of Chicago Press. Gans, Herbert. 2004. Deciding What’s News: A Study of CBS Evening News, NBC Nightly News, Newsweek and Time, 25th anniversary edition. Evanston, IL: Northwestern University Press. Graber, Doris. 2006. Mass Media & American Politics. Washington, DC: CQ Press. Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago, IL: The University of Chicago Press. Iyengar, Shanto and Jennifer McGrady. 2007. Media Politics: A Citizen’s Guide. New York: W.W. Norton & Company. Jacobs, Lawrence and Robert Shapiro. 2000. Politicians Don’t Pander: Political Manipulation and the Loss of Democratic Responsiveness. Chicago, IL: The University of Chicago Press. Kernell, Samuel. 1997. Going Public: New Strategies of Presidential Leadership, 3rd edition. Washington, DC: CQ Press. Kingdon, John. 2003. Agendas, Alternatives, and Public Policies, 2nd edition. New York: Longman Press. Lakoff, George. 2004. Don’t Think of an Elephant. White River Junction, VT: Chelsea Green Publishing. Lippman, Walter. 1922. Public Opinion. New York: Macmillan. Nelson, T.E., Rosalee A. Clawson and Zoe M. Oxley 1997. “Media Framing of a Civil Liberties Conflict and Its Effect on Tolerance.” American Political Science Review 93: 567–583. Pan, Zhondang and Gerald M. Kosicki. 2001. “Framing as a Strategic Action in Public Deliberation.” In Stephen D. Reese et al., editors, Framing Public Life. Perspectives on Media and Our Understanding of the Social World. Hillsdale, NJ: Lawrence Erlbaum Associates. Riker, William. 1986. The Art of Political Manipulation. New Haven, CT: Yale University Press. Schattschneider, E.E. 1960. The Semisovereign People. New York: Holt, Rinehart, & Winston. Sellers, Patrick. 2010. Cycles of Spin: Strategic Communication in the U.S. Congress. New York: Cambridge University Press. Tversky, Amos and Daniel Kahneman. 1974. “Judgment under Uncertainty: Heuristics and Biases.” Science 185, 4157 (September 27): 1124–1131.

Part I

Origins

2

Framing and Value Recruitment in the Debate Over Teaching Evolution Thomas E. Nelson, Dana E. Wittmer, and Allyson F. Shortle

Introduction1 August 2, 2005, was a triumphant day for intelligent design (ID), the latest and most powerful challenge yet mounted to the standard model of evolution through natural selection derived from Charles Darwin and Alfred Russell Wallace. During a press conference, President Bush himself endorsed the teaching of ID in the public schools. “Both sides ought to be properly taught . . . so people can understand what the debate is about,” he said (Garofoli 2005). The President chose his words carefully that day. The reporter (Ron Hutcheson of Knight-Ridder Newspapers) actually asked the President two questions: first, about his personal views on the theory, and, second, whether or not it should be taught in the public schools. The President limited his initial answer to the teaching question. When pressed by Hutcheson about whether endorsing the teaching of the theory implied that he “accepts the validity of intelligent design as an alternative to evolution,” the President again sidestepped the scientific issue, merely reiterating “Part of education is to expose people to different schools of thought . . . You’re asking me whether or not people ought to be exposed to different ideas, and the answer is yes.” The response from the scientific community to this apparent endorsement of ID was swift and scornful, so much so that in the following days presidential science adviser John H. Marburger III assured the nation’s scientists and teachers that Bush believed that evolution is the reigning idea in biology, and that intelligent design “is not a scientific concept” (Garofoli 2005). Still, by muddling the scientific issues and harnessing his views on the education question to a liberal value like open-mindedness, the President affirmed the political rehabilitation of creationism as intelligent design: a far more palatable and effective proposition for public education. In the debate over intelligent design, simple ground-floor descriptive terms become bones of contention, and so just to begin with an even-handed layout of this issue’s conceptual territory is itself daunting. Intelligent design advocates bristle at the comparison of ID to creationism; even the suggestion that ID “evolved” from creationism is objectionable. The theory, in its simplest terms, contends nothing more than that an intelligent force deliberately guided the

12 Origins course of evolution (Dembski 2004). According to ID, there is no plausible alternative account for the complexity of life as we know it. Who, or what, this force is, ID proponents cannot say, nor is there a consensus on when or how this guidance was applied. This vagueness has many virtues, chief among them that it avoids the legally fatal connection between traditional creationism and biblical literalism, at least superficially. ID is not, its proponents insist, creationism. For ID opponents, the link between creationism and ID is not only clear and obvious, but is the principal reason for their objection to the theory, so much so that they insist on using the more awkward synthetic label “intelligent design creationism” (Pennock 2001b). The senior author interviewed staffers at the National Center for Science Education, the front-line opponents of ID, just days after they finished their testimony at the pivotal trial of Kitzmiller v. Dover Area School District (2005). Tammy Kitzmiller and ten other parents had brought suit against the Dover, Pennsylvania, school board. The board mandated that high school biology students learn ID as an alternative theory of evolution. The parents argued that ID is religiously inspired and directed, and the mandatory teaching of it amounts to an unconstitutional government establishment of a narrow, sectarian religious perspective. In other words, it is creationism, despite what its promoters say. The NCSE staffers gleefully showed me a piece of damning evidence, in their view. Via subpoena, the plaintiffs obtained early drafts of the ID textbook Of Pandas and People (Davis et al. 1993). The word “design” had been substituted for “creation” in several passages that were otherwise unchanged. What could be clearer evidence that ID is simply “creationism in a cheap tuxedo,” they asked. Judge John E. Jones III apparently agreed, for he wrote in his opinion, “The evidence at trial demonstrates that intelligent design is nothing less than the progeny of creationism” (Jones III 2005, p. 31). The ruling in Dover is clearly a setback for the movement, but it remains that ID has enjoyed more success in less time than “creation science” could ever claim. President Bush is not the only high government official to express support for ID, and the Dover school board is not the only body to advance this cause. Many skirmishes between ID supporters and opponents have flared up at the state and local level, and ID has left a noticeable, albeit temporary, impact on state school policy in Ohio and Kansas (Scott 2004). Senator Rick Santorum proposed an amendment to the 2002 “No Child Left Behind” legislation that would have encouraged the “critical analysis” of evolution. Santorum’s amendment struck all the right liberal democratic chords, as it lauded “open mindedness” and “critical thinking.” Even Edward Kennedy was moved to praise the amendment, and it passed 91–8 (Forrest and Gross 2004, p. 242).2 During a May 3, 2007, debate, three of the ten candidates for the Republican nomination for President raised their hands when asked if they did not believe in evolution. Seeming to read from the ID playbook, Senator Sam Brownback of Kansas explained, “I find there are too many complexities in the cell and wonders in the mind.”3

Framing and Value Recruitment 13 The alleged continuity between creationism and intelligent design, although critical for the ultimate legal viability of the movement within public school classrooms, is but one dimension of this complex legal, scientific, and political phenomenon. In truth, much of the American public believes in creationism (Miller et al. 2006). This label is therefore not necessarily damaging in the court of public opinion, regardless of how deeply it makes scientists and civil libertarians shudder. The political success of ID depends not so much on its advocates covering the tracks that lead back to creationism but on successfully claiming bedrock values such as free speech, fairness, and open-mindedness as their own. There is no doubt that ID promoters recognize the strategic importance of such values, and have crafted their public appeals accordingly (Fish 2005). Framing is an important tool that ID promoters wield to advance this strategic purpose. Everyone seems to recognize the political potency of values. Pundits and political operatives give much of the credit for Republican electoral success in the early part of the twenty-first century to the party’s relative skill at appealing to “moral values” (Kristof 2004; Lakoff 2002). Scholars across diverse fields also recognize the importance of values for organizing and directing political activity (Feldman and Steenbergen 2001; Goren 2001; Inglehart 1990). In some ways, values are the epitome of stable individual political predispositions. As core principles and ideals, values like egalitarianism, individualism, humanitarianism, and environmentalism offer general criteria by which the citizen can evaluate the constantly shifting landscape of actors and issues. And yet, while ample evidence confirms the individual stability and political centrality of values, the translation between abstract values and specific opinions is not as straightforward and immediate as we might expect. Whether or not any given person will perceive the relevance of a general value for a specific issue is a matter of personal reflection as well as social influence. We see the ID debate as a vivid case study of value recruitment: the harnessing of social and political values in persuasive speech. Value recruitment describes the efforts of political persuaders to influence the connections individuals make between broad social values and particular political issues. We use the term recruitment because values, like star athletic or academic prospects, have qualities that would enhance just about any organization’s performance. Values are consensus desiderata—they are ideal conditions, dispositions, or behaviors that just about everyone admires to some degree (Feather 1995; Maio and Olson 1998; Schwartz 1996). Who finds conditions like freedom objectionable? Who would like to see less personal responsibility or honesty in our world? There is much to gain, and little to lose, in convincing the public that your side of the debate commands such admirable values. This chapter continues with a general description of value recruitment theory. We then turn to a synopsis of intelligent design’s heyday in Ohio, roughly 2002 through 2006. We point out how values were recruited by the contestants in the ID debate, focusing specifically on how the two sides used framing to advance their claims on key values. We present the results of a content analysis

14 Origins that tracked specific values in mass media coverage of the issue. We then present a laboratory experiment that examined the impact of framing on opinion about teaching evolution, as well as judgments about the importance of different values. We conclude by discussing how our findings relate to other work presented in this volume, and how our research enhances framing theory. What are persuaders trying to accomplish through value recruitment, and how do they go about it? We argue that there are four broad strategic goals that motivate value recruitment, with numerous rhetorical tactics that further the pursuit of these goals. • • •

•

Value association. Communicators must convince the public that their side of the debate holds the most legitimate claim to a cherished value. Value disassociation. Communicators must convince the public that their opponents do not hold legitimate claims to a cherished value. Value enhancement. Communicators must boost the esteem or importance the public attributes to a value already associated with the communicator’s position. Value diminishment. Communicators must discredit a value associated with the opposition, if it cannot be dissociated.

Just as political parties may be said to “own” different issues (Petrocik 1996), so can competing sides in a public debate come to be seen as representing or embodying different values. Likewise, just as parties work hard to manage issue ownership, so can competing groups try to manipulate public perceptions as to who owns the most coveted values in a debate. Values may be categorized as contested or exclusive. Both sides strive to gain control over contested values, while exclusive values have been ceded by one side to the other. The strategic path chosen and the consequent rhetorical tactics applied will therefore depend on the type of value. Contested values will inspire tactics that support the strategies of value association and disassociation, while the side that owns a value exclusively will try to boost its esteem among the public as its opponents try to diminish its perceived worth. A number of issues generate or expose conflicts among treasured values. The public will often view a change in policy direction as advancing one value while diminishing another. President Clinton’s health care reform initiative of 1994, for example, was judged by many as advancing the cause of social welfare by providing universal coverage, but at the cost of free choice by (allegedly) limiting individual selection of physicians. Awareness of such conflicts can vex citizens who treasure both values; complex and ambivalent attitudes often result (Feldman and Zaller 1992; Lau and Redlawsk 2006). Value enhancement and diminishment could be deployed in such situations. If a communicator views an important segment of the public as torn between a value her side owns and a value held exclusively by the other side, she will concentrate her rhetorical skill on elevating the priority of her value relative to the rival’s.

Framing and Value Recruitment 15 To illustrate, consider the rhetoric surrounding the values of scientific discovery and fairness in the ID debate. Both sides covet the value of scientific discovery, and much of their public rhetoric projects an image of themselves as legitimate scientists, and their opponents as scientific frauds. The principal ad hoc organization favoring ID in Ohio was named Science Excellence for All Ohioans, while the principal opposition group was named Ohio Citizens for Science (emphasis added). Fairness, however, is a value claimed almost exclusively by ID promoters. They argue that it is “only fair” to include intelligent design in school science curricula, since it is a legitimate competing alternative theory of life’s development on earth. With a few exceptions, evolutionists do not try to steal “fairness” from ID promoters and claim it for themselves; rather, they suggest it is irrelevant or unimportant to the issue. Science itself is not fair, they claim, so why should we impose that condition on science classes? Does fairness require that we teach geocentrism as an alternative to heliocentrism in astronomy classes, or that we should provide equal time for Holocaust deniers in history classes?

ID in Ohio: An Exploration of Value Recruitment Rhetoric Communicators accomplish the four general value recruitment strategies by drawing on specific rhetorical tactics, such as framing (Nelson 2004). The public debate over teaching ID in Ohio’s public schools suggested an important role for value recruitment. To be sure, much of the rhetoric concerned disagreements on factual questions, such as the strength of the scientific evidence supporting the theory of evolution via natural selection. Still, a substantial amount of discussion appeared to be directed toward the four value recruitment goals. It therefore seemed an ideal case study to examine the diversity, dynamics, and impact of value recruitment rhetoric. Amended Substitute Senate Bill 1, passed by the 124th General Assembly of the Ohio Legislature, directed the State Board of Education to adopt statewide academic standards for each of grades kindergarten through twelve in reading, writing, and math by December 31, 2001, and in science and social studies by December 31, 2002. It also “requires the State Board to adopt a model curriculum aligned with the academic standards which school districts may (but are not required to) use for instruction” (Am. Sub. HB1 Final Analysis, LSC, 2001). To fulfill this legislative mandate, the Ohio State Board of Education enacted a public process to author the Academic Content Standards. Writing teams included kindergarten through undergraduate collegiate educators from across the state, as well as representatives from business and industry, parents, and community leaders. Draft documents were also made available to the general public, and some forty focus groups and twelve regional meetings were held in various locations across the state. This effort to codify what Ohio students should know about science attracted considerable input from activists and organizations inside and outside the state. John Calvert, head of the Intelligent Design Network, which had actively

16 Origins participated in the Kansas Board of Education’s decision to remove evolutionary theory from the high school science curriculum, was invited to make a presentation to the Ohio board. Ad hoc organizations formed within Ohio to lobby for or against the inclusion of ID tenets in the Ohio standards. Ohio Citizens for Science, headed by Case Western Reserve University biology and philosophy professor Patricia Princehouse, became the most active public opponents. Robert Lattimer, a chemist working in private industry, helped form Science Excellence for All Ohioans, which became the most visible proponents. Lattimer, who had worked for years on behalf of conservative causes in public education, also served on the Science Standards Writing Committee. One of the most dramatic public events was a March 2002 debate staged between proponents and opponents of ID in front of the Board of Education and hundreds of spectators in Columbus (Mangels 2002). Speaking in favor of ID were Stephen Meyer and Jonathan Wells, both of the Discovery Institute, a Seattle-based think tank and the foremost organization advocating for ID at the national level. Opposed were Lawrence Krauss, a Case Western physics professor, and Kenneth Miller, a Brown University biology professor. Although we will discuss this in more detail at a later point, it is important to note that this debate was significant in part because Meyer used his presentation to propose the “teach the controversy” compromise. On December 11, 2002, the Science Academic Content Standards were adopted unanimously. The standards added a new definition of science: “Recognize that science is a systematic method of continuing investigation, based on observation, hypothesis testing, measurement, experimentation, and theory building, which leads to more adequate explanations of natural phenomena.” The standards also included a statement in Life Sciences, Grade 10, Indicator 23 which reads “Describe how scientists continue to investigate and critically analyze aspects of evolutionary theory. (The intent of this indicator does not mandate the teaching or testing of intelligent design.)” The same statement was also added to Benchmark H in Life Sciences, Grade 10. In this way, Board members resolved any question as to their stance on the subject of Intelligent Design. Only positive comments stand on record: one Board member publicly called the resolution a “win-win situation.” Critics, however, saw this as a troubling advancement for the ID movement. They noted that the critical analysis of evolution was a fundamental ID tenet, and that critical analysis had not been expressly recommended for any other scientific theory or concept. The second requirement of Amended Substitute Senate Bill 1 was the development of model curricula in each area. This sparked further debate as to the most effective and proper way to implement the standards set in December 2002 pertaining to the critical analysis of evolution. According to Board of Education minutes, this process remained on schedule throughout the end of 2003. At a Standards Committee meeting on February 9, 2004, the Board considered the model curriculum framework for science and social studies. Most of the discussion focused on the lesson entitled “Critical Analysis of Evolution.” Board members were strongly divided over whether the proposed curriculum

Framing and Value Recruitment 17 drew too heavily from the ID canon. One board member objected specifically to the inclusion of Jonathan Wells’s Icons of Evolution (Wells 2002). Wells, a Discovery Institute research fellow who had participated in the March 2002 debate, has written that standard biology textbooks fraudulently promote Darwinian evolution with falsified data, suspect demonstrations, and concocted illustrations. Committee Co-chair Michael Cochran contested the claim that the lesson promoted intelligent design, and defended the process that developed it as impartial and inclusive. Deborah Owens Fink noted that there was no mandate for a teacher to use any particular lesson. Following this discussion, the motion to eliminate the lesson was defeated by a vote of 2–6. A resolution of intent to adopt lesson Set A (science and social studies) was approved by a 6–2 vote. The full State Board considered the “resolution of intent” to adopt Set A at its meeting the next day—February 10, 2004. During the Public Participation portion of the meeting, witnesses spoke on the science lessons: most of the comments focused on the “Critical Analysis of Evolution” model lesson in particular. Of the 16 speakers, half opposed the lesson, claiming that it was “bad science,” that it contained intelligent design, or that it was religiously motivated. The other eight speakers were supportive of the lesson, arguing that it did not contain intelligent design concepts, that it supported public polling results, and that it developed critical thinking in students, as well as being “good science.” The overall theme was that the “Critical Analysis of Evolution” lesson was the only one that reflects the intent of Benchmark H and Indicator 23 in the Science Standards. The resolution of intent to adopt lesson Set A was approved by a 13–4 vote. The March State Board of Education meeting minutes reflect the continued escalation of the debate. During the meeting, 42 witnesses spoke to the Board: 16 citizens supported some aspect of the curriculum or science standards, while 23 expressed some level of opposition. At the meeting’s end, Mr. Hovis moved that the “Critical Analysis of Evolution” lesson plan be removed altogether, so that the committee could “wait and do a better job.” The motion was defeated in a vote of 7–10, with one abstention. The Science Model Curriculum Set A, including the “Critical Analysis” model curriculum, was adopted by a 13–5 vote, temporarily ending the debate in Ohio. ID’s privileged position in Ohio education policy was short-lived, however. On February 14, 2006, in the wake of the Dover decision, the school board voted to remove the “critical analysis of evolution” language from the curriculum guidelines and lesson plans (Candisky 2006). The following November, Deborah Owens Fink, one of the two most passionate advocates for ID on the Ohio Board, was soundly defeated by a challenger who made Owens Fink’s support for ID the centerpiece of his campaign. Our investigation of ID value recruitment rhetoric began to take shape with a set of interviews conducted by the senior author with prominent activists on both sides of the issue, as well as members of the Ohio Board of Education. These interviews, along with a close reading of the voluminous scholarly and

18 Origins popular writings by advocates and critics of ID, yielded a finite, recurrent set of themes and arguments. We extracted from this set a working list of prominent values, both contested and exclusive, as well as rhetorical devices deployed to recruit those values. This list, revised many times, served as the coding scheme for a content analysis of newspaper coverage of the ID issue in Ohio. We examined 96 newspapers articles from January 2000 through July 2004 that focused on the intelligent design debate.4 Each article was coded twice, once for values and once for tactics. The coding unit for the most part was the sentence, but we quickly realized that a number of sentences incorporated multiple values and/or tactics. Complex sentences were then broken down further into discrete “thought units” (Gottman 1979). The coding scheme was refined until two separate coders reached an acceptable level of reliability.5 The next section describes the seven values that dominated news coverage, including the rules we followed to recognize and code these values. Their relative frequencies are presented in Figure 2.1. We then describe some of the key value recruitment tactics that presented themselves in the debate, focusing in particular on framing. We conclude by presenting an experiment on the impact of framing science on attitudes and values respecting the issue.

Democratic process

Scientific discovery

Educational excellence

Religion and spirituality

Intellectual integrity Free speech, fairness, academic freedom Church–state separation

Development and progress 0

20

40

60

% Articles including value (N = 96)

Figure 2.1 Frequencies of Values in Content Analysis

80

100

Framing and Value Recruitment 19

Values in the ID Debate Eight values dominated discussion of the issue in the newspaper articles. For convenience, we group them into three master categories: (1) “Enlightenment” values pertaining to learning, discovery, and progress; (2) “liberal” values pertaining to democratic government and individual rights; and (3) religious sensibilities. Enlightenment Values Scientific Discovery To our reckoning, this was the prize value in the debate. Both sides worked strenuously to portray themselves as more scientifically advanced and honest than the opposition. In the content analysis, this category included: all descriptions of the specifics of a theory; discussions about the value of knowledge and having an accurate, truthful understanding about life on Earth; claims about the relationship between science and religion; and statements concerning how the scientific evidence favors or disproves one side of the debate. Claims for scientific values were often couched in efforts to frame science in inclusive or exclusive terms, e.g., “For intelligent design to be science, its proponents must state testable ideas about how it works and conduct experiments or make systematic observations to test their hypotheses.”6 In Kansas and Ohio, this framing contest extended to official state definitions of science. ID opponents sought to define science as focused exclusively on natural explanations for natural phenomena, while ID supporters backed broader definitions that left room for supernatural causation (i.e., the actions of a powerful, intentional “designer”). Educational Excellence As the debate’s tangible focus was public-school education policy, it is perhaps not surprising that educational values also received prominent mention. This coding category includes: discussions about the value of passing knowledge and understanding on to our children; supporting their intellectual growth; and nurturing curiosity and mastery. Furthermore, it encompasses claims about what children should learn and what should be included in classroom discussion of biology. Rival claims over education values also revealed the strategic acumen of the ID movement. Aware that much of the educational and scientific establishment would dismiss their efforts to recruit scientific values, ID supporters played up educational values such as “critical thinking” that would appeal to their critics. The pinnacle of this strategy was the aforementioned “teach the controversy” proposal, with its claims about the pedagogical advantage of presenting students with a variety of perspectives. This strategy put ID critics in the difficult position of having to reject this seemingly commonsensical idea.

20 Origins They did so by making comparisons between ID and other non-scientifically validated theories, e.g. “Being open-minded does not mean that adding two plus two equals five should be taught in math class . . . or teaching witchcraft in health is a good idea.”7

Intellectual Integrity The rivals for this debate also strove to portray themselves as more intellectually rigorous and honest than their opponents. We saw many instances of each side trying to define itself as open and sincere, while at the same time attempting to color the opposition as close-minded, deceitful, and manipulative. We coded discussions of the virtues of honest, objective discussion, and accusations that one side is not being truthful or forthright in some way within this category. For example, “ID should stand for insidious deception, as it is a disguise for creationism, meant to sneak particular religious views into our science classrooms.”8

Development and Progress Within this broad category fall all discussions of economic development, progress, and economic competitiveness. This is just about the only value that the evolution side can realistically claim as exclusively its own. Although ID has gone far in discarding the “bumpkin” image of young-Earth creationism, opponents have made some mileage out of predicting that official endorsement of intelligent design would not only set up the state for immediate ridicule but also would yield grave long-term economic consequences. “They also warn of dire consequences for economic development and the state’s image should Ohio become the first state in the nation to include intelligent design in grade-bygrade guidelines of what students should learn in science class.”9

Liberal Democracy Values Free Speech, Fairness, and Academic Freedom This cluster of values worked almost exclusively to the advantage of ID supporters. ID opponents, because they were trying to exclude the theory from science classes, were often portrayed as being on the wrong side of these popular democratic values. Our coding included statements concerning the importance of allowing both sides to make their case; claims about academic freedom; and references to equality of opportunity. Moreover, within this category are included accusations that the mainstream scientific establishment is censoring the debate and the teaching of alternative perspectives, and, therefore, indoctrinating students by not allowing them to hear important ideas. For example, ID supporters would claim, “New ideas are rejected today by the reigning establishment, just as they were in Copernicus’ time.”10 Moreover,

Framing and Value Recruitment 21 ID supporters would attempt to negate the credibility of ID opponents by explaining that “it all seems persuasive when you only hear one part of the argument.”11 Church–State Separation This was the main constitutional principle on the side of ID opponents, as in previous battles over creationism. It is not entirely clear that this principle rises to the level of a consensus value, however. A fringe movement within Christian conservatism portrays church–state separation as a historical myth, and argues that the founders intended the United States to be a “Christian nation” (Barton 1993). This group is naturally drawn to ID, and is enthused, not troubled, by the prospect of privileging one religion’s creation story above all others. Prominent ID leaders, however, whatever they might privately think about the establishment clause, do not publicly endorse teaching the Old Testament view of creation in science classes. Recognizing the legal power of this principle, they have always maintained that ID is religiously neutral, and therefore poses no challenge to the First Amendment. Church–state separation is not a value exclusive to ID critics. Some ID intellectuals have advanced the view that “evolutionism” is a kind of “secular religion,” and its privileged position in the public schools does just as much damage to the establishment clause as teaching blatant creationism. Others argue that government promotion of evolution “undermines” students’ religious faith, and thus amounts to government interference in the free expression of religion. Our scheme therefore coded claims that government should not endorse any particular religion, nor should government interfere with the free expression of religion. The following is an example of a sentence that would fall within this category: In 1987, the U.S. Supreme Court stated that teaching religiously motivated pseudoscience, creationism, intelligent design, whatever you want to call it, in a public-school science class violates the separation of church and state and it is clearly illegal.12 Democratic Process This category included discussions of the process by which policy decisions are made, with an emphasis on public participation, the rule of law, and the consent of the governed. Statements regarding public opinion, public control over education, and compromise are all included in this category. A common argument made by ID supporters is that public opinion is on their side, which to them signifies that ID should be included in the school curriculum. For example, “Supporters of intelligent design cite poll results showing that nearly two-thirds of Ohioans support teaching evolution and presenting the scientific evidence against it.”13

22 Origins Conversely, ID opponents argue that some decisions, such as what to teach in science classes, are not a matter to be decided by the public at large. Furthermore, in this category we also see reference to one side accusing the other of violating the proper “democratic” procedure. This turned out to be one of the most common arguments, to our surprise.

Religious Sensibilities Religion and Spirituality In contrast to the unabashed recruitment of scientific discovery and educational excellence, claims for religious values on both sides, but especially for ID proponents, have been cautious and nuanced. To minimize their legal and political risks, ID promoters must be far cagier about the centrality of faith to their activism. Most ID supporters are, naturally, devout Christians, but they explain this as their personal belief, not an ID tenet. Specifically, they claim that “Design theory is compatible with a belief in God and the Bible, but it does not require adherence to any particular faith or doctrine.”14 Evolutionists do not speak with one voice on religious values. They range from the deeply religious to the unapologetically atheist. Richard Dawkins has drawn notoriety for his claim that God is a “delusion,” and that science can answer virtually all important questions about how the universe and life within it came to be (Dawkins 2006). “Theistic evolutionists”—who fully embrace evolution and belief in God—appreciate Dawkins’s spirited and eloquent defense of evolution, but many wish he would keep quiet when it comes to theology. They blame Dawkins and other outspoken atheistic evolutionists for contributing to the public’s perception that evolution is anti-God. Most American critics of ID strike an accomodationist stance, arguing that science and religion are fully compatible (Gould 1997). A few, such as Kenneth Miller and Francis Collins (director of the Human Genome Project), claim that evolution actually strengthens their faith in God (Miller 1999). Conversely, ID proponents claim that evolution is not neutral with respect to religion, but decidedly hostile to it (Ruse 2005). Moreover, ID proponents claim that “Darwinism” qualifies as a religion of a sort. Figure 2.1 shows the percentage of articles that mentioned a specific value at least once. There is a high degree of variation in how frequently each value category is used. At the low end is progress, appearing in fewer than one-quarter of the articles. At the high end are democratic process and scientific discovery, making appearances almost 90 percent of the time. The importance of science is not surprising; this was, after all, a debate over the content of public school science standards. We were surprised by the frequency of mentions of democratic process. However, this number might be inflated. While there were numerous complaints by activists about their opponents twisting the democratic process in some way, this category also includes more mundane descriptions of the procedures for changing statewide educational requirements.

Framing and Value Recruitment 23

Tactics Value recruitment tactics are rhetorical tools that seek to create, or undo, a perceived association between a position on the ID issue and a treasured value. If such an association already exists, value recruitment tactics can enhance or diminish the perceived importance of the value. A concise listing can be found in Table 2.1. The eleven tactics fall into three broader categories: framing tactics, value qualification tactics, and social cueing tactics. We will briefly describe the latter two categories before turning our attention to the focus of this chapter: framing. Value qualification tactics recognize, explicitly or implicitly, the multiplicity of values that apply to an issue. To manage this value pluralism and inevitable value conflicts, communicators may qualify certain values to change their priority relative to others. Such tactics are especially prominent, we believe, when communicators address exclusive values. Values that the communicator’s side already “owns” will be assigned a high priority, while the opponent’s values will be portrayed as unimportant or irrelevant. Institutional role assignment (Nelson 2004), value ranking (Nelson 2004), value expansion, value definition and delimitation, and value conflict all achieve this purpose. In value ranking, for example, the communicator asserts that, whatever importance we may attribute to a pair of values in general, a specific priority must hold for this issue. Scientists acknowledge, for example, that classroom discussions about evolutionary theory pose difficult challenges for religious fundamentalists, but they stress that scientific truth should take precedence over respecting the religious views of students (Dawkins 2006; Pennock 2001a). Values are vague and abstract notions, often lacking clear material reference (Lau and Redlawsk 2006). One way to concretize these ideas is to associate them with prominent social cues (Grant and Rudolph 2003; Gwiasda and Nelson 2001; Nelson and Kinder 1996). People might “embody” social values by thinking about groups that represent them. In both the Scopes and Kitzmiller cases, the value of church–state separation became indelibly linked, for better or worse, with the American Civil Liberties Union (Humes 2007). The respect accorded these values within any particular policy may depend on the feelings about these associated groups. Social cueing tactics recruit values by appropriating the reputations of social actors. Savvy political communicators try to restructure the linkages we make between groups and values, or to build up or tear down groups they know we associate with values. They manipulate the group’s reputation in order to legitimize or discredit the values associated with it. Extremism, dogmatism, and compromise; exposing true motives ; and credentials and public opinion are all social cuing tactics. ID supporters deftly applied the extremism, dogmatism, and compromise tactic when they boldly announced at the March 2002 public debate a proposal for resolving the impasse: teach the controversy. Instead of exposing children exclusively to the evolutionist or ID view, why not make children aware of the whole controversy by presenting both sides in a head-to-head, debate-style format? It is perhaps not coincidental that

“The question now appears to be: Should the state ask schools to ensure that students are taught about scientific challenges to Darwin’s theory of evolution?” “Meanwhile, Republican legislative leaders, such as Governor Bob Taft, want to leave it up to the state board to decide what students should be taught about life on Earth.” “ ‘I know that this is an issue and I prefer to bring it forward now,’ Sheets said afterward. ‘I just prefer to do it in a credible, responsible and honest manner.’ ” “Evolutionists complain that intelligent design does not deserve to be taught alongside evolution because it is untested and hasn’t been published in scientific journals or subjected to rigorous peer review, as is the norm in the sciences.” “They shouldn’t teach a theory as a fact when it is not.” “Not believing in some kind of intelligent design would be like believing that Scrabble pieces have an inherent ability to arrange themselves into words . . .” “We don’t teach that the sun might revolve around the earth.”

Summary claims about the qualities, characteristics, and elements of science or intelligent design; statements concerning what qualifies as true science, including testability, falsifiability, replicability, and the need for independent, empirical data. Claims that evolution is a dead and obsolete idea that is a type of religion itself. Attempts to link intelligent design to creationism, or to refute the link.

Comparisons of the intelligent design debate to other political or scientific conflicts.

Framing science and ID

Analogy

Example text

Attempts to provide a summary label for the issue, or to ascribe it to a particular category; descriptions of a policy as favoring one side; characterizations of one side’s position, goals, or aims, and attempts to describe one side’s motivations. Mentions of previous or contemporary legislative or legal battles over teaching creationism and descriptions of the content of the science standards. Assertions about who should properly make a decision or policy change.

Description

Issue labeling/categorization

Framing tactics

Tactic

Table 2.1 Tactics from Content Analysis

“Public schools should not be in the business of religious instructions; that is up to the parents.” “But the topic is suitable only for theology or philosophy classes—not science.”

Assertions that certain institutions have an obligation to uphold particular values. Statements about what schools should do/teach and claims about what belongs in science classrooms.

Institutional role assignment

Extremism, dogmatism, compromise

(Continued overleaf)

“Mr. Calvert said that by not allowing intelligent design to be taught with evolution, ‘you are essentially indoctrinating students in a dominant view.’ ” “But supporters of the Intelligent Design Network say the scientific establishment is censoring debate.”

“Miller, noting that he is a religious person, said he worries that by forcing intelligent design into the classroom, students will feel pressured to choose between science and religion.”

Claims that science does, or does not conflict with religious beliefs are included in this category.

Value conflict (religion v. science)

Portrayals of an individual or group as being extreme, radical, or out of touch with ordinary people and mainstream values. Assertions that a group is fixed in their ways of thinking and does not deal in an objective and evenhanded manner with the evidence and/or they refuse to compromise.

“But Meyers said a broader definition of science to include the controversy would enliven classroom teaching and improve science education in Ohio.”

Satements that make one value seem more important by associating with another.

Value expansion

Social cueing tactics

“Teachers should be free to talk about intelligent design, Fink and Turner said, regardless of arguments from most scientists who say the concept does not have merit because it has not been tested or subjected to extensive peer review.”

Explicit comparisons between values, including the assertion that one value is “more important” than other values and arguments. Claims that a certain value or principle is “fundamental” or “important” versus “trivial” or “incidental,” without explicit comparison to some other value.

Value ranking

Value qualification tactics

Description

Descriptions of the “true” motives of a group, in contrast to their publicly stated motives or goals. Statements accusing one side of manipulating the process and arguments that the other side is shifting tactics because of a political defeat.

Statements that attach a label to an individual to categorize that person with a specific group.

Descriptions of the current status of public opinion, or the opinion of specific groups. References to official bodies or organizations, particularly when they are associated with an opinion or position on the issue.

Tactic

Exposing true motives

Group labels and credentials

Public opinion; group opinion

Table 2.1 Continued

“Ohio scientists are angry that a proposed lesson for the new science curriculum . . .” “Polls show that many Ohioans believe in intelligent design and that other ideas should be taught.”

“ACLU lawyer” “Biology Professor at Ohio State University”

“ID is a way of sneaking creationism into the public schools.” “The strategy of intelligent design proponents is simple: falsely argue that there are only two alternatives (theistic and atheistic) to explain origins and diversity. Deny your religious motives.”

Example text

Framing and Value Recruitment 27 this strategy received its first promotion during a widely attended public debate between ID advocates and critics. The implicit message was: if open debate between alternative views is good enough for the adult citizens of Ohio, why is it not good enough for their children? Framing Tactics For the remainder of this chapter, we concentrate on a tactic that has been crucial for both sides in this issue: framing. A common complaint in the framing literature concerns the expanding, unruly population of framing concepts, definitions, and operationalizations. We will have more to say about this state of affairs in the conclusion; for now, at the risk of adding to the confusion, we are compelled to point out some key variations in the manifestation of framing in the ID issue. Apart from framing itself, we also identify issue labeling and categorization and analogy as important framing tactics. Issue Labeling and Categorization Some frames provide comprehensive definitions and narratives that explain the entire problem—how it came to be, and how we should resolve it (Gamson and Modigliani 1989; Goffman 1986). Gray refers to these as “whole story frames” (Gray 2003); we prefer the less adroit name issue labeling and categorization. An issue label or category is a comprehensive, synthetic proposition that explains the current predicament, accounts for the behavior of major protagonists, and recommends a course of action. Schaffner and Atkinson in this volume, for example, examine an instance where strategies to win over public support on the inheritance tax issue centered on the Republicans’ success at applying the “death tax” label, versus the Democrats’ inability to employ successfully the competing label of an “estate tax.” While there are several competing labels in the ID debate, two of the most prominent are “Dogmatic Darwinism” and “The Trojan Horse.” The “Dogmatic Darwinism” label, favored by ID supporters, explains that conventional scientists are hostile to ID because it doesn’t just question Darwinism but undermines the bedrock materialist foundation of modern biology—indeed, of all science. ID isn’t just an alternative theory, but an entirely new paradigm that promises to revolutionize all of the life sciences, and possibly other disciplines as well. According to this label, biology is waging a fierce, but in the long-run fruitless, battle against obsolescence. It will not go down without a fight, however: biologists allegedly resort to all kinds of nefarious tactics, including censorship and professional ostracism, to marginalize ID scholarship. The documentary film Expelled: No Intelligence Allowed (Frankowski 2008) is a feature-length exposition of the Dogmatic Darwinism issue label. The “Trojan Horse” label, favored by evolutionary biologists, tells a very different story, to say the least (Forrest and Gross 2004). It charges that ID is an innocent-appearing notion with an apparently modest and worthy goal (fair

28 Origins treatment) that is used as a weapon to advance a far more insidious goal of supplanting humanist ideas and materialist science with biblical fundamentalism. Furthermore, this nutty, fringe idea would never be taken seriously but for the well-funded, religiously inspired propaganda machine promoting it. It should not be regarded as an “alternative theory” of evolution, but rather as fundamentalist nonsense. We drew directly from these two issue labels when we designed the manipulations for our experiment. These competing labels seek to wrest the plum value of scientific discovery away from the adversary and secure it for one side. For example, a great challenge for ID supporters is to somehow account for the near-universal contempt that the biological establishment holds for their theory. The “Dogmatic Darwinism” label explains that this disregard is based not on the objective scientific merits of ID but rather on the less lofty motive of jealous turf protection. Similarly, the “Trojan Horse” label reveals ID as a piece of the fundamentalist Christian political agenda, not an “alternative theory.” Likewise, these issue labels make different cases for the importance of the democratic values free speech, fairness, and open-mindedness. “Dogmatic Darwinism” elevates these values by accusing the scientific establishment of “censoring” alternative points of view. The “Trojan Horse” label downplays these values, by portraying this “alternative point of view” as a thin cover for religious fundamentalism. Framing We reserve the term framing for rhetoric that seeks to define and construct key concepts at the sub-issue level. In other words, we will use framing to refer to rival interpretations of contested ideas, rather than comprehensive explanations for the entire issue. One of the most important contested ideas was science itself. ID supporters use a “marketplace of ideas” frame, suggesting that scientific debate should be modeled on political debate. In this frame, one’s preferred scientific theory is a personal expression, like one’s political party or religion. Lay people—scientific consumers—should be exposed to all competing theories, and then allowed to decide for themselves which they prefer. ID supporters frequently remind us that “Darwinism” is “just a theory, not a fact,” implying that, since it can never be conclusively proved, it is no better than any other theory. Just as political systems should tolerate diversity of viewpoints, science should tolerate diversity of theories. ID opponents favor a “progress of ideas” frame that recognizes scholarly debate, but stresses that scientific advancement comes about when stronger, more plausible theories supplant weaker ones. Evolution supporters will of course admit (through clenched teeth) that evolution rightly deserves the label theory, but bristle at the suggestion that its stature is equal to that of rival accounts for life’s development, especially ID. A theory, they say, is not a personal opinion to which we are entitled, but rather a set of ideas that explain empirical observations while organizing research activity. As such, there clearly are better and worse theories, and evolution via natural selection is about as good as

Framing and Value Recruitment 29 theories get. Our experimental treatments were closely modeled on these competing frames for science. Analogy It might seem like a stretch to consider analogies as a kind of framing tactic, but in the ID debate they serve a common purpose: to shape perceptions about the issue and its key concepts (Braman and Nelson 2007; Holyoak and Thagard 1995). New or difficult political issues are often compared to more familiar social problems, where value applications and priorities are more certain. Analogies are powerful tools for explanation, instruction, and persuasion, and so it is no surprise that they figure prominently in rhetoric surrounding the ID issue. Analogies clarify difficult ideas, and so are a popular ingredient in science instruction (think of the analogy between the solar system, with planets orbiting the sun, and the atom, with electrons orbiting the nucleus). Evolutionists and ID supporters regularly use analogies to convey the essence and finer points of their respective theories. Perhaps the most popular is the “mousetrap” analogy, originally suggested by Michael Behe to explain his notion of “irreducible complexity”— a cornerstone of ID theory (Behe 2006). The mousetrap, a relatively simple device, cannot (allegedly) function if any part is missing. A mousetrap without a spring or trigger is not just a less effective mousetrap, but a worthless collection of metal and wood. It is difficult to envision how a “better mousetrap” could have evolved through the gradual accumulation of parts; a board with a spring is no more effective at catching mice than a board alone. The mousetrap must therefore have been designed in its entirety. Similarly, according to Behe, complex microbiological processes such as blood clotting could not have developed in a gradual, stepwise fashion. Incomplete blood clotting doesn’t improve our reproductive fitness relative to no blood clotting at all, so how could the gradual evolution of blood clotting ever get off the ground? Analogies teach different kinds of lessons, and so suit different communicative purposes. In the ID debate, analogies not only explain difficult concepts but also make arguments about the applicability and priority of values. In each case, the goal is to highlight a relevant value that governs the evaluation of the familiar (analog) issue, and apply it to the novel case (ID) by arguing that the target and analog share common, critical features. When ID critics compare the theory to other discredited notions like geocentrism and a flat earth, they are trying to diminish the importance of fairness and equal rights. ID defenders, wary of such comparisons, try to make their theory seem cutting-edge rather than oldfashioned by referencing esoteric and seemingly advanced topics like information theory, complexity theory, and cybernetics.

Experiment To examine the effects of value recruitment on attitudes about ID, we conducted a laboratory experiment. We presented subjects with a pro-ID or anti-ID

30 Origins message, drawing on rhetoric that had repeatedly appeared in our qualitative analysis. Our manipulations incorporated two rhetorical tactics: framing and issue labeling. ID proponents energize their claim on values such as fairness and free speech through the “marketplace of ideas” frame. In this frame, science is similar to a political debate, whereby researchers and students consider a number of contending theories, review the evidence relevant to each one, and draw their own conclusions about which theory is most plausible. This frame puts a high premium on free speech and fairness, as they uphold the goal of exposing students to the broadest variety of alternative theories. ID opponents, including the vast majority of life scientists, offer an alternative “progress of ideas” frame. This frame does not deny the importance of competing ideas, but emphasizes that scientific fields have communal standards and procedures for selecting among theories, and that better theories eventually win out over inadequate theories. Free speech and fairness for every alternative are less crucial from this perspective, as they imply simply wasting students’ time with outdated and discredited ideas. Issue labels provide an all-encompassing account for the issue, explaining how it has gained prominence, and revealing the sometimes-hidden motives of relevant actors. Like the alternative frames for science, these labels give further impetus to a specific value, or in some cases demonstrate how a value is not relevant to the issue. ID supporters offer a “Dogmatic Darwinism” label to describe how self-interested biologists and educators have high-handedly cut off legitimate debate about life’s origins and development. ID opponents have used a “Trojan Horse” label to de-emphasize free speech and fairness, while reinforcing church–state separation. The “Trojan Horse” label emphasizes the continuity between old-style creationism and ID, and warns that ID is simply the “thin edge of the wedge” of Christian fundamentalism in the public schools. Manipulations Subjects read an article about the ID issue that purportedly was excerpted from an online newspaper. Subjects were randomly assigned to a pro-ID, anti-ID, or control version of the article. The pro-ID article included the “marketplace of ideas” frame for science; the anti-ID article contained the “progress of ideas” frame. Some articles also contained language that established the issue label. So, for about 60 percent of the subjects in the pro-ID condition, the marketplace of ideas frame was accompanied by “Dogmatic Darwinism” rhetoric. Similarly, for about 60 percent of the subjects in the anti-ID condition, the progress of ideas frame was accompanied by “Trojan horse” rhetoric. Table 2.2 describes the experimental treatments in detail. Procedures The sample for the experiment was drawn from the Human Subjects pool at the Ohio State University, which consisted of undergraduate students who

Framing and Value Recruitment 31 Table 2.2 Experimental Treatments Condition

Content

Control

Educators in Ohio are considering whether to modify science requirements in public high schools by including a theory known as “Intelligent Design” (ID). This theory contends that life is too complex to have arisen through the process of evolution by natural selection. Supporters of ID say that some other force, perhaps God, created some or all of life in its present form. They further argue that intelligent design should be taught in public schools as an alternative to Darwin’s theory of evolution. Most scientists endorse Darwin’s theory, and do not believe that ID is valid.

Marketplace of ideas

Eric Brewster, 42, has two children in the Pickerington school district. He attended a public meeting of the Ohio Board of Education to express his support for teaching ID. “This is how science is supposed to work. You put all the ideas on the table, and then let people decide which one makes most sense to them.”

Marketplace + Dogmatic Darwinism

Includes the marketplace paragraph as described above, as well as the paragraph: Another parent who spoke in favor of teaching ID was Margaret Brady, whose children attend the Marietta schools in southeast Ohio. “Scientists are opposed to this because they don’t want anyone to hear about any alternatives to their point of view. They have a monopoly on science teaching, and don’t want anyone to challenge it.”

Progress

Eric Brewster, 42, has two children in the Pickerington school district. He attended a public meeting of the Ohio Board of Education to express his opposition to teaching ID. “This is not how science is supposed to work. You don’t just put any idea on the table. You stick with the ones that have been tested and verified, and you get rid of the mistaken ones.”

Progress + Trojan Horse

Includes the progress paragraph as described above, as well as the paragraph: Another parent who spoke against teaching ID was Margaret Brady, whose children attend the Marietta schools in southeast Ohio. “People are pushing ID because they think it’s a way to bring God into the classroom. They’re really only interested in forcing their religion on everyone else.”

32 Origins previously agreed to participate in exchange for extra credit in a political science course. Once the participants arrived at the laboratory, they were placed at a computer station, where all of their responses were collected using an interactive computer program. The computer program instructed the participants to answer an online survey that would take approximately 20 minutes to complete. They were informed that the experiment concerned their thoughts on the quality of Internet news sources, and were then given one of the online news stories concerning the debate over teaching ID alongside evolution in the schools. After the survey, the participants were given an explanation as to the true nature of the experiment. Measures The participants were told before beginning the experiment that the purpose of the study was to assess the quality of online news sources. To sustain this cover story, subjects first answered a few questions about the format and content of the news story. Then they answered questions concerning their opinions about science instruction in public schools. The first was a multiple response item which asked them to check all of the theories that should be included in public high school biology classes (evolution, creationism, intelligent design, something else). A second item asked them to indicate the strength of their support or opposition to including ID in high school science classes, on a six-point scale. These two questions served as the dependent variables in our analysis. We predicted that support for including intelligent design and perhaps also creationism in high school curricula would be higher in the marketplace/Dogmatic Darwinism conditions and lower in the progress/Trojan Horse conditions, relative to the control group. Several questions followed that measured the importance of scientific and democratic values in relation to the ID issue. We predicted that these value importance measures would serve as potential mediators of any effect of the experimental treatments on opinion toward teaching ID and/or creationism (Nelson 2004). Subjects rated the importance of eight values: educational quality, fairness and equality, tolerance and open-mindedness, religion and spirituality, scientific discovery and integrity, government neutrality on religion, freedom of speech, and community control. Other questions presented participants with paired statements that pitted one value against another, forcing respondents to indicate which of the two was closer to their own personal views. For example, one question asked subjects to choose between the following two options: “(1) The content of high school science classes should be guided by community norms, standards, and preferences” or “(2) The content of high school science classes should be guided by scientific and educational experts.” The questionnaire ended with a set of standard demographic and political indicators.

Framing and Value Recruitment 33 Results We examined the effect of our experimental manipulation on subjects’ opinions about the teaching of ID and creationism in public schools. For simplicity’s sake, our analysis combines all of the participants who received the Dogmatic Darwinism and/or marketplace of ideas messages into a “pro-ID” group; all participants who received the Trojan Horse and/or progress of ideas messages were combined into an “anti-ID” group. Support for teaching ID was strongest in the pro-ID condition (about 0.5 points higher than the control group on a six-point scale—see Table 2.3). Surprisingly, support was slightly higher in the anti-ID condition than in the control group, although the difference was trivial. It thus appears that the marketplace condition was somewhat effective in moving opinion, but the progress condition was not. Further examination of demographic and political indicators showed something else: that subjects in the pro-ID condition were also significantly more conservative than subjects in the other two groups. This presents a serious problem for the internal validity of our study, since ID activists largely come from the political right. We think it unlikely that our manipulations affected such a stable disposition as ideology, so the most reasonable interpretation is that political ideology was not randomly distributed across conditions. Sure enough, when ideology is controlled, the once significant difference between the proand anti-ID groups becomes nonsignificant. The story does not end there, however. Looking at the multiple response measure, we found that participants in the pro-ID condition also favored including creationism in high school biology classes more than subjects in the other two groups (see Table 2.3). This significant effect (p = 0.03, two-tailed) persisted even after controlling for ideology. Thus, describing science as a “marketplace of ideas” while simultaneously condemning evolutionists as “dogmatic” resulted in greater support for including creationism in public school biology classes. Table 2.3 Experimental Results

Support teaching ID (1 = strongly oppose; 6 = strongly support)1 Teach creationism?2 Who should decide content of science curriculum? (1 = local community; 6 = experts)1 N

Control

Pro-ID

Anti-ID

2.75 (0.27)

3.29 (0.24)

2.97 (0.21)

15.00

43.10

23.00

4.30 (0.15)

3.53 (0.18)

3.97 (0.16)

42

58

62

Notes 1. Entries are mean values for the variable within each treatment group, with standard errors in parentheses. 2. Entries are the percentage of participants within each treatment group who favored teaching creationism in public schools.

34 Origins This particular statistic comes from an OLS regression using the two treatment conditions as dummy variables (control group excluded) and controlling for political ideology. Results are similar when logit is used. None of the individual value importance ratings showed a significant effect from our experimental treatment. Subjects in the pro-ID condition were, however, significantly more likely to hold the opinion that curricular content should be decided by the community, rather than scientists or education experts (p = 0.04, two-tailed). We interpret this finding as consistent with the claim that the pro-ID rhetoric prioritizes democratic, egalitarian values. To establish that such values mediate the effect of framing and issue labeling rhetoric on support for teaching creationism, we estimated a causal model. We regressed the dichotomous measure of support for teaching creationism on a dummy variable representing the pro-ID treatment. The unstandardized coefficient was positive and significant (0.19, p < 0.05). When attitudes about community control over education curriculum are included as a predictor, the coefficient for the dummy variable representing the experimental treatment decreases by about one-third (0.12, n.s.). This pattern of results suggests that opinions about the importance of community norms relative to expert opinion help to mediate the impact of the pro-ID rhetoric on support for teaching creationism (Baron and Kenny 1986). Summary The experiment has shown that the framing of the ID issue and key concepts under contention can have a noticeable impact on attitudes. Our experimental treatments combined labels for the entire ID issue with frames for science. In one treatment, the issue was labeled as a case of “dogmatic Darwinists” ruthlessly suppressing rival theories; the accompanying frame portrayed science as a “marketplace of ideas” where give-and-take among rival ideas is welcome. The other treatment portrayed the controversy as a surreptitious imposition of Christian fundamentalism; the accompanying frame cast science as a “progress of ideas” in which outdated notions are superseded by fresher perspectives. Although opinions about teaching intelligent design did not noticeably change, support for teaching creationism was greater in the ID-friendly condition. Moreover, judgments about the importance of relevant values corresponded to these treatments. In the ID-friendly condition, subjects were more likely to endorse the view that it is important to leave curricular decisions in the hands of local communities, rather than relying on “experts.” Such a view fits in well with the populist themes of the “Dogmatic Darwinism” and “marketplace of ideas” frames. Indeed, other fringe theories have been promoted by playing upon suspicions about intellectual elites (Garwood 2008). The causal analysis, furthermore, showed that the impact of treatments on support for teaching creationism could at least be partly attributed to their impact on populist values. These results fit with our argument that framing serves the strategic goals of value recruitment. Frames change the relevance and importance of values with

Framing and Value Recruitment 35 respect to issues; these changes can affect individual opinions about policy change. Our experimental treatments were drawn directly from actual rhetoric surrounding this issue. Broadly speaking, the kinds of egalitarian values featured in the pro-ID condition appeared in roughly half of the newspaper articles we examined. There is much we can do, however, to provide a more nuanced and fine-grained analysis of public rhetoric surrounding the ID issue. We need to classify, for example, the source of the rhetoric as partisan or objective. This would enable us to tell how many of the references to democracy, for instance, were made in the course of a persuasive argument versus a more pedestrian description. We would also, obviously, be able to tell whether the two sides to this dispute reference all of the values in roughly equal proportion. Our theoretical categorization of values as “contested” and “exclusive” would suggest otherwise. We also need to examine contingencies in the appearance of values. In other words, are particular values more likely to be mentioned when others have appeared in the same article? Preliminary analysis suggests that that educational excellence appears relatively frequently with scientific discovery and intellectual integrity, but is also contingent on the democratic process category. Both sides have tried to define what educational excellence should mean, with ID supporters stressing critical thinking and ID opponents emphasizing the value of peer-reviewed scientific findings. Therefore, much of the debate about educational excellence also touches upon the best way to establish what the process will be for determining whose notion of educational excellence should ultimately make it into the standards.

Discussion A common refrain among framing researchers is the lack of a common definition of framing. The discussion during the concluding session of our Washington DC conference revealed just how unsettled this question remains. Our two distinguished discussants, renowned for their own work on framing, expressed radically different views about this core concept. Robert Entman favored an expansive and inclusive definition of framing that has no trouble accommodating all of the diverse applications of the concept presented during the conference and included in this volume. Shanto Iyengar, however, advocated a narrower conception, and pointedly asked whether a term that is applied so broadly can hold its analytic power. Diversity in definition and application partly reflects the several disciplinary strains that gave rise to framing scholarship. We can think of at least six: the symbolic interactionist framework in sociology (Goffman 1986); social movement theory (Gamson and Modigliani 1989); legislative politics (Kingdon 1984; Riker 1986; Schattschneider 1960); mass media studies (Entman 2004; Iyengar 1991); behavioral decision theory (Kahneman and Tversky 1984); and organizational behavior (Lewicki et al. 2003). Modern political scholarship (e.g., Kinder and Sanders 1996) tends to emphasize, or reflect, perspectives two through five. That’s a lot of conceptual territory to manage.

36 Origins Some order is achieved by partitioning framing into types. Druckman helpfully distinguished between equivalence and emphasis frames (Druckman 2001): the former exemplified by Kahneman and Tversky’s seminal studies on the effect of manipulating a decision maker’s reference point; the latter referring to messages that focus an audience’s attention on a narrow aspect of a complex issue. We can also distinguish between issue frames and media frames (Price and Tewksbury 1997): the former referring to expressly persuasive depictions of a controversy; the latter referring to habitual styles of news coverage. We suggest a further partitioning of issue frames into issue categories and labels (summary depictions of the entire controversy), frames (tendentious definitions of key concepts in dispute), and analogies (comparison of the issue or disputed ideas to more familiar concepts). While this move further expands framing jargon, we hope that it also sharpens and extends our understanding of framing form and consequence. Thankfully, the other chapters of this volume do no damage to this conceptual framework. Much like Chapters 4, 10, and 11, our study focuses upon the aspect of framing whereby key concepts are defined. In terms of showing the impact of both issue categorization and framing by defining concepts, Chapter 4 explores the emergence of the partial-birth abortion debate by examining the efforts of a pro-life group—especially, their success in shifting debate by coining the term “partial-birth abortion.” In exploring the life course of a framing campaign by an interest group, Gerrity demonstrates how groups attempt to categorize an issue through their own summarizing of the issue at hand. Further, she demonstrates how these groups defined the debate in terms of “brutality,” and how this definition was more effective than the opposition’s “tragic and rare” terms. Similarly, Chapter 7 shows the success of another issue labeling campaign—the shift from the “inheritance tax” to the “death tax.” Chapter 9 focuses on the effectiveness of defining capital punishment in “innocence” terms, which is heightened after controlling for real-world factors that aren’t always taken into consideration. In addition, Chapter 8 demonstrates how the terms in which a president defines abortion can even have an impact on partisan identification of the public (given consistency with other elites’ definition of those terms). Other chapters focus less on what framing is than on what it does. Harris’s chapter concerns itself with the origins of frames, and how they serve elites’ political purposes, as does Ansley and Sellers’s chapter. The Wagner chapter might also be considered a “macro” conception of framing, not in its examination being focused solely on elites but rather in its findings that presidential and partisan framing of abortion have an impact on long-term partisan identification trends. Framing, like campaigning, is a dynamic and unfolding process, with multiple stages. An important goal of our project is to study the life course of value recruitment, from its origins in the ideological and strategic thinking of policy entrepreneurs, through its middle passage in the mass media, to its final impact on public attitudes. We have learned an awful lot about this issue from talking

Framing and Value Recruitment 37 to activists, attending public hearings, reading the scholarship and propaganda, and closely examining news media coverage. This qualitative enterprise has suggested an important role for framing within this debate, and the experiment lends support to that impression. At the same time, we believe we have learned something about the relation of framing to other styles of political communication. Among its other merits, framing helps communicators stake their claim to some of the most important resources our political culture offers: values.

Notes 1. An earlier version of this paper was presented to the 2007 annual meeting of the Midwest Political Science Association. This research is supported by National Science Foundation Grant #0352261 to the first author. We thank Maureen Beaver for her research on the history of the ID debate in Ohio, and Gifty Adutwum for help with data collection. Address all correspondence to: Thomas E. Nelson, Department of Political Science, Ohio State University, Columbus, Ohio 43210-1373. E-mail: [email protected]. 2. The Senate reconsidered its support for this amendment after critics pointed out veiled references to ID tenets. The amendment was deleted, and only included in the final bill language as a “Joint Explanatory Statement.” 3. Seelye (2007). 4. To determine relevant content for the purposes of this study, we used every Lexis Nexis article that was written about the Intelligent Design debate in Ohio between January 2000 and the middle of 2004. The January 2000 start date came from the rise of the issue in the public consciousness as a result of Am. Sub. HB 1, the bill that required the creation of academic standards. We chose the middle of 2004 as our end date as there was a notable lull in the debate after the passage of the lesson plan in March. This gave us a set of 96 articles from as close to home as Columbus, Ohio, and as far away as London and Ireland. 5. Our reliability was 0.73 using the Scott Pi formula or the Kappa formula. 6. “Schools Shouldn’t Mix Science, Creationism.” Columbus Dispatch (Ohio), January 23, 2002. 7. Candisky (2002a). 8. “Schools Shouldn’t Mix Science, Creationism.” 9. Candisky (2002b). 10. Lore (2002) 11. Clines (2002) 12. “Intelligent-Design Folly Is Going to Cost Plenty.” Columbus Dispatch (Ohio), March 14, 2004. 13. Columbus Dispatch, May 27, 2002. 14. “Intelligent Design Merits Equal Time.” Columbus Dispatch (Ohio), January 26, 2002.

Works Cited Baron, Reuben M. and David A. Kenny. 1986. “The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations.” Journal of Personality and Social Psychology 51(6): 1173–1182. Barton, David. 1993. The Myth of Separation, 3rd edition. Aledo, TX: WallBuilder Press. Behe, Michael J. 2006. Darwin’s Black Box: The Biochemical Challenge to Evolution, 10th anniversary edition. New York: The Free Press.

38 Origins Braman, Eileen and Thomas E. Nelson. 2007. “Mechanisms of Motivated Reasoning? Analogical Perception in Discrimination Disputes.” American Journal of Political Science 51: 940–956. Candisky, Catherine. 2002a. “Design vs. Evolution Discussion Monday.” Columbus Dispatch (Ohio) (March 8), sec. News. 1A. Candisky, Catherine 2002b. “Governor Sidesteps Evolution Argument.” Columbus Dispatch (Ohio) (10 February). Candisky, Catherine. 2006. “State Drops Analysis of Evolution; Reversal Represents a Setback to Backers of Intelligent Design in Science Classes.” Columbus Dispatch (Ohio) (February 15). Clines, Francis X. 2002. “Ohio Board Hears Debate on an Alternative to Darwinism.” New York Times (March 12), sec. A; National Desk. 16. Davis, P. William, Dean H. Kenyon, and Charles B. Thaxton. 1993. Of Pandas and People: The Central Question of Biological Origins, 2nd edition. Dallas, TX: Haughton Pub. Co. Dawkins, Richard. 2006. The God Delusion. Boston: Houghton Mifflin Co. Dembski, William A. 2004. The Design Revolution. Downers Grove, IL: InterVarsity Press. Druckman, James N. 2001. “The Implications of Framing Effects for Citizen Competence.” Political Behavior 23(3): 225–256. Entman, Robert M. 2004. Projections of Power: Framing News, Public Opinion, and U.S. Foreign Policy. Chicago, IL: The University of Chicago Press. Feather, Norman T. 1995. “Values, Valences, and Choice: The Influences of Values on the Perceived Attractiveness and Choice of Alternatives.” Journal of Personality & Social Psychology 68(6): 1135–1151. Feldman, Stanley and Marco R. Steenbergen. 2001. “The Humanitarian Foundation of Public Support for Social Welfare.” American Journal of Political Science 45(3): 658–677. Feldman, Stanley and John Zaller. 1992. “The Political Culture of Ambivalence: Ideological Responses to the Welfare State.” American Journal of Political Science 36(1): 268–307. Fish, Stanley. 2005. “Academic Cross-Dressing.” Harper’s Magazine 311 (1867): 70–72. Forrest, Barbara and Paul R. Gross. 2004. Creationism’s Trojan Horse: The Wedge of Intelligent Design. New York: Oxford University Press. Frankowski, Nathan. 2008. Expelled: No Intelligence Allowed (DVD). Premise Media. Gamson, William A. and Andre Modigliani. 1989. “Media Discourse and [US] Public Opinion on Nuclear Power: A Constructionist Approach.” American Journal of Sociology 95: 1–37. Garofoli, Joe. 2005. “Bush Pushes Very Hot Button; President’s Comments Embolden Anti-Evolutionists.” San Francisco Chronicle (August 8). Garwood, Christine. 2008. Flat Earth: The History of an Infamous Idea. New York: Thomas Dunne Books. Goffman, Erving. 1986. Frame Analysis: An Essay on the Organization of Experience. Boston: Northeastern University Press. Goren, Paul. 2001. “Core Principles and Policy Reasoning in Mass Publics: A Test of Two Theories.” British Journal of Political Science 31: 159–177. Gottman, John M. 1979. Marital Interaction: Experimental Investigations. New York: Academic Press.

Framing and Value Recruitment 39 Gould, Stephen Jay. 1997. “Nonoverlapping Magisteria.” Natural History 106 (March): 16–22. Grant, J. Tobin and Thomas J. Rudolph. 2003. “Value Conflict, Group Affect, and the Issue of Campaign Finance.” American Journal of Political Science 47(3) (July): 453–469. Gray, Barbara. 2003. “Freeze Framing: The Timeless Dialogue of Intractability Surrounding Voyageurs National Park.” In Roy J. Lewicki, Barbara Gray, and Michael Elliott, editors, Making Sense of Intractable Environmental Conflicts: Frames and Cases. Washington, DC: Island Press. Gwiasda, Gregory W. and Thomas E. Nelson. 2001. “Extremists in Our Midst: How Opponent Portrayals Influence the Balance of Values in Opinion.” Presented at the annual meeting of the Midwest Political Science Association, Chicago, IL. Holyoak, Keith James and Paul Thagard. 1995. Mental Leaps: Analogy in Creative Thought. Cambridge, MA: MIT Press. Humes, Edward. 2007. Monkey Girl: Evolution, Education, Religion, and the Battle for America’s Soul. New York: Ecco. Inglehart, Ronald. 1990. Culture Shift in Advanced Industrial Society. Princeton, NJ: Princeton University Press. Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago, IL: The University of Chicago Press. Jones III, John E. 2005. Tammy Kitzmiller, Et Al., Vs. Dover Area School District, Et Al. www.findlaw.com: US District Court, PA Middle District. Kahneman, Daniel, and Amos Tversky. 1984. “Choices, Values, and Frames.” American Psychologist 39(4): 341–350. Kinder, Donald R. and Lynn M. Sanders. 1996. Divided by Color: Racial Politics and Democratic Ideals. Chicago, IL: The University of Chicago Press. Kingdon, J.W. 1984. Agendas, Alternatives, and Public Policies. Boston: Little, Brown. Kristof, Nicholas D. 2004. “Living Poor, Voting Rich.” New York Times (November 3). Lakoff, George. 2002. Moral Politics: How Liberals and Conservatives Think, 2nd edition. Chicago, IL: The University of Chicago Press. Lau, Richard R. and David P. Redlawsk. 2006. How Voters Decide: Information Processing during Election Campaigns. Cambridge and New York: Cambridge University Press. Lewicki, Roy J., Barbara Gray, and Michael Elliott. 2003. Making Sense of Intractable Environmental Conflicts: Frames and Cases. Washington, DC: Island Press. Lore, David and Dispatch Science Reporter, 2002. “Evolution Debate Raises Issue: What’s ‘Science’?” Columbus Dispatch (Ohio) (17 March). Maio, Gregory R. and James M. Olson. 1998. “Values as Truisms: Evidence and Implications.” Journal of Personality & Social Psychology 74(2): 294–311. Mangels, John. 2002. “Darwin, ‘Design’ Face Off at Hearing.” Cleveland Plain Dealer (March 12). Miller, Jon D., Eugenie Carol Scott, and Shinji Okamoto. 2006. “Public Acceptance of Evolution.” Science 313 (August 11): 765–766. Miller, Kenneth R. 1999. Finding Darwin’s God: A Scientist’s Search for Common Ground between God and Evolution. New York: Cliff Street Books. Nelson, Thomas E. 2004. “Policy Goals, Public Rhetoric, and Political Attitudes.” The Journal of Politics 66(2): 581–605. Nelson, Thomas E. and Donald R. Kinder. 1996. “Issue Frames and Group-Centrism in American Public Opinion.” The Journal of Politics 58(4): 1055–1078.

40 Origins Pennock, Robert T. 2001a. “Why Creationism Should Not be Taught in Public Schools.” In Robert T. Pennock, editor, Intelligent Design Creationism and Its Critics: Philosophical, Theological, and Scientific Perspectives. Cambridge, MA: MIT Press. Pennock, Robert T., editor, 2001b. Intelligent Design Creationism and Its Critics: Philosophical, Theological, and Scientific Perspectives. Cambridge, MA: MIT Press. Petrocik, John R. 1996. “Issue Ownership in Presidential Elections, with a 1980 Case Study.” American Journal of Political Science 40(3): 825–850. Price, Vincent and David Tewksbury. 1997. “News Values and Public Opinion: A Theoretical Account of Media Priming and Framing.” In G. Barnett and F.J. Boster, editors, Progress in the Communication Sciences. Greenwich, CT: Ablex. Riker, William H. 1986. The Art of Political Manipulation. New Haven, CT: Yale University Press. Ruse, Michael. 2005. The Evolution-Creation Struggle. Cambridge, MA: Harvard University Press. Schattschneider, E.E. 1960. The Semisovereign People: A Realist’s View of Democracy in America. New York: Holt, Rinehart & Winston. Schwartz, Shalom. 1996. “Value Priorities and Behavior: Applying a Theory of Integrated Value Systems.” In The Psychology of Values: The Ontario Symposium, Vol. 8. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Scott, Eugenie Carol. 2004. Evolution Vs. Creationism: An Introduction. Westport, CT: Greenwood Press. Seelye, Katherine Q. 2007. “At G.O.P. Debate, Candidates Played to Conservatives.” New York Times (May 5). Wells, Jonathan. 2002. Icons of Evolution: Science or Myth? Why Much of What We Teach about Evolution Is Wrong. Washington, DC: Regnery Publishing, Inc.

3

Partisan Framing in Legislative Debates Douglas B. Harris

On September 28, 2007, President George W. Bush used his weekly radio address to explain his veto of the reauthorization of the State Children’s Health Insurance Program (SCHIP). Avowing support for SCHIP’s original intent to provide medical care to poor, uninsured American children, the President nevertheless objected that congressional Democrats’ expansion of SCHIP had transformed the program to include citizens well above the poverty line, some of whom already had private insurance. The Democratic response to the President’s radio address was given not by a member of Congress or other Democratic officeholder as is the norm but instead by a 12-year-old boy from Baltimore, Maryland. Graeme Frost, who had received medical coverage under SCHIP after a serious car accident, gave a personal account of his injuries, the difficulty of his recovery, and the assistance the SCHIP program provided his family throughout the ordeal. Wondering “why President Bush wants to stop kids who really need help from getting CHIP,” the boy said, “I just hope the President will listen to my story and help other kids to be as lucky as me.”1 This dramatic exchange between the President and the 12-year old boy came at a pivotal time in a long legislative saga that involved not one but two presidential vetoes, two failed override attempts in the House, and, eventually, passage when, in early 2009, new President Barack Obama signed expanded SCHIP into law. Amidst SCHIP’s legislative process was an intensely political story as Democrats and Republicans vied for public attention and approval. Public opinion, it seems, supported the Democrats’ SCHIP bill when it was thought to merely expand the successful program to more children and families in need. Still, President Bush and Republicans portrayed the bill as a far too costly attempt to stretch the program beyond its original mission to provide government-funded coverage for not only children but also immigrants (legal immigrants but not US citizens), adults, and even middle-class families who could afford (and some who already had) private health insurance. Although most observers believed that Democrats had the upper hand in the battle for public opinion, Republican pollster David Winston contended that public support was not so clear when voters could hear the Republicans’ side of the story (Drucker 2007). Both sides had calculated that success—political,

42 Origins legislative, or both—hinged on the terms of the debate and how the issue was framed for the press and the public. Message framing, as it is understood in the political communications literature, involves putting an issue or event into a broader rhetorical or symbolic context potentially consequential for attitude activation or change and, indeed, “value recruitment” (see Nelson, this volume) in the mass public (Entman 1993). In some respects, all political communication involves condensing a complex reality into a handful of thoughts, arguments, or symbols that affix, interpret, and convey meanings (Edelman 1964, 1971). But this is especially the case in legislative politics where concise and persuasive explanations are integral to winning electoral support among attentive publics back home (Fenno 1978) and where selective interpretations are key to building internal legislative coalitions (Smith 1984). Media-oriented congressional party leaders gain influence by shaping the opinion context within which their colleagues act. As such, an increasing amount of congressional leadership involves constructing and honing message frames and communicating them to the public and in legislative debate. Examining the legislative debates on SCHIP from Fall 2007 to early 2009 reveals three key dimensions of the Democratic and Republican arguments, each of which reveals an embedded framing strategy. First, tapping Graeme Frost to give the Democratic response to the President’s radio address was part of an overall Democratic effort to highlight many specific, personal stories of constituents who had or might benefit from SCHIP. Using such human impact or episodic frames is good for garnering press coverage as well as for putting issues into commonly understood storylines that help the public “determine the personal relevance of issues” (Neuman et al. 1992, pp. 61–62). Still, although human impact frames are good at gaining media coverage, their effects on public opinion are less clear. In early framing research, Shanto Iyengar (1987, 1991) demonstrated that news stories that portray issues in such isolated, often personalized, “episodes” rather than employing broader thematic frames tend to “shift responsibility from society to the poor” (Iyengar 1991, p. 67), potentially weakening arguments for policy change. Although most efforts to employ episodic frames use them in tandem with arguments about the broader impacts of a particular policy, to the extent that a policy argument becomes about a particular individual, it can distract from the main issue. Indeed, one unfortunate consequence of Democrats’ effort at dramatizing the issue by focusing on Graeme Frost was that, to discredit the Democrats’ position on SCHIP, conservative bloggers examined his family’s personal finances, arguing that they could afford private medical insurance. Second, perhaps the most important frame employed in the SCHIP fight was an assessment of who stood to benefit most from the program. Nelson and Kinder (1996) argue that public evaluations of the beneficiaries of policy proposals affect public support for that proposal, particularly when political elites employ these “group-centric” frames. “Americans,” they say, “reach political decisions on matters of policy as if they had first determined the moral qualifications of the intended beneficiaries” (Nelson and Kinder 1996, p. 1071).

Partisan Framing in Legislative Debates 43 Thus, Democrats were on politically advantageous ground when arguing that they wanted to help (and Republicans were willing to neglect) children. Because children, unlike adults, are more likely to be considered blameless for their social and economic situations, they are considered most deserving of public help. Democrats’ children-centered focus left many Republicans, like John Linder (R-GA), in the difficult political position of having “to point out that nobody on this side opposes children” (Congressional Record February 4, 2009, H970). Republicans focused instead on what they perceived as expansions and “loopholes” in the legislation whereby adults and recent immigrants could receive coverage under the legislation. In the Senate, Charles Grassley (R-IO) defended his amendment to take away states’ option to cover legal immigrants arguing that it was “all about priorities”: “The difference is that the additional low-income kids who get health coverage with my amendment are U.S. citizens” (Congressional Record January 27, 2009, S854). Taking on the elements of the bill that might allow for adults to get coverage, Senator Orrin Hatch (R-UT) said, “there is no ‘A’ in the CHIP program. There is only a ‘C’ and we all know what that ‘C’ stands for and it is not adults” (Congressional Record January 27, 2009, S855). Building on these arguments and trying to blunt Democrats’ criticism that they were insensitive to children, the Republican alternative offered by Senate Minority Leader Mitch McConnell (R-KY) was dubbed the Kids First Act. Third, sometimes legislators employ “linkage” frames that portray the connectedness of two or more issues that, without the frame, might otherwise be perceived as separate (Evans 2001). In the SCHIP debates, when the White House and congressional Republicans would complain about the cost of expanding the program, some congressional Democrats were quick to put the spending in light of the cost of the Iraq War. Particularly prevalent during the legislative debates in Fall 2007, this “linkage” frame led to some of the SCHIP debates’ most rancorous exchanges on the House floor and in the media (Armstrong 2007). Although linking SCHIP to Iraq had largely fallen out of the debate by 2009, Representative Jim McDermott (D-WA) nevertheless concluded his speech in support of the bill saying, “How anybody could vote against this, I have no idea, after you’ve [Republicans] wasted a trillion dollars on a war in Iraq, and have the real estate industry totally out control, and then you say to the children, you can’t see a doctor” (Congressional Record February 4, 2009, H970). Incredulous at the connection between SCHIP and Iraq, House Republicans applied a different linkage frame by arguing that expanded SCHIP should be seen in the context of some Democrats’ aims to move toward universal nationalized health insurance. Most notably, Republicans argued that expanding the program to include middle-class (not poor) children would be too costly and might lead to “crowd out” wherein families that had private insurance would opt instead for government coverage. Following this logic, Representative Pete Sessions (R-TX) said “this legislation moves us closer and closer to . . . a government-run health care program” (Congressional Record February 4, 2009,

44 Origins H934). More bluntly, Representative Steve King (R-IA) thought SCHIP “a foundation stone for socialized medicine in the United States” (Congressional Record February 4, 2009, H968). On SCHIP as with many other legislative fights, victory, defeat, and the longterm political costs thereof would be determined by the parties’ abilities to maintain message unity and the extent to which the press and public adopted competing frames. Despite the importance of message framing and public opinion strategies to contemporary legislative politics, we know very little about how congressional leaders seek to frame policy debates (but see Sinclair 1995, pp. 266–268; Evans 2001; Sellers 2000, 2002). Leveraging the “behind-thescenes” insights that can be gleaned by consulting archival collections, this chapter pulls back the curtain on the framing strategies of party leaders in the US House of Representatives.2 How does a framing strategy work in the legislative context? Do elites conceive of their strategies to explain and publicize legislative proposals as framing per se? To what extent do the strategic considerations and behaviors of congressional leaders comport with scholarly assumptions about elite framing? Examining internal poll and focus group reports, strategy memoranda, and notes from internal leadership meetings from House Democratic and Republican leaders from the 1980s and 1990s, this chapter sheds light on these questions by documenting some of the otherwise obscure behaviors of how House party leaders use polls to create and deploy message frames.

Polling and Framing in the House Parties Contemporary congressional party leaders employ polling and focus group techniques to monitor the public opinion environment, to assess their own party images, and to test language to aid in their efforts to frame media messages and legislative debates. But these public opinion techniques are of recent vintage in congressional politics. Traditionally, members of Congress, including leaders, tracked public opinion only informally by using opinion “surrogates” and/or tabulating letters, phone calls, and editorials (Herbst 1993). It was not until the 1980s that public opinion polling became a regular staple of congressional party leadership planning. Like increased attention to media politics (Sinclair 1995; Harris 1998), public opinion politics took hold under Speaker Tip O’Neill and increased in sophistication on both sides of the aisle thereafter reaching a climax in the mid-1990s (Jacobs et al. 1998; Harris 2005b). Thus, by the time of the Republican takeover of the House and Newt Gingrich’s innovative integration of public opinion politics, media messages, and legislative leadership, both House Democrats and Republicans had been polling somewhat regularly for over a decade. The advent of regular polling provided legislative leaders access to the kinds of precise opinion information (i.e., “market research”) necessary to “sell” and to frame issues and policy proposals. The Republican takeover of Congress was, in part, a result of Gingrich-directed efforts at wordplay. As early as the

Partisan Framing in Legislative Debates 45 Republicans’ January 1994 “message meeting,” then-Minority Whip Gingrich recommended to all House Republicans a memorandum written by public opinion expert Frank Luntz on the “key words, phrases and ideas” needed to achieve majority status.3 Indeed, one top Republican press aide claimed that Luntz was brought in regularly, specifically because he “specialized in how to focus group phrases” and because he “recommended language.” Building on this, the wording of the provisions of the “Contract with America” was famously poll- and focus-group tested by, among others, Luntz who continued to be integral to House Republican marketing efforts including, for example, on tax messaging into the late 1990s (see Schaffner and Atkinson, this volume).4 Although Republicans’ rhetorical efforts are better known, Democrats, too, had used public opinion techniques to determine media messages and word choice. As early as 1981, a focus group report advised Speaker O’Neill and Democratic leaders to describe “Social Security as a contract between the government and the people” and another pollster warned them of the negative public response to uses of the word “program” in describing their policy aims.5 Although these strategies and tactics, known as “crafted talk” (Jacobs and Shapiro 2000), are most frequently associated with Presidential politics, it is noteworthy that congressional leaders, too, “craft talk” on policies (Jacobs et al. 1998, p. 27; Harris 2005b). One Republican press aide described “market research” and “find[ing] the right way to phrase” policy proposals as a predominant use of polling in Congress, and a Democratic leadership aide agreed that “The polling was less for what we should do [than it was for] ‘Here’s the policy, but how can we explain it?’ ” Even more notably, congressional party insiders consciously use the “framing” metaphor to describe these efforts to explain policies to the press and the public. In the mid-1980s, for example, an increasingly media-conscious Speaker O’Neill tapped Representative Don Edwards (D-CA) to head a group to provide media strategy for the leadership and coordinate Democratic member contacts with reporters. Looking to more thoroughly integrate media and legislative strategies, the Edwards Media Group recommended that whenever “a key amendment is won or lost, the floor leader should promptly communicate with the press so that the news story can be framed with our point of view in mind.”6 Not only did Democratic leaders seek to integrate framing tactics with floor strategy but they also sought to proliferate framing efforts throughout the Democratic Caucus. By the late 1980s, the House Democratic leadership had established the Democratic Message Group, a leadership-selected group of media-savvy House Democrats who met daily to plan and coordinate media and legislative strategies. Future Clinton adviser Paul Begala, then a staffer for Majority Leader Gephardt, wrote a memorandum to Message Group staff entitled “Our Friend, the Message Frame” in which he explained the concept of framing (he wrote, “Think of the issues of the day as a colossal mural, and you have a two-foot by two-foot frame”) and he aimed “to take a crack at pushing some representative issues through the message frame.”7 Not merely a “talking points” document, Begala’s memo amounted to a training exercise—an “illustrative” model—for

46 Origins party communicators of how to frame legislative issues consistent with the party’s overall message. Republicans, too, use the framing metaphor to describe their efforts at integrating legislative debates with political communications. Whereas reframing broader ideology (e.g., moving America from a “liberal welfare state” to a “conservative opportunity society”) was central to Republicans’ efforts to take majority control of the House, this was accomplished day-to-day with a reframing of individual issues. For example, Newt Gingrich’s “Strategy Whip” organization looked to “develop language, code words and rhetoric which frame our themes on our terms.”8 In striking distance of the majority in 1993, House Republicans institutionalized a “Leadership Management Model” that built on Gingrich’s successes by “focus[ing] on key issues (possibilities are taxes, regulation, crime, labor, etc.) where the Leadership will be charged with developing a strategy and plan to frame and manage the issues to Republicans’ advantage.” Observing that when Republicans controlled the White House they had “to ‘play’ on every issue moving through the House,” one silver lining of losing the Presidency in 1992 was that it allowed for a new freedom in messaging: “Now, we can pick our fights better and gear up for the important fights, framed our way, rather than constantly playing on their timetable and their turf.”9 Even after they had achieved majority status, House Republicans continued to search for code-words and frames as components of both electoral and legislative strategy. In September 1997, Speaker Gingrich’s Planning Advisory Team, specifically designed to integrate legislative and media/political strategy, met to “To plan the development of those issues which will frame the 1998 to the year 2000 national debate.”10 More generally, when asked to describe House Republican leaders’ reliance on public opinion information during the 104th and 105th Congresses, a Republican communications aide said that a primary use of polls was to “frame arguments . . . to help us form arguments.” House leaders and other insiders also see framing as a necessary strategy to compete with a more media-savvy and media-capable White House, particularly during divided government (Kernell 1997; Maltese 1994). In early 1981, pollster Peter Hart told the House Democratic Caucus that because “Ronald Reagan has changed the dialogue on you,” Democrats needed to “re-focus and reframe the dialogue on the whole economic area.”11 Once Reagan’s honeymoon was over, Democrats looked for opportunities to reassert their legislative priorities and to do so on their own terms. Planning party strategy in 1981, an aide to Speaker O’Neill wrote, “In framing the debate and defining the stakes on key issues, there are a number of themes the Democratic Party should consistently stress. As legislative issues unfold and opportunities to communicate with the electorate arise, Democrats should be prepared to forcefully and graphically focus the dialogue on these themes.”12 During the Clinton years, Gingrich used his extraordinarily high media visibility to frame public opinion. Indeed, even when Republican members pressured Speaker Gingrich to suspend his televised daily press conference, for fear that they were detrimental to the Republican image, some in Gingrich’s inner circle believed

Partisan Framing in Legislative Debates 47 that the Speaker’s arguments were making inroads in public opinion. One top Republican press aide argued (unsuccessfully) for maintaining the Speaker’s daily press conference thinking “they worked wonderfully” citing the Speaker’s success in “framing the national news from the House of Representatives.” Though most are skeptical that Congress can out-communicate the President, some congressional party insiders believe out-framing the White House is possible and, usually, worth attempting. One important consequence of all of this attention to opinion information is that it impresses upon congressional leaders the volatility of public opinion and the sensitivity of public support to framing effects. When pollsters and consultants show leaders and other legislators rhetorical paths to victory, those presentations often juxtapose the “good” messages with the “bad” and demonstrate how word choice is often the difference between victory and defeat.13 For example, when Democratic leaders planned strategy on the trade issue in the mid-1980s, a DCCC poll revealed that a majority (“about 54 or 55 percent”) of Americans favored “protectionism” but that support “drop[s] off [on] the issue” when Americans heard free trade arguments.14 DNC polling similarly stressed the importance of “how [trade] is framed for the voters” and that “it is very easy to see the ground shift under your feet if the context of the debate changes.”15 By the same token, when attempting to reform Medicare in 1995, House Republicans were advised by pollsters of the dangers of employing the word “cut” or even “change” when describing their Medicare efforts. Instead they were advised to say that they would “slow the rate of growth” of Medicare spending in their efforts to “preserve, protect, and improve” the program (Maraniss and Weisskopf 1996). Citing the “slow the rate of growth” frame as the most important element of the rhetorical battle over Medicare (adding that “the polling data was stunning” on this point), one Republican press aide noted that “for a period of some months, every time a reporter anywhere in the country used the word ‘cut,’ [Gingrich, Tony Blankley, and Ed Gillespie] got on the phone” to correct the coverage. Increasingly sensitive to the volatility of public opinion and, specifically, to framing effects, contemporary congressional leaders not only hone their words but they also police media coverage seeking to influence what frames are communicated to the public.

The Strategic Use of Message Frames in Congress Not only do congressional leaders and leadership staff think of their communications efforts as “framing” per se but they also understand (both intuitively and from their access to opinion information) many of the potential impacts of framing. Whereas political communications scholars’ examinations of media coverage and public opinion have identified myriad framing effects, comparatively little attention has been paid to whether or not elites consciously employ these frames. Confirming that legislative leaders do, in fact, think of their communications efforts in these terms, instances of congressional party leaders’ efforts to employ what the political communications literature refers to

48 Origins as causal frames, group-centric frames, episodic frames, and conflict frames follow. Causal Frames Policy alternatives can be framed in ways that offer interpretations and theories about the causes of social problems (Iyengar 1987, 1991), thereby implicitly suggesting policy solutions. Take, for example, the issue of drug abuse and drug trafficking. In an analysis of how congressional committees handled these issues, Baumgartner et al. (1993, p. 669) suggest that elites might “differentiate between enforcement (including interdiction and border controls) and treatment (including education, rehabilitation, and health topics).”16 Implicit in a politician’s selection of these competing frames is, first, an assumption about the underlying causes of drug abuse and drug trafficking and, second, prescriptions for appropriate policy responses that play to the divergent strengths in public opinion of Republicans and Democrats, respectively. In the late 1980s when drug abuse was a particularly salient issue, Democrats’ internal polling suggested a slight advantage over Republicans in handling the drug issue especially when the issue was framed as a social issue (e.g., in terms of treatment) rather than as a crime issue.17 Consistent with this advice, the Democratic Caucus and the Democratic Message Board organized a coordinated eventsand-message campaign on the issue in which they sought to frame the drug issue more broadly, playing to both Democrats’ ideological commitments and public image strengths. Observing “the critical importance waging the drug war has in many of our districts,” the Democratic Caucus encouraged members to “undertake a coordinated ‘drug war’ event”; suggested events included: “Tour a drug rehabilitation facility; meet with at-risk youth in school drug programs; view hospital treatment for cocaine-addicted babies; and travel your districts with drug enforcement professionals.”18 Another example of congressional parties using a causal frame was in the area of social policy. Nelson et al. (1997, p. 568) write: “a social policy such as welfare may be framed as a key element of the ‘cycle of poverty’ that reinforces degenerate morals and behavior patterns or as a ‘helping hand’ or ‘ladder out of poverty’ for those willing to better their condition.”19 Confronting social policies in the mid-1980s, House Republican leadership staffers considered advancing what they termed “symbolic issues” (“welfare . . . or abortion or school prayer?”) or reframing the causal relationship between social conditions and their direst consequences. One staffer recommended that congressional Republicans “Redefine assumptions: Poverty causes crime. Maybe crime causes poverty. Maybe we quit talking about money [for program beneficiaries] and start talking about values.” Echoing conservative thinker Charles Murray’s (1984) thesis that, rather than helping recipients, the welfare state creates a permanent underclass, it is notable that party leaders had recommended some of Murray’s work to House Republican press secretaries as early as 1982.20 Republicans found this causal formulation so effective that they were still using it a decade later. Assessing

Partisan Framing in Legislative Debates 49 the “central[ity]” of “Gingrich’s scientifically honed oratory,” one columnist reported, “Ronald Reagan attacked poor people for abusing programs; Gingrich attacks the programs for abusing the poor. Asked about the tactic, Gingrich smiles and replies, ‘You cracked the code’ ” (DeParle 1996). Group-centric Frames As exemplified in the debate over SCHIP, public opinion regarding policies is often “group-centric” and contingent upon public support for the policy’s likely beneficiaries (Nelson and Kinder 1996). A good example of congressional leadership sensitivity to “group centrism” in public opinion is the House Democratic efforts throughout the 1980s to advocate policies to win back middle- and working-class voters they had lost to Reagan and the Republicans in the 1980 elections. Pollsters and consultants told Democrats that many voters increasingly perceived them as defenders of the poor and not of the middle class. At a time when President Reagan was apt to cite examples of “welfare queen” cheats and when Democrats were trying to regain ground with middle- and working-class voters, pollsters advised Democratic leaders “to structure more of the arguments along the 90/10 fashion that works for us instead of just being the defender of the poor; we need to be able to set it up as their government and their [the Republicans’] government is the government of the rich.”21 Framed this way, another observed, Democrats could turn political battles into the “rich vs. everyone else,” forcing Republicans to defend the wealthy beneficiaries of their policies rather than the Democrats defending the poor.22 In the words of another Democratic consultant, party leaders should talk about policies in “big guy/little guy terms” and frame issues in regards to “who benefits.”23 Of course, such word choice and selective “group-centric” framing appeals have been major elements of Democrats’ efforts to court working-class and, especially, middle-class voters from the Reagan to the George W. Bush Administrations. Refocusing on the middle class became a dominant theme in Democratic policy and rhetoric throughout the 1980s and into the 1990s. During the 1989 fight over a Bush Administration-supported capital gains tax cut, new Majority Leader Richard Gephardt (1989) sought to frame the argument in terms of the “rich vs. the rest,” writing in a Washington Post editorial that the passage of the capital gains cut would help to “fulfill Leona Helmsley’s dream that ‘only the little people pay taxes.’ ” Distilling the complex legislation into a familiar Democratic frame, Gephardt wrote: “After you sort through the technical jargon and the economic gobbledygook, the choice facing Congress comes down to this: under the Bush proposal, the rich get a tax cut paid for by the middle class. Under our proposal, the middle class gets an IRA paid for by the rich.” By the 101st Congress, Democratic leaders and the Democratic Message Board repeatedly encouraged members to focus on family security issues and what they called the “Middle Class Squeeze.”24 In fact, nearly every letter the Democratic Message Board sent House Democrats in the 101st

50 Origins Congress ended with a reminder of the party’s overarching message: “As always, our objective is to communicate the message in the context of the basic theme that Democrats are fighting for working families, economic security, and rebuilding America.”25 These efforts to become the “party of working, middle-income, middle class people”26 dovetailed with middle-class anxieties about the Bush Administration’s inattention to domestic policies and “kitchen table economics.” A House Democratic leadership aide (one of many who later joined the Clinton Administration) believed that these reframing “efforts from 1989 to 1992 laid the groundwork for Clinton in ’92.” Indeed, Democrats continued to employ this “rich vs. the rest” formulation during the George W. Bush Administration, particularly as they argued against Bush’s 2001 tax cuts. Episodic Frames To what extent are issues framed as discrete insular cases and to what extent are they considered broader social conditions that might require public attention? Examining the impact of “episodic” and “thematic” frames in television news stories, Iyengar (1991, p. 14) found that television’s emphasis on the “case study or event-oriented report” (“episodes”) as opposed to stories that “plac[e] public issues in some more general or abstract context” (“themes”) impacts the individual’s sense of who is responsible for social problems. In the early battles between House Democrats and the Reagan Administration, Speaker O’Neill and the top leadership made the effects of Reagan’s budget cuts the centerpiece of an effort to refurbish the Democrats’ image and advance themes that redefined the “us and them” divisions (see below) between the two parties.27 The substance of the fight involved Democrats’ efforts to highlight individual cases to illustrate “hardships” and “human losses” as “the real story of Reaganomics.”28 Throughout 1982, House Democrats were invited to spotlight individual “hardships” in their districts in one-minute speeches and media events while committee and subcommittee chairs were asked to hold “road show” hearings where they would invite testimony from citizens harmed by budget cuts in the hopes of getting “maximum coverage” from print and broadcast outlets.29 Democrats used these individual “episodes” to question the “fairness” of the Reagan program as they ran a coordinated message campaign for the 1982 elections, the tag-line of which was “It’s Not Fair, It’s Republican.” If the press prefers dramatic episodes, emotional appeals, and compelling personal stories, then these efforts made sense as a strategy to obtain media coverage. But, if the broader aim was to move public opinion and affix responsibility for these conditions to the Reagan Administration and its budget policies, experimental studies of the impact of “episodic” framing throw doubt on the effectiveness of this strategy. Iyengar (1991) found, specifically, that “the predominant news frame for poverty has the effect of shifting responsibility from society to the poor. Were the networks to increase the level of thematic framing in their coverage of poverty, Americans might be more apt to consider society

Partisan Framing in Legislative Debates 51 or government rather than the poor responsible” (p. 67). Of course, House Democrats easily could (and did) make their case in more systematic and abstract ways (e.g., by reporting increases in the number of Americans below the poverty line), but they thought these personal stories (episodic frames) more likely to get press, particularly broadcast, attention. House Democrats and Republicans alike thought the “episodic” framing strategy successful (perhaps because both equated media coverage with public opinion). Whereas Democrats continued these efforts to highlight the personal effects of Reagan’s policies (see Barry 1990, p. 615), House Republican leaders were convinced that media coverage of such episodes were devastating to their party’s image. For example, when Newsweek ran a cover story on “Reagan’s Polarized America” in April 1982, Minority Leader Michel wrote a letter to the magazine’s editor complaining that the article’s “‘rich-poor’ theme” was “simplistic, distorted and unacceptable in a publication with a well-deserved reputation for trying to portray complex situations clearly but in all their complexity.” Michel also wrote a Dear Colleague letter to all House Republicans in which he acknowledged the effectiveness of the article: “The Newsweek cover and the illustrations inside . . . will do more than our Democratic friends could ever do toward reinforcing the rich vs. poor imagery that I find so repugnant and unfair.”30 When CBS broadcast Bill Moyers’s similarly themed special, People Like Us, a few weeks later, Michel’s office warned Republican Press Secretaries that the Moyers program “surely had an impact” on the public “that will have a very adverse effect on all Republicans.”31 Whether or not they were correct, Republicans believed that these and other Democratic efforts were taking a toll on the party image, and they sought to answer in kind. Citing the Democrats’ “initial success” in opposing the Reagan budget through the “personalization of the impact of spending limitations” (“President Reagan in particular and Republicans in general were cast as personally responsible for each crying baby, each aging oldster in America”), a Republican Policy Committee working group headed by Dick Cheney recommended that Republicans “use the same approach on Speaker O’Neill and House Democrats . . . Each member ought to be able to locate ‘horror stories’ in his or her own district that are tied in theme or general subject matter to our issues—immigration, bankruptcy, reform, education, public employees vs. taxpayers, etc.”32 Although the ultimate effects of such “episodic” frames are unclear, Republicans nevertheless moved to mirror this strategy out of a perception (perhaps a misperception) that Democratic efforts had been effective. Conflict Frames Another common framing technique is the conflict frame. Considerably more attractive to reporters and more prevalent in media stories than they are among the mass public, frames that divide “protagonists into ‘us’ and ‘them’ ” exist in both public opinion and elite communications (Neuman et al. 1992, pp. 62, 64–66). Indeed, in the competitive party context of Congress and especially in

52 Origins a closely divided country in which both parties must appeal to a large percentage of independent “swing voters,” party leaders’ efforts to redefine both parties’ images—to frame “us and them”—are especially critical. Both Democrats and Republicans employ “conflict frames.” Characterizing how Republicans explain political conflicts—“They explain the world in terms of who belongs in which party: If you’re a Big Government bureaucrat, a welfarequeen, a liberal elite who has contempt for middle class values and absolute faith in permissive social policy and fuzzy-headed, weak foreign policy, you belong to Them [Democrats]. But if you believe in rugged individualism, national strength and Old Fashioned American Values, you’re one of US [Republicans]”— Gephardt staffer Paul Begala described “re-draw[ing] the lines between Us and Them” as “the essence of message-making.”33 A perennial concern of legislative party leaders, one of the avowed goals of House Democrats’ 1994 Message retreat, for example, was to “develop a message frame to accomplish [party message] efforts” including “what do we want communicated about Democrats vs. Republicans; our agenda vs. theirs.”34 By the same token, the Gingrich-led Conservative Opportunity Society (COS) 1987 agenda included finding “wedge” and “magnet” issues and “polarizing on our terms” in part by “defining the left in the House and in the country.”35 At a 1988 COS strategy session, Gingrich looked to highlight legislative battles and develop media themes to reconstruct the two parties’ images in ways that “the news media and the voter finds true and compelling the longer it is studied.” Asserting that there is “a humanitarian Republican Party” and that the Democrats constitute “a loony left,” Gingrich argued that “A listening, solution oriented, humanitarian Republican Party can beat a corrupt, Labor boss, loony left, big city machine Democratic Party. Repetition, debate and definition are the keys to communicating our message.”36 Although achieving majority status was the primary aim of these early Gingrich efforts, the framing of inter-party conflicts “in the House and in the country” continued after the Republican takeover. Looking to the 1996 elections and believing that “people need large themes around which they can orient their own particular needs and emphases,” the House Republican Conference’s “public opinion research agenda” included a search for the “major definitions of the differences between us and them [that have] the greatest power to help people understand what is at stake.”37

Conclusions and Future Directions Media and framing strategies are increasingly prominent, if heretofore underappreciated, aspects of contemporary congressional politics. As research proceeds in this area, additional themes identified by archival and interview research confirm the value of, at least, pursuing two new directions in the political communications literature on framing. First, any close look at congressional framing and message politics should add fuel to scholars seeking to understand the competitive nature of elite framing, a particularly important new direction in the political communications literature.

Partisan Framing in Legislative Debates 53 In recent research, Sniderman and Theriault (2004; see also Druckman 2001 and in this volume) point out that elite-sponsored frames are frequently countered by the frames of other elites, thus moderating or negating the overall effects of framing on public opinion. Although the presence and competing impacts of multiple frames deserve attention in most communications contexts, such competition is a defining characteristic of congressional party communications as majority and minority leaders must compete not only regularly with one another but often with competing voices from within their parties and not infrequently with a White House which has more media expertise and commands far more media attention. Archives and interviews reveal that internal congressional messaging efforts both account for the competitive nature of framing and embed the dynamism of message-making into message strategies. Understanding that their messages, if they will be heard at all, will be heard in a comparative context, congressional message-makers have pollsters and focus group experts test their best arguments against those of their opponents before “going public.” For example, when seeking to commit House Republicans to the party-honed message “playbook” just prior to the August 1998 recess, leaders reported on polling that contrasted the “positive messages” of both parties and revealed the superiority of the Republican message. Showing Republican legislators not only evidence of the overall success of the messages but also the impact of adding both parties’ “negative messages,” the “Playbook” provided “talking points” that were “designed to offer our Members a taste of what the Democrats will say, and suggested rebuttals.”38 Similar dynamic and competitive message-making is a constant consideration in congressional party communications. One Republican communications aide said, You’re always dealing with a matrix of four: . . . what I say about my proposal, what they will say about my proposal; what I will say about their proposal, what they will say about their proposal . . . Message is often “us on us” and knowing what they are going to say about us . . . so we can get ready. To be sure, the difficulty of congressional message-making mounts exponentially when the “they” is the President. Even during the 104th Congress, when Gingrich’s visibility took the public Speakership to its highest levels (see Harris 1998), the Speaker could not match the President’s visibility (Seelye 1995). As another Republican communications aide said, “Most Americans, God bless them, don’t really care what Congress thinks . . . The megaphone of the White House is . . . consistently underestimated by people here on the Hill . . . We could do anything [but] . . . the White House would trump, define and redefine our message for us. We never get the last word.” A second new direction might be toward a better appreciation of the visual aspects of congressional media politics. Recent political communications research advocates broadening the research focus “to include symbolic

54 Origins elements” of framing (Nelson and Kinder 1996, p. 1073). Archival evidence confirms that not all of congressional parties’ efforts at framing debates involve merely “crafting talk,” but there also are elements of symbolism and non-verbal gestures. Republicans, for example, spent considerable time and attention planning and making gestures aimed at overcoming their party’s image as “insensitive” to the problems of the poor. To blunt the impact of the Reagan budget cuts and the emergent Republican image as “uncaring,” Dick Cheney’s RPC task force advised House Republican members to “form ‘partnerships’ within [their local] community to take care of social/welfare needs through private initiative” and even to personally participate in “Foster Grandparents” programs.39 As late as 1997, GOPAC was still trying to help Republicans repair the damage, and, notably, the image makeover was to be accomplished not only through GOPAC’s better known wordplay but also through the careful deployment of “visuals” and “symbols.” Thinking in terms of “wedge” and “magnet” visual images, GOPAC coached Republicans on how to visually and symbolically communicate their way to a better party image, including extending the “charity” and “volunteerism” efforts that dated back to the Cheney task force.40 Similar innovations on the Democratic side were inspired, in part, by the need to compete with the White House. One Democratic leadership aide referred to the visual successes of Mike Deaver in the Reagan White House and asserted the House Democrats made numerous “attempts to get backdrops” to dramatize their policy issues. Another Democratic press aide told of a “message retreat” held at the Greenbriar resort where Democratic members and staff watched a video of Ronald Reagan making a televised appeal as a former Reagan media adviser “talked about crafting the message and about how it was more than just the words you said.” One place where the Democrats sought to compete with visuals and backdrops was by turning the Speaker’s “ministerial” function of signing bills into what Speaker Wright described as “good visible, unique photo opportunities”; an internal planning document for such a ceremony for Wright’s signing of a “GI Bill” in Statuary Hall in May 1987 reveals that the guest list included members of veterans organizations and the military, and a handwritten notation on the document read “50 persons; Flags!”41 Believing Republicans to be “highly successful in organizing their message around themes laden with emotion,” Democrats were advised by pollsters to “advance Democratic symbols to communicate our belief in God, strength of family, and law and order, and create opportunities every day to associate Democratic Members with these values.”42 Although they are much more difficult to analyze, these non-verbal efforts to frame debates likely are no less important to party image construction and debate framing. As Nelson and Kinder observe, “Visual images, icons, analogies, slogans, and catchphrases all occupy important positions in a frame’s architecture” (1996, p. 1073). Describing the comprehensive approach Newt Gingrich took to framing, a Republican press aide noted that “Newt tried to give them the whole package: the visual as well as the verbal.”

Partisan Framing in Legislative Debates 55 With the rise of legislative partisanship in the last three decades and particularly during divided government, media politics has grown in importance in Congress and has become a primary function of legislative party leadership. And, to be sure, contemporary congressional communications has grown in sophistication during that time as well. What has lagged, however, is our understanding of these developments and their consequences. As scholars of Congress and of political communications turn more attention to this subject, the archival evidence cited in this chapter provides some initial “behind-the-scenes” confirmation that legislative party leaders engage in framing as a conscious strategy. Frames exist not only in media coverage and in public opinion but also as explicit components of political elites’ media and legislative strategies.

Notes 1. The text of the President’s Radio Address is available in the Weekly Compilation of Presidential Documents, http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi? dbname=2007_presidential_documents&docid=pd01oc07_txt-3; the radio response is available at http://www.democrats.org/a/2007/09/twelve-year-old.php, accessed February 7, 2009. 2. Because revelations of media efforts diminish their effectiveness and subject political elites to charges of manipulation, politicians are apt to discount their poll use and to hide their media efforts from public view. The few studies that have examined presidents’ and interest groups’ framing efforts have used archival resources to uncover evidence of the strategic considerations normally hidden from public view (Jacobs and Shapiro 1994; Druckman et al. 2004; Gerrity, this volume). Only recently have political scientists begun leveraging the unique insights and data that can be gleaned by consulting congressional archives collections (Frisch and Kelly 2003; Harris 2005a). The collections cited herein include: the Thomas P. O’Neill Papers, Special Collections, John J. Burns Library, Boston College, Chestnut Hill, Massachusetts (TPO); the Jim Wright Collection, Mary Couts Burnett Library, Texas Christian University, Fort Worth, Texas (JCW); the Papers of Thomas S. Foley, Special Collections, Washington State University, Pullman, Washington (TSF); Records of the House Democratic Caucus, Manuscript Division, Library of Congress, Washington, DC (HDC); the Robert H. Michel Collection, Dirksen Congressional Research Center, Pekin, Illinois (RHM); the Richard A. Gephardt Papers, Missouri Historical Society, St. Louis, Missouri (RAG); and the Papers of Representative Newt Gingrich, Special Collections, University of West Georgia (NLG). I wish to acknowledge the support of the Everett McKinley Dirksen Congressional Research Center and the Caterpillar Foundation and Loyola College in Maryland which supported travel to archives and interviews with congressional staff. Thanks also to Speaker Newt Gingrich and the staff at newt.org who allowed me early access to the Gingrich papers as well as to Speaker Jim Wright and congressional staff members interviewed under the condition of anonymity; unattributed quotations are from these interviews. 3. Frank Luntz, “From Minority to Majority: A Strategy for the Republican House Leadership,” F “Planning/Strategy Memos” Box 2629, NLG. 4. Frank Luntz to Interested Parties, “Re: Talking About Taxes,” June 24, 1997, Box 2211, NLG. 5. See Hart and Associates, “A Report on a Focus Group Discussion Session Conducted in Portland, Oregon,” October 1981, O’Donnell Files Box 12; Vic Fingerhut to Congress Leaders “Subject: What Democrats Should Do Tonight, Tomorrow, the

56 Origins

6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19. 20.

21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

31. 32.

Day After, and Next Week in Response to the President’s Speech and the GOP PR Follow-Up,” September 24, 1981, F “Democratic Response to the State of the Union Address, December 1983-January 1984,” Press Relations Box 18, TPO. Edwards Discussion Group Memo, July 15, 1986, F “Don Edwards Discussion Group” Kirk O’Donnell Files, Box 1, TPO. Paul Begala to “Message Staff,” “Re: Our Friend, the Message Frame,” May 13, 1991, Press Files Box 418, F “Press Files/1990/Democratic Strategy,” TSF. F “Whip Organization: ‘Staying Ahead of the Curve,’ ” Box 1063, NLG. F “Legislative. Special 103rd Reform. Republican Leadership Management Model” Legislative—Special Subject, Other Special Subjects Box 6, RHM. Agenda, The Speaker’s Planning Advisory Team, F “Management Meeting 9/18/1997,” Box 2422, NLG. Transcript, Democratic Caucus Meeting March 25, 1981, Party and House Leadership Box 193, TSF. Undated memorandum (clearly from 97th Congress), F “Leadership Political Consultants,” Kirk O’Donnell Files, Box 12, TPO. See, for example, Vic Fingerhut to Congressional Leaders, September 24, 1981 and November 15, 1983, F “Democratic Response to the State of the Union Address, December 1983–January 1984,” Press Relations Box 18, TPO. Transcript, Democratic Caucus Meeting August 13, 1986, F “99th Meetings 13 August 1986” Box 25–4, HDC. “Principal Findings of the 1985 DNC Public Opinion Survey,” F DCCC 1986, Box 30 “DCCC, Tony Coelho 1982,” TPO. For a discussion of the “oscillation” in the use of these frames regarding drug abuse over time, see Baumgartner and Jones (1993, pp. 152–156). See, for example, Democrats for the 1990s, “Overview of Key Findings: A Survey of Voter Attitudes” Press Files Box 417, F “1989/Polling Information,” TSF. Steny H. Hoyer and Vic Fazio to Press and Scheduling Coordinators, September 7, 1989, F “Press Files/1985/Drugs,” Press Files Box 416, TSF. This characterizes Gamson and Lasch (1983). Handwritten notes from “Leadership Staff Meeting, 3/7/86,” F “Subject: Republican Leadership Staff,” Press Series Box 30, see Johnson to Republican Press Secretaries April 7, 1982, F “RPS 4/7/82 Economic Conditions,” Press Series Box 25, RHM. Transcript, Democratic Caucus Meeting March 25, 1981. Patrick Caddell, Cambridge Survey Research, February 20, 1982, F “Political Agendas for House Democrats, 1982,” Press Files Box 13, TPO. Fingerhut, “Themes and Symbols for a Communications Strategy,” TPO. Democratic Caucus “Dear Colleague” on “Middle Class Squeeze” September 17, 1990, Press Files Box 419, TSF. Vic Fazio to Democratic Message Board, April 23, 1990, F “1990 Budget,” Press Files Box 418, TSF. “1990 Speech Themes,” Box 420, TSF. 1981 Media Plan Notes, TPO. Tip O’Neill, Jim Wright, Tom Foley, and Bill Alexander, “Dear Colleague” July 28, 1982, F “Reagan, 1981–86,” Press Relations Box 20, TPO. “Hearing Checklist” Press Files, TPO. Robert H. Michel, “Letter to the Editor” March 31, 1982, and “Dear Colleague Letter” April 1, 1982; F “Leadership 97th Dear Colleague April 1, 1982 Newsweek Article,” Leadership Files Box 4, RHM. F “Republican Press Secretaries: 4/26/82 Bill Moyers’ CBS Special,” Leadership Files Box 4, RHM. Tim Wyngaard, Policy Committee to Gregg/Cheney Working Group October 18, 1983 “Re: Arrogance of Power,” F “Subject. Republican Policy Committee,” Press Series Box 30, RHM.

Partisan Framing in Legislative Debates 57 33. Begala, “Re: Our Friend, the Message Frame.” 34. Talking Points Document, F 5 “House Democratic Communications Meeting [Message Retreat/Strategy Retreat 1994 1 of 2],” Box 417, RAG. 35. Proposed COS Vision Agenda June 2, 1987, F “COS 87,” Box 2353, NLG. 36. “House GOP Themes,” May 10, 1988, F “Strategy Group Spring 1988,” Box 1063, NLG. 37. Public Opinion Research Agenda for 1996 [dated April 15, 1996], F “Strategy and Planning 1996 [1 of 3],” Box 2601, NLG. 38. “Communications Playbook,” Draft July 20, 1998, and “Impact of the Playbook: Public Opinion Reaction in Targeted Districts,” F “Communications Playbook,” Box 2423, NLG. 39. Johnson and Wyngaard to Michel and Cheney April 14, 1982, “Subject: Rich/Poor Issue, Teeter Poll and House Republican Responses,” F “Memoranda, 1981–88 (2),” Press Series Box 1, RHM. 40. “GOPAC: Communications Strategies for our Governing Majority” July 1997, F “Communication GOPAC,” Box 2420, NLG. 41. F “GI Bill Ceremony,” CM Files Box 3–44, JCW; see also Sinclair 1995. 42. “Leadership Options” June 29, 1989, F “1989/Leadership Transition,” Press Files Box 417, TSF.

Works Cited Armstrong, Drew. 2007. “Democrats to Retry SCHIP Expansion.” CQ Weekly (October 22): 3091–3092. Barry, John M. 1990. The Ambition and the Power. New York: Penguin. Baumgartner, Frank R. and Bryan D. Jones. 1993. Agendas and Instability in American Politics. Chicago, IL: The University of Chicago Press. Baumgartner, Frank R., Bryan D. Jones, and Jeffery C. Talbert. 1993. “The Destruction of Issue Monopolies in Congress.” American Political Science Review 87: 657–671. DeParle, Jason. 1996. “Rant, Listen, Exploit, Learn, Scare, Help, Manipulate, Lead.” New York Times Magazine (January 26). Drucker, David M. 2007. “DCCC: SCHIP Hurts GOP.” Roll Call (November 6). Druckman, James N. 2001. “The Implications of Framing Effects for Citizen Competence.” Political Behavior 23(3): 225–256. Druckman, James N., Lawrence R. Jacobs, and Eric Ostermeier. 2004. “Candidate Strategies to Prime Issues and Image.” The Journal of Politics 66: 1180–1202. Edelman, Murray. 1964. The Symbolic Uses of Politics. Urbana: University of Illinois Press. Edelman, Murray. 1971. Politics as Symbolic Action: Mass Arousal & Quiescence. New York: Academic Press. Entman, Robert M. 1993. “Framing: Toward Clarification of a Fractured Paradigm.” Journal of Communication 43(4): 51–58. Evans, C. Lawrence. 2001. “Committees, Leaders, and Message Politics.” In Lawrence C. Dodd and Bruce I. Oppenheimer, editors, Congress Reconsidered, 7th edition, 217–243. Washington, DC: CQ Press. Fenno, Richard F., Jr. 1978. Home Style: House Members in Their Districts. Boston: Little, Brown. Frisch, Scott A. and Sean Q. Kelly. 2003. “Don’t Have the Data? Make Them Up! Congressional Archives as Untapped Data Sources.” P.S. Political Science and Politics 36: 221–224.

58 Origins Gamson, William A. and Kathryn E. Lasch. 1983. “The Political Culture of Social Welfare Policy.” In Shimon E. Spiro and Ephraim Yuchtman-Yaar, editors, Evaluating the Welfare State. New York: Academic Press. Gephardt, Richard A. 1989. “Hurray for IRAs.” Washington Post (September 25). Harris, Douglas B. 1998. “The Rise of the Public Speakership.” Political Science Quarterly 113: 193–212. Harris, Douglas B. 2005a. “Recovering History and Discovering Data in the Archives: An Alternative ‘Mode of Research’ for Congressional Scholars.” Extension of Remarks 28(2): 3–10. Harris, Douglas B. 2005b. “House Majority Party Leaders’ Uses of Public Opinion Information.” Congress & the Presidency 32(2): 133–155. Herbst, Susan. 1993. Numbered Voices: How Opinion Polling Has Shaped American Politics. Chicago, IL: The University of Chicago Press. Iyengar, Shanto. 1987. “Television News and Citizens’ Explanations of National Affairs.” American Political Science Review 81: 815–832. Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago, IL: The University of Chicago Press. Jacobs, Lawrence R. and Robert Y. Shapiro. 1994. “Issues, Candidate Image, and Priming: The Use of Private Polls in Kennedy’s 1960 Presidential Campaign.” American Political Science Review 88: 527–540. Jacobs, Lawrence R. and Robert Y. Shapiro. 2000. Politicians Don’t Pander: Political Manipulation and the Loss of Democratic Responsiveness. Chicago, IL: The University of Chicago Press. Jacobs, Lawrence R., Eric Lawrence, Robert Y. Shapiro, and Steven S. Smith. 1998. “Congressional Leadership of Public Opinion.” Political Science Quarterly 113: 21–42. Kernell, Samuel. 1997. Going Public: New Strategies of Presidential Leadership, 3rd edition. Washington, DC: CQ Press. Maltese, John Anthony. 1994. Spin Control: The White House Office of Communications and the Management of Presidential News, 2nd edition, revised. Chapel Hill: University of North Carolina Press. Maraniss, David and Michael Weisskopf. 1996. Tell Newt to Shut Up! New York: Simon & Schuster. Murray, Charles A. 1984. Losing Ground: American Social Policy, 1950–1980. New York: Basic Books. Nelson, Thomas E. and Donald R. Kinder. 1996. “Issue Frames and Group-Centrism in American Public Opinion.” Journal of Politics 58(4): 1055–1078. Nelson, Thomas E., Rosalee Clawson, and Zoe M. Oxley. 1997. “Media Framing of a Civil Liberties Conflict and Its Effect on Tolerance.” American Political Science Review 91(3): 567–583. Neuman, W. Russell, Marion R. Just, and Ann N. Crigler. 1992. Common Knowledge: News and the Construction of Political Meaning. Chicago, IL: The University of Chicago Press. Seelye, Katherine Q. 1995. “Clinton Edges Congress in Amount of Television Coverage, Study Says.” New York Times (April 4). Sellers, Patrick J. 2000. “Manipulating the Message in the U.S. Congress.” Harvard International Journal of Press/Politics 5: 21–30. Sellers, Patrick J. 2002. “Winning Media Coverage in the U.S. Congress.” In Bruce I.

Partisan Framing in Legislative Debates 59 Oppenheimer, editor, U.S. Senate Exceptionalism, 132–153. Columbus: Ohio State University Press. Sinclair, Barbara. 1995. Legislators, Leaders, and Lawmaking. Baltimore, MD: Johns Hopkins University Press. Smith, Richard A. 1984. “Advocacy, Interpretation, and Influence in the U.S. Congress.” American Political Science Review 78: 44–63. Sniderman, Paul M. and Sean M. Theriault. 2004. “The Structure of Political Argument and the Logic of Issue Framing.” In Willem E. Saris and Paul M. Sniderman, editors, Studies in Public Opinion. Princeton, NJ: Princeton University Press.

4

Building a Framing Campaign Interest Groups and the Debate on Partial-birth Abortion Jessica C. Gerrity

Interest groups engage in framing campaigns in order to influence the way the public and decision makers prioritize and conceptualize policy issues. While a framing strategy may entail years of lobbying members of Congress as well as the public, interest groups’ efforts to shape perceptions of a policy issue can yield great rewards. Issue framing can lead people to alter their preferences, and the way politicians and the public understand an issue can be vital to its political outcome (Iyengar 1991; Kahneman and Tversky 1982; Sears 1993; Nelson et al. 1997; Druckman 2001). It thus becomes essential to examine the origins of framing campaigns and unpack how interest groups attempt to shape definitions of issues in the media and the minds of the public. Drawing on elite interviews and the archival records of interest groups involved in the debate over The Partial-birth Abortion Ban Act (1995–2003), this chapter illustrates how interest groups think strategically about framing political issues. Indeed, the groups examined in this chapter, Concerned Women for America (CWA), The Family Research Council (FRC), The National Right to Life Committee (NRLC), and National Abortion Rights Action League (NARAL) Pro-Choice America, refer to framing implicitly, using phrases such as “message development,” “message promotion,” “educating the public,” and “shifting the dialogue,” as well as more explicitly, referring to a “frame” or a “frame strategy” in their internal and public communication. The title of an article in a November 18, 1997, issue of the National Right to Life News reads “How to Frame the Issue by the Language We Use.” In the article, NRLC coaches readers on the most effective ways to “persuasively talk about life issues.” Similarly, a NARAL representative I interviewed spoke of “fighting a framing war,” explaining that “the framing wars are very important, they are vital. People retain very little, so having the right phrase or way of talking about an issue is key.”1 To address the gap in our understanding of elite framing, this chapter explores interest groups’ framing efforts in the debate on partial-birth abortion, a ban on a medical procedure used to perform a particular kind of abortion. The debate on partial-birth abortions has become one of the most important battlegrounds for abortion in the past ten years. Interest groups on both sides of the debate engaged in efforts to shape the public’s perception of the issue. Using the debate on partial-birth abortion as a lens through which to examine interest groups’

Building a Framing Campaign 61 framing strategies, I ask, how do interest groups conceptualize framing as a political strategy? What are the factors that influence their framing strategies? This chapter demonstrates that both sides of the abortion debate recognize the importance of framing and view framing as a political strategy. Evidence from archival research and interviews demonstrates that groups tailor their frames in order to realize different goals; in particular, groups consider how their frames will resonate with the public, members of the media, and elected officials. Groups also take advantage of opportunities created by those venues to engage in political framing.2 I begin this chapter by discussing the partial-birth abortion issue and the primary frames promoted by each side of the debate. Second, drawing on previous research, I discuss the reasons why interest groups may choose to employ framing as a political strategy. Third, I present findings from elite interviews and archival research in the partial-birth abortion debate. Finally, I draw some conclusions about interest groups’ use of framing as a political strategy and propose some areas for further research.

The Ban on Partial-birth Abortion: Issue Emergence and Response The first ban on partial-birth abortions was introduced in the House of Representatives in 1995. President Clinton vetoed it as well as a subsequent bill in 1997. In both instances, an attempt to override the veto passed in the House but failed in the Senate. A third version of the bill, the Partial-birth Abortion Ban Act of 1999, was introduced in the Senate and passed in both chambers, despite President Clinton’s veto threat. Congressional action stalled in 2000 when the Supreme Court ruled in Stenberg v. Carhart that Nebraska’s partialbirth law was unconstitutional because it was too vague and did not include a health exception. The proposed federal ban did not include a health exception, and pro-ban members were not eager to include it. As a result, the issue was temporarily tabled until President George W. Bush signed the ban into law in 2003, making good on his campaign promise to sign the ban. The Supreme Court, in Gonzales v. Carhart, upheld the law in 2007. All versions of the ban sought to prohibit an abortion procedure involving dilation of the cervix and the vaginal extraction of an intact fetus, medically known as “intact dilation and extraction,” as well as to establish criminal penalties and civil liability for those who perform the procedure. The procedure is generally used late in the second trimester or in the third trimester of pregnancy, but the actual number and timing of abortions using this procedure has been fiercely debated. The Alan Guttmacher Institute estimates that approximately 1 percent of all abortions in 1992, the year for which complete data are available, were estimated to have been performed beyond 20 weeks, late in the second trimester. The NRLC estimates that the procedure is used far more frequently and by women in their second and third trimesters of pregnancy.

62 Origins The debate on banning what came to be known as the partial-birth abortions can be traced back to a research report on a late-term abortion procedure called “dilation and extraction,” or more commonly “D&X,” written by Cincinnati physician Martin Haskell, and distributed and presented in September, 1992, at a National Abortion Federation annual meeting in Dallas, Texas (Gorney 1998). The NRLC obtained a copy of the report and played an important role in circulating this report in pro-life circles in Dallas and in other states and in naming the procedure “partial-birth abortion.” To encourage people to think about the procedure in a more emotional way, the NRLC produced an advertisement, which ran in several newspapers across the country including the Minneapolis Star-Tribune on May 12, 1993, which illustrated the procedure in a series of drawings with the caption. “Do these drawings shock you? We’re sorry, but we think you should know the truth.” According to Cynthia Gorney (1998), the label partial-birth abortion came about as a result of a meeting between the Florida Republican congressman Charles Canady, who sponsored the first ban on partial-birth abortions in 1995, and the longtime NRLC lobbyist Douglas Johnson. The goal was to frame the issue in a way that was more visceral than was believed possible with vague medical terminology. The creation of the label “partial-birth abortion” is an important aspect of the debate; yet, the NRLC does not have much incentive to draw attention to this notion of “labeling” since doing so implies that the issue is selectively and strategically defined. Rather, according to CWA and FRC, the NRLC did not take credit for the label and proceeded as if the procedure has always been around under the name partial-birth abortion. NRLC and other pro-life groups, such as CWA and FRC, used the label “partial-birth abortion” exclusively, focused on the “brutality and inhumanity” of the partial-birth abortion procedure, and focused attention almost entirely on the procedure and the pain felt by the “unborn baby.” Hence, I refer to this frame as the “brutality frame.” NRLC consistently used visual imagery early in the debate. NRLC made a tremendous effort to educate its members and leadership on how to talk about abortion, advising people to use the term partialbirth abortion and to emphasize the brutality of the procedure, and urging others to do so as well. The NRLC, CWA, and FRC consistently promoted the “brutality frame” in the debate on partial-birth abortion. According to the NARAL representative I spoke with, the pro-choice community was caught off guard by the partial-birth abortion issue in large part because they were unfamiliar with the term. When the bill was first introduced in the House in 1995, NARAL struggled to “decipher exactly what the bill was trying to ban.”3 In response, NARAL’s campaign featured stories of women who faced tragic circumstances late in their pregnancies, including cases where the fetus was severely malformed or when the procedure was necessary to preserve the woman’s own health and future reproductive capacity. According to NARAL, they “took the bait” and started talking about the procedure, which only drew more attention to what happens in an abortion and focused attention on areas of the debate that do not traditionally favor reproductive rights, namely a

Building a Framing Campaign 63 focus on fetuses and not mothers. Moreover, NARAL felt that, by letting ban supporters frame the issue first, NARAL and other pro-choice groups were at a distinct disadvantage. Timing is an important aspect of winning a framing war (Leech et al. 2002). However, in order to successfully frame, actors need to be prepared with a well-developed frame. NARAL initially thought that it was just as risky to put forth an ill-conceived frame as it was to let the other side frame the issue first. NARAL shifted its framing strategy soon after it perceived that its “tragic and rare” frame—its attempt to focus on the tragic and rare circumstances under which a woman may need the partial-birth procedure—was failing to generate opposition to a federal ban on the procedure. NARAL tried to refocus the debate after 1996 and shift its message back to its traditional arguments of non-governmental interference, privacy, and constitutionality. However, as the NARAL representative I spoke with put it, “the media and people form a position about something in a moment in time and then they do not go back and revisit it, so it is vital that they frame it first and early on in the debate.” NARAL believes that its inability to frame the debate first was one reason why it lost the framing war on partial-birth abortion. The other primary reason, according to NARAL, was that the other side’s “sales pitch was more convincing.” The majority of the public also believed the sales pitch was more convincing. A Gallup Organization telephone poll in March 1997, right before the Senate vote in the spring of 1997, showed that 55 percent of Americans would vote to make the procedure illegal. Support for the ban increased to 60 percent in 1998 and 1999 and rose to 66 percent by 2000. The question was not asked before 1997, and the same question was not asked again after 2000, but this gives us a general idea of the public’s support for the ban during some of the key moments in the debate. Experimental research conducted by Freedman (1999) supports the notion that framing mattered in the partial-birth abortion debate. Freedman (1999) found significant framing effects occurred among respondents who were initially ambivalent about the abortion issue, or “people who report finding opposing arguments somewhat or very persuasive” (1999, p. 227). While abortion attitudes are generally stable over time (Wlezien and Goggin 1993; Wilcox and Norrander 2002) it appears that opinions on partial-birth abortion did shift in the short term. The partial-birth abortion issue may have also contributed to general abortion attitudes becoming more conservative in 2000 than was the case in 1989 (Wilcox and Norrander 2002).

Interest Group Framing According to Riker (1986, 1990), interest groups mastering the art of “heresthetics” would be able to affect policy outcomes by convincing decision makers to emphasize new dimensions of an issue, or by altering the context in which policy choices are made. The art of heresthetic is about “structuring the world so that you can win” (Riker 1986, p. ix). Interest groups have a strong

64 Origins motivation to frame issues in ways that make it difficult for key actors to ignore them. In debates over issues such as abortion, interest groups may target three audiences: the public, other political elites, and the news media. Moreover, different considerations inform interest groups’ targeting of these different audiences. First, interest groups may use framing to shape public opinion and mobilize members of the public to contact elected officials (Kollman 1998). Research on framing and public opinion suggests that this can be a fruitful strategy for interest groups since framing has been shown to influence public opinion (Iyengar 1991; Kahneman and Tversky 1982; Nelson and Kinder 1996). A powerful example is the Republican redefining of the inheritance tax as the “death tax” (see Schaffner and Atkinson’s chapter in this volume). By most accounts opponents of the inheritance tax succeeded in reframing the issue since 77 percent of the public now believe that they will be affected by the “death tax” when they die, when in fact the percentage of Americans who will pay an inheritance tax has consistently remained about 2 percent (Graetz and Shapiro 2005). The effect of a framing strategy can be significant, given that a citizen’s decision to support an issue, or a legislator’s decision to support a bill, may differ dramatically depending on which dimension of an issue is emphasized. Abortion policy falls into a category of policy known as morality policy (Mooney 2000). Morality policies are generally highly salient to the public, easily understood, and attract a higher than normal level of citizen participation (Mooney 2000, p. 9). Policies that are classified under morality policy can be thought of as “easy issues” because technical and information barriers do not limit who can understand them (Carmines and Stimson 1989). All morality policies contain technical elements; however, interest groups frame political issues in order to highlight the emotional aspects of an issue and downplay any technical meaning. Because morality policies are highly salient to the public and the news media, a significant number of political actors will attempt to frame morality policies. Hence, while framing “easy issues” is generally more difficult than framing more technical and obscure issues (Leech et al. 2002), the rewards of doing so are great. Second, in addition to targeting the public, interest groups target other election officials such as members of Congress, state legislators, and federal agency officials. In order to effectively frame issues in the public’s mind, interest groups need credible and trustworthy sources to promote their frames (Lupia and McCubbins 1998; Alverez and Brehm 1995; Druckman 2001). Druckman explains that “perceived source credibility appears to be a prerequisite for successful framing. Framing effects may occur, not because elites seek to manipulate citizens, but rather because citizens delegate to credible elites for guidance” (2001, p. 1061). Of course, for some members of the public interest groups are highly credible sources. However, for many more, interest groups are thought to represent the extreme viewpoints of an issue and hence are not perceived as credible, persuasive, or trustworthy (Gunther 1992). In addition to source credibility, other political elites may have an easier time attracting and

Building a Framing Campaign 65 keeping the news media’s attention because they are a go-to source for journalists or they may garner more attention because of their policy positions, leadership, or party role. Aside from the incentives to target highly credible sources and elites who seem to have a direct line to the news media, interest groups know that the more messengers they have delivering their message, the more likely they are to reach a broad audience. Interest groups need other elites in order to reach the public and the media, and must craft frames that appeal to them; however, groups also target the media directly. The third audience targeted by interest groups is the news media. The news media are an important conduit between interest groups and the mass public, and the messages that interest groups develop must appeal to the media before interest groups can hope to influence public opinion. Several interest group scholars have noted that interest groups (particularly citizen groups or public advocacy groups) that wish to engage the public have much to gain by using media coverage to communicate their policy goals and arguments to citizens (Leech et al. 2002; Danielian and Page 1994; Schlozman and Tierney 1986; Schattschneider 1960). While there are some drawbacks to interest groups’ use of free media, namely that the group has less control over the content and tone of the coverage, paid media advertisements such as direct mail advertisements and issue advocacy ads are often the most costly aspect of an interest group’s advocacy strategy, making the opportunity for free media coverage quite desirable. In order to effectively frame political issues, interest groups need to reach multiple audiences—the public, media, and other political elites. Yet, we know very little about how interest groups target these different audiences. The next section begins to unpack interest groups’ appeals to these three audiences, again in the context of the partial-birth abortion debate.

Appealing to the Public How do interest groups satisfy these three audiences—the public, media, and elites—while staying true to their core position? Much like a corporation developing a new product line consistent with its brand, interest groups rely on consumer data, in this case public opinion polls. Both sides of the abortion debate strive to figure out which messages will “resonate.” Messages with resonance build on a core belief, grab the public’s attention, or cause the public to think about an issue in a new way. In order to find out what resonates, interest groups, such as the large, well-funded groups I examine here, spend money hiring pollsters and public relations specialists (West and Loomis 1999). According to financial reports filed by NARAL and NRLC in 2006, both groups spent approximately 10 percent of their overall budgets on public relations consultants, public opinion polling, and message development. As they considered the framing of partial-birth abortion, interest groups on both sides of the debate considered the way the public currently thinks about abortion as well as the strategies that had worked or failed in the past. NARAL

66 Origins describes its initial attempts to address the partial-birth abortion issue as largely unsuccessful. It promoted several frames the year after partial-birth abortion had made it onto the congressional agenda. Most notable among them, according to a NARAL representative, was NARAL’s decision to focus on the “tragic and rare” circumstances under which women may need the procedure. NARAL talked about particular women who needed to undergo abortive procedures that may be categorized as partial-birth abortions late in their pregnancy as a result of grave danger to their own health or because their baby was severely malformed and would not live outside of the womb. In discussing these cases, NARAL drew attention to what happens in an abortion—something it had traditionally avoided. NARAL discovered that it was doing exactly what the NRLC and other pro-life groups wanted them to do, by focusing on what happens in an abortion and focusing on the procedure. “We inadvertently played into it by going along with it,” according to NARAL. Moreover, pro-life supporters countered that the procedure was in fact never truly necessary and more often used for secondary, non-health-related reasons. In a NARAL 1996 Fiscal Year Plan, under a section heading titled “Articulate Values That Resonate,” Executive Director Kate Michelman writes of the partialbirth abortion issue, “At this time we must regain control over the debate and present an alternative vision which resonates with America’s pro-choice majority.”4 NARAL felt that it had lost control of the debate on partial-birth abortions and as a result lost the support of many traditional pro-choice supporters. According to a NARAL representative, NARAL realized too late, after partial-birth abortion was on the national agenda for a year, that its strategy of focusing on when women might need a late-term abortion was not working as more and more pro-choice supporters in Congress and in the public supported the ban on partial-birth abortions. NARAL surmised that in order to shift the debate in a more favorable direction it needed to draw attention to the broad values and principles that supported NARAL’s opposition to the ban. NARAL changed its course about a year after the ban was introduced in Congress and reverted to a strategy that had proved successful for them in the past. In their public statements and materials, NARAL focused on themes of individual liberty and privacy that they promoted in their Who Decides campaign, when they discussed partial-birth abortion. NARAL’s Who Decides campaign, crafted in response to the Supreme Court’s 1989 Webster decision, was a theme that NARAL discovered would resonate with the public in the early to mid-1990s. The theme tapped into the public’s support for individual liberty and privacy. NARAL identified a trend in the polling on abortion—both men and women were more likely to get behind a campaign centered on limited government intrusion than one grounded in reproductive rights. In Chapter 2 in this volume Nelson et al. refer to this phenomenon as “value recruitment,” or the linking of a specific issue to a social value about which there is broad societal consensus. NARAL and others recognized the need to capture what the NARAL representative I spoke with referred to as “middle America.” The phrase refers to the majority of Americans

Building a Framing Campaign 67 who are ambivalent about abortion and who support legalized abortions but also support some restrictions. NARAL believed that linking reproductive rights to the public’s sense of individual privacy and resistance to the government’s intrusion into one’s social life would allow them to reach “middle America.” According to a NARAL representative, after a year of unsuccessfully challenging the pro-life framing of the ban on partial-birth abortions, the battle-tested Who Decides frame made a lot of sense: The Who Decides theme played an important part in the partial-birth abortion debate. You have to talk about broader values and not just specifics if you want to reach men and women on this issue. It’s easier to communicate a universal message. Medical privacy is easier to communicate.5 CWA and FRC also reported relying on public opinion for cues about how best to structure their frames. They too, are very interested in capturing what CWA describes as “middle America,” or those Americans CWA believes will come to support restrictions on abortion if those restrictions are framed in the right way.6 CWA recognizes that many Americans are torn between women’s rights and the rights of the “unborn” and believe that partial-birth abortion allowed them to circumvent this dilemma. CWA argued that it was difficult to prove that a partial-birth abortion was ever medically necessary. According to Wendy Wright of CWA: People like the idea of individual autonomy, but they have come to understand that it is not only a woman involved in this, but an unborn baby—this is one of the great things about partial-birth abortion, that there is an actual baby. People will recognize this, but they will feel torn and think that the woman has some priority rights, or her health dilemmas need to be taken into consideration. In response we highlight the fact that there is never a case when the procedure is medically necessary.7 The partial-birth abortion issue also created other opportunities to talk about abortion in ways that pro-life supporters believed furthered their cause. The pro-life groups I examined stressed that talking about partial-birth abortion in graphic detail—in describing what the procedure entails and the pain that the baby experiences—created the opportunity to talk about other types of abortion procedures. According to a representative of the FRC, “Partial-birth abortion is important because a lot of people started to realize that you are taking about killing a baby that is in the process of being born, and is literally inches and seconds away from being born. If you kill it you may go to prison, you can’t kill a baby that is being born.” According to CWA, its framing of partial-birth abortion encouraged the public to ask the question, “If partial-birth abortion looks like this, what do other procedures look like?” The pro-life groups I spoke with believed that, if they could frame partial-birth abortions in the right way, they could change the public’s mind about abortion.

68 Origins Groups on both sides of the partial-birth abortion debate, according to the interviews and the archival research I conducted, think about how to frame issues in ways that resonate with the public. But to reach the public, frames must first attract coverage from the news media.

Appealing to the News Media The pro-life interest groups I studied felt that until the partial-birth abortion debate came along that they were not able to describe what an abortion entails in the mainstream media and therefore unable to communicate their message to the mass public. This may be surprising since the first bill banning the procedure was introduced in the House of Representatives in 1995, just over twenty years since Roe v. Wade was decided—the pro-life movement had ample time to educate the public on what an abortion entailed. Yet, the pro-life groups I studied all agreed that, until partial-birth abortion came around, the pro-life movement did not have the opportunity to truly reach the public. The following excerpt from an interview with Wendy Wright at CWA demonstrates this point: When we tried to describe what would happen in an abortion prior to PBA we were cut off, we were not allowed to describe it in the media. I would try to use descriptive language like dismemberment, and describe how the abortionist has to put the pieces of the baby together afterwards to make sure they got all the pieces because if they left some of it in the woman it may cause a fatal infection. I would describe things like that but it wouldn’t make it on the TV newscast or in print. The partial-birth abortion bill and the debate that ensued focused on the technical aspects of abortive surgery, and on describing exactly what happens in abortive surgery. As a result, the debate shifted from the traditional broad framework of women’s rights and morality to new and often uncomfortable territory, introducing a new dimension of the abortion debate into mainstream coverage of abortion. This differed from any other past abortion issue that Congress has taken up—parental notification, funding for abortions overseas, constitutional amendments, clinic protests, and Medicaid-funded abortions. While the details of an abortive procedure were relevant in each of these issue debates, for the first time, according to the FRC, a discussion of what goes on in an abortion was germane to the discussion. The pro-life groups I examined felt that, because the media were not interested in hearing their side of the debate, they were forced to wait until an issue or event forced the media into covering abortion the way pro-life groups wanted them to. Partial-birth abortion, according to pro-life groups, was an issue that forced the media to talk about abortion in an unprecedented way. The NARAL representative I spoke with agreed: “the other side won because they got a platform for their message that they didn’t have previously.”8

Building a Framing Campaign 69 Wendy Wright of CWA maintained, “even though there is not a partial-birth ban in effect in America [at the time the interview was conducted there was not] it has done more to influence public opinion than probably any other aspect of the abortion issue to date.”9 The representatives I spoke with at CWA and FRC both felt that the partial-birth abortion debate gave them the opportunity to insert their point of view into the debate because the media covered the issue so intensely. While all agreed that other abortion issues, like parental notification laws and the early pregnancy abortifacient, RU-486, are just as important to their members and to their organization, they all agreed that they were unable to attract the same levels of attention from the media or the public, regardless of how graphic or emotional their portrayal of the issues’ effect on women and babies. William Saunders of FRC put it this way: “we are in the business of persuading, but we can’t persuade if we can’t get people’s attention or you can’t get them to think about it once you have their attention.”10 Without the media, interest groups only have limited potential for attracting the public’s attention. In Chapter 8 in this volume, Wagner’s analysis of presidential and congressional issue framing demonstrates that when political actors present consistent, competing, inter-party frames over time they have the potential to affect long-term shifts in partisan identification. Similarly, interest groups look for opportunities to consistently shape the abortion debate in the news media and frame abortion issues in the minds of the public and decision makers. Interest groups on either side of the issue approached the media differently. NARAL expended effort to get free coverage for its positions on partial-birth abortion by issuing press releases, media kits, and press conferences, as well as paying for advertising in national newspapers, television markets, and radio stations. Kate Michelman, NARAL president at the time of the debate, cultivated relationships with the press, as evidenced by the many memos she wrote to reporters and news anchors to praise them for the way they discussed partialbirth abortion in their news report, or to suggest future lunch dates to discuss even better ways to talk about the issues. In one memo sent to “talk show hosts and correspondents” dated October 22, 1999, and titled, “Coverage of Socalled Partial-birth Abortion Ban,” Kate Michelman outlines the implications that “two key developments in the debate” have for “how the bans are described by the media.” In the memo, Michelman describes how “the courts have consistently stated that the vague, medically inaccurate language used to describe a ‘partial-birth abortion’ procedure would ban not one obscure late-term abortion method, but the safest and most common procedures used throughout pregnancy.” This more symbiotic relationship stands in contrast to the way the pro-life groups view the media. While NARAL worked hard to cultivate opportunities to appear on morning talk shows, speak to various audiences, and contribute opinion pieces to newspapers, CWA, on the other hand, described employing more grassroots strategies and cultivating relationships with religious news media outlets since it could not count on a relationship with the journalists in the mainstream media. CWA, FRC, and NRLC worked hard to stay on message and appeal to the media’s reporting norms.

70 Origins For example, CWA was sensitive to the media’s need for “good quote” and punchy sound bites.11 To that end, CWA tried to craft brief messages that also appealed to reporters’ desire to report within a framework of conflict. According to Wright, the outcry over liberal bias has forced reporters to insert the “obligatory conservative” into their stories and reports. And while this often works to further CWA’s goals, there are also times when this works against them. To illustrate this point, Wright described her frustration with a recent inundation of media calls about statements made by pro-choice groups and legislators suggesting that CWA and other pro-life groups would oppose the HPV vaccine because pro-life groups always believe “religion trumps science.” According to CWA, because news reporters assume conflict, they always assume that pro-life groups are opposed to the stands of pro-choice groups, even when they are not taking an opposing position. In order to shape the media’s coverage of partial-birth abortion, the NRLC made a tremendous effort to educate its members and leadership on how to talk about abortion to the media. In an August 21, 1996, issue of the National Right to Life News, the NRLC membership newsletter, in an article titled “How to Talk About Partial-Birth Abortion” NRLC describes the following talking points to its readers—points that would be repeated in almost all the newsletters that addressed partial-birth abortion: • • •

Always use the term partial-birth abortion. When possible, urge that interviewers and reporters also use the term partial-birth abortion. Always describe the procedure before discussing it. Be sure to stress that the babies are alive and experience great pain when they are subjected to a partial-birth abortion.

All of the pro-life groups I examined expressed frustration with the media and the sense that the media favor the pro-choice position. Surveys of journalists demonstrate that the media do overwhelmingly support a woman’s right to decide whether to have an abortion (Lichter 1990). It makes sense, then, that NRLC is concerned with message discipline. If it believe that its efforts to get its point of view across will always be reduced to a single quote, then the language it uses becomes very important. While the NRLC and other pro-life groups remained skeptical of the media, NARAL’s relationship with the media was based on the belief that there are opportunities to work together with reporters and the news media. As a result, NRLC has pursued more grassroots strategies to educate their members on communication with the media and the public. NRLC has worked hard to reach out to religious media outlets in addition to the mainstream news media. NARAL, on the other hand, has had more success communicating their messages to the mainstream media and as a result has not spent as much time coaching its members on framing techniques. While pro-life and pro-choice groups approach the media differently, interest groups on both sides of the partial-birth abortion issue worked hard to construct

Building a Framing Campaign 71 messages that they believed would resonate with the media. Next, I consider how groups frame their messages to appeal to elected officials.

Appealing to Elites Both sides of the abortion issue believe that effective issue framing can compel traditional opponents to support an issue. This is particularly the case for politicians who must defend their positions to constituents. NARAL believed that it lost the support of many of its traditional congressional allies because of successful pro-life framing of partial-birth abortion. The NARAL representative I interviewed described why she thought some pro-choice members in Congress voted for the ban and against pro-choice interests: The sales pitch was more convincing. The title of the procedure had a lot to do with it. It was impossible for some offices to walk out on it; although some people ended up changing their mind over time. Rep. Jim Moran (D-VA) ended up changing his mind after he spoke with a woman who needed the procedure for health reasons and as a result changed his vote.12 One important difference that sets partial-birth abortion apart from any other abortion issue is that it stayed on the congressional agenda for a sustained period of time.13 It was introduced in 1995 and signed into law by President George W. Bush in 2003. The level of congressional attention is important, according to Wendy Wright of CWA, because the longer it stays on Congress’s agenda, the more members of Congress talk about the issue, the more media attention the issue gets and the more opportunities interest groups have to get their voices heard. Moreover, according to CWA, the media are more likely to cover the issue when Congress is debating it on the floor of the chamber or discussing the issue in other public venues. Indeed, all of the interest group representatives I spoke with emphasized how valuable it is when members of Congress take up the issue in some way, either by sponsoring legislation or by debating the issue on the chamber floor and creating news-making events. When members of Congress talk about partial-birth abortion, sponsor legislation, or hold a hearing, they are, according to NARAL, “providing a platform” upon which groups can build in the media. In a NARAL internal memo titled “Strategic Plan, 1997–2000” it talks about “increasing NARAL’s political influence” by “repositioning the choice issue in the public debate.” However, interest groups also recognize their need for legislators to help them achieve these goals, not only through sponsoring legislation and voting on legislation but by attracting media attention to an issue, which in turn creates opportunities for interest groups to assert their frames in the public debate. According to the NARAL representative I interviewed: It is absolutely vital that you have surrogates in Congress. . . . Nancy Keenan [president of NARAL] can go on a talk show and talk about the

72 Origins ban but people know her as an advocate and she is already known to have a point of view. It makes a huge difference who says it—gives more validity and voice to their position.14 Interest groups benefit from the potential legitimacy that results when elected officials draw attention to issues and frame them in ways that further their interests. While rarely successful in contemporary debates on abortion, interest groups also attempt to win over members of Congress who do not typically agree with their pro-life position. The ban on partial-birth abortion received wide bi-partisan support in the House and Senate, which further placed pro-choice supporters in a defensive position and NRLC and other pro-life groups in an advantageous position. Pro-ban groups gained support from a high profile, pro-choice Democrat, the late Senator Pat Moynihan (D-NY). In May 1996, Senator Moynihan concluded, “This is just too close to infanticide. A child has been born, and it has exited the uterus, and what on earth is this procedure?” (Saletan 2004, p. 234). NRLC widely adopted the infanticide language and used the Moynihan comment to underscore bi-partisan support for the ban. Interest groups lobby members of Congress to get the issues they care about on the congressional agenda as well as try to convince members of Congress to promote their frames in their public statements. Both sides of the issue report working hard to develop a framing strategy that appeals to their supporters in Congress in addition to crafting frames that are difficult for their opponents to ignore. Interest groups monitor how their allies discuss the issues they care about via the Congressional Record, a record of debate and speeches on the chamber floors. According to Bill Saunders, it is important to the FRC that members of Congress—particularly their allies—promote the FRC’s way of looking at an issue. The FRC believes that it is important that their allies in Congress “focus on the brutality of the partial-birth abortion procedure or focus on what abortion does to women.”15 When interest groups succeed in getting members of Congress to promote their frames, they may also satisfy an important condition in the framing literature regarding source credibility as an influence on the effectiveness of framing (Druckman 2001). For some members of the public, interest groups like the NRLC and NARAL will be credible sources and capable of framing the issue in their mind, or highlighting a new dimension of an issue; however, for many more, interest groups on either side of the abortion issue are thought to represent the extreme viewpoints of an issue and hence are not perceived as credible, persuasive, or trustworthy. Interest groups know that they need others to promote their frames in order to reach the public; however, they also realize that the more Congress talks about the issue of interest the more likely the news media will be to cover the issue. The FRC representative described a typical situation where the FRC might work together with a member of Congress to develop a framing strategy:

Building a Framing Campaign 73 If we had a member who said, for instance, I think this is a good issue, work with us on that, we would certainly do that. If we get resonance on an issue we can work with them about how to frame it or whatever else. If we find a congressman who is interested in an issue, we would say to them, this is the way we think this issue should be framed. Interest groups recognize the importance of having credible elites and authority figures promote their positions in the media and as a result consider the needs and constraints of legislators when they craft frames. In some cases, this relationship may become institutionalized, such as the weekly meeting between pro-life groups and the pro-life Value Actions Team (VAT), a profamily, pro-life group of members of Congress, led by Joe Pitts (R-PA) in the House and Sam Brownback (R-KS) in the Senate. The weekly meetings allow pro-life interest groups and members of Congress to meet to “divvy up work, decide on an agenda and approach, and develop a message,” according to Wendy Wright of CWA.16 Interest groups supply elites with frames and arguments in an attempt to subsidize their congressional framing efforts. They may do this by funding focus groups to test out frames or they may test a frame at the state level first to provide members of Congress with some evidence of a frame’s potential success or failure in achieving their legislative goals. In some cases interest groups and members of Congress work together to develop a framing strategy that furthers both of their goals. In the case of partial-birth abortion, pro-life members of Congress were enthusiastic advocates of the issue. In addition to sponsoring a ban on partialbirth abortions, pro-life members of Congress worked with interest groups to develop a successful framing strategy. NARAL, on the other hand, discovered that many of its traditional supporters in Congress were uncomfortable with the partial-birth abortion issue as a result of constituent pressure to support the ban. In some cases, traditional pro-choice supporters voted for the ban. However, more worrisome for NARAL was that their traditional allies did not challenge the pro-life framing of partial-birth abortion. This was in part because the pro-choice framing strategy was ineffective, but also in part due to the uncomfortable position in which the partial-birth abortion issue placed many members of Congress, regardless of their traditional stand on abortion. In order for interest groups to count on congressional support, they must craft frames that members of Congress find compelling.

Conclusion In the partial-birth abortion debate, interest groups considered three audiences when they built their framing campaigns. First, when crafting frames, interest groups considered how the public currently thought about abortion and attempted to craft messages that would resonate with the public. Groups also considered frames that had proved successful in the past. Second, interest groups considered the types of messages they believed the news media would find

74 Origins compelling and newsworthy. Third, groups thought about crafting messages that elites, in particular members of Congress, would find appealing enough to use in their public statements. This chapter has explored the way that interest groups think about political framing and the considerations that influence groups’ framing strategies. The evidence from the interviews and archival research presented in this chapter underscores interest groups’ high level of commitment to framing as a political strategy. Both sides of the partial-birth abortion debate report devoting time and resources to developing frames that resonate with the public, media, and members of Congress, and to countering frames from opposing groups. Moreover, this chapter demonstrates that interest groups approach these three different targets with varying expectations and degrees of trust. Prior to the partial-birth abortion debate, pro-life groups felt frustrated by what they perceived as a lack of opportunities to educate the public on abortion. However, the pro-life label “partial-birth abortion” and their framing campaign appealed to the public, media, and elites and opened up new possibilities for pro-life groups. NARAL and other pro-choice groups were unable to successfully counter the powerful pro-life frame, but nonetheless considered how their framing activity would resonate with multiple audiences. Within the broader debate over partial-birth abortion, each side of the debate offered frames that registered varying degrees of success. NARAL, while unsuccessful in their efforts to counter the partial-birth label, was able to shift from what they perceived as a losing framing strategy to promoting frames that had worked for them in the past. At the heart of research on framing lie questions about the ability of political elites to affect the role and structure of issue debates. These are longstanding theoretical questions in the political science literature, but empirically underexamined by interest group scholars. While scholarship often looks for interest group influence on policy outputs, a more pervasive form of influence may be found in how interest groups affect the framing of the policy debate, the problem definition, and the construction of policy alternatives. Future research on interest group framing should attempt to address the origins of frames and the strategies that underpin framing campaigns, as well as the impact of frames. New modes of communication, in particular web-based strategies such as those discussed by Ansley and Sellers in Chapter 5, may create more and different framing opportunities for interest groups. Further research is needed to determine if interest groups are innovating in ways that Ansley and Sellers find campaigns are innovating by adopting strategies that include a more dynamic role for group members. The Internet provides interest groups with ways to test new frames on multiple audiences before they directly target the public, media, and political elites. Future research should attempt to tease out whether and how interest groups are employing web-based strategies to engage in political framing.

Building a Framing Campaign 75

Notes 1. Phone interview with National Abortion Rights Action League (NARAL) Pro-Choice American representative on June 16, 2006. 2. The chapter draws on interviews with two pro-life groups and one pro-choice group about their framing efforts on the issue of partial-birth abortion and abortion more generally. I also examined NARAL and the National Right to Life Committee (NRLC) archived records. The time period I examine in the archives is from 1994 to 2003. I interviewed two pro-life interest group representatives, Wendy Wright, Senior Policy Director of Concerned Women for America (CWA) and William Saunders, Senior Fellow at The Family Research Council (FRC), at their headquarters in Washington, DC, in December 2005. I conducted one phone interview with a NARAL Pro-Choice America representative in June 2006. I also conducted archival research at the Schlesinger Library at the Radcliffe Institute for Advanced Study at Harvard University in Cambridge, MA. The library is the official archive for several major women’s rights organizations, such as the National Organization for Women and NARAL Pro-Choice America. The library also collects materials on pro-life interest groups, including NRLC, making it a valuable resource for interest group scholars. The archive contained NARAL’s media packets, press releases, public opinion reports, memoranda, and correspondence with the media as well as NRLC’s newsletters. 3. Phone interview with NARAL Pro-Choice American representative on June 16, 2006. 4. Schlesinger Library, internal memo, 1996. 5. Phone interview with NARAL Pro-Choice American representative on June 16, 2006. 6. Interview with Wendy Wright of CWA on December 12, 2005. 7. Interview with Wendy Wright of CWA on January 3, 2006. 8. Phone interview with NARAL Pro-Choice American representative on June 16, 2006. 9. Interview with Wendy Wright of CWA on December 12, 2005. 10. Interview with William Saunders of FRC on December 14, 2005. 11. Interview with Wendy Wright of CWA on December 12, 2005. 12. Phone interview with NARAL Pro-Choice American representative on June 16, 2006. 13. The length of time a policy issue stays on the congressional agenda may be both a consequence and a cause of political actors’ successful attempts to define a public policy issue (Baumgartner and Jones 1993; Kingdon 1984). The longer the issue remains on the congressional agenda, the more time policy actors have to define the problem and possible solutions (Kingdon 1984) as well as focus the media’s attention on particular aspects of an issue which they find advantageous to their political goals (Baumgartner and Jones 1993). 14. Phone interview with NARAL Pro-Choice American representative on June 16, 2006. 15. Interview with William Saunders of FRC on December 14, 2005. 16. Interview with Wendy Wright of CWA on December 12, 2005.

Works Cited Alverez, R. Michael and John Brehm. 1995. “American Ambivalence Toward Abortion Policy: Development of a Heteroskedastic Probit Model of Competing Values.” American Journal of Political Science 39(4): 1055–1082. Baumgartner, Frank R. and Bryan D. Jones. 1993. Agendas and Instability in American Politics. Chicago, IL: The University of Chicago Press.

76 Origins Carmines, Edward G. and James A. Stimson. 1989. Issue Evolution: Race and the Transformation of American Politics. Princeton, NJ: Princeton University Press. Danielian, Lucig H. and Benjamin I. Page. 1994. “The Heavenly Chorus: Interest Group Voices on TV News.” American Journal of Political Science 38(4): 1056–1078. Druckman, James. 2001. “On the Limits of Framing Effects: Who Can Frame?” The Journal of Politics 63(4): 1041–1066. Freedman, Paul. 1999. “Framing the Abortion Debate: Public Opinion and the Manipulation of Ambivalence.” Unpublished Dissertation, University of Michigan. Gorney, Cynthia. 1998. Articles of Faith: A Frontline History of the Abortion Wars. New York: Simon & Schuster. Graber, Doris. 2002. Mass Media and American Politics, 6th edition. Washington, DC: Congressional Quarterly Press. Graetz, Michael J. and Ian Shapiro. 2005. Death by a Thousand Cuts: The Fight over Taxing Inherited Wealth. Princeton, NJ: Princeton University Press. Gunther, Albert C. 1992. “Biased Press or Biased Public? Atttitudes Toward Media Coverage of Social Groups.” The Public Opinion Quarterly 56(2): 147–167. Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago, IL: The University of Chicago Press. Kahneman, Daniel and Amos Tversky. 1982. “The Psychology of Preferences.” Science 246: 135–142. Kingdon, John W. 1984. Agendas, Alternatives, and Public Policies. Boston: Little, Brown. Kollman, Ken. 1998. Outside Lobbying: Public Opinion and Interest Group Strategies. Princeton, NJ: Princeton University Press. Leech, Beth, Frank Baumgartner, Jeffrey M. Berry, Marie Hojnacki, and David C. Kimball. 2002. “Organized Interests and Issue Definition in Policy Debates.” In Allan J. Cigler and Burdett A. Loomis, editors, Interest Group Politics, 6th edition. Washington, DC: Congressional Quarterly Press. Lichter, Robert S. 1990. The Media Elite: America’s New Powers Brokers. Winter Park, FL: Hastings House. Lupia, Arthur and Mathew D. McCubbins. 1998. The Democratic Dilemma: Can Citizens Learn What They Need to Know? New York: Cambridge University Press. Mooney, Christopher Z. 2000. The Public Clash of Private Values. New York: Chatham House Publishers. Nelson, Thomas E. and Donald R. Kinder. 1996. “Issue Frames and Group-Centrism in American Public Opinion.” The Journal of Politics 58(4): 1055–1078. Nelson, Thomas E., Rosalee A. Clawson, and Zoe M. Oxley, 1997. “Media Framing of a Civil Liberties Conflict and Its Effect on Tolerance.” American Political Science Review 91(3): 567–583. Riker, William H. 1986. The Art of Political Manipulation. New Haven, CT: Yale University Press. Riker, William. 1990. “Heresthetic and Rhetoric in the Spatial Model.” In James Enelow and Melvin Hinich, editors, Advances in the Spatial Theory of Voting. Cambridge: Cambridge University Press. Saletan, William. 2004. Bearing Right: How Conservatives Won the Abortion War. Berkeley: University of California Press. Schattschneider, E.E. 1960. The Semisovereign People: A Realist’s View of Democracy in America. New York: Holt, Rinehart & Winston.

Building a Framing Campaign 77 Schlozman, Kay Lehman and John Tierney. 1986. Organized Interests and American Democracy. New York: Harper & Row. Sears, David O. 1993. “Symbolic Politics: A Socio-Political Theory.” In Shanto Iyengar and William J. McGuire, editors, Explorations in Political Psychology. Durham, NC: Duke University Press. West, Darrell M. and Burdett A. Loomis. 1999. The Sound of Money: How People Get What They Want. New York: Norton & Company. Wilcox, Clyde and Barbara Norrander. 2002. Understanding Public Opinion. Washington, DC: Congressional Quarterly Press. Wlezien, Christopher B. and Malcolm L. Goggin. 1993. “The Courts, Interest Groups, and Public Opinion about Abortion.” Political Behavior 15(4): 381–405.

5

Mobilizing to Frame Election Campaigns Taylor Ansley and Patrick J. Sellers

Following the example of many earlier candidates and campaigns, Barack Obama and his presidential campaign attempted to frame the 2008 election around issues favorable to the candidate. For Obama’s team, that goal required linking John McCain to the unpopular incumbent president, focusing on an economic recession, and emphasizing a desire to bring the war in Iraq to a close. Obama’s campaign deployed many typical methods for framing elections. Campaign spokespeople and high-profile surrogates logged countless hours speaking on radio and television shows and to crowds across the country. The campaign submitted its candidate to interviews with the press, sent out official statements when current events mandated a reaction, and crafted numerous advertisements for various media. These tactics attempted to frame the election campaign—both what issues would be discussed and how those issues would be debated—in a top-down, disciplined, and controlled manner. The Obama campaign’s efforts parallel those presented elsewhere in this volume. In the debate over intelligent design described in Nelson et al.’s chapter, proponents and opponents of the concept each worked to frame that debate around values and questions favorable to their interests. Harris’s chapter describes how leaders from both parties in Congress regularly employed competing frames to define terms of congressional debate. Interest groups competed in a similar manner to frame the partial-birth abortion debate, as recounted in Gerrity’s chapter. Across all these examples, the frames originated among political elites, such as campaign managers, congressional party leaders, or interest group staffers. The Obama campaign recognized that such a top-down, centralized process could hinder the credibility and broader acceptance of frames and messages, particularly if the public and potential supporters viewed that information as coming solely from the campaign. As Druckman’s chapter suggests, the perceived credibility of a source strongly impacts the public’s willingness to accept frames from that source. Recognition of this link led the Obama campaign to supplement its top-down approach to messaging with a more decentralized reliance on volunteers. The campaign hoped that, if its message partially originated among these volunteers, that message might prove more credible and persuasive to volunteers and the general public.

Mobilizing to Frame Election Campaigns 79 To this end, the Obama campaign cultivated a network of grassroots supporters and allowed them to help set the terms and tone of the debate, both collectively and individually. Building a large group of committed volunteers did not represent a seismic shift in campaign strategy; the two presidential campaigns of George W. Bush, for example, famously utilized an extensive network of canvassers. But, Team Obama’s volunteers went well beyond canvassing; they utilized an unprecedented set of tools to create, disseminate, and amplify frames. The Obama campaign extended a wealth of resources to engaged supporters, which in turn helped those individuals craft unique appeals to their own personal networks. One post-election analysis observed: The [MyBarackObama.com] web site contained videos, speeches, photos and how-to guides that gave people the raw materials they needed to create their own compelling content in support of Obama. In return, supporters created more than 400,000 pro-Obama videos and posted them to YouTube. They also wrote more than 400,000 blog posts on the MyBO Web site. The campaign could not possibly have generated this much content on its own. And it was better that it didn’t . . . The authentic usergenerated video is more compelling and elicits more support than official productions because we are more trusting of information that is from people who hold similar beliefs, share the same politics or religion, or are the same age or gender as us. (The Social Pulpit 2009) Each of those blog posts (and thousands of others, beyond the campaign’s own site), YouTube videos, emails, phone calls, and in-person conversations helped frame the election campaign in Obama’s favor. By Election Day his campaign enjoyed twice the web traffic, four times the YouTube viewers, and five times the Facebook friends of the rival McCain campaign. The Obama campaign’s massive online advantage led observers to laud the campaign for revolutionizing electoral strategies (The Social Pulpit 2009). But, many of the advances in decentralized messaging and framing had appeared in earlier election contests. In the 2006 US Senate election in Ohio, for example, Democrat Sherrod Brown’s campaign relied on YouTube to test-market a campaign ad. “Within the first 48 hours” after releasing the ad on Brown’s website, “over 20,000 people watched the video” according to an email to supporters (Brown email, October 17, 2006). Such a positive reaction encouraged the campaign to invest in a “statewide television buy” the next week to air the same ad (Brown email, October 17, 2006). Even more intriguing than the television ad was another multimedia feature that Brown’s campaign introduced to its website a week before Election Day. An email to supporters encouraged “average Ohioans to record messages outlining their reasons for supporting Sherrod Brown.” Those audio messages were in turn posted to Brown’s website, and communications staffers quoted voter endorsements in campaign literature (Brown email, October 30, 2006).

80 Origins These tactics were not possible ten years ago and therefore raise important questions about the evolving role of framing, mobilization, and technology in election campaigns. Exactly what innovative strategies do contemporary candidates employ in their attempts to frame election campaigns? Which candidates embrace these new strategies? How effective are the new strategies? In addressing these questions, we argue that the most innovative campaigns are increasingly adopting strategies that simultaneously empower their volunteers and help frame the election campaign. These strategies require the campaigns to give up some control over their message, but the resulting promotion of that message is more credible and effective. To support this argument, we first compare traditional and new strategies of volunteer mobilization and message promotion. The second part of the chapter examines which of the 2006 Senate candidates most frequently embraced the new strategies, and why. In the third section of the chapter, we begin to illustrate the effectiveness of the new strategies, using the 2006 Senate contest in Ohio as a case study.

From Mobilizing Volunteers to Framing the Election Collective action has always played a central role in election campaigns (Sandler 1992; Hardin 1968; Aldrich 1995; Beck et al. 2002). Candidates build coalitions of voters through giving strong stump speeches, effectively “work[ing] a room,” and campaigning door-to-door (Duquin 2003a, 2003b; Sheehan 2003). To amplify the themes and arguments made during these contacts, campaigns turn to earned media, as well as paid advertising on television, on radio, and in newspapers (Mendelson and St. Hilaire 2003; Arnold 2003; Shimmel 2003; White 2003; Faucheux 2003b). The goal of earned and paid media alike is to disseminate the campaign’s message to existing and potential supporters. This task has grown increasingly difficult in recent decades. Media markets have become progressively more expensive. The television audience grows more divided with expanded channel options and competition from online entertainment; digital recorders like TiVo allow viewers to tune out to advertising altogether (Trippi 2004). Facing these challenges of a fragmented media environment, each campaign must still work to build a winning coalition of voters by framing the election around issues and arguments that favor the candidate (Riker 1986; Iyengar and Kinder 1987). When crafting these coalitions, campaigns have long worked to motivate and empower volunteers to act on behalf of the campaign. These individual volunteers register voters, distribute yard signs, call prospective voters, canvass neighborhoods, work polling places on Election Day, and collate campaign literature (Elam 2003; Trent and Friedenberg 1991; Herrnson 2003). Personal appeals among neighbors and friends, relying on pre-existing relationships and trust, are more credible, and thus more likely to recruit other volunteers. Through personal contact with potential voters, enthusiastic volunteers can help deliver their campaign’s message, potentially more effectively than television ads or the

Mobilizing to Frame Election Campaigns 81 candidate himself or herself. Personal contact is “the single best means to win votes” (Shea and Burton 2001, p. 184). Volunteer support and enthusiasm are contagious, not simply within the campaign but in terms of media coverage and attracting more volunteers: “[a]n air of grassroots popularity can be a self-fulfilling prophecy” (Shea and Burton 2001, p. 185). In sum, campaigns have long recognized the value in peer-to-peer, decentralized mobilization efforts. The campaigns facilitate and strengthen personal outreach by entrusting their supporters with the tools of success (be that information, logistical support, or simply the passion of an exciting campaign), and by supporting their efforts in any way possible. In mobilizing volunteers to perform these tasks, campaigns have always had to address significant barriers discouraging such participation (Kanazawa 1998; Riker and Ordeshook 1968). Labor-heavy activities like phone-banking and canvassing typically occur in the evenings or weekends, forcing volunteers to forfeit their free time. Other barriers include “alienation” from the political system and the “high information cost” of political participation—the personal effort required for an individual to understand political issues (Buchanan 2001). This problem is worse among young voters who are demonstrably less interested in current events and political debates than older generations (Mindich 2005). Campaigns also experience opportunity costs in their efforts to motivate collective action. Engaging supporters to do more than vote takes scarce time and resources away from other aspects of the campaign operation. While dedicated supporters are valuable campaign assets, they are hard to identify and recruit when information like voter rolls and databases are flawed and incomplete (Shea and Burton 2001). Campaigns must invest time and resources in grooming volunteers, ensuring that canvassers do not visit homes at inappropriate times, for instance, or that phone bank volunteers do not call a given voter too many times (Webb and Mockus 2003; Shea and Burton 2001). Finally, these grooming efforts may not always work; volunteers may ignore directions and the campaign’s desired message, creating embarrassment and controversy. Even after overcoming all of these costs to volunteers and campaigns, there remains a final obstacle to effective collective action: free riding. From the perspective of campaign volunteers, the victory of their candidate is a nonexcludable benefit, whose consumption cannot be confined solely to those supporters who volunteered or even voted. While major financial contributors may derive some private benefit of access once the candidate enters office, the electoral victory is fundamentally a collective benefit for a campaign’s supporters. As with any such collective good, potential volunteers who desire that good may recognize that they can reap the benefits without contributing any action or effort. How can campaigns overcome the challenge of free riding? One possible solution is to provide private social benefits only to volunteers who contribute to the campaign. These benefits include non-monetary intangibles such as pride in feeling important and useful, inspiration from a charismatic candidate, and

82 Origins trust in fellow volunteers and the overall campaign (Shea and Burton 2001). Campaigns motivate supporters by valuing their efforts and unique contributions, while empowering them to participate in concrete ways. By using the campaign’s tools and information to contact and persuade their friends, neighbors, and coworkers, the volunteers develop a greater sense of efficacy and satisfaction. These benefits are available only to individuals who actually volunteer and help the campaign. The excludability thus combats free riding. While campaigns have long worked to overcome these various challenges and provide social benefits, recent technological innovations have helped contemporary campaigns perform the tasks significantly more efficiently and effectively (Hathaway 2003; Blaemire 2003; Malchow 2003). Email makes volunteer contact and recruitment easier and more efficient. Campaigns can make telephone campaigning more effective by relying on micro-targeting and more comprehensive records of voter contact (Desmone 2003). More accurate (and instantly updatable) voter registration databases and contact lists improve a campaign’s capacity to organize operations by county or precinct (Schneider 2003). Streamlining these activities through technology lowers the campaigns’ opportunity costs of mobilizing volunteers. Technology also helps the campaigns utilize volunteers more effectively. Using email and the web, campaigns can engage volunteers on their own schedules and help supporters reach out to their social networks with knowledge of, or enthusiasm for, a candidate (or issue). Volunteers rarely confine their advocacy efforts to the web and instead use online communications in concert with offline interactions (Wellman et al. 2001). Such outreach is demonstrably effective: when one member of a social group volunteers for a campaign, that person’s friends are also more likely to participate (Kenny 1992). The campaigns also use the new technology to create new opportunities for social interaction, by strengthening weak social networks and attracting previously inexperienced volunteers from diverse backgrounds (Haythornthwaite 2002; Hindman 2005). Technological innovations and the accompanying expansion of social interaction also create new ways for contemporary campaigns to draw attention to their preferred issues and arguments and frame the election campaign. Motivated volunteers have always talked about their candidate and campaign in private conversations with family and friends. Technological innovations allow volunteers to make these conversations more public, potentially distributing the campaign’s message to many more people. Through campaign email, dynamic website content, and blogging, campaigns empower their supporters to promote issues and arguments to preexisting and new social networks (Swanson 2003; Connell 2003; Drezner and Farrell 2004). Activists spread videos virally, bypassing the gatekeepers of the traditional media (Darr and Barko 2004). Supporters forward emails to friends, lending a personal touch to the campaign message and reaching a new audience; “two-thirds of those who receive political email forward it to someone else” in their address book (Darr et al. 2004). Campaigns can rely on sympathetic bloggers to dig up dirt on their opponent’s record, a task previously performed almost exclusively

Mobilizing to Frame Election Campaigns 83 by professional researchers (Bovee 2003). Whether identifying “citizen leaders” and enlisting them with the campaign, or engaging otherwise anonymous bloggers in the effort, campaigns use technology to encourage collective action by amateur supporters (Faucheux 2003a). These supporters receive enhanced social rewards from appreciative campaigns, social groups, and blog audiences (Germany 2006). By sharing their support and enthusiasm for the campaign with more people, the volunteers both enjoy the social benefits of participating in the campaign and assist the candidate by championing his or her issues and arguments. In this manner, new technologies help provide both the private benefits of social interaction and the public benefits of promoting the campaign’s message. The Risks of Decentralization The electronic empowerment of volunteers also presents important risks for a campaign. Traditionally, campaigns have maintained tight control over efforts both to mobilize activists and to promote the campaign’s message. In this centralized, top-down framework, supporters can act collectively by working though the formal campaign. The campaign office provides and tabulates petitions and volunteer forms, while providing yard signs, bumper stickers, press releases, and position papers. For many campaigns, this centralized approach characterizes their initial step into online campaigning. The campaign headquarters tightly controls email distribution lists, while crafting the campaign website from press releases, descriptions of issue positions, and candidate biographies. The centralized production of this information helps campaigns control the distribution of their message (hopefully leading to news coverage of that exact same message). These online promotional efforts closely resemble the more traditional, centralized strategies that candidates have followed for decades. As noted in this chapter’s introduction, however, campaigns are increasingly experimenting with a decentralized organizational structure, involving the mobilization of volunteers and the distribution of the campaign’s message. The Internet plays a central role in both strategies. These decentralized campaigns attract volunteers by providing personalized resources for activists and empowering supporters operating outside the formal campaign. Personalized resources include postcard or letter-to-the-editor tools that encourage supporters to use their own voices in support of the campaign, and voter endorsements that promote self-expression and individual appeals. The campaign website empowers individuals outside the formal campaign by providing resources to host house parties or mount fundraising campaigns of their own. While the traditional, centralized campaign would send a uniform postcard to all supporters, the more decentralized approach would allow a visitor to the campaign website to send a postcard with a personalized message to close friends. The former approach relies on the campaign to initiate the mailing, while the latter empowers supporters to reach out to friends in their own words.

84 Origins The postcard example also illustrates the second focus behind a decentralized campaign: the distribution of information about the campaign and its message. Campaign spokespeople, and more recently websites, have long provided biographies, press releases, and position papers to journalists. More recently, bloggers have made extensive use of this information. Desiring positive news coverage and exposure, campaigns work hard to make this information accessible. Traditionally, however, the campaign authors all this information and makes it available on the campaign’s website. This authorship reduces the credibility and persuasiveness of the information, because viewers and readers expect the campaign to frame the election in a self-interested and non-objective manner. A more decentralized campaign, in contrast, seeks to diversify the information that voters receive about the campaign. The campaign website will include information authored from outside the official campaign. The website may aggregate stories from newspapers or unaffiliated blogs. The website can also operate its own blog on which visitors can provide their comments about the election. In addition, the campaign can make it even easier to place campaign information and material elsewhere on the Internet. Campaign advertisements posted as YouTube videos are in the public domain and are easily posted on blogs or outside websites. A shared component of these efforts is a reduction in control over the campaign’s message. Such a move holds great risk: visitors to the campaign website may find critical or embarrassing information about the candidate on the campaign blog. Another website may post and portray one of the campaign’s ads in a negative light. Yet, the benefits of reducing control are also great (Schaffner and Sellers 2007). By posting stories from unaffiliated and therefore credible news outlets, the campaign may validate its own claims that appeared suspect when coming only from the campaign. By posting YouTube videos from the campaign, bloggers spread the videos to blog readers independent of campaign-driven communications. The move toward decentralized campaigns is important for two reasons. First, traditional centralized campaigns rely heavily on promoting a single, consistent message in all ways at all times (Shea and Burton 2001). Candidates and their staff receive encouragement and training to stick to the campaign message and relate every issue to a central theme. The decentralized campaign moves away from this narrow focus on “staying on message.” Instead, a community of volunteers and bloggers help develop and promote the campaign’s message, through comments and suggestions on the campaign website and supportive blogs. The decentralized process may prove harder for a campaign to guide and manage. The campaign must allow supporters to use their own words and arguments, while working to ensure that the themes and frames do not diverge too far from those preferred by the campaign. The broader involvement of supporters means that the campaign’s message is less likely to consist of a fixed combination of words, such as “preserve, protect, and strengthen Medicare”

Mobilizing to Frame Election Campaigns 85 (Jacobs and Shapiro 2000). But, the main themes and arguments of the campaign may receive wider distribution and acceptance, because more people outside the campaign leadership are involved in developing and promoting those themes and arguments. Second, the supporters’ expanded role in developing and promoting campaign messages tends to blur the distinction between these activities and the mobilization of supporters. This mobilization becomes more important to framing the campaign and its dialogue. Mobilized volunteers recruit new supporters, contribute to the campaign’s messages, and help distribute those messages. Jerome Armstrong, founder of the popular liberal blog MyDD and paid consultant to several Senate campaigns in 2006, summarized the newly prominent role of bloggers in particular: [Early in the campaign] a candidate reaches out to the base, and specifically over the internet, reaching out to bloggers, especially those most local . . . the candidate aims to get buzz from the activists on the ground, and recognizes the local blogosphere as the pre-field organization of the campaign. At this stage, it’s a very decentralized operation, and the candidate’s message works because he’s seen the movement rally around it, and everything going forward builds off of this base. [Later in the campaign], when the campaign brings on field staff, they are integrated with the grassroots. The communication staff are in daily [contact] with the bloggers for message and tactics, and rapid response and oppo[sition] research. (Armstrong 2006) Another leading blogger described the impact of this close interaction: The media messaging created a context for conversation, the internet supplied that conversation, and the field campaign organized around that conversation. If you hear compelling ideas from enough sources, and then a friend asks you to act on those ideas, you do. (Stoller 2006) In sum, the decentralized strategies can both mobilize supporters and frame the election campaign. Implications These arguments suggest several specific patterns of behavior in election campaigns. First, most campaign websites are likely to provide the more traditional opportunities for supporter involvement (e.g., email signup lists, financial contributions, etc.), as well as candidate biographies, press releases, and other types of information produced by the campaign staff. At the same time, the websites of more decentralized campaigns will also allow volunteers more control over their activities in support of the campaign (organizing house

86 Origins parties or personal fundraising events). These campaigns will also be more likely to loosen their control of the campaign message, by allowing non-official voices on the campaign website and making it easier to use campaign materials outside the website. Second, certain types of candidates will be more likely to run decentralized campaigns. As detailed below, this decision may hinge on factors such as extent of Internet usage in the candidate’s state and the characteristics of the actual candidate (such as party). Finally, if the strategies of a decentralized campaign are effective, the candidates who adopt these strategies will be more effective in shaping the campaign dialogue. The next section of the chapter provides initial empirical support for these expectations, drawing evidence from the 2006 Senate races.

The 2006 Senate Campaigns In 1998 Jesse Ventura (I-MN) organized much of his gubernatorial campaign through email. During the 2000 primary Republican presidential candidate John McCain demonstrated the potency of Internet fundraising (Larson and Paystrup 2005). Howard Dean’s 2004 presidential campaign often receives credit as the first truly innovative online candidacy. Building on these earlier examples, candidates for the US Senate in 2006 embraced online campaigning and web-based innovations to a degree previously unseen in American politics. Websites emerged as clearinghouses for news, issue positions, and candidate biographies, providing information for journalists and bloggers. At the same time, candidate websites continued to grow into engines for grassroots mobilization and coordination. This section examines which Senate candidates employed which tools, and why. A first step requires listing the possible tools. As discussed above, the tools varied by whether the campaign controlled their use in a top-down, centralized manner, or the campaign used the tools to empower supporters and distribute information in a decentralized manner. To assess how frequently the 2006 Senate campaigns employed the various tools, we turned to the campaign websites. We downloaded all candidate websites one month in advance of the general election, on October 5, 2006.1 With few exceptions,2 we placed the elements3 present on the homepages into the centralized and decentralized categories. Table 5.1 lists the various possible tools and their categorization. Certain decentralized tools are particularly noteworthy. Although inherently fundraising events, “house parties” represented an extremely visible decentralization of campaigning, empowering dedicated volunteers to host events in their own homes. Utilized by seven campaigns in 2006, the tool allows supporters to schedule, plan, advertise, and host events without centralized control from the campaign itself. Claire McCaskill (D-MO) emphasized the importance of the volunteer-driven events: “Successful House Parties are driven by dedicated hosts. These hosts are people just like you, people willing to welcome their friends and neighbors to discuss the importance of electing Claire to the United States

Mobilizing to Frame Election Campaigns 87 Table 5.1 Campaign Website Tools Centralized

Decentralized

• • • • • • • • • • • • • • • • • • • • •

• • • •

Event calendar Email list signup Volunteer signup Coalitions Contact the campaign Donate money Campaign store Petition Yard sign/bumper sticker request Register to vote Register others to vote Contact district office Vote by mail Debate schedule Voter survey/online poll Downloads (literature, walk lists, etc.) Candidate biography Issue positions Accomplishments Press releases Endorsements

• • • • • • • • • • •

House parties Postcard tool Letter to the editor tool Password-protected volunteer resources “Tell a friend” feature Wikipedia incorporation Online chats with the candidate Voter endorsements/“Tell your story” feature “Raise money for the candidate” feature “Share ads with friends” feature County-by-county fundraising challenge Invite candidate to a local event Aggregated news stories Campaign blog with comments or diaries Multimedia

Senate” (http://claireonline.com/action/houseparties.jsp.html, October 10, 2006). Campaigns that encouraged house parties provided an easily searchable database of planned parties, enabling supporters to connect with others in their geographic vicinity. Bob Casey (D-PA) urged interested individuals to participate: “Attending a Team Casey event is great way to meet people in your community and find out how you can help in this critical election. Both volunteer-hosted and official campaign events from all over the state are listed—simply enter your zip code and find out what is happening in your area” (www.bobcasey.com/ actioncenter/event/index.html, October 2, 2006). House parties raised funds for the campaigns and encouraged social interaction and efficacy among supporters. Among the information tools, campaign blogs served as rapid-fire response mechanisms for campaigns, while providing immediate information for activists, bloggers, and reporters alike. Along with news and event updates, the campaign blogs frequently served as a source for blog-friendly media, providing videos of candidate stump speeches, photos of campaign events, and links to articles that bloggers might find interesting. While many campaign blogs allowed comments on their entries, Sen. Bob Menendez’s (D-NJ) blog went even further, providing a forum for supporters to post their own blog entries (“diaries” in blogging parlance) and thus draw the campaign’s attention to their concerns (http://www.blog.menendez2006.com, October 2, 2006). Campaigns also used email and websites to draw reporters’ and bloggers’ attention to new campaign advertisements, attacks against the opposing

88 Origins candidate, or immediate responses to attacks on their candidate. As one staffer explained, campaigns assume that reporters, bloggers, and opposition researchers subscribe to their email lists; therefore, emails target supporters, sympathetic bloggers, and journalists alike (Vellis 2006). Claire McCaskill (D-MO) sent an email granting a “Sneak Preview” of her ad featuring Michael J. Fox discussing stem cell research (McCaskill email, October 20, 2006). Similarly, Ben Cardin’s (D-MD) opponent Michael Steele (R-MD) responded to a “Cardin smear campaign” with an ad released via email (Steele email, October 27, 2007). When campaigns wanted their message and ads promoted in traditional media and blogs, they emailed journalists, bloggers, and supporters, often giving those groups the first glimpse of new ads or lines of attack.4 Beyond these individual tools, how frequently did the campaigns rely upon the different types of tools? Across the 61 campaign websites that we accessed, the mean number of centralized tools was 8.46. Maryland candidate Michael Steele (R) used the greatest number of these tools (16), while challengers Alan Schlesinger (R-CT) and Jean Hay Bright (D-ME) used the fewest (4). Significantly, all 61 campaigns used at least some of these tools on their websites, illustrating how widely the Internet and web-based technology have spread through election campaigns. The campaigns employed the decentralized tools less frequently, averaging only 3.70 tools per website. Nevada Democrat Jack Carter used the most tools (10). Two candidates used none of the decentralized tools: Senator Edward Kennedy (D-MA) and challenger Robert Lorge (R-WI). The lower usage likely reflects the newness of the decentralized campaign strategies and a corresponding lack of knowledge of them among many candidates. Wikipedia, for example, began in 2001, and YouTube came online only in 2005. The Use of Website Tools In light of the varying reliance on different types of tools, it is important to consider which candidates and campaigns are using the tools more extensively. In states with greater use of the Internet, candidates may be more likely to embrace new campaign technologies, particularly the decentralized strategies. These approaches may more effectively motivate supporters and shape campaign dialogue if a higher proportion of the electorate is familiar with the Internet. In a similar vein, more populous states contain more potential volunteers and voters, many of whom may be amenable to Internet-based appeals. Similarly, these larger states have more extensive news organizations more capable of taking advantage of the decentralized information tools offered by campaigns. In addition to the context of the state, it is important to consider the context of the election. At the start of the campaign, Democrats held a minority of seats in the Senate. Their minority status may have encouraged them to pursue new technologies more vigorously; the GOP, in contrast, may have been more comfortable in its majority and therefore less motivated to innovate in their election campaigns (Jacobson 2001). In addition, Democratic candidates often

Mobilizing to Frame Election Campaigns 89 cater to a “base and clientele” more adaptable to emerging technologies (Swanson 2006). A similar dynamic may have resulted from the competitiveness of individual races. Candidates in competitive races may have been the most motivated to try innovative campaign strategies, such as the various decentralized website tools. In a close race, failing to use these tools could make the difference between winning and losing. Uncompetitive contests did not create the same incentive, so candidates in these races may have embraced the new strategies less extensively. A final factor to consider is the extent to which a candidate has served in Congress. Senior senators such as Robert Byrd (D-WV) have won reelection for decades, which increased their confidence in their previous campaign strategies and decreased their interest in trying new strategies. In a similar manner, candidates such as Ben Cardin (D-MD) had served in the House for years and may have had confidence in the campaign strategies developed in their successful reelection campaigns in that chamber. In contrast, candidates with little or no experience in running for Congress may have had little or no reason to be confident in their campaign strategy, particularly if 2006 marked their first run for the Senate. These individuals may have been more open to and interested in adopting new campaign strategies. To examine whether these factors help explain reliance on the tools in Table 5.1, we collected measures of state Internet penetration, state population, candidate party, competitiveness, and experience in Congress.5 We then assessed the five factors’ impact using a statistical technique called regression. This technique can assess whether a “dependent variable” such as reliance on website tools is significantly related to a variety of factors (each called an independent variable). The analysis can yield regression coefficients, indicating the strength and direction of relationship, and predicted values, suggesting the value of the dependent variable associated with particular values of an independent variable. Using the technique of regression, we compared the five independent variables above to each of two dependent variables: the campaigns’ reliance on the centralized and decentralized tools in Table 5.1. The estimated coefficients of the regression models appear in Table 5.2. In the model explaining use of the centralized tools, none of the independent variables produced significant coefficients. The use of these tools appears so widespread that any variation in their use is random and not linked to any of the explanatory factors that we identified. Many of the tools are relatively easy to implement technologically. The tools are also consistent with the centralized, top-down approach to campaigning that has characterized electoral strategy for decades. In contrast, reliance on newer tools and a more decentralized strategy were less uniform across the campaigns, and two factors help explain this variation. Partisanship exerted the strongest influence on use of the decentralized tools. The model predicts that the average Republican candidate included 2.55 of these tools on the campaign website (the 95 percent confidence interval around this prediction ranges from 2.00 to 3.18). The corresponding prediction for the

90 Origins Table 5.2 Models of Campaign Tools (Standard Errors) Centralized Internet access Households Party Congressional experience Competitiveness of race

–1.05

–1.04

(0.98)

(1.50)

0.03

0.02

(0.02)

(0.02)

–0.03

–0.63***

(0.09)

(0.14)

0.001

–0.008

(0.004)

(0.006)

0.02

0.16**

(0.03) Constant

Decentralized

2.66***

(0.05) 1.80*

(0.57)

(0.88)

N

61

61

LL

–138.67

–119.21

LR χ2

2.08

Pseudo R2

0.01

33.72*** 0.12

* p < 0.05, ** p < 0.01, *** p < 0.001.

average Democrat was 4.75 (with a confidence interval of 3.99 to 5.56). The separation of the confidence intervals gives us great confidence that reliance on the tools varied significantly by party. The Democratic candidates used nearly twice as many of the decentralized tools as the Republicans did. The other significant influence came from the competitiveness of the individual Senate race. The average candidate in the least competitive races used 2.81 decentralized tools on the campaign website (with a confidence interval of 2.27 to 3.51). In the most competitive races, the average candidate used 4.61 tools (with a confidence interval of 3.74 to 5.60). The non-overlapping confidence intervals again indicate that the competitiveness of the race was significantly tied to use of the decentralized tools. The other independent variables were not significantly related to the use of the decentralized tools. Campaigns appeared to use these tools with little regard to the candidate’s campaign experience or the number of people or use of the Internet in the candidate’s state. The insignificance of these factors points to the growing use of the tools. Even in small, rural states, candidates like Jon Tester (D-WY) find and embrace the new technologies as an effective way to reach out to voters and shape the campaign dialogue. Dedicated supporters, journalists, and bloggers are online everywhere, even in states with low percentages of Internet access. Overall, Democrats and candidates in competitive races were most likely to include the decentralized tools on their campaign websites. Their use of the

Mobilizing to Frame Election Campaigns 91 tools indicates that they believe the tools could help them electorally. The next section of the paper examines the accuracy of that belief.

Decentralization and Dialogue Many of the decentralized website tools offer the promise of greater social benefits for volunteers, as well as more extensive promotion and discussion of a campaign’s arguments and issues. To assess whether these tools have the desired impact on campaign dialogue, we focus on the 2006 Senate race in Ohio. Incumbent Senator Mike DeWine (R) and challenger US Representative Sherrod Brown (D) differed sharply in their use of the decentralized strategies and tools. As a result, our arguments above suggest that the candidates had varying success in framing the campaign. DeWine and Brown both used many of the same centralized tools. Both campaign websites offered opportunities to volunteer, contribute, and purchase campaign merchandise. Available information included a candidate biography, press releases, and an issues page. Overall, our coding found seven centralized tools on DeWine’s website, and ten on Brown’s site. The candidates differed more widely in their use of decentralized tools. We found only two such tools on DeWine’s website: aggregated news stories about the campaign, and an encouragement to write a letter to the editor of a local newspaper (including email addresses for state papers). Brown, in contrast, used eight decentralized tools. His website provided details for hosting or attending a house party, as well as raising money for the candidate. The campaign’s main page linked to a separate page containing “Tools for Bloggers and Webmasters,” including links to campaign ads in YouTube and blog postings about the campaign (by Brown and his wife, among others). Our argument suggests that these differences between the campaigns impacted the candidates’ ability to frame public discussion of the election. Brown provided numerous opportunities for volunteers to engage the campaign and each other socially. These supporters could also contribute to the content of Brown’s campaign, while bloggers and journalists found it easy to find and utilize Brown’s multimedia content in their own stories and posts. DeWine offered almost no such opportunities for interaction and exchange of campaign material. The campaigns’ different use of decentralized strategies appeared to shape their influence over public dialogue about the election. Capturing the Dialogue Our examination of the Ohio election dialogue focused on the campaigns’ ability to emphasize preferred policy issues and thereby frame public discussion of the election. We first determined the issues that each campaign discussed most frequently, by examining the policy issues discussed on their campaign websites, in their television advertisements, and in emails sent to supporters.6 This examination yielded seven issues, each of which was among the top five issues

92 Origins discussed by at least one candidate. We also captured how frequently these issues entered discussions of the campaign in blogs and two Ohio newspapers (The Columbus Dispatch and The Cleveland Plain Dealer).7 Table 5.3 lists how frequently the campaigns, blogs, and newspapers discussed the seven issues.8 The issue of the economy received the most attention, with the blogs, newspapers, and candidates all mentioning the issue most or second-most frequently. Other issues drew more varied attention, and their varation suggests that Brown was more successful in framing public dialogue about the election. For example, the Democrat referred to trade more frequently than any other issue (68 mentions), while the topic tied for fifth in frequency among DeWine’s mentions (13). The issue ended up receiving more than 170 mentions in the newspapers and blogs. This issue thus appeared relatively more important to Brown than DeWine, and the Democrat’s frequent mention of the issue corresponded with its fairly extensive coverage in newspapers and blog. In contrast, DeWine’s public relations efforts appeared less successful. For example, his campaign made more mentions (33) of the issue of families and children than any other topic except the economy. Brown’s campaign mentioned the issue only 19 times, which ranked sixth across his discussion of the seven issues. Despite the apparent importance of the issue for DeWine, the newspapers and blogs made only 120 references (combined) to the issue. While important to the incumbent senator, this issue received less public attention than trade, the topic receiving the most emphasis from Brown. Overall, the two campaigns differed substantially in their adoption of new campaign tools, with Brown adopting many more decentralized tools than Dewine did. This contrast between campaigns corresponds with an apparent difference in their ability to frame public dialogue about the Ohio election. Issues important to Brown appeared more prominent in public discussions of the election; DeWine’s preferred issues received less attention. We argue that this correspondence of patterns is not coincidental: Brown’s willingness to Table 5.3 Mentions of Issues in the 2006 Ohio Senate Race Issue

Campaigns Blogs

Newspapers

Brown

DeWine

Economy

262 (2)

112 (1)

61 (2)

48 (1)

Trade

122 (5)

50 (4)

68 (1)

13 (tied 5)

National security

135 (4)

38 (5)

24 (3)

13 (tied 5)

Iraq

335 (1)

69 (2)

21 (5)

16 (tied 3)

Health care

83 (6)

30 (7)

22 (4)

3 (7)

Families/children

54 (7)

66 (3)

19 (6)

33 (2)

159 (3)

36 (6)

5 (7)

Terrorism

16 (tied 3)

Note In each cell, the first number is the number of mentions of the particular issue by the particular outlet or campaign. The number in parentheses is the ranking (by number of mentions) of the issue, across all seven issues by the outlet or campaign.

Mobilizing to Frame Election Campaigns 93 decentralize his promotional efforts likely helped his campaign frame the election around issues more favorable to him. Likewise, DeWine’s reliance on more traditional, centralized campaign tools may have limited his own ability to frame the election. These pieces of evidence are admittedly correlational. We are unable to demonstrate that volunteers’ usage of the decentralized web tools directly led to more discussions of campaign messages in blogs and newspapers. Such a linkage would require detailed information on website traffic, email readership, and volunteer involvement. Campaigns are understandably hesitant to provide this confidential information. But, the evidence suggests that the decentralized strategies had an important impact. Brown embraced the strategies more extensively and appeared more able to frame campaign dialogue in ways consistent with those strategies.

Conclusion The decentralized tools and strategies described above merge two traditionally distinct campaign activities: mobilizing supporters and framing the election. Recognizing that empowered volunteers will be more invested in the success of the campaign and more persuasive in promoting the campaign message, campaigns can effectively target motivated supporters to work as surrogate campaigners in their own communities and peer groups. Normatively, these trends appear to be a welcome development. While the Internet is far from a perfectly democratic medium (Internet access is still prohibitively expensive for many), the further inclusion of engaged and excited voters in political campaigns should lead to a more informed and active public as a whole. When dedicated supporters engage campaigns online and then approach friends and co-workers with the passion they gain from participating, the enthusiasm and interest spread. This development is positive for campaigns hoping to mobilize voters and distribute their message, as well as for activists who access new tools for engaging friends and gaining social benefits from political participation. Since the 2006 elections, the decentralized tools of communication and mobilization have grown even more popular. Phil de Vellis, Brown’s Internet Communications Director, made a splash online with a homemade advertisement posted on YouTube shortly after the start of the 2008 Democratic presidential primary season. The video, using a famous Apple Computer advertisement from its “Think Different” campaign, inserted images of Sen. Hillary Clinton (D-NY) with the tag line “Vote Different” and then-Clinton rival Sen. Barack Obama’s (D-IL) campaign website address (http://youtube.com/watch?v=6h3GlMZxjo. April 30, 2007). This video received over three million views in just under two months.9 De Vellis, who initially released the video anonymously, later claimed credit and used the success of his project as a demonstration of the power of grassroots campaigning. “I wanted to show that an individual citizen can affect the process,” de Vellis wrote on a popular blog. “This ad was not the first citizen ad, and it will not be the last. The game has changed” (Vellis 2007).

94 Origins As this chapter’s introduction suggested, Obama and his supporters continued to develop innovative campaign strategies throughout the 2008 election campaign. Volunteers produced YouTube videos touting the candidate10 and sent “counterviral emails” to offset widespread email rumors about Obama’s religion and background. Supporters developed an application for the iPhone and iPod Touch “that enabled owners to mobilize their friends and contacts in battleground states through the Apple devices” (Stirland 2008). These innovations furthered the goals of framing the election in a favorable manner and mobilizing supporters to promote the campaign’s messages in credible, persuasive ways. Despite its many successes, the 2008 Obama campaign also illustrates the potential dangers facing campaigns that embrace decentralized tactics. In July 2008, the largest self-organized group of supporters utilizing the Obama campaign’s social networking and mobilization tool (my.barackobama.com) was a group of supporters protesting the then-Senator’s stance on the controversial Foreign Intelligence Surveillance Act (Melber 2008b). The protest prompted a response from the Obama campaign and received widespread coverage in the press. The controversy created a detour from the agenda the campaign likely would have preferred to emphasize (Melber 2008a). In decentralized campaigns, individual volunteers help craft the campaign’s message and promote it persuasively. These efforts help the campaigns frame the election in favorable ways. We argue that campaigns appear increasingly willing to accept the risks that accompany decentralized messaging and mobilization, because the new strategies can help the campaigns win.

Notes 1. We downloaded almost all websites in their entirety using a program called SiteCrawler. We downloaded websites from Ohio on a weekly basis, from early September to the end of the campaign. Unfortunately, Jim Webb’s (D-VA) website did not download properly, forcing us to omit his campaign from the analysis. 2. We disregarded foreign-language options, for example, since offering a translated site usually did not involve different content. We did not code charitable links or graphics counting down the days until the election, displaying the ever-rising trade deficit, or totaling the cost of the Iraq war. 3. “Elements” refers to links on a candidate’s homepage (“News Center,” “Issues,” etc.) and information present on the page itself (such as a campaign video or email signup option). We also coded any page linked from a candidate’s homepage that offered volunteer activities, following links with labels like “Get Active,” “Get Involved,” “Take Action,” “Action Center,” and “Supporter Resources.” 4. See Ansley (2007) for more details on the specific tools. 5. The measure of Internet penetration is the percentage of a state’s households that have Internet access. The population measure is the number of households in a state (in millions). For the party measure, Democratic candidates receive a 0 and Republicans a 1; in Connecticut both Ned Lamont and Joseph Lieberman received a 0. The measure of competitiveness comes from Charlie Cook; the least competitive races receive a 1, while the most competitive ones receive a 4. Finally, the measure of congressional experience equals the number of years that a candidate had served

Mobilizing to Frame Election Campaigns 95

6.

7.

8. 9. 10.

in either chamber of Congress. In the models assessing these effects, the dependent variables are counts. Testing suggested that over- and under-dispersion were not a problem, so we used Poisson regression to estimate the impact of the independent variables on the two dependent variables. We found substantively identical results using OLS regression. Each week we downloaded the entire content of both websites. If a website mentioned an issue on the day of a weekly downloading, we assumed that the website mentioned that issue for the entire week. We obtained transcripts of 29 campaign ads from National Journal Group’s Policy Central website. We assumed that each ad ran on the day that it was published and then for three days afterwards (this assumed length of each ad buy is arbitrary, but using different lengths does not appear to affect the conclusions of the analysis below). In August 2006 one of the authors signed up (via campaign websites) to be on email distribution lists from both campaigns. For each campaign, we included emails sent between September 6, 2006 and November 7, 2006 (Election Day). We used Google’s Blogsearch to determine the daily mentions of each campaign issue. This engine aims to include every blog in the world which produces a site feed (http://blogsearch.google.com/). We used Lexis-Nexis to search the two newspapers. In the searches for each issue, we used the phrase ((“Sherrod Brown” OR “Mike DeWine”) AND Senate AND Ohio), plus the following words or phrases for specific issues: economy (economy OR jobs OR taxes OR “minimum wage”), trade (trade), national security (“national security” OR “homeland security” OR “national defense”), Iraq (Iraq), health care (“health care”), families and children (families OR children), and terrorism (terrorists OR terrorism). The calculations for the campaigns combine mentions on the campaign website, in campaign ads, and in emails. As of May 4, 2007, the video had been viewed 3,213,370 times. Perhaps most famous among these supporter-produced YouTube videos was the “Yes We Can” video produced by and starring singer will.i.am and other celebrities. By the end of November 2008, that video had been viewed more than 14 million times. http://www.youtube.com/watch?v=jjXyqcx-mYY.

Works Cited Aldrich, John A. 1995. Why Parties? The Origin and Transformation of Political Parties in America. Chicago, IL: The University of Chicago Press. Ansley, Taylor. 2007. “From Blogs to the Ballot Box: Online Campaigning in 2006 U.S. Senate Races.” Honors Thesis, Department of Political Science, Davidson College. Armstrong, Jerome. 2006. “How the Blogs and the Internet Change Campaigning.” MyDD, 10 August. http://mydd.com/story/2006/8/10/95711/6854. Arnold, Marguerite. 2003. “TV Ad Production: A Political Campaign Primer.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Beck, Paul Allen, Russell J. Dalton, Steven Greene, and Robert Huckfeldt. 2002. “The Social Calculus of Voting: Interpersonal, Media, and Organizational Influences on Presidential Choices.” The American Political Science Review (March). Blaemire, Bob. 2003. “Database Management and the Permanent Campaign.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Bovee, John. 2003. “Using the Internet for Opposition Research.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc.

96 Origins Buchanan, Bruce I. 2001. “Mediated Electoral Democracy: Campaigns, Incentives, and Reform.” In W. Lance Bennet and Robert M. Entman, editors, Mediated Politics: Communication in the Future of Democracy. New York: Cambridge University Press. Connell, Mike. 2003. “Designing Lively Web Sites.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Darr, Carol and Julie Barko. 2004. “Under the Radar and Over the Top: Online Political Videos in the 2004 Election.” Institute for Politics, Democracy, and the Internet (October 20). Darr, Carol, Julie Barko, and Brandon Robinson. 2004. “Putting Online Influentials to Work for Your Campaign.” Institute for Politics, Democracy, and the Internet (July 21). Desmone, Roseanne. 2003. “How Technology Is Making Telephone Campaigning More Effective.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Drezner, Daniel W. and Henry Farrell. 2004. “The Power and Politics of Blogs.” Presented at the Annual Meeting of the American Political Science Association, August. Duquin, Lorene Hanly. 2003a. “Door to Door Campaigning.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Duquin, Lorene Hanly. 2003b. “How to Work a Room.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Elam, Doug. 2003. “Increasing Turnout with Signs and Volunteers.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Faucheux, Ronald A. 2003a. “Citizen Leader Programs: How to Build a Grassroots Support Network.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Faucheux, Ronald A. 2003b. “Newspaper Ads: An Underrated Campaign Tool.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Germany, Julie Barko. 2006. “Person-to-Person-to-Person: Harnessing the Political Power of Online Social Networks and User-Generated Content” [Executive Summary]. Institute for Politics, Democracy, and the Internet (September 15). Hardin, Garrett. 1968. “The Tragedy of the Commons.” Science 162: 1243–1248. Hathaway, T.R. 2003. “Electronic Field Operations: A Case Study.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Haythornthwaite, Caroline. 2002. “Strong, Weak, and Latent Ties and the Impact of New Media.” The Information Society 18. Herrnson, Paul. 2003. Congressional Elections: Campaigning at Home and in Washington. Washington, DC: CQ Press. Hindman, Matthew. 2005. “The Real Lessons of Howard Dean: Reflections on the First Digital Campaign.” PS 3(1) (March). Iyengar, Shanto and Donald R. Kinder. 1987. News That Matters. Chicago, IL: The University of Chicago Press. Jacobs, Lawrence and Robert Y. Shapiro. 2000. Politicians Don’t Pander. Chicago, IL: The University of Chicago Press. Jacobson, Gary. 2001. The Politics of Congressional Elections, 5th edition. New York: Longman. Kanazawa, Satoshi. 1998. “A Possible Solution to the Paradox of Voter Turnout.” The Journal of Politics 60(4) (November).

Mobilizing to Frame Election Campaigns 97 Kenny, Christopher B. 1992. “Political Participation and Effects from the Social Environment.” American Journal of Political Science 36(1) (February). Larson, Suzanne and Patricia Paystrup. 2005. “The Internet and Democracy: John McCain’s New Approach to Politicking.” In Mitchell S. McKinney, Lynda Lee Kaid, Dianne G. Bystrom, and Diana B. Carlin, editors, Communicating Politics: Engaging the Public in Democratic Life. New York: Peter Lang Publishing. Malchow, Hal. 2003. “New Campaign Targeting Techniques.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Melber, Ari. 2008a. “Online Activists Keep the Pressure on Obama.” The Nation (July 7). Melber, Ari. 2008b. “Surveillance Protest Group Tops Obama Web Site.” State of Change (a blog of The Nation), July 1. http://www.thenation.com/blogs/ state_of_change/333805. Mendelson, Adam and Chris St. Hilaire. 2003. “Winning the Earned Media Game: A New Campaign Model.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Mindich, David T.Z. 2005. Tuned Out: Why Americans Under 40 Don’t Follow the News. New York: Oxford University Press. Riker, William. 1986. The Art of Political Manipulation. New Haven, CT: Yale University Press. Riker, William H. and Peter C. Ordeshook. 1968. “A Theory of the Calculus of Voting.” The American Political Science Review 62(1) (March). Sandler, Todd. 1992. Collective Action: Theory and Applications. Ann Arbor: University of Michigan Press. Schaffner, Brian and Patrick Sellers. 2007. “Winning Coverage in the U.S. Senate.” Political Communication 24: 377–391. Schneider, Dick. 2003. “Precinct Organization: From the Ground Up.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Shea, Daniel M. and Michael J. Burton. 2001. Campaign Craft: The Strategies, Tactics, and Art of Political Campaign Management. Westport, CT: Praeger. Sheehan, Michael. 2003. “On the Stump: Fundamentals of a Campaign Speech.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Shimmel, Howard. 2003. “Cable TV Spots: Reaching Swing Voters.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. “The Social Pulpit: Barack Obama’s Social Media Toolkit.” 2009. Edelman Digital Public Affairs Team Insights. (January). http://www.edelman.com/image/insights /content/Social%20Pulpit%20-%20Barack%20Obamas%20Social%20Media%20 Toolkit%201.09.pdf. Stirland, Sarah Lai. 2008. “Propelled by Internet, Barack Obama Wins Presidency.” Wired.com, November 4. http://blog.wired.com/27bstroke6/2008/11/propelled -by-in.html. Stoller, Matt. 2006. “Turnout, Turnout, Turnout.” MyDD, August 9. http://mydd. com/story/2006/8/9/123937/5061. Swanson, Cory, Deputy Campaign Manager, Conrad Burns for US Senate. 2006. Personal interview, Billings, Montana, August 8. Swanson, Eric R. 2003. “Permission Marketing: New Campaign Tool.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc.

98 Origins Trent, Judith S. and Robert V. Friedenberg. 1991. Political Campaign Communication: Principles and Practices. New York: Praeger. Trippi, Joe. 2004. The Revolution Will Not Be Televised: Democracy, the Internet, and the Overthrow of Everything. New York: HarperCollins. Vellis, Phil de. 2006. Internet Communications Director, Sherrod Brown for US Senate. Personal interview, Cleveland, Ohio, July 27. Vellis, Phil de. 2007. “I Made the ‘Vote Different’ Ad.” The Huffington Post, March 21. http://www.huffingtonpost.com/phil-de-vellis-aka-parkridge/i-made-the-votedifferen_b_43989.html. Webb, Catherine and Joseph Mockus. 2003. “Volunteer Recruitment.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc. Wellman, Barry, Anabel Quan Haas, James Witte, and Keith Hampton. 2001. “Does the Internet Increase, Decrease, or Supplement Social Capital? Social Networks, Participation, and Community Commitment.” American Behavioral Scientist 45(3). White, Joe Slade. 2003. “Radio Spots That Pop.” In Ronald A. Faucheux, editor, Winning Elections. New York: M. Evans & Company, Inc.

Part II

Impact

6

Competing Frames in a Political Campaign James N. Druckman1

Nearly all framing studies examine whether contrasting frames, when considered individually, can shift public opinion on an issue.2 The typical experiment randomly assigns individuals to receive one of two alternative representations of an issue. For example, in studies of people’s willingness to allow hate groups such as the Ku Klux Klan to conduct a rally, individuals learn of the issue framed either in terms of free speech (e.g., all groups have a right to speak) or in terms of public safety (e.g., rallies often lead to violent confrontations between the hate group and counter-demonstrators). In this case, the relevant comparison is the difference of opinion between individuals in the two conditions. Such studies employ one-sided designs insofar as the exclusive focus is on how exposure to a single frame affects opinions. Most framing studies find that contrasting frames nearly always have a statistically significant impact when compared to one another. For instance, individuals exposed to the free speech frame are significantly more willing to allow Klansmen to rally than are individuals who receive the public safety frame (e.g., Nelson et al. 1997). This research therefore suggests that, if one side can establish the relevant terms of debate over an issue, it can successfully persuade individuals to support its position. Politics, however, is typically competitive, fought between parties or ideological factions, and debated issues are framed in opposing terms. The strategic use of framing to mobilize public opinion on a contested issue is a tactic available to all sides. Surprisingly, social scientists have little to say about which of many competing frames (e.g., free speech, public safety, opposition to racism) will shape public opinion. Sniderman and Theriault (2004, pp. 141–142) explain that “framing studies . . . have neglected the fact that frames are themselves contestable. They have instead restricted attention to situations in which citizens are artificially sequestered, restricted to hearing only one way of thinking about a political issue” (also see Entman 1993; Riker 1995; Wittman 1995).3 In this chapter, I build on the work of Sniderman and Theriault (2004) and Chong and Druckman (2007a, 2007b, 2007c) to explore how opinion formation works in competitive mass communication (framing) environments. I do so by first reviewing Chong and Druckman’s (2007a, 2007b) approach to studying competitive framing. I then present a new study of competitive framing

102 Impact that focuses on a publicly funded casino proposal. The results of the study show that not all frames are effective, that competition plays a critical role when it comes to framing, and that the key to being an effective frame lies in a frame’s “strength.”

Framing and Campaigns A large number of studies over the last 20 years show that framing effects have the potential to fundamentally shape public opinion. These studies demonstrate that for a given issue (e.g., a hate group rally request) the opinions of participants (randomly) exposed to one frame (e.g., free speech frame) significantly differ from those of participants (randomly) exposed to another frame (e.g., public safety frame). The problems with these studies, which have been carried out on a wide range of issues including affirmative action, welfare, and gun control, are that they (1) focus exclusively on frames that are effective (i.e., that successfully impact public opinion), and (2) expose individuals to just a single frame (e.g., either free speech or public safety). These features bear little resemblance to the reality of most political contexts. A number of factors have been shown to make a given frame ineffective. For example, frames from non-credible sources (e.g., the National Enquirer) or that contradict strongly held values (e.g., a public safety frame to free speech advocates) do not affect public opinion (e.g., Brewer 2003; Druckman 2001a, 2001b). Not all frames work. On the second point—concerning exposure to just a single frame—even a cursory glance at real-world political campaigns makes clear that individuals receive various frames from competing sides (e.g., the ACLU promotes a free speech frame while the NAACP emphasizes safety concerns or a racial prejudice frame). More systematic evidence on this point comes from Chong and Druckman’s (n.d.) study of the framing of 14 distinct issues. The authors analyzed coverage of these issues in major newspapers over time, counting the number of frames put forth on each issue (as well as other features of the frames). They then computed a score to capture the “total effective number of frames,” which amounts to a weighted score of the number of frames used on a given issue (e.g., frames employed more often receive greater weight).4 (The term “effective” here does not refer to the success of the frame in affecting public opinion; but rather, to the extent of its presence in newspaper coverage on the issue.) Across the 14 issues, the average effective number of frames is 5.09 (standard deviation = 1.19). The issue for which the fewest effective frames were employed was coverage of a 1998 Ku Klux Klan rally in Tennessee (with 3.03 effective frames including free speech, public safety, and opposing racism). The issue with the most frames was coverage of the 2004 Abu Ghraib controversy concerning prisoner abuse by members of the armed forces (with 6.9 effective frames, including military responsibility, presidential administration responsibility, individual responsibility, military commander responsibility, negative consequences for international relations, positive consequences for international relations, negative domestic consequences, positive domestic consequences, etc.).

Competing Frames in a Political Campaign 103 Importantly, on each issue—including the Ku Klux Klan rally and Abu Ghraib—many of the frames employed competed with another, meaning they came from opposing sides. As mentioned, a free speech frame of a hate group rally likely increases support while a public safety frame decreases it. Similarly, the Abu Ghraib individual responsibility frame suggests that fault lies with the individuals involved whereas the administration or military commander frames put the bulk of the blame on the culture established by higher-level actors. That opposing sides simultaneously employ competing frames also is evident in several of the chapters from the first part of this volume. Gerrity reveals how different sides offer competing interpretations of the partial-birth abortion debate, Harris shows how parties in Congress compete by providing alternative frames of the same issue, and Nelson describes how supporters of neo-creationist theories of evolution reinvented themselves by incorporating new values (e.g., a marketplace of ideas) in an effort to counter supporters of conventional evolutionary theories. These chapters suggest that some frames succeed and others fail in affecting public opinion. Yet, as mentioned, little work has more systematically—that is, beyond offering impressionistic case studies—explored how competition between frames affects public opinion. One notable exception is Sniderman and Theriault (2004) who argue and demonstrate, with two experimental surveys, that when competing frames are presented with one another (e.g., a free speech and a public safety), they cancel out such that the frames do not affect individuals’ opinions. Chong and Druckman (2007a, 2007b, 2007c) build on Sniderman and Theriault (2004). Specifically, they conceive of “competition” as the presence of frames aimed at supporting different sides of an issue (a “pro” side and a “con” side)— that is, the frames have distinct positional directions. For example, the free speech frame promotes the right to rally (“a pro frame”) while the public safety frame implies opposing that right (“a con frame”). They then identify two dimensions of competition. One dimension concerns the number of frames offered from each side. Continuing with the example, the free speech frame (and/or other “pro frames”) may be presented one time, two times, ten times, etc., while the public safety frame (and/or other “con frames”) could be presented the same or any other number of times. There is variation in the repetition of each side’s frame(s). The other dimension is the strength of the frames—this gets at the likely effectiveness of the frame in actually influencing public opinion. (In this context, “effective” refers to a frame’s impact on opinions and not its presence in media content.) Chong and Druckman (2007b, p. 640) explain that “(perceived) strength refer[s] to the extent to which a frame emphasizes relatively available and applicable considerations.” Availability means that individuals are able to connect a given consideration (e.g., free speech) to the issue at hand (e.g., the hate group rally); they understand that it is a potentially relevant consideration. Availability is assessed by asking respondents to list what considerations come to mind when they think of the issue. Applicability refers to how compelling the frame is perceived to be, and is assessed by asking respondents to rate the relative “effectiveness” of a frame. (While this leaves unclear exactly why a given

104 Impact frame is perceived as effective or not, it is nonetheless analogous to the widely used technique in psychology to assess argument strength.) Overall strength is a relative construct; a frame is deemed relatively stronger than another if it registers significantly higher availability and applicability scores (for detailed discussion, see Chong and Druckman 2007a, 2007b, 2007c). Indeed, different frames on each side might be relatively strong or weak compared to one another. For example, individuals likely perceive the public safety frame to be stronger (both in terms of availability and applicability) than an alternative “con” frame that argues the rally should not be held because it will result in litter in the streets. In short, competition between contrasting frames varies by how often each side is repeated and the relative strength of the frames used on each side. The result is a potentially infinite number of combinations that differ in terms of frequency of repetition and the strengths of frames from opposing sides. Chong and Druckman (2007b) test 16 combinations in two laboratory experiments (one on the issue of urban sprawl and the other on a hate group rally request), finding that frame strength plays the most decisive role; a frame’s relative strength matters more than its repetition (regardless of the side of the argument endorsed by the frame). In what follows, I build on Chong and Druckman’s (2007a, 2007b, 2007c) efforts in two ways. First, I move outside of the controlled laboratory setting to implement an experimental framing study in the context of an Election Day exit poll (on the need to move outside the lab, see Kinder 2007). Second, I look at various combinations of frames that Chong and Druckman (2007b) do not explore. In sum, I explore framing outside of the lab with a heterogeneous sample of voters, exposing them to a novel combination of frames. In the next section, I describe the issue on which I focus: a publicly funded casino. I present the relevant frames and present evidence about their relative strengths. I then turn to the study design which involved experimentally embedding various combinations of frames in an Election Day exit poll, implemented in the Chicago, IL, area. Next, I present the results, which show how various frames, in different combinations, impact public opinion. I conclude with a discussion of what the results imply for future studies of framing.

Framing a Casino I focus on a proposal for a state-funded land-based gambling casino with profit to be used for funding education and property tax relief. The overall attitude of interest, then, is the extent to which an individual supports or opposes a stateowned and state operated gambling casino. I put this issue in the context of the 2006 Illinois Gubernatorial election between Democrat Incumbent Rod Blagojevich, Republican Judy Topinka, and Green party candidate Richard Whitney. The campaign’s initial focus concerned the state of the economy. Illinois’ economy was trending in a negative direction—as prominent state economist

Competing Frames in a Political Campaign 105 J. Fred Giertz explained at the time, “the last three months are cause for concern . . . In 2005, Illinois made up some of the lost ground that took place in the three years that followed the 2001 national recession. 2005 was an encouraging year. So far, 2006 is not” (Reutter 2006). An August 2006 poll showed the economy was by far the most important issue on voters’ minds with 37 percent citing it as the central issue followed by only 19 percent mentioning the war in Iraq and 14 percent saying national security (Rasmussen Report 2006). How the candidates would combat the failing economy and raise revenue appeared as if it would dominate the campaign. Topinka’s economic plan—as enunciated on August 23—revolved around the creation of a land-based state-owned Chicago casino with profit to be used for state spending on education and property tax relief. In contrast, Blagojevich proposed leasing the state lottery to generate revenue. (Both candidates opposed income or sales tax increases; Whitney’s positions were nominally covered throughout the campaign.) An August 24 Chicago Tribune editorial stated that the candidates are “framing this contest for governor just as it needs to be framed: How can a grossly overcommitted state government bend the financial trend lines that point inexorably toward ruin?” Topinka’s casino idea split the public—a mid-September Tribune polled 54 percent in opposition to the casino plan—and cut across partisan lines.5 Indeed, while Blagojevich opposed the proposal, he had just a year earlier proposed to double gambling positions, and Chicago Democratic Mayor Daley was open to the plan (although would presumably have prefered that the city own the casino so as to receive most of the revenue).6 Just as in-depth discussions of the candidates’ competing revenue plans began, however, the campaign took an unexpected turn. With little forewarning, a rash of corruption allegations were launched including accusations that Blagojevich traded state jobs for personal payoffs and improperly spent state money. Topinka also received scrutiny for her role in the administration of previous Governor George Ryan who was on trial for charges of corruption. These events became the focus of a fairly vicious negative campaign which led to the ascendance of corruption as the key campaign issue, overtaking the economy and casino plan, which virtually disappeared from the agenda. A content analysis of Chicago Tribune coverage of the campaign (from the date of the casino proposal until Election Day)7 showed that the casino ended up receiving only 3 percent of the coverage while the budget in general received only 5 percent.8 Initial expectations of more coverage were proven incorrect with the corruption charges becoming the major issue in the campaign; corruption received 45 percent of all coverage. Thus, by the time the experimental survey was implemented on Election Day, the bulk of voters presumably had scant knowledge of the casino proposal, perhaps at most possessing a vague memory of the proposal.9 As I will discuss, the Election Day exit poll experiment entailed (randomly) exposing voters to a description of the proposal using different combinations of frames. A first step in designing the experiment entailed identifying the possible frames for the casino issue and assessing their strengths.

106 Impact I identified the set of possible frames for the casino proposal by examining coverage of proposals for state-owned casinos in various states and exploring advocacy group propaganda on both sides of casino proposals. This led to the identification of seven often-used frames, as presented in Table 6.1. On the pro side (i.e., support the casino), the frames include the economic benefits from the casino, the entertainment value of the casino, and the positive community effects. The con (i.e., oppose the casino) frames include the social costs of the casino, the morality of casinos, and the corruption and/or patronage that could come from the casino. Finally, the politics behind implementing a casino proposal served as a directionally neutral frame. As mentioned, three key elements of frames include their positional direction, the number of times they appear (i.e., repetition), and their strength. To assess strength, I follow the pre-test approach used by Chong and Druckman (2007b); I also use the pre-test to confirm that the directional leanings of the frame (i.e., pro/con) are consistent with what I just described. The pre-test involved a sample of 63 voting-age adults who did not reside in Illinois. It took place approximately one month prior to Election Day. 10 As explained, strength entails two dimensions—availability and applicability—and so I asked pre-test respondents to assess the availability and applicability of each of the seven frames. Like Chong and Druckman (2007b), I assessed the availability of different considerations by randomly asking about half of the respondents (n = 32) the following open-ended question: “Many states are considering proposals for the state to operate land-based gambling casinos. When it comes to such proposals, what types of ideas or considerations do you think people consider? List any ideas or considerations, even if you personally Table 6.1 Casino Frames Frame

Description

Economic benefits

Finances gained from state-owned casino will have positive economic benefits in terms of education, tax relief, job creation, development, etc.

Entertainment

Casinos are harmless entertainment, and people have been doing it for centuries.

Community building

Casinos give a community an identity and help to promote social capital.

Social costs

Casinos have severe social costs (e.g., addiction, debt, suicide, family impact, health). They also mostly adversely affect the poor.

Morality

Gambling is biblically immoral since it is a form of theft.

Corruption

A state-owned casino provides many patronage possibilities with concomitant corruption and crime problems.

Politics

A state-owned casino would need other political support (e.g., from the state legislature).

Competing Frames in a Political Campaign 107 do not think they are particularly important.” The average respondent listed 1.78 ideas. I coded responses focusing on the aforementioned frames. I randomly asked the other group of respondents (N = 31) to evaluate the applicability and direction of the considerations emphasized in each of the frames. Again, like Chong and Druckman (2007b), I listed each consideration and asked respondents to judge its effectiveness (i.e., perceived persuasiveness) (on a 1–7 scale with 1 being definitely not effective, 4 being not sure, and 7 being definitively effective), and its directionality (on a 1–7 scale with 1 being definitely opposed or “con,” 4 being neither, and 7 being definitely supportive or “pro”). In Table 6.2, I report the availability, perceived applicability, and directionality scores. As discussed, strength requires both availability and applicability and is a “relative” concept—such that a frame might be relatively stronger than another if it has higher availability and applicability. The results suggest a fair degree of availability of the economic benefits and social costs frames, and some availability for the entertainment and morality frames. The other frames appear to generally not be available: corruption is significantly less available than entertainment, and neither community building nor politics is listed by any respondents. Clearly, the economic benefits frame and the social costs frame constitute compelling frames in opposing directions. Both display significantly greater perceived applicability than all other frames (at the 0.05 level for all comparisons), but are nearly indistinguishable from one another in terms of applicability (t60 = 1.33, p < 0.20 for a two-tailed test).11 They also significantly differ from one another in terms of direction (t60 = 11.75, p < 0.01)— social costs is a frame opposed to the casino proposal and economic benefits is a supportive frame. Thus, if I find differential effects of these frames, it clearly stems from direction and not strength variations. In sum, I use economic benefits as the pro-strong frame and social costs as the con-strong frame. I also use three weak frames. On the weak pro side, I use the entertainment frame; it is significantly less applicable than both the economic benefits and social costs frames (respectively, t60 = 3.30, p < 0.05; t60 = 4.30, p < 0.01) and is significantly more pro than the social costs frame (t60 = 8.03, p < 0.01).12,13 I use two weak con frames. One is the corruption frame, which has very low availability (6 percent), does not significantly differ from the weak entertainment Table 6.2 Pre-test Results (std. dev. in parentheses) Frame (consideration)

Availability percentage (N = 32)

Perceived applicability score (N = 31)

Directionality (support) score (N = 31)

Social costs Economic benefits Entertainment Morality Corruption Community building Politics

59 56 25 31 6 0 0

4.55 (1.89) 3.97 (1.54) 2.74 (1.39) 2.23 (1.15) 3.00 (1.67) 1.97 (1.20) 1.84 (1.39)

2.52 (1.03) 5.81 (1.17) 4.97 (1.28) 1.90 (0.94) 2.23 (1.15) 4.81 (1.28) 3.65 (1.43)

108 Impact Table 6.3 Experimental Casino Frames Supportive (pro)

Opposed (con)

Strong

Economic benefits

Weak

Entertainment

• Morality • Corruption Social costs

frame in terms of applicability (t60 = 0.67, p < 0.55), and is significantly less applicable than the strong economic benefits and social costs frames (respectively, t60 = 2.38, p < 0.05; t60 = 3.42, p < 0.01).14 In terms of direction, it significantly differs from the pro entertainment and economic benefits frames (respectively, t60 = 8.87, p < 0.01; t60 = 12.15, p < 0.01). The other weak con frame is morality, which is significantly more opposed than the entertainment and economics frames (respectively, t60 = 10.76, p < 0.01; t60 = 14.51, p < 0.01), and does not substantially differ in applicability from the weak entertainment frame (t60 = 1.57, p < 0.12 for a two-tailed test).15,16 Table 6.3 summarizes the frames I employ in the experiment.

Experimental Procedure and Design I tested the impact of various combinations of competing frames by embedding experimental conditions in an Election Day exit poll. The exit poll makes for a relatively realistic context in which to assess framing effects since the respondents had just voted in an election where the issue at hand (i.e., the casino proposal) had at least some relevance.17 I implemented the exit poll survey experiment by assembling 24 teams of two student pollsters. I then randomly selected polling locations throughout the northern part of Cook County, Illinois. Each polling team spent a randomly determined 2–3 hour daytime period at their polling place. A pollster asked every third voter to complete a selfadministered, anonymous questionnaire in exchange for $5. The main dependent variable asked “A proposal is being considered for the Illinois state government to operate a land-based gambling casino. What do you think—do you oppose or support the proposal for a state-run gambling casino? Circle one number on the following 7-point scale (where 1 = oppose strongly, 4 = not sure, 7 = support strongly).” I incorporated the framing conditions—which I will soon discuss—by altering the wording of this question. The surveys included various other items meant to capture potential individual level correlates of casino support.18 One question measured a respondent’s values toward government regulation of business, with the expectation that increasing support for regulation would correlate with higher support for the casino. The precise question asked “In general, do you feel that government regulation of business: usually does more harm than good; or is necessary to keep businesses from engaging in practices that are harmful to the public?” with higher scores indicating increased support for regulation. Another measure asked respondents:

Competing Frames in a Political Campaign 109 “How many times have you ever been to a casino?” Response options included “Never,” “1–2 times,” “3–5 times,” “6–10 times,” and “>10 times” (see Donahue and Miller 2006). Given the prominence of the corruption theme in the gubernatorial campaign, I also asked respondents: “In your opinion, to what extent, if any, has the Blagojevich administration engaged in corrupt practices?” with higher scores indicating increased perceptions of corruption. I imagine voters who see the current administration as more corrupt will be less likely to involve state government in running a casino. Along similar lines, I included a standard trust in government item, asking “How much of the time do you think you can trust the government in Washington to do what is right?” (with choices being “just about always,” “most of the time,” or “some of the time”). The survey included standard demographic measures that asked for respondents’ party identification (on a 7-point scale with higher values indicating more Republican), gender (0 = male, 1 = female), minority status, age, and a few other items.19 In Table 6.4, I report the descriptive statistics for the sample. (The Ns vary by variable because of non-responses.)20 Impressively, the vote totals of 61 percent for Blagojevich, 19 percent for Topinka, and 18 percent for Whitney match almost perfectly the actual totals that the candidates received in the universe of the polling area (where they received 61 percent, 21 percent, and 18 percent respectively).21 The table also shows that the respondents come from fairly diverse backgrounds; although, as would be expected in northern Cook County, the sample is skewed towards Democrats and voters who are both politically interested and involved.22 Given the experimental approach, along with my ability to control for these variables, the focus on these voters is not problematic; moreover, it is a representative sample of actual, heterogeneous voters from the area (rather than being composed of the more homogeneous samples typical in laboratory experiments). Experimental Conditions The experimental conditions introduced various mixes of frames that replicate and extend the conditions used in Chong and Druckman’s (2007b) experiments. Table 6.5 lists the full set of conditions, to which respondents were randomly assigned. The first listed condition is the control group that simply answered the main dependent variable question described above (N = 61). This group serves as a baseline for evaluating the effects of the frames in the other conditions. I randomly assigned another group to the strong pro economic benefits frame such that respondents answered the main dependent variable question with the added statement: “. . . Some say that the revenue from the casino would provide tax relief and help to fund education . . .” (N = 54).23 A similarly one-sided group randomly received the strong con social costs frame that included: “. . . some say that a state-run casino will have severe social costs, such as addiction and debt . . .” A final one-sided condition randomly exposed respondents to the weak con corruption frame: “. . . Some say that a state-run casino would increase

110 Impact Table 6.4 Demographic and Political Profile of Sample Variable

Scale (overall distribution %)

Average (std. dev.)

Education (N = 555)

1 = less than high school (1) 2 = high school (8) 3 = some college (21) 4 = year college degree (28) 5 = advanced degree (42) 1 = < $50,000 (32) 2 = $50,000–$100,000 (29) 3 = > $100,000 (39) 1 = 18–24 (18) 2 = 25–34 (11) 3 = 35–44 (19) 4 = 45–54 (21) 5 = 55–64 (18) 6 = 65–74 (9) 7 = 75+ (4) Male (43) Female (57) White (77) African American (10) Asian American (4) Hispanic (2) Other (3) Prefer not to answer (4) 1 = strong Democrat (29) 2 (27) 3 (15) 4 = Independent (18) 5 (4) 6 (4) 7 = strong Republican (3) 1 = not interested (1) 2 (3) 3 (4) 4 = moderately interested (19) 5 (18) 6 (27) 7 = extremely interested (28) 1 = just about always (2) 2 = most of the time (14) 3 = some of the time (84) 0 days a week (14) 1 (10) 2 (10) 3 (11) 4 (8) 5 (13) 6 (7) 7 (27)

4.03 (1.00)

Household income (N = 504) Age (N = 548)

Sex (N = 558) Ethnicity (N = 560)

Party identification (N = 553)

Interest in politics (N = 572)

Trust in government (N = 561) Number of days a week watch the news (N = 555)

2.07 (0.85) 3.53 (1.71)

n/a n/a

2.63 (1.59)

5.44 (1.41)

2.83 (0.41) 3.93 (2.54)

Number of days a week 0 days a week (9) 1 (5) 2 (8) 3 (9) 4 (12) read paper during campaign (last two months) 5 (12) 6 (5) 7 (40) (N = 544)

4.66 (2.41)

Gubernatorial vote choice (N = 403)

n/a

Blagojevich (Democrat) (61) Topinka (Republican) (19) Whitney (Green) (18) Other (2)

Competing Frames in a Political Campaign 111 Table 6.5 Experimental Conditions Condition

Frames

Predicted effect on overall opinion Baseline

Control

None

Pro-strong

Economic benefits

Increase support for the proposal

Con-strong

Social costs

Decrease support for the proposal

Con-weak

Corruption

No effect

Con-strong– pro-strong

Social costs– economic benefits

No effect (i.e., cancel out)

Con-strong– pro-weak

Social costs– entertainment

Decrease support for proposal

Con-weak– pro-strong

Corruption– economic benefits

Increase support for the proposal

Con-weak– pro-weak

Corruption– entertainment

No effect

Con-weak– con-weak

Corruption–morality

No effect

Con-weak– pro-strong– con-weak

Corruption–economic benefits–morality

Increase support for proposal

patronage and corruption in state government . . .” (N = 58). The theory suggests that the strong economic benefits and social costs frames will significantly push respondents in pro and con directions, respectively, while the weak corruption frame will have no effect. The six other conditions include various combinations of frames exposing respondents to two or more competing frames.24 I explore the predicted canceling effect of two opposing strong frames by pairing the social costs and economic benefits frames (N = 55) (Druckman 2004; Sniderman and Theriault 2004). I also include a condition with the strong con social costs frame along with the weak pro entertainment frame, with the expectation that the strong social costs frame will exhibit a greater effect (N = 58) (Chong and Druckman 2007b). The other conditions include the weak con corruption frame. One of these pairs the corruption frame with the strong pro economic benefits frame (N = 55). This is again a strong-versus-weak condition, and, thus, only the economic benefits frame should significantly affect respondents. I also included a condition that paired the weak con corruption frame and the weak pro entertainment frame (N = 57), and a condition that grouped the weak con corruption frame with a weak con morality frame (N = 60). In both these cases, I expect no effect since weak frames are predicted to not influence opinions. The final condition adds the strong pro economic benefits frame to the weak corruption and morality frames with the expectation of increased support (N = 60). The final column of the table lists the predicted effect on overall opinion (relative to the control group).25

112 Impact

Results I focus the analysis on how the different combinations of frames affect overall attitudes toward the casino proposal. Specifically, I estimate the relative impact of the frames by using a statistical technique called an ordered probit model; in so doing, I treat the control group as the point of comparison (Chong and Druckman 2007a), and include separate dummy variables for each of the nine treatment conditions.26 I list the condition means in appendix A,27 and report the specific regression results in Appendix B. Using the results reported in Appendix B, I generated the probabilities that an average respondent from each condition would support the casino proposal. For example, holding all the demographic and political (control) variables at their means, I find that the average control group respondent has a 24 percent chance of supporting the casino proposal (I construe support as reporting a score of 5, 6, or 7 on the 7-point scale). In Figure 6.1, I graph the probabilities, relative to the control group, that an average individual from each condition would support the casino proposal. The graph presents the difference between the probability of support in each group and the control group’s aforementioned 24 percent likelihood of support.28 The graph thus reveals the specific impact of the frame(s), relative to the control group, in moving respondents to be more or less supportive of the casino proposal. In the graph, each condition (abbreviated) variable name describes the specific frames to which respondents were exposed. (The abbreviations Cn = con, Pr = pro, Str = strong, and Wk = weak; the other abbreviation should be self-evident.) The figure shows that, as predicted, only strong frames influenced opinions. Specifically, the two single strong frame conditions show substantial effects— the strong pro economic benefits frame moved respondents to be 24 percent more supportive the casino (compared to the control group—thus, the average individual in this group was 24 percent + 24 percent = 48 percent likely to be supportive), while the strong con social costs frame moved respondents to be 18 percent less supportive (or in total, 24 percent – 18 percent = 6 percent likely to be supportive). These single strong frame results replicate conventional framing experiments that expose respondents to one strong frame or another (e.g., free speech or public safety). The other single frame condition used the weak con corruption frame and, consistent with expectations, this had virtually no effect relative to the control group. Similarly, simultaneous exposure to the two strong frames—social costs and economic benefits—did not significantly move respondents (the 5 percent decline is not statistically significant), suggesting that the frames counteract one another (consistent with Sniderman and Theriault’s (2004) prediction). By contrast, when the social costs frame is matched with the weak pro entertainment frame, respondents react, as predicted, by becoming less supportive, with the probability of support dropping by 15 percent. The strong frame wins out. This is an important caveat to Sniderman and Theriault’s (2004) theory that competing frames cancel out: frames cancel out only when both frames are strong.

–30

–20

–10

0

10

20

30

40

50

Econ (Pr-str)

24**

Soc costs (Cn-str)

–18*

1

Soc costs– econ (Cn-str– pr-str)

–5

Soc costs– entert (Cn-str– pr-wk)

–15**

Corr-econ (Cn-wk– pr-str)

22**

Corr–entert (Cn-wk– pr-wk)

–5

Frame condition (**p < 0.01; *p < 0.05 for one-tailend tests)

Corr (Cn-wk)

Figure 6.1 Likelihood of Casino Support

Relative impact of frames (compared to control) (%)

Corr–moral (Cn-wk– cn-wk)

3

Corr–econmoral (Cn-wk– pr-str– cn-wk)

41**

114 Impact The other three dual conditions match a weak con corruption frame with other frames. As expected, respondents who received the strong pro economic benefits frame along with the corruption frame become significantly more supportive (by 22 percent). The insignificance of both the weak con corruption–weak pro entertainment condition, and the weak con corruption–weak con morality condition further highlights the lack of impact from weak frames (neither condition significantly moves opinions), even when there are multiple weak frames in the same direction (e.g., in the case of the con corruption–morality combination). The final condition shows that the one strong pro economic benefits frame overwhelmed the weak con corruption and weak con morality frames. This again accentuates that strength is more important than repetition. Moreover, notice the probability of support in this condition exceeded the control group by 41 percent which is substantially greater than the impact of the strong pro economic benefits frame alone (24 percent). This is evidence of what Chong and Druckman (2007a, 2007b) call a contrast effect, where the presence of a weak contrary frame (or in this case two weak frames) can backfire, pushing respondents even further in the opposite direction. That is, the con side would have been better off offering no frame (as in the economic frame only condition where there was no con frame and opinion increased by 24 percent) instead of offering the weak con frames (as in this last condition where opinion increased by 41 percent).

Conclusion Competing sides regularly seek to offer alternative frames for making sense of issues and events. Incorporation of competition into studies of framing is critical if we are to understand how frames work. The results concerning frame strength and contrast effects support Chong and Druckman’s (2007a, 2007b) theory. Strength is more important than repetition, and the impact of a frame in isolation may differ from its effect when mixed with other frames (as evident from the contrast effect). The results also provide one of the few examples of framing effects among a heterogeneous group of voters in a campaign setting. The obvious unanswered question concerns what makes for a strong frame (Chong and Druckman 2007c). Unfortunately, extant work on persuasion provides little guidance on the conditions of strength—for example, dual process models of attitude change (e.g., Petty and Wegener 1998) distinguish the origins of strength (e.g., cues or argument quality) but say little about what factors matter when, and what makes for a high-quality argument (or frame). There are, however, some recent studies in political science that isolate strengthpromoting elements—for example, strong frames tend to comport with cognitive biases (Arceneaux 2007), highlight specific emotions (Aarøe 2008; Petersen 2007), include multiple, frequently appearing, arguments (Baumgartner et al. 2008), and/or have been used in the past (Edy 2006). This research offers a compelling baseline for future research that needs to identify not only the

Competing Frames in a Political Campaign 115 elements of strength but also the conditions under which a frame might be strong or weak. This will enhance our understanding of public opinion formation, enable us to make sense of why some campaigns succeed while others fail, and reveal what makes for the most effective campaign strategy (e.g., Schaffner 2008).

Appendix A

Table 6.A1 Experimental Condition Means Condition

Frames

Mean (std. dev.; N)

Control Pro-strong Con-strong Con-weak Con-strong–pro-strong Con-strong–pro-weak Con-weak–pro-strong Con-weak–pro-weak Con-weak–con-weak Con-weak–pro-strong– con-weak

None Economic benefits Social costs Corruption Social costs–economic benefits Social costs–entertainment Corruption–economic benefits Corruption–entertainment Corruption–morality Corruption–economic benefits–morality

3.10 (1.93; 59) 4.20 (1.68; 54) 2.04 (1.28; 57) 3.07 (1.97; 58) 2.91 (2.01; 55) 2.05 (1.49; 58) 4.20 (2.04; 55) 2.37 (1.77; 57) 3.09 (2.05; 59) 4.42 (2.13; 58)

Appendix B Table 6.B1 reports the regression result; it includes control variables that might plausibly affect casino attitudes.29 The experimental condition variables appear on the left side and the control variables appear on the right. The text contains a discussion of the impact of the experimental conditions. Also of note are some of the effects of the demographic and political variables. Those who have a greater distrust of government are less supportive of the proposal, presumably reflecting their general lack of trust in expanding governmental control. Also, older people and women exhibit significantly less support while minorities show more support. Politically, those who voted for Topinka are more supportive, which, even though it was her initial proposal, is mildly surprising given the lack of coverage of the casino (as discussed). The other variables including perception of corruption in the current administration, values regarding state regulation of business, frequency of going to a casino, and partisanship had no significant effects. Overall, it is interesting that demographics play a larger explanatory role than political values and partisanship, perhaps reflecting political ambivalence about casinos, which indeed makes people susceptible to framing effects in the first place. (That is, people are susceptible to framing in the first place because they are not sure how to weight alternative considerations; see Druckman 2001b).

116 Impact Table 6.B1 Experimental Framing Effects on Support for Casino Proposal (1 to 7 Scale) Experimental condition Econ (Pr-str) Soc Costs (Cn-str) Corr (Cn-wk) Soc costs–econ (Cn-str–pr-str) Soc costs–entert (Cn-str–pr-wk) Corr–econ (Cn-wk–pr-str) Corr–entert (Cn-wk–pr-wk) Corr–moral (Cn-wk–cn-wk) Corr–econ–moral (Cn-wk–pr-str–cn-wk) t1 through t6 Log likelihood Number of observations

Control variable 0.78** Administration corruption (0.21) –0.37* Regulation value (0.22) 0.13 Casino visits (0.21) –0.05 Distrust government (0.21) –0.55** Partisanship (Republican) (0.23) 0.88** Age (0.21) –0.30 Minority (0.22) 0.08 Female (0.21) 1.00** Vote for Topinka (0.22) See below –815.86 478

–0.41 (0.26) 0.04 (0.23) 0.22 (0.16) –0.50* (0.26) –0.38 (0.24) –0.39* (0.19) 0.44** (0.15) –0.22* (0.11) 0.26* (0.16)

Note Entries are ordered probit coefficients with standard errors in parentheses. **p ≤ 0.01; *p ≤ 0.05 for one-tailed tests. The coefficient and standard errors for t1 through t6 are as follows: –1.41 (0.39), –0.82 (0.39), –0.47 (0.39), –0.10 (0.39), 0.34 (0.39), 0.93 (0.40).

Notes 1. This chapter grew out of an ongoing collaborative project with Dennis Chong, who offered sage advice at all stages of the research presented here. I thank Jason Reifler for his generous contributions, and Toby Bolsen, Jaclyn Cheron, Caitlin Chester, Thomas Leeper, Lauren Matecki, Tommy Szalasny, Cara Walsh, and Jonathan Weber for research assistance. Support was provided by Northwestern University and the AT&T Research Scholar Fund. 2. This introduction was co-written with Dennis Chong. 3. Exceptions include Sniderman and Theriault (2004), Brewer and Gross (2005), Hansen (2007), and Chong and Druckman (2007a, 2007b, 2007c). 4. They also report a count of the total number of frames; however, this is of limited interest since some frames appear only a few times (and a strict count does not adjust for frequency of use). 5. The casino plan was not an issue on which voters would directly vote (e.g., an initiative) but initially appeared to be a critical campaign issue (e.g., Pearson 2006). 6. Interestingly, after being re-elected, Blagojevich expressed support for a casino plan (Meitrodt and Garcia 2007). 7. This covers August 24 through November 6. 8. Details on the content analysis are available from the author. 9. The main exception would be attentive voters who formed opinions early in the campaign, during the brief period when the casino proposal received substantial coverage. Indeed, these voters did behave differently, although I do not focus on

Competing Frames in a Political Campaign 117

10.

11.

12.

13.

14. 15.

16.

17.

18.

19.

these differences here. Detailed analyses of these voters as a distinct sub-group are available in Druckman (2009). I used non-Illinois voters to ensure that the respondents had not been influenced by the ongoing campaign. It turned out that such exposure probably would not have mattered, given the lack of coverage of the casino proposal. However, when planning for the pre-test, I was not certain how the campaign would develop. The pre-test respondents closely resembled the Exit Poll respondents in terms of demographic backgrounds. The “t-statistic” generates a probability indicating the likelihood of the two scores (e.g., economic benefits frame average and social costs frame average) being compared are the same—that is, that the differences between them is due to random chance and not “real” systematic differences. The p < 0.20 means that there is around a 0.20 chance that they are “really” the same (and the difference is just due to arbitrary chance). This is typically seen as too high to conclude “real differences.” Typically, real differences are accepted as evident when the probability at least approaches 0.10. I did not consider using the community building frame since it registered 0 percent availability and never appeared in the coverage, suggesting it may not be relevant to the issue (despite occasional references to it by advocacy groups). I also did not consider using the politics frame since it is directionally ambiguous. One potential problem, however, is that, while entertainment is significantly more supportive than all others except economic benefits, I also find that economic benefits is significantly more supportive than entertainment (t60 = 2.70, p < 0.01 for a twotailed test). This is a potential confound insofar as I will not necessarily know based on this pre-test if variance in the impact of economic benefits and entertainment stems from strength or directionality. However, it is perceived as significantly more applicable than morality (t60 = 2.11, p < 0.04 for a two-tailed test). The morality frame is significantly more opposed than the social costs frame (t60 = 2.50, p < 0.05 for a two-tailed test). This is in another potential confound in that differential effects of morality and social costs could stem from strength or directionality. This is similar to the confound between economics and entertainment. The two potential confounds are between morality and social costs differing in terms of both strength and degree of opposition (although both are significantly more opposed than the support frames), and between entertainment and economics significantly differing in terms of both strength and support. In both cases however, I have confidence that differences between these frames will reflect strength variations rather than directional variations since in both cases the differences in terms of strength are much greater than the differences in direction. Also, overall, the set of supportive frames are significantly different from the opposed ones, and the set of strong frames are significantly stronger than the weak ones. Even if voters did not recognize the connection to the election itself, they were still voting on state office which would ultimately decide whether or not to pursue the state-owned casino proposal. The survey included a series of belief importance measures that read “Listed below are several considerations that people might weigh when thinking about the proposal for a state run gambling casino. Please rate how unimportant or important each consideration is to you where: 1 = very unimportant . . . 5 = very important.” I listed considerations that correspond with the main frames: economic benefits, corruption, morality, social costs, and entertainment. In this chapter, I do not analyze the belief measures; however, the results garnered by using various belief importance measures as the dependent variable (instead of overall opinion) are consistent with the results reported below. The survey also included various other items pertaining to the gubernatorial campaign that are not relevant to the casino proposal.

118 Impact 20. A total of 575 respondents were assigned to the main experimental conditions, as described below; not surprisingly, given the context of the survey, a number of respondents did not answer all questions. 21. Data on vote totals come from Cook County Election Department website at http://www.voterinfonet.com/. 22. In 2004, 82 percent of these voters opted for Kerry. I also note that the apparent high degrees of cynicism presumably reflects the times and location (i.e., a Republican president with low approval, among largely Democratic respondents). 23. I faced severe space limitations since participation was solicited on the spot, meaning the survey had to be kept short (see Traugott and Lavrakas 2000). To ensure a reasonable number of respondents, then, I could not ask them to read lengthy, framed articles about the casino issue and then answer various questions. I followed typical practice on survey experiments, which is to embed alternative frames in the question itself. 24. I did not implement an exhaustive set of conditions because of the nature of using an exit poll. Specifically, all data had to be collected in one day and I could only estimate the number of respondents (given the number of pollsters I had hired). Also, I could not use the experiment to test repetition effects directly, since it would have meant repeating the same consideration in very close succession which would be awkward (i.e., I could not embed them in alternative longer articles, given time limits). 25. There were two other conditions in the exit poll that I do not discuss here. These conditions asked exit poll respondents to assess the availability and applicability of the frames (similar to the pre-test). Based on these results, there is some evidence that the corruption frame was stronger than the pre-test indicated, and that the constant corruption theme of the campaign in general may have affected voters’ interpretation of the casino. This evidence, however, is speculative and, for the analyses reported below, I continue to infer from the pre-test data that the corruption frame is weak. Further details are available from the author. 26. This differs from the Chong and Druckman’s (2007b) analysis because, unlike their data, these data include a non-exhaustive set of framing conditions. 27. Across conditions, the mean casino support score is 3.14 (2.01). 28. I compute these probabilities using “Clarify” (Tomz et al. 1999). I do not report standard deviations because “Clarify” provides probabilities for each dependent variable value (1 through 7), and I sum the probabilities for 5, 6, and 7 (which signify support). The results are consistent using different breakdowns. 29. This was not necessary given successful random assignment (which was confirmed); however, it is worthwhile to have a better understanding of casino attitudes. Also note that when I include other control variables, as listed in Table 6.4, none is significant and none of the results presented below changes. In the presented analysis, I include variables that either had theoretical relevance or were found to be significant in preliminary analyses.

Works Cited Aarøe, Lene. 2008. “Investigating Frame Strength: The Case of Episodic and Thematic Frames.” Unpublished paper, Aarhus University. Arceneaux, Kevin. 2007. “Cognitive Biases and the Strength of Political Arguments.” Unpublished paper, Temple University. Baumgartner, Frank R., Suzanna L. De Boef, and Amber E. Boydstun. 2008. The Decline of the Death Penalty and the Discovery of Innocence. New York: Cambridge University Press.

Competing Frames in a Political Campaign 119 Brewer, Paul R. 2003. “Values, Political Knowledge, and Public Opinion About Gay Rights: A Framing-based Account.” Public Opinion Quarterly 67: 173–201. Brewer, Paul R. and Kimberly Gross. 2005. “Values, Framing, and Citizens’ Thoughts about Policy Issues.” Political Psychology 26: 929–948. Chong, Dennis and James N. Druckman. 2007a. “A Theory of Framing and Opinion Formation in Competitive Elite Environments.” Journal of Communication 57: 99–118. Chong, Dennis and James N. Druckman. 2007b. “Framing Public Opinion in Competitive Democracies.” American Political Science Review 101: 637–655. Chong, Dennis and James N. Druckman. 2007c. “Framing Theory.” Annual Review of Political Science 10: 103–126. Chong, Dennis and James N. Druckman. n.d. “Identifying Frames in Political News.” In Erik P. Bucy and R. Lance Holbert, editors, Sourcebook for Political Communication Research: Methods, Measures, and Analytical Techniques, Routledge. Forthcoming. Donahue, Amy K. and Joanne M. Miller. 2006. “Experience, Attitudes, and Willingness to Pay for Police and Fire Protection.” American Review of Public Administration 36: 395–418. Druckman, James N. 2001a. “On the Limits of Framing Effects: Who Can Frame?” Journal of Politics 63(4): 1041–1066. Druckman, James N. 2001b. “The Implications of Framing Effects for Citizen Competence.” Political Behavior 23(3): 225–256. Druckman, James N. 2004. “Political Preference Formation.” American Political Science Review 98: 671–686. Druckman, James N. 2009. “Pre-Treatment Effects in Framing Experiments.” Unpublished paper, Northwestern University. “Editorial—The Next Governor’s Challenge.” 2006. Chicago Tribune (August 24), Section 1. Edy, Jill A. 2006. Troubled Pasts: News and the Collective Memory of Social Unrest. Philadelphia, PA: Temple University Press. Entman, Robert M. 1993. “Framing: Toward Clarification of a Fractured Paradigm.” Journal of Communication 43(4): 51–58. Hansen, Kasper M. 2007. “The Sophisticated Public: The Effect of Competing Frames on Public Opinion.” Scandinavian Political Studies 30: 377–396. Kinder, Donald R. 2007. “Curmudgeonly Advice.” Journal of Communication 57: 155–162. Meitrodt, Jeffrey and Monique Garcia. 2007. “Governor Might Take Gamble.” Chicago Tribune (May 26), Section 1: 16. Nelson, Thomas E., Rosalee A. Clawson, and Zoe M. Oxley. 1997. “Media Framing of a Civil Liberties Conflict and Its Effect on Tolerance.” American Political Science Review 91: 567–583. Pearson, Rick. 2006. “Topinka Gambling on Casino.” Chicago Tribune (August 24), Section 1: 1. Petersen, Michael Bang. 2007. “Causes of Politcial Affect: Investigating the Interaction Between Political Cognititions and Evolved Emotions.” Unpublished paper, Aarhus University. Petty, Richard E. and Duane T. Wegener. 1998. “Attitude Change.” In Daniel T. Gilbert, Susan T. Fiske, and Gardner Lindzey, editors, Handbook of Social Psychology, 4th edition, Vol. II. New York: McGraw-Hill.

120 Impact Rasmussen Report. 2006. “Illinois Governor: Blagojevich Still Ahead.” August 10, http://legacy.rasmussenreports.com/2006/August/IllinoisGovernor.htm. Reutter, Mark. 2006. “U. of I. Economic Index Indicates Economy Growing, But At Slower Pace.” News Bureau: University of Illinois at Urbana-Champaign. http://news.illinois.edu/news/06/0501flash.html. Riker, William H. 1995. “The Political Psychology of Rational Choice Theory.” Political Psychology 16: 23–44. Schaffner, Brian F. 2008. “Understanding the Origins of Campaign Strategies.” Unpublished paper, University of Massachusetts, Amherst. Sniderman, Paul M. and Sean M. Theriault. 2004. “The Structure of Political Argument and the Logic of Issue Framing.” In Willem E. Saris and Paul M. Sniderman, editors, Studies in Public Opinion. Princeton, NJ: Princeton University Press. Tomz, Michael, Jason Wittenberg, and Gary King. 1999. “Clarify.” Version 1.2.1, Cambridge, MA: Harvard University, June 1. http://gking.harvard.edu/. Traugott, Michael W. and Paul J. Lavrakas. 2000. The Voter’s Guide to Election Polls, 2nd edition. New York: Chatham House Publishers. Wittman, Donald. 1995. The Myth of Democratic Failure. Chicago, IL: The University of Chicago Press.

7

Taxing Death or Estates? When Frames Influence Citizens’ Issue Beliefs Brian F. Schaffner and Mary Layton Atkinson

Introduction Many of the chapters in this volume focus on how the way a policy proposal is framed can influence public support or opposition for that proposal. For example, in Chapter 2, Nelson et al. show that framing science as a “marketplace of ideas” leads to more support for the teaching of intelligent design than when science is framed as the “progress of ideas.” In Chapter 6, Druckman shows that “strong” frames influenced support for a publicly funded casino. Framing effects are also prominent when it comes to citizens’ evaluations of tax policies. For example, citizens express significantly less support for tax cuts when the question is framed in a way that emphasizes the services that would have to be sacrificed to pay for those cuts (Welch 1985). While studies of framing effects have demonstrated that frames can influence the public’s support or opposition for policies, less attention has been paid to whether frames can also influence the public’s beliefs about the content of policy proposals. In this chapter, we document just such an effect in the debate over whether to repeal the inheritance tax. While this tax applies to only about 2 percent of Americans, Republicans framed it as a “death tax” to suggest that the tax affected a much larger share of the public. Using a survey experiment conducted in 2003, we demonstrate the effectiveness of this strategy by demonstrating that citizens exposed to the “death tax” were more likely to think that most Americans were subject to the tax. We begin this chapter by introducing the inheritance tax debate and outlining how Republicans came to employ the “death tax” frame to change the way the public viewed the tax and ultimately attract support for a repeal. We then introduce the survey experiment we use. We find that employing the “death tax” frame led some respondents to think that the tax applied more widely than with those who were not exposed to that frame. Furthermore, these beliefs had important effects on whether citizens supported or opposed a permanent repeal of the inheritance tax. The chapter concludes with our discussion of the consequences of these findings for broadening our understanding of framing effects.

122 Impact

The Emergence of the “Death Tax” Frame On issues related to spending and taxation, Democrats have traditionally emphasized which groups gain and lose from particular tax policies while Republicans tend to make more sweeping statements about the fairness of the tax code and the need to reduce spending and taxes (Jacoby 2000; Petrocik 1997; Popkin 1997; Sears and Citrin 1985; Feldman and Zaller 1992). Take for example an episode during the most recent presidential campaign when Barack Obama and John McCain tried to frame Obama’s tax plan in different ways. In defending his tax plan, Obama stressed the importance of giving the middle class a tax cut, even if it meant that taxes on the wealthiest Americans would increase. The strategy behind this framing of the issue was to emphasize that only wealthy Americans would have to pay higher taxes and that the middle class would benefit by having their taxes cut. McCain, on the other hand, attempted to frame Obama’s tax plan in an unfavorable light by arguing that increasing taxes on the wealthy would be unfair. McCain referred to “Joe the plumber” as a symbol of this unfairness to demonstrate how people who worked hard to attain more success (and larger incomes) would ultimately be penalized by having to pay a higher tax rate.1 The case of the “estate tax” frame versus the “death tax” frame is another example of two frames that comport with the overarching message themes typically advanced by the Democratic and Republican parties. Democrats historically used the “estate tax” label to draw attention to the groups that would benefit from its elimination—wealthy families privileged enough to own estates. The “estate tax” frame conveys that most Americans will not have to pay this tax. The “death tax” label, on the other hand, is a broad frame employed by the Republican Party to induce the public to think more broadly about the tax (Graetz and Shapiro 2005). As with the example from the 2008 campaign, Republicans hoped to use the “death tax” label to highlight their view that the tax was unfair. Since fairness is an important value for most Americans, an appeal on these terms might be persuasive to an individual even if that particular individual would not be subject to the tax. Republicans also hoped to use the “death tax” frame to imply that the tax could apply to anyone. While most Americans view an estate as something owned only by the very wealthy, death is a reality regardless of one’s income bracket. Thus, ultimately the “death tax” frame provided Republicans with a tool for broadening the relevance of the issue—Americans might suddenly feel as though the tax could apply to them and they would also be able to connect the principle of fairness to their own views on taxation. The development of the “death tax” frame and the movement for the tax’s repeal has a long history that has been well documented by Ian Shapiro and his collaborators (Graetz and Shapiro 2005; Birney and Shapiro 2005). While a diverse coalition of Washington outsiders, from “farmers to florists, from cattle ranchers to newspaper owners,” advocated for an inheritance tax repeal during

Taxing Death or Estates? 123 the 1970s and 1980s, the issue was regarded as a non-starter and received little support from national political elites until the early 1990s (Graetz and Shapiro 2005, p. 4). Only in 1994, when the Republicans took over the House did such a repeal begin to seem more realistic. It was also during the early 1990s that advocates recognized the need to make the issue of the inheritance tax repeal more palatable to the general public. The favored tactic for accomplishing this was bringing the issue under the traditional GOP umbrella strategy of highlighting fairness in taxation. In Chapter 4, Jessica Gerrity documents how interest groups played a central role in creating and advancing the “partial-birth abortion” frame. In a similar way, interest groups appeared to be the catalysts for the “death tax” frame. In the early 1990s, Jim Martin, president of the 60 Plus Association, a non-profit organization committed to shoring up Social Security and ending the inheritance tax, was the first to use the death tax label (Graetz and Shapiro 2005). Martin’s phrase began to catch on among conservatives around the same time that Pat Soldano, a financial planner turned lobbyist, hatched the idea of casting the inheritance tax repeal as an issue of central importance to small and medium-sized businesses (Graetz and Shapiro 2005). Soldano argued that while the “ultra-wealthy can take care of themselves,” it is small business owners who have the most reason to be concerned with the expense of the inheritance tax (Graetz and Shapiro 2005, p. 21). Along with Frank Blethen, owner and publisher of the Seattle Times, and other advocates, Soldano targeted her efforts at groups not traditionally associated with the Republican Party, such as women, minorities, gays, and lesbians. She did this by promoting stories of clients who battled illness while finalizing their estates in an effort to humanize and popularize the issue. By 1995, House Speaker Newt Gingrich, Ways and Means Committee Chairman Bill Archer, and other leading Republicans were encouraging their colleagues to use the “death tax” label (Graetz and Shapiro 2005). Figure 7.1 indicates just how quickly the “death tax” label became part of the congressional debate. The phrase “death tax” showed up just 11 times between 1989 and 1996 while the “estate tax” label was used 454 times during this same period. However, during the 105th Congress, the “death tax” label was used 163 times and by the 106th Congress, “death tax” showed up almost as frequently in the congressional record as “estate tax.” As Harris notes in Chapter 3, congressional party leaders use coordinated strategies to frame policy issues in advantageous ways. By the 105th Congress, both sides were promoting their own frames of the inheritance tax debate. Republican lawmakers did not adopt the “death tax” frame simply because interest groups were promoting it. They were further encouraged about the effectiveness of the frame from focus groups and polling. Republican strategist Frank Luntz advocated for the use of the “death tax” label because it conveyed a sense that the tax was unfair. Luntz’s focus groups confirmed the effectiveness of the “death tax” label. For example, a reporter from the New Yorker describes a Luntz focus group conducted during the 2000 presidential campaign:

124 Impact 400 Estate tax 350 Death tax

Number of mentions

300 250 200 150 100 50 0

Figure 7.1 Use of “Estate Tax” and “Death Tax” Labels in Congressional Record, 1989–2004 Source: Compiled by the authors using thomas.loc.gov

The sheet of paper was an exercise on inheritance taxes. He asked people what they would most want to eliminate: an estate tax, an inheritance tax, or a death tax. Death tax won big. They vented for a while about how deeply unfair it was: you work hard your whole life and the government takes it all away at the end. Then Luntz asked them how much they thought you were allowed to pass on after your death without incurring a tax. All the non-accountants guessed way too low. He told them that the actual figure was six hundred and seventy-five thousand dollars. “Now that you know that,” said Luntz, “would anyone not want to abolish the tax?” Nobody raised a hand. (Lemann 2000, p. 108) This exercise illustrated how the “death tax” label could effectively frame the inheritance tax in a way that led citizens to believe that it applied more widely than it actually did. By the late 1990s, Republicans employed a coordinated strategy to promote the repeal of the inheritance tax, which featured the use of the “death tax” label and emphasized the unfair effect of the tax on real families. This effort followed

Taxing Death or Estates? 125 the course charted by Soldano and other advocates, focusing on popularizing the issue, transforming it from an issue of concern only to wealthy, white men, into an issue of fairness behind which Americans of all stripes could rally. For instance, one unnamed staffer to a Republican Congress member described to Graetz and Shapiro why the death tax label became such an integral part of the campaign to repeal the inheritance tax, stating, “‘Estate tax’ sounds like it only hits the wealthy, but ‘death tax’ sounds like it hits everyone. They focus grouped this a lot, and people viewed a death tax as very unfair. You don’t have to be really rich to be worried about a death tax” (2005, p. 77). In 2000, just 2.2 percent of all American adults who died left estates that were subject to the inheritance tax; this figure dropped to just 1.2 percent in 2002 when the threshold for administering the tax increased to $1,000,000.2 Thus, the tax was not applied nearly as widely as Republicans hoped to insinuate with their use of the “death tax” frame. Claims that the inheritance tax might force Americans who owned farms and small businesses to lose those establishments were also highly exaggerated. Using 2002 data, a Congressional Research Service report documented that “most farmers and business owners are unlikely to encounter estate tax liability” and “only a tiny fraction . . . would be at risk of being forced to liquidate the family business to pay estate or gift taxes” (Gravelle and Maguire 2005, pp. 13–14). In fact, a New York Times reporter was unable to locate a single family that had been forced to sell a farm to pay the inheritance tax (Birney and Shapiro 2005). Yet, these facts did not appear to undermine the usefulness of the “death tax” frame. As Luntz himself expressed, “‘a compelling story, even if factually inaccurate, can be more emotionally compelling than a dry recitation of the truth’ ” (Graetz and Shapiro 2005, p. 81). By the last year of President Bill Clinton’s term, a Republican-controlled Congress passed an inheritance tax repeal which was promptly vetoed by Clinton. However, when similar legislation passed a year later with a Republican president in the White House, the inheritance tax repeal was successfully signed into law. But this was only a temporary victory for opponents of the inheritance tax since a sunset provision was included in the law. This meant that even though the tax will be fully phased out by 2010, unless Congress acts, it will be fully reinstated in 2011. Thus, after 2001, the debate was focused on whether Congress should pass new legislation to make the inheritance tax repeal permanent.

Framing by Credible Sources Despite all of their efforts, it was unclear how effective the “death tax” frame was in affecting public opinion towards the inheritance tax. Indeed, Birney and Shapiro argue that the impact of this framing may have been more limited than is widely believed (2005, p. 7). They examine several examples of surveys (including the survey analyzed here) that used both the “estate tax” and “death tax” frames and found little difference in attitudes depending on the frame given to the respondent.

126 Impact Framing effects are often limited by the fact that citizens are discerning when it comes to the cues they take from elites about policy issues. Even if elites attempt to craft messages to their advantage, citizens may still be able to cope with this strategic behavior by lending more credibility to some sources relative to others. Scholars have taken a variety of approaches to understanding how citizens decide which sources are credible. Lupia and McCubbins (1998) argue that a source is credible when citizens think the source possesses knowledge and can be trusted to reveal information. For example, Druckman (2001) examines credibility as a function of the perceived expertise held by a source and applies the concept to framing effects. He finds that frames promoted by the New York Times were more effective than those appearing in the National Enquirer. While citizens will defer to sources that they believe have more knowledge on an issue to take advantage of this expertise, they also tend to favor sources that they believe share their general beliefs (Lupia and McCubbins 1998; Downs 1957; Popkin 1991). Because of the central role they play in organizing American politics, parties provide a particularly useful way for citizens to determine which sources to take cues from (Carmines and Kuklinski 1990; Mondak 1993). In many cases, an individual may believe that one of the two major parties is better at handling a particular issue (Petrocik 1996). If citizens trust a particular party on an issue, they are more likely to accept messages from that party’s elites. In the case of the inheritance tax debate, citizens who lend more credibility to Democrats would likely be receptive to the “estate tax” frame that Democratic elites promote and they would likely reject Republican efforts to frame it as the “death tax.” On the other hand, the “death tax” frame is more likely to be accepted by those who find the Republicans, who were advancing that frame, more credible. Because citizens are discriminating when it comes to which side’s frames they will be receptive to, elites may be less able to manipulate political debate. Druckman notes that because citizens will delegate to credible sources, “under certain conditions, framing may be akin to a competent learning process rather than manipulation” (2001, p. 244). However, just because citizens believe that they are relying on credible sources does not always mean that they are receiving credible cues. As Kuklinski and Hurley (1994) note, if citizens always attribute credibility to messages that come from supposedly credible sources, they are likely to be misled at least some of the time. Page and Shapiro (1992) outline numerous instances when public officials have misled the public during policy debates. Indeed, the inheritance tax debate provided elites with a particularly fruitful issue on which to use frames to influence what the public believed about the issue. After all, Americans tend to be relatively uninformed about the minutiae of tax issues such as this one and, when faced with little information about political issues, citizens tend to rely more on cues and signals sent by elites. Thus, when it came to framing the inheritance tax debate, political elites had an opportunity to influence not just which side of the issue people were on but also their beliefs about the policy itself.

Taxing Death or Estates? 127

The Survey Experiment To determine what effect the “death tax” frame had on what citizens believed about the tax, we examine a survey conducted in 2003 by International Communications Research for the Kaiser Family Foundation, National Public Radio, and the Kennedy School of Government. The survey firm interviewed 1,339 adults by phone about matters of tax policy. Most important for our study is a survey experiment in which, randomly, half of the respondents were asked whether they favor or oppose eliminating the “federal estate tax” while the other half were asked about the “federal estate tax that some people call the death tax”3 (NPR012 2003). Specifically, the questions read as follows: Version A: “There is a federal estate tax—that is, a tax on the money people leave when they die. Do you favor or oppose eliminating this tax, or don’t you know enough to say?” Version B: “There is a federal estate tax that some people call the death tax. This is a tax on the money people leave when they die. Do you favor or oppose eliminating the tax, or don’t you know enough to say?” Two points about this survey experiment deserve mention. First, both versions of the question include the “estate tax” frame. However, half of the respondents also received the “death tax” frame, allowing us to capture the effect that this frame has on survey respondents while holding the inclusion of the estate tax frame constant in both questions. Often in framing studies, participants are exposed to one frame or another, but not both (for example, Nelson and Kinder 1996; Kinder and Sanders 1990; Iyengar 1991). In one notable exception, Chong and Druckman (2007) use an experiment to present one group of respondents with competing frames on an issue while other groups received just one frame or the other. They find that when citizens are exposed to competing frames, the effects of these frames are more limited (also see Sniderman and Theriault 1999). Our study compares respondents exposed to one frame (the “estate tax” frame) to those who receive both competing frames. Based on prior research examining the effects of competing frames (Chong and Druckman 2007; Sniderman and Theriault 1999), we might expect to find only limited framing effects under the latter condition. Second, while the survey experiment only introduces the label “death tax” to the respondent, the extensive use of that label by Republicans during the inheritance tax debate means that the term was loaded with the arguments it was meant to convey. Indeed, in Chapter 2, Nelson et al. argue that labeling of this nature is an important tool used by elites when framing issues. Since the “death tax” label was used almost exclusively by Republicans during the debate, the effect of the frame should be limited to those citizens who believe that Republicans are a credible source of information on tax issues. Likewise, those who believe Democrats are more credible on the issue would be more likely to reject the “death tax” frame.

128 Impact To capture credibility, we use a question on the survey that asks respondents which party would best deal with the federal tax system. Presumably, citizens are more likely to accept messages on tax issues from the party that they think is better equipped to handle those issues. Respondents were offered the option of saying either the Democrats or Republicans were better on taxes or they could say that there was not much difference between the parties. Nearly 40 percent of the respondents chose the latter option, while the others were split evenly between choosing Democrats and Republicans. In the analyses below we divide respondents into three groups—those who say Democrats are better on the issue, those who think Republicans are better, and those who say there is no difference between the parties. Since the “death tax” frame tends to be promoted by Republicans, it should have the biggest effect on those who see Republicans as most credible on tax issues and should have little or no effect on those who view Democrats as more credible on these issues. Furthermore, those who believe Democrats are more credible on tax issues should demonstrate the lowest support for repealing the inheritance tax while those who see Republicans as most credible should be the most supportive of the repeal. Individuals who thought both parties were equally credible on the issue should hold a position between these two extremes.

Results We begin by examining how adding the “death tax” frame affects whether respondents believe the inheritance tax should be repealed. Among all groups, 61 percent favored repealing the tax, 18 percent opposed the repeal, and 22 percent said they did not know enough about the issue. Despite the fact that only 1.2 percent of Americans who died in 2002 had estates that were subject to the inheritance tax, an overwhelming majority of respondents wanted to repeal the tax. To compare, less than 50 percent of the respondents to the same survey favored eliminating the tax on stock dividends. There were important partisan differences on the issue, as 76.8 percent of Republicans favored the repeal of the inheritance tax while just 46.9 percent of Democrats felt the same way. Nevertheless, support for eliminating the inheritance tax was widespread among respondents to this survey, a testament to Republican success in framing the issue. Table 7.1 presents the distribution of answers to this question for those who did and did not receive the “death tax” frame.4 Respondents are divided into groups depending on whether they believe Democrats, Republicans, or both parties are stronger on tax issues. Interestingly, the influence of the “death tax” frame was small and lacked statistical significance for each of these groups. The effect was smallest among respondents who think Democrats are better on tax issues; respondents who received the “death tax” frame held nearly identical positions on the issue as those who did not receive the frame. In fact, this group became slightly more supportive of repealing the inheritance tax when they received the “death tax” frame. However, the difference between the support

Taxing Death or Estates? 129 Table 7.1 Effect of “Death Tax” Frame on Support for Inheritance Tax (%) Group

Question answer

No “death tax” frame

“Death tax” frame

Difference

Democrats credible (N = 383)

Favor repeal Oppose repeal Don’t know

47.8 26.9 25.4

48.4 31.3 20.3

+0.6 +4.4 –5.1

Republicans credible (N = 388)

Favor repeal Oppose repeal Don’t know

76.1 10.3 13.6

79.4 9.3 11.3

+3.3 –1.0 –2.3

Both parties credible (N = 482)

Favor repeal Oppose repeal Don’t know

55.2 17.8 27.0

61.8 15.8 22.4

+6.6 –2.0 –4.6

Note None of the differences reported in this table are statistically significant at p < 0.1.

in either condition was not statistically significant, indicating that we cannot be confident that the effect would exist in the population of American adults. The “death tax” frame also appeared to have only a small effect on respondents who thought Republicans or both parties were more credible on tax issues. Among those who found Republicans more credible, receiving the “death tax” label led to a 3.3 percent increase in support for repealing the tax. For the group who thought both parties were credible on tax issues, receiving the “death tax” frame led to a 6.6 percent percent increase in support for the repeal. However, as with the first group, the difference in support was not statistically significant, indicating that we cannot be confident that the “death tax” frame had an effect on support for the repeal among any of the three groups. As noted above, one reason for the negligible effects is that respondents offered the “death tax” frame were also given the competing “estate tax” frame. In addition, citizens tend to have notably widespread preferences for reducing or repealing all types of taxes, particularly when they are not faced with the government spending tradeoffs necessary to cut those taxes (Page and Shapiro 1992; Welch 1985). Thus, support for repealing the inheritance tax was already quite high, even without the “death tax” frame. While the “death tax” frame did not have a strong effect on support for repealing the inheritance tax, we are primarily interested in learning whether the frame did affect the public’s beliefs about the issue. As noted above, Republicans attempted to use the “death tax” frame to suggest that a wide range of families would be affected by the tax, not just the very wealthy. Thus, citizens who were exposed to the “death tax” frame and lend credibility to Republicans on tax issues should be more likely to think that a larger share of the public has to pay the tax. Fortunately, the survey allows us to capture this effect. Following the question about support for the inheritance tax, respondents were asked the following: Do you think that most families have to pay the federal estate tax when someone dies or only a few families have to pay it?

130 Impact The responses to this question reveal just how successful Republicans were in framing the inheritance tax issue—47 percent of respondents answered that most families have to pay the tax, while 37 percent said only a few families pay and 16 percent did not know. Thus, a large share of the population appeared convinced by Republican efforts to present the inheritance tax as something that applies to a large share of the population. Interestingly, this belief was pervasive among respondents from both parties. While 46.1 percent of Republicans believed that most families had to pay the tax, 47.1 percent of Democrats held that same belief. We are primarily interested in determining whether invoking the “death tax” frame led respondents to believe that most families had to pay the tax. Though the wording in this question was not itself manipulated, the question was asked after the one analyzed above, allowing us to compare responses among those who received the “death tax” frame in the earlier question and those who did not. Table 7.2 presents these results. As expected, citizens who thought Democrats were more credible on tax issues were not significantly affected by the addition of the “death tax” frame. Respondents in this group who received the “death tax” frame appeared to have similar beliefs as those who did not receive the frame. Since these citizens believed Democrats to be more credible on tax issues, they were less susceptible to framing effects by Republican elites. Among the second two groups, however, the “death tax” frame had a significant effect on how widely respondents thought the tax applied. When respondents who viewed Republicans as better on tax issues were told that the “estate tax” is also called the “death tax,” they were 9.4 percent more likely to think that most families pay the tax and 9.3 percent less likely to answer that few families pay it. In fact, respondents who found Republicans credible on the issue were nearly evenly divided when the “death tax” frame was not offered; however, when this same group did receive the frame, they were about 17 percent more likely to say that the tax applied to most families. Statistical tests indicate that

Table 7.2 Effect of “Death Tax” Frame on Beliefs About Who Pays Tax (%) Group

Question answer

No “death tax” frame

“Death tax” frame

Difference

Democrats credible (N = 383)

Most families pay Few families pay Don’t know

43.3 39.8 16.9

44.5 37.4 18.1

+1.2 –2.4 –1.2

Republicans credible (N = 388)

Most families pay Few families pay Don’t know

43.5 44.6 12.0

52.9 35.3 11.8

+9.4* –9.3* –0.2

Both parties credible (N = 484)

Most families pay Few families pay Don’t know

45.0 42.1 12.8

52.5 31.0 16.5

+7.5 –11.1* +3.7

Note * Indicates two-tailed difference of proportions test statistically significant with p < 0.1.

Taxing Death or Estates? 131 we can be 90 percent confident that this difference in support among survey respondents would exist in the population of American adults. A similar pattern was also evident with respondents who viewed both parties as credible on the issue. Respondents in this group who did get the “death tax” frame were 7.5 percent more likely to think that most families had to pay the tax and 11.1 percent less likely to think that few families had to pay it. Thus, while this group was almost evenly divided between saying that most families pay and few families pay when they did not receive the “death tax” label, they were over 20 percent more likely to say that few families paid the tax if they received the “death tax” frame. These differences were also statistically significant, indicating that we can be confident that this effect would exist in the population of American adults. Thus, our findings indicate that simply mentioning the “death tax” label leads to significant framing effects on what citizens believe about the tax and more limited effects on their support for repealing it. However, this effect was mediated by the credibility of the frame’s source. The “death tax” frame was not effective in influencing the beliefs of citizens who thought Democrats were a credible source of information on tax issues. However, when the other groups received the frame, they were more likely to think that most families had to pay the inheritance tax.

Discussion Our findings reveal that even the minor addition of the “death tax” frame to a question on a survey can significantly influence how respondents think about the issue and, to a lesser extent, what position they take. While these framing effects are somewhat less substantial than those often documented in experimental settings, they are notable given the minimal nature of the “treatment” and the uncontrolled environment in which the survey was administered. Because the inheritance tax issue had been a source of considerable political debate for several years prior to the administration of the survey we analyze, most citizens had already been exposed to significant framing efforts by elites. As a result, the fact that we uncover even limited framing effects under these conditions is notable. The goal of the “death tax” frame is to imply that the tax applies more widely than it actually does and our findings indicate that the frame does work in this way, particularly among citizens who are predisposed to lend credibility to Republicans on the issue. In this way, the framing effects observed here are somewhat different from typical framing effects. Rather than using the frame to direct citizens’ attention toward some considerations and away from others, the “death tax” frame changes what the public believes about the content of the inheritance tax policy. Thus, our findings indicate that some frames can be effective tools not only for influencing how citizens weigh different aspects of an issue but also for changing their beliefs about that issue. But what consequence does framing of this nature have for the role of public opinion in the policy process? By being able to point to overwhelming public

132 Impact support, Republicans successfully repealed the inheritance tax (albeit temporarily) in 2001 (Birney and Shapiro 2005). Was the policy action taken truly supported by public opinion, or was this support merely an artifact of significant misunderstanding of the issue? In an extensive examination of public opinion on the inheritance tax, Birney and Shapiro find that “surveys consistently show that the number of people in favor of repeal drops when respondents are given information on exemption levels or how many people pay” (2005, p. 5). Indeed, the survey we analyze here provides a final question that allows us to determine how citizens’ beliefs about the policy may have affected their positions on whether to repeal the tax. The survey asked each respondent who favored repealing the tax if they would still feel that way if it only applied to estates of $1 million or more. At the time the survey was administered, the tax code specified that only estates of $1 million or more were subject to the tax. Thus, the question provides a way of determining respondents’ “true” opinions about the inheritance tax; that is, whether respondents would still have favored the repeal if they knew which estates were actually subject to the tax. Of those who originally supported repealing the estate tax, 34.5 percent said that they would change their minds and oppose repealing the tax if it applied only to estates of $1 million or more. In other words, over one-third of the support for the estate tax was predicated on the belief that it applied to a larger number of Americans than it actually did. How important was the misconception that the inheritance tax applied to estates of less than $1 million? Recall that 61 percent of the survey respondents originally supported repealing the inheritance tax while just 18 percent opposed it. But when we adjust the original answers to this question to account for those who changed their minds when they found out about the $1 million threshold for paying the inheritance tax, a very different picture emerges. With more information about how narrowly the inheritance tax applies, respondents were evenly divided between those supporting (37.0 percent) and opposing (38.5 percent) the repeal.5 Thus, our findings indicate not only that Republicans were able to influence the public’s beliefs about how widely the inheritance tax applied but also that the misperceptions promoted by the “death tax” frame have important consequences for public support for repealing the tax. By extension, these framing effects had important implications for the policy debate on the issue as well. As Birney and Shapiro (2005, p. 32) note: In the case of estate tax repeal, before repeal advocates refashioned the issue’s public image, the conventional wisdom had indeed been that an avalanche of public outrage would blast politicians who supported repeal. Yet, the coalition for repeal eventually did such an effective job at convincing legislators that public opinion was on their side, and could reliably be maintained that way, that the conventional wisdom was all but reversed. Many legislators then wondered if it would be acceptable to uphold the estate tax.

Taxing Death or Estates? 133 By using the “death tax” frame to lead a significant share of the population to believe that the tax applied to most families, Republicans were able to craft public opinion in a way that allowed them to temporarily repeal the tax. While increased pressure for a permanent repeal has stalled in the face of growing budget deficits, the “death tax” frame provides an important example of how some frames can actually affect the public’s beliefs about the scope of policy proposals.

Conclusion Public opinion scholars have typically celebrated the fact that citizens use elite cues to help shape their behavior, particularly since citizens tend to be discriminating about whom they accept cues from (Sniderman et al. 1991; Popkin 1991). Our analysis did present evidence of such discrimination. Framing effects were mediated by the source of the frame, a fact that would presumably lead citizens to make more reasonable decisions on the issue. However, Kuklinski and Hurley warn that “one who processes heuristically might assume the veridicality of a trusted leader’s statement and therefore not spend much time evaluating the statement itself ” (1994, p. 732). They go on to note, “It is very possible that citizens-as-cue-takers focus so heavily on the ‘who’ that the ‘what’ recedes to the background” (1994, p. 732). In this chapter, we present evidence that this did occur in the debate over the inheritance tax. Thus, this chapter sounds a note of caution about the extent to which a reliance on credible sources can be equated with the reception of credible information. Most issue framing leads individuals to place more emphasis on some aspects of an issue relative to other aspects. However, in this chapter, we have demonstrated that frames can also influence what people believe to be true about a policy proposal. And as the inheritance tax debate demonstrated, a frame that influences what the public believes about an issue can be a very powerful tool for moving public opinion.

Note 1. Unfortunately for McCain, “Joe the plumber’s” story did not quite match that narrative, which ultimately helped to undermine McCain’s argument. 2. Based on calculation by the Internal Revenue Service (http://www.irs.gov/pub/ irs-soi/histab17.xls). 3. It should be noted that both groups of respondents are told the tax in question is “a tax on the money people leave when they die” (NPR012 2003). While this statement is vague, it is necessarily so. In order to test what people know about who is affected by the estate tax in the questions that follow, these early survey questions must be vague enough so as not to lead respondents to answers on later questions (Zaller and Feldman 1992). 4. We use difference in proportions tests to examine framing effects in this analysis. Since respondents were randomly assigned a question with or without the “death tax” frame, controls for other factors should not be necessary. Nevertheless, when we used a multinomial logit model to control for other factors we uncovered similar, though slightly stronger framing effects.

134 Impact 5. The number of people responding “don’t know” increases to 24.5 percent under this condition. However, the survey also posed the question about $1 million estates to people who originally answered “don’t know.” Unexpectedly, 45.0 percent changed their answers to favoring the repeal and only 26.1 percent opposed the repeal.

Works Cited Birney, Mayling and Ian Shapiro. 2005. “The Political Uses of Public Opinion: Lessons from the Estate Tax Repeal.” Unpublished paper. Carmines, Edward G. and James H. Kuklinski. 1990. “Incentives, Opportunities, and the Logic of Public Opinion in American Political Representation.” In John A. Ferejohn and James H. Kuklinski, editors, Information and Democratic Processes. Urbana: University of Illinois Press. Chong, Dennis and James N. Druckman. 2007. “Framing Public Opinion in Competitive Democracies.” American Political Science Review 101: 637–655. Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper & Row. Druckman, James N. 2001. “The Implications of Framing Effects for Citizen Competence.” Political Behavior 23(3): 225–256. Feldman, Stanley and John Zaller. 1992. “The Political Culture of Ambivalence: Ideological Responses to the Welfare State.” American Journal of Political Science 36: 268–307. Graetz, Michael J. and Ian Shapiro. 2005. Death by a Thousand Cuts: The Fight over Taxing Inherited Wealth. Princeton, NJ: Princeton University Press. Gravelle, Jane G. and Steve Maguire. 2005. “Estate and Gift Taxes: Economic Issues.” Congressional Research Service. Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago, IL: The University of Chicago Press. Jacoby, William G. 2000. “Issue Framing and Public Opinion on Government Spending.” American Journal of Political Science 44(4): 750–767. Kinder, Donald R. and Lynn M. Sanders. 1990. “Mimicking Political Debate with Survey Questions: The Case of White Opinion on Affirmative Action for Blacks.” Social Cognition 8: 73–103. Kuklinski, James H. and Norman L. Hurley. 1994. “On Hearing and Interpreting Political Messages: A Cautionary Tale of Citizen Cue Taking.” Journal of Politics 56(3): 729–751. Lemann, Nicholas. 2000. “The Word Lab.” The New Yorker (October 16): 100–117. Lupia, Arthur and Mathew D. McCubbins. 1998. The Democratic Dilemma: Can Citizens Learn What They Need to Know? New York: Cambridge University Press. Mondak, Jeffrey. 1993. “Source Cues and Policy Approval: The Cognitive Dynamics of Support for the Reagan Agenda.” American Journal of Political Science 37(1): 186–212. Nelson, Thomas E. and Donald R. Kinder. 1996. “Issue Frames and Group-Centrism in American Public Opinion.” The Journal of Politics 58(4): 1055–1078. Page, Benjamin I. and Robert Y. Shapiro. 1992. The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago, IL: The University of Chicago Press. Petrocik, John R. 1996. “Issue Ownership in Presidential Elections, with a 1980 Case Study.” American Journal of Political Science 40(3): 825–850.

Taxing Death or Estates? 135 Petrocik, John R. 1997. “Campaigning and the Press: The Influence of the Candidates.” In Shanto Iyengar and Richard Reeves, editors, Do the Media Govern? Thousand Oaks, CA: Sage Publications. Popkin, Samuel L. 1991. The Reasoning Voter. Chicago, IL: The University of Chicago Press. Popkin, Samuel L. 1997. “Voter Learning in the 1992 Presidential Campaign.” In Shanto Iyengar and Richard Reeves, editors, Do the Media Govern? Thousand Oaks, CA: Sage Publications. Sears, David O. and Jack Citrin. 1985. Tax Revolt: Something for Nothing in California. Cambridge, MA: Harvard University Press. Sniderman, Paul M. and Sean M. Theriault. 1999. The Dynamics of Political Argument and the Logic of Issue Framing. Chicago, IL: Midwest Political Science Association. Sniderman, Paul M., Richard A. Brody, and Phillip E. Tetlock. 1991. Reasoning and Choice: Explorations in Political Psychology. New York: Cambridge University Press. Welch, Susan. 1985. “The ‘More for Less’ Paradox: Public Attitudes on Taxing and Spending.” Public Opinion Quarterly 49: 310–316. Zaller, John and Stanley Feldman. 1992. “A Simple Theory of the Survey Response: Answering Questions versus Revealing Preferences.” American Journal of Political Science 36(3): 579–616.

8

Great Communicators? The Influence of Presidential and Congressional Issue Framing on Party Identification Michael W. Wagner

Introduction Conventional political wisdom and empirical political science research suggest that Democrats in the electorate are as pro-choice as Republican voters are pro-life—where both groups are mirroring the positions taken by elected officials in each party1 (Adams 1997; Gerrity et al. n.d.). However, it was not always this way. Even though the electorate sent pro-life Republican Ronald Reagan to the White House after his victory over the pro-choice Jimmy Carter in the 1980 election, the August 30, 1982, edition of Newsweek’s story, “Abortion: The Debate Begins” referred not to a GOP—Democratic divide on abortion but to Republican senator Jesse Helms’s bill that human life begins at conception and Republican senators Bob Packwood and Lowell Weicker’s filibuster of Helms’s proposal. At that time, there was not a significant difference between Republican and Democratic voters’ views about abortion (Gerrity et al. n.d.). The Democrats were no more transparent in their views on abortion in the years following Roe v. Wade, the decision that legalized abortion in the United States. A March, 1975, Newsweek article chronicles Idaho Democrat Frank Church’s efforts at passing anti-abortion legislation while stories appearing in issues of the same magazine in 1976 chronicled Jimmy Carter’s pro-choice political position on abortion and pro-life personal position on the same issue. More than a decade later, the November 16, 1992, Newsweek reported that the abortion issue represented the “great dividing line” in American politics; even though Republican leaders were divided themselves on whether to frame the abortion issue in ardently pro-life terms to appeal to religious conservatives or to frame the issue in top strategist Lee Atwater’s preferred nomenclature of the “big tent,” where people with diverse views on abortion were welcome in the GOP. Four years later, Newsweek’s pre-election issue labeled pro-choice Republicans “unusual” and highlighted the pro-choice stances of every Democratic senate candidate in a close election. By 1999, the parties’ abortion positions (which had been long standing pieces of their official party platforms) were solidified in conventional political wisdom connecting partisan politicians’ positions on abortion to the preferences of Republicans and Democrats in the electorate. Newsweek columnist David Brooks noted in 1999 that, while he felt

Great Communicators? 137 that Republican candidates needed to take on some section of the Republican establishment to appeal to moderates, attacking the pro-life movement would be “political suicide” (Brooks 1999, p. 35). In other words, in contemporary American politics, an individual’s opinion about abortion can be used to reliably predict their party identification (Adams 1997; Stimson 2004). How did we get to this point? Did people change their minds on the abortion issue to match their party’s position? Or, did prominent partisan politicians take competing positions on the abortion issue, framing the issue in ways that slowly nudged pro-choice voters to the Democrats while simultaneously inching pro-life voters to the GOP? If so, does partisan framing affect party identification on all major political issues, or just hot-button, cultural issues like abortion? In this chapter, I examine how presidents and members of Congress framed their arguments on taxes and abortion, two prominent issues that have helped shape modern partisan alignments, in order to demonstrate that the consistency with which partisan elites, especially presidents, frame issues has long-term consequences on partisan alignments in the electorate (Sundquist 1983; Carmines and Stimson 1989; Adams 1997). This chapter argues that the role of presidential issue framing, and elitepartisan framing more generally, is a missing link that informs our understanding of how partisan change occurs in American politics. Combining an original content analysis of 25 years of partisan issue framing (1975–2000) in Newsweek and the New York Times with public opinion data from the American National Election Study (ANES) from the same time period, this chapter demonstrates that the consistency with which presidents frame particular kinds of issues in the marketplace of ideas helps to explain the conditions under which issue-based party change occurs in the electorate (Carsey and Layman 2006; Wagner 2007)2. In short, when presidents (and nominated candidates for president) frame issues in ways that are consistent with their own party platform and when there is partisan competition over the way particular types of issues are framed by elites in the news media, presidential issue framing influences party identification. This claim comes with two important caveats: we should expect changes in partisanship only for those citizens for whom the issue is important, and we should expect more substantial shifts when it comes to “cultural” issues like abortion than for “social-welfare” issues like taxes. The analysis presented here advances our understanding of the role that partisan issue framing has on partisan change by revealing the unique role that presidents play in communicating to the American people in the rough and tumble world of partisan political communication. This chapter proceeds as follows. First, I explain how presidential issue framing’s influence on party identification synthesizes a broad theoretical framework; second, I explain the chapter’s research design, which takes advantage of an original dataset containing a comprehensive content analysis of the frames political elites used to describe the issues of taxes and abortion from 1975 to 2000; third, I examine how the frames used by presidents, presidential candidates, and members of Congress influenced long-term trends in party

138 Impact identification; and finally, I consider the implications of the findings on our knowledge of the presidency, partisan change, and framing.

Conflict Extension, the Media, and Partisan Issue Framing Scholarly accounts treating partisanship as strikingly stable and able to influence political preferences dominate modern political science (Campbell et al. 1960; Green et al. 2002; but see Fiorina 1981). Party identification is the chief engineer of political opinion for the vast majority of preferences—that is, those that are not deeply held (Carsey and Layman 2006; see also Zaller 1992). So when should we expect one’s political partisanship to affect her opinions? And when should we expect one’s opinions to affect his partisanship? When issues clearly divide the two major political parties, party-based issue change (people develop issue positions in line with their party’s) and issue-based party change (people change parties to more closely match their issue positions) should be likely to occur under different, particular circumstances (Carsey and Layman 2006). This “conflict-extension” perspective argues that when political elites, like party leaders and candidates, take competing positions on issues with elites of the other major political party, signals are sent to the public explaining which views go with each party (Carsey and Layman 2006; Carmines and Stimson 1989). Thus, some people could react to these signals by changing their party identification to match their issue preferences while other citizens could change some issue preferences to better match the positions taken by elected officials within the party. The principal question is, who should undergo issue-based party change and who should undergo party-based issue change? First, people who are not aware that the two major parties have taken diverging positions on an issue should not have any reason to alter either their partisanship or their issue preferences accordingly. The crucial element of the “conflict extension” perspective is individual awareness of party differences on issues (Layman and Carsey 2002). Issue-based party change should occur for issues that are important to people. In other words, when one is a member of a party and becomes aware that her party is taking a different position on an issue than she would prefer, and that issue is of the utmost importance to her, the conditions are ripe for issue-based party change (Carsey and Layman 2006; see also Brody and Page 1972; Schattschneider 1960; Carmines and Stimson 1989; Carmines and Wagner 2006; Tormala and Petty 2002, 2004). On the other hand, individuals who do not find an issue to be all that significant to them should not be expected to undergo issue-based partisan change. Rather, these individuals would likely amend their preferences on the issue to match the position of their party, undergoing party-based issue change (Campbell et al. 1960). One element missing from the theoretical accounts of issue-based party change and party-based issue change is the dimension of conflict on which the

Great Communicators? 139 issue exists (or, indeed, cross-cuts). That is, does issue-based party change happen for all types of issues? Cultural issues (or “easy issues”) appear to be more deeply held than other issues (Carmines and Stimson 1989; Alford et al. 2004). Presidents and other partisan elites may find it very difficult to persuade individuals to alter their preferences (or even preference weights) for issues that are held at this “gut” level. These issues are the prime candidates, then, for issue-based party change. Shafer and Claggett’s (1995) account of two “deep” issue dimensions of conflict that exist in the American public provide a useful description of the major types of issues that dominate the American political landscape.3 The first dimension focuses on the economic and social-welfare issues that became prominent during the New Deal. These are issues that deal with the government’s involvement in the economy, such as taxes, spending on health care, social security, and welfare. The linchpin connecting these issues is that they focus on distribution, “tapping arguments over the appropriate (re)distribution of economic benefits to the less fortunate” (Shafer and Claggett 1995, p. 24). The second dimension is concerned with cultural values including issues like abortion, gay rights, and prayer in public schools; “easy issues” in the Carmines and Stimson (1989) parlance. These issues are connected by their focus on “the implementation of American values—values that define appropriate social behavior” (Shafer and Claggett 1995, p. 23). These two issue dimensions have become incorporated into a single broad ideological dimension for party elites (Poole and Rosenthal 1997), but they remain largely separate and distinct for the American electorate. The issues that have been prominently featured in accounts of aggregate partisan change fall almost exclusively into the cultural issues category. Carmines and Stimson (1989) demonstrate how racial issues transformed American politics before they fused onto the social-welfare dimension (Kellstedt 2003). Adams (1997) provides a similar account of “issue evolution” for abortion, explaining how congressional elites took different positions on abortion before the public responded to this difference by slowly following the direction of elite debate. To summarize, when individuals are aware of party differences on an issue, we should expect individuals with preferences that are incongruent with positions staked out by their party to undergo party-based issue change when the issue at hand is less important to people (especially social-welfare issues) and issue-based party change for issues of import (especially cultural issues). This chapter investigates the issue-based party change hypothesis by examining the impact of presidential issue framing on a cultural issue (abortion) and a social-welfare issue (taxes). Public Awareness of Party Differences: Partisan Elites, Framing and the Media Before issue-based party change can occur, we must know more about how individuals become aware of party differences on issues. If we can better

140 Impact understand the conditions under which people will learn about party differences on issues, we improve our ability to predict, understand, and explain instances of conflict extension and issue-based party change. The mass media are a crucial intermediary between elite actors and the public. Journalists’ professional norms of objectivity regularly lead media outlets to provide arguments to “both sides” of an issue4 (Graber 2006). Since political parties provide structure for political debate (Schattschneider 1942; Sniderman 2000; Wright and Schaffner 2002), they are often key players in media stories that provide two “sides” to an issue. Elite partisans make up a significant portion of the “official sources” that are routinely relied upon in political news stories (Cook 1998). The claim that people’s evaluations of issues are influenced by how they are characterized in elite discourse is not new in the study of politics, or even social science more generally (see Introduction, above). Scholars from as diverse perspectives as E.E. Schattschneider (1960), Murray Edelman (1988), and William Riker (1990) have theorized about how issues can be portrayed in ways that encourage specific opinions and/or actions. Empirical examinations demonstrating how opinion systematically and intelligibly depends on the way issues are framed by elites abound (Iyengar 1991; Nelson and Kinder 1996). What is new to the framing discussion is the notion that partisan issue frames, when consistently reported in the media, provide individuals the opportunity to become aware of party differences on issues, creating the possibility for either issue-based party change or partisan-based issue change. Political parties provide citizens the opportunity to choose between these competing values. Parties have an interest in providing different, competing choices to the public in order to win public power (Schattschneider 1942; Sniderman 2000). What the majority of framing studies have not considered, then, are two key points. First, citizens may be more susceptible to framing effects under some conditions more than others (Sniderman and Theriault 2004; Druckman 2001; Chong and Druckman 2007). Second, since both major political parties in American politics tend to contest issues and because the news media’s norm of providing objective coverage results in the reporting of “two sides” of the issue at hand, citizens are regularly exposed to alternative ways of framing issues, and thus have a choice regarding how they might think about them. Too often, framing studies fail to account for this important fact, leading to potentially misleading conclusions about framing’s effects in the political world (Druckman in this volume; 2004). We should expect individuals to become aware of party differences on an issue after the news media have provided prominent coverage to that issue and reported consistent, competing partisan frames defining the issue (see Iyengar and Kinder 1987; Sniderman and Theriault 2004). Indeed, as I have previously demonstrated, when partisan elites present consistent frames on issues like taxes and abortion over time, people are more likely to believe there are important differences between the two major parties. When parties provide inconsistent frames on issues like energy policy, people are not more likely to notice important differences between the parties (Wagner 2007).

Great Communicators? 141 Once the parties are competing over an issue and important differences are noticed, we should expect presidential framing to affect party identification for cultural issues that are important to people while party identification should be less influential on economic issues like taxes. Those who have a position on the cultural issue in question that is inconsistent with their party’s position as described consistently by the president (or the out-party’s standard-bearer) are the most likely to undergo issue-based party change over time.

Coding Presidential Framing To develop a dataset of issue frames used by partisan elites when framing taxes, a social-welfare issue, and abortion, a cultural issue, I coded every article in Newsweek magazine containing any mention of either issue.5 Using Lexis-Nexis, I engaged in a “full text” search of every Newsweek magazine article and editorial from 1975 to 2000 that contained the word “abortion.”6, 7 I coded abortionrelated frames for every story in Newsweek that mentioned the word abortion. A total of 1,236 stories were coded. Those stories contained a total of 2,258 frames such as “abortion is murder,” “abortion is about choice,” “the government shouldn’t pay for abortion,” and so forth. For taxes, I again engaged in a Lexis-Nexis “full text” search of every Newsweek magazine article and editorial from 1975 to 2000 that contained any one of the following phrases: “tax policy,” “tax plan,” “tax cut(s),” “cut taxes,” “raise taxes,” or “tax hike.” Initially, I searched only for the words “taxes” or “tax”; however, the result was that the Lexis-Nexis search of Newsweek produced a very high percentage of articles that were unrelated to tax policy in the United States and thus were not coded. After trying several combinations of various words and phrases, I settled on the above search to provide a comprehensive list of stories about US tax policy without forcing me, or the other coders, to sift through large amounts of unrelated stories.8 A total of 1,259 stories were coded. Those stories contained a total of 2,659 frames that were coded into 208 framing categories such as “taxes are too high,” “the wealthy pay too much in taxes,” “tax cuts are unfair to the poor,” and so forth. Each story was coded for a wide variety of elements.9 Most crucial for this chapter, each story was coded for up to six different frames present in the story, the source of the frames, whether that source was the president or the politician who became his party’s nominee for president,10 whether the sources were identified with partisan labels, and whether the partisans’ frames matched the official party platform position11 on the issue in question. Since the Eisenhower Administration, Newsweek has been delivered to the mailboxes of over two million subscribers per year. While the New York Times is the national “paper of record” and has been shown to be a reasonable, if imperfect, proxy of media coverage in general (Brians and Wattenberg 1996), the Newsweek audience is comprised of less, as Kellstedt (2003) claims, “highbrow” readers than the readers of the New York Times. This reduces worries over the generalizability of New York Times findings to less elite

142 Impact members of society. Further, Newsweek’s circulation has been quite consistent from 1975 to 2000, the years I examine in this analysis. Thus, it is reasonable to conclude that Newsweek’s impact on society has been relatively stable. To supplement his Newsweek data, Kellstedt (2003) also coded a random sampling of articles about racial issues that appeared in the New York Times. As Gerrity et al. (n.d.) found with respect to coverage of abortion and interest groups, Kellstedt found no significant differences in the way racial issues were framed in the New York Times and Newsweek. This reduces concerns about the generalizability of the Newsweek data to other national media sources.

Issue Framing and Party Identification: The Cases of Taxes and Abortion12 Before examining the role that presidential issue frames play in long-term trends in party identification, it is useful to have an idea of the various sources employed by reporters in their stories about tax and abortion policies. If elite partisans such as presidents and members of Congress do not play a prominent role in framing these issues, it is not reasonable to expect their frames to affect longterm trends in party identification. As Table 8.1 shows, Republicans are the most common source of tax policy frames, providing 41.3 percent of them. Democrats provide 26.9 percent of the frames, providing face validity to Petrocik’s “issue ownership” hypothesis that Republicans own the tax issue in the public’s eyes (1996). Illustrating that the partisan nature of the framing of

Table 8.1 Distribution of Frame Sources for Taxes, 1975–2000, % (no. of frames) Source

Frequency of source providing a frame

Republican Democrat Journalist Expert Public opinion Other House/Senate, no party Interest group, business Interest group, citizen Federal Reserve Chair Religious leaders Ross Perot Senate member, no party Person Other politician, no party House Member, no party Court

41.3 (1,093) 26.9 (711) 17.7 (468) 5.8 (153) 3.0 (80) 2.5 (66) 2.4 (62) 1.9 (51) 1.1 (28) 0.8 (21) 0.1 (3) 0.6 (17) 0.4 (10) 0.3 (9) 0.2 (5) 0.2 (6) 0.04 (1)

Note Percentages add up to more than 100 because frames are allowed to have more than one source.

Great Communicators? 143 tax policy, fully 68.2 percent of all issue frames about tax policy had a party label attached to them. Of the Republican-sourced frames, 67.7 percent came from GOP presidents, the Republican Party’s nominee for president, or presidential administration staff. Republicans in Congress provided nearly 22 percent of all Republican-sourced frames. Turning to Democratic sources of frames regarding tax policy, members of Congress edge out Presidents Carter and Clinton as the most widely used Democratic source on tax frames when considering the entire time period. During the respective times that Jimmy Carter and Bill Clinton were in office, presidents are the most prominent category among Democrats. Indeed, the presidents, nominees for the White House, and Democratic presidents’ administration officials, and eventual presidential nominees account for 57 percent of Democrats’ frames on taxes overall. Democratic members of Congress provided nearly 35 percent of tax-related frames, while non-nominated candidates for president and “other” Democrats provided the rest. Figure 8.1 illustrates the general relationship between the consistency of presidential and congressional issue frames from 1975 to 2000. Consistency is defined as the frame being consistent with the party’s platform. Thus, if Republican Ronald Reagan said that a constitutional ban on abortion was unnecessary, the frame would be coded as inconsistent since the Republican platform called for such an amendment. The values presented in the figure

Framing consistency percentage

100 90 80 70 60 50 40 30 20 10

00 20

98 19

96 19

94 19

92 19

90 19

88

86

19

19

84

82

19

19

80

78

19

19

19

76

0

Year Congressional Democrats

Presidential Democrats

Congressional Republicans

Presidential Republicans

Figure 8.1 The Variability of Presidential and Congressional Framing on Tax Issues, 1975–2000

144 Impact represent the percentage of frames offered by a particular source that was consistent with the source’s party platform in a two-year period. Presidents, their administration officials, and nominated candidates for president of both parties exhibited variance but were fairly consistent in their framing of tax issues, with Democratic presidents often achieving perfect consistency. Democratic presidents never fell below 76 percent consistency in their frames while Republicans, with the exception of Gerald Ford (66 percent), never fell below 71 percent consistency with their party platform. Consistent with the themes outlined by Schaffner and Atkinson in the previous chapter, Republican presidents framed tax issues in terms of individual freedom and lower taxes for the citizenry and businesses. Democratic presidents framed tax issues in terms of equality and the use of taxes for government programs. While Republican and Democratic members of Congress generally used the same themes as their party’s presidents, congressional framing, perhaps unsurprisingly, varied much more considerably. Democratic congressmembers exhibited an “off-year” election pattern where non-presidential election seasons brought with them more consistent framing of tax issues. On the Republican side, the two years with the lowest levels of consistency (1982 and 2000) were also the years when they were quoted the least in Newsweek, likely artificially inflating the importance of frames they made that did not match the party position on tax policy. Table 8.2 shows that while abortion frames were provided by a partisan source over 38 percent of the time, abortion was reported in a much less partisan manner than tax issues were as Republicans and Democrats provided 68 percent of tax frames from 1975 to 2000. Interestingly, from 1975 to 1980, partisan elites provided only 20 percent of the total number of frames on the abortion issue. By 1990, over 35 percent of abortion frames contained partisan sources. In 2000, 53 percent of all abortion frames had a party label attached to them. This increase in the partisan nature of abortion frames comports with the issue evolution model (Gerrity, this volume; Adams 1997; Gerrity et al. n.d.). Table 8.2 Distribution of Frame Sources for Abortion, 1975–2000, % (no. of frames) Source

Frequency of source providing a frame

Republican Interest group Democrat Court Journalist Other Religious community Medical community Public opinion Politician, no party

23.4 (528) 17.6 (398) 15.2 (344) 14.1 (318) 12.6 (254) 9.5 (215) 6.1 (137) 3.0 (67) 2.7 (60) 1.9 (42)

Note Percentages add up to more than 100 because frames are allowed to have more than one source.

Great Communicators? 145 Overall, though, Republicans provided the most abortion frames of any general source (23.4 percent). Interest groups were reported as giving 17.6 percent of abortion frames, while Democrats were linked with 15.2 percent. Court decisions, statements by judges, and the like were tied to 14.1 percent of abortion issue frames, and 12.6 percent came from journalists. Focusing specifically on Republican sources of abortion frames, congressional sources slightly edged out Gerald Ford, Ronald Reagan, George H.W. Bush, Bob Dole, and George W. Bush’s use of abortion frames. However, when those Republicans were in office, they dominated the amount of frames used by members of the GOP. This was especially true for Ronald Reagan, who personally provided 40 percent of all Republican frames on abortion from 1980 to 1988. Members of his administration provided another 18 percent of frames during his presidency. On the other side of the aisle, congressional sources framed the abortion issue most often for the Democrats. However, during the Clinton years, the president provided over 41 percent of Democratic frames on abortion, and other members of his administration provided another 20 percent. Nominated presidential candidates like Walter Mondale and Michael Dukakis supplied 18.6 percent of Democratic abortion frames while non-nominated presidential candidates gave 15.7 percent of frames in Newsweek from 1975 to 2000. Figure 8.2 shows that presidential issue frames took some time to become consistent on the abortion issue. Remarkably, presidents and other elite partisans

Framing consistency percentage

100 90 80 70 60 50 40 30 20 10

20 00

19 98

19 96

19 94

19 92

19 90

19 88

19 86

19 84

19 82

19 80

19 78

19 76

0

Year Congressional Democrats

Presidential Democrats

Congressional Republicans

Presidential Republicans

Figure 8.2 The Variability of Presidential and Congressional Framing on Abortion, 1975–2000

146 Impact were about as likely as not to express party doctrine on the abortion issue in the late 1970s and early 1980s. President Carter had a famously inconsistent view regarding abortion. On occasion, he framed the issue as a moral one—opposing it. On other occasions, he framed the issue as one of women’s rights—supporting it. Ronald Reagan became increasingly consistent in his opposition to the abortion issue; George H.W. Bush struggled with consistency at times with his own view, but regularly framed the abortion issue in ways consistent with his party’s platform. President Clinton was very consistent with his frame that abortion should be “safe, legal, and rare.” Congressional Democrats’ and congressional Republicans’ framing of the abortion issue essentially mirrored the consistency with which their party’s standard-bearer framed the issue. In both cases, the consistency with which partisan elites framed the abortion issue became increasingly consistent, on average, over time with the Democrats in government providing a slightly more consistent case than Republicans did from 1975 to 2000. Variables used in the Time-Series Analysis of Framing and Party Identification Figures 8.1 and 8.2 demonstrate that partisan frames on taxes and abortion vary over time. While we often presume in framing studies that politicians pick the right frame and stick to it, the data demonstrate that the real political framing process includes bouts of inconsistency, appeals to voters who aren’t in the party “base,” and time spent searching for an effective argument. Thus, partisan frames vary, but do they matter? In order to systematically explain how presidential and congressional issue framing has mattered over time, I use party identification as the dependent variable in my analyses. Party ID 7pt is a 7-point measure of respondents’ self-identification with a political party, measured 1 = Strong Democrat to 7 = Strong Republican (see Appendices A and B for tables presenting the results of the time-series analyses on taxes and abortion). The major independent variables of interest measure the appearance of particular dimensions of frames and the percentage of party matches on frames over time. Democratic Presidential Match, Republican Presidential Match, Democratic Congressional Match, and Republican Congressional Match are all variables that measure the level of consistency with which political actors provided issues frames on tax policy during each two-year period between American National Election Studies for the years examined in this chapter (as seen in Figures 8.1 and 8.2).13 In “off years” for the party that was out of the White House, I used frames from the last candidate for president and the next candidate for president in the two-year interim to achieve a value for that year. Not surprisingly, the number of frames during these years is lower than years in which a candidate is running for president or actually is president. As checks on this measure, I also ran the analyses by coding the non-White House holding party’s “off year” score as the

Great Communicators? 147 congressional score for that year and as an average of the year preceding and following the off year. The results of the analysis presented here are entirely consistent with the results of the additional analyses. The value for each of the presidential and congressional “matching” variables is determined by taking the percentage of time that frames matched the official party position on taxes at time 2 (Frame2) minus the percentage of frames in that category at time 1 (Frame1) divided by the percentage of frames in that category at time 1 (Jones and Baumgartner 2005): (Frame2 – Frame1) Frame1 For example, Figure 8.2 shows that Ronald Reagan and his administration framed abortion consistently 81 percent of the time from 1985 to 1986 as compared to 72 percent of the time from 1983 to 1984. Thus, the Republican Presidential Match score for 1985–86 is calculated as: (81 – 72) 72 Number of Stories is simply a measure of the number of stories in any given two-year period about the issue in question. In addition to the media variables detailed above, several control variables must be put in any model examining party identification. Ideology is a similar measure, coded 1 = Strong Liberal to 7 = Strong Conservative. Party Strength is a four-point “folded” measure of the Party ID 7pt variable with 1 = Independent and 4 = Strong Partisan.14 Religiosity measures how much guidance one’s religion plays in one’s life from 1 = No role to 4 = a great deal of guidance (Layman 2001). Age is a variable representing the respondent’s age. Gender is coded 1 = male and 0 = female. Political Knowledge is coded on a three-point scale where 0 = low knowledge, 1 = moderate knowledge and 2 = high knowledge. Race is coded 1 = white/non-black and 0 = black. Education is a four-point scale coded 1 = no high school degree to 4 = college degree. South is coded 1 if the respondent lives in the former Confederacy and 0 otherwise. Income is a categorical variable that increases as the respondent’s income category increases. MIP Tax is coded 1 if taxes or tax policy was named as the “most important problem” facing the nation by the respondent and 0 otherwise while MIP Abortion is coded 1 if abortion was named the “most important problem.” The most important problem variable is a blunt measure as it does not specifically measure instances where respondents named taxes as the most important problem. The ANES collapses taxes with similar issues to create a “social-welfare” category of important issues and does the same with “cultural” issues like abortion. This, unfortunately, prevents absolutely direct tests of the hypothesis that issue-based party change occurs for issues important to the

148 Impact survey respondent. However, these variables represent the best-available proxy for the kinds of issues the respondent feels are important. Results The results of three time-series analyses including the variables concerning the consistency with which Republican and Democratic presidents and members of Congress framed tax issues moderately support Carsey and Layman’s issuebased party change hypothesis and moderately support my claim that presidential and partisan framing are less likely to engineer partisan change on social-welfare issues (see Appendix A). The full model fails to support the hypothesis. The variable measuring the Democratic presidents’ framing of taxes is negative, suggesting party identification moving towards the Democrats. The Republican presidential framing variable is positive, implying party identification moving towards Republicans. However, neither variable reaches conventional levels of statistical significance, indicating that we cannot be confident in these effects. Both variables capturing the role of congressional issue framing also failed to achieve levels of statistical significance, preventing us from being confident in their effects. Restricting the analysis to those for whom taxes is a highly important issue,15 the presence of Republican presidents’ frames that consistently matched the GOP’s contemporary positions on tax policy significantly and positively affected party identification (Republican Presidential Match, p < 0.05). The Democratic Presidential Match variable does not approach conventional levels of statistical significance indicating that we cannot be as confident in the influence of this variable. Once again, this points to the issue-ownership hypothesis (Petrocik 1996) that suggests that Republicans own taxes and stand to grow their coalition when Republican presidents and presidential nominees espouse frames in favor of low taxes. Both of the variables measuring the consistency of all partisan congressional frames about taxes were not significant, pointing to a special role for Republican presidents in issue-based party change on tax issues. The taxes-as-important model also contains a positive, significant coefficient for Gender (p < 0.05) that was not present in the full model. The rest of the variables in the middle column model have identical direction and significance as the model in the left-hand column with the exception of Party Strength. For those people for which taxes are important, Party Strength does not influence party identification over time in the cross-sectional ANES surveys. For those who find taxes, or issues like taxes, to be important, consistent partisan frames on tax issues positively affect identification with the Republican Party. The model restricting the analysis to those who did not name taxes as the most important problem facing the nation looks nearly identical to the full model. Consistently with my expectations, no issue-based party change occurs for those who do not find taxes to be an important issue. By setting the values of the partisan matching variables to their highest and lowest values, Table 8.3 reports various likelihoods that an ANES respondent

Great Communicators? 149 Table 8.3 Effects of Presidential and Congressional Framing of Taxes on Party Identification (%) Consistency High DP/CD High RP/CR High DP/RP Low CD/CR Low DP/CD High RP/CR High DP/CD Low RP/CR Low DP/CD Low RP/CR

Strong Weak Democrats Democrats

Independents Weak Strong Republicans Republicans

29

13

10

17

31

29

12

11

16

32

28

16

11

14

31

24

21

11

15

29

28

16

15

13

28

Note High = the highest matching variable score for the framing source, Low = the lowest matching variable score for the framing source. DP = Democratic Presidents, CD = Congressional Democrats, RP = Republican Presidents, CR = Congressional Republicans.

would identify as a strong Democrat, weak Democrat, Independent, weak Republican, or strong Republican. In each case, respondents are highly politically aware and believe that taxes are the most important issue facing the nation. Other variables are held at their respective means. Table 8.3 shows that when Democrats and Republicans are framing tax issues consistently within their party, but competitively between parties, Republicans benefit the most. This comports with the issue ownership hypothesis which suggests that Republicans fare better in elections where taxes are a prominent issue (Petrocik 1996). Republicans do the best when presidents frame taxes at a high level of consistency, but GOP members of Congress frame taxes with low consistency. Democrats similarly benefit the most when presidents are framing tax issues consistently as compared to congressional Democrats. Both of these results imply that presidential framing is more consequential than congressional framing of issues, perhaps yet another indicator of the media’s focus on presidential news as compared to congressional news. Interestingly, when everyone exhibits low consistency in their framing, the electorate is predicted to be much more moderate, which is consistent with the idea that issue-based party change can occur only when there is awareness of elite differences on important issues. When the parties are not clashing, awareness of differences is harder to come by, even if the differences are spelled out in party platforms. The results of the three cross-sectional, random effects time-series analyses examining how presidents and Congress members framed abortion strongly confirm the issue-based party change hypothesis (see Appendix B). The full model is in the left-hand column. There, Democratic Presidential Match has a statistically significant (p < 0.05) and negative (toward the Democrats) influence on party identification while Republican Presidential Match has a significant

150 Impact (p < 0.05) positive (toward the Republicans) affect on party identification over time. While the results for the Republican Congressional Match and Democratic Congressional Match variables both suggest that elite partisan framing matters in general as well, the statistical significance does not quite reach conventional levels. Thus, as the presidents or nominees for president of the two major political parties communicated in increasingly consistent issue frames over time, the public reacted in a predictable way to the new information filling in the political landscape. The variable measuring the importance of the abortion issue is positive and statistically significant (p < 0.01), suggesting that the more important abortion is to people, the more likely they are to move towards the Republican Party (see also Layman 2001). Number of Stories is also positive and statistically significant (p < 0.05), implying an agenda-setting effect on abortion that favors the Republican Party. When the model is restricted to those who find abortion or similar “culture wars” policies to be the most important problem facing the country, the Democratic Congressional Match and Republican Congressional Match variables achieve statistical significance in the expected directions (p < 0.05) while the Number of Stories variable fails to approach conventional levels of statistical significance. The presidential framing variables’ coefficients are statistically significant once again. Thus, when people believe abortion is an important issue, and when the political parties provide increasingly consistent, competitive issue frames about abortion, party identification moves in a way consistent with the issue evolution and conflict extension models. When the analysis is restricted to those who do not find abortion to be the most important problem facing the nation, framing does not have a statistically significant effect on party identification. By setting the values of the partisan matching variables on abortion to their highest and lowest values, Table 8.4 reports various likelihoods that an ANES respondent would identify as a strong Democrat, weak Democrat, Independent, weak Republican, or strong Republican. In each case, respondents are highly politically aware and believe that abortion is the most important issue facing the nation. Other variables are held at their respective means. Here, the advantage seems to be with the Democrats. When both Democratic and Republican presidents and members of Congress frame abortion in internally consistent, but externally competitive ways, Democrats benefit, with 37 percent of respondents predicted to place themselves at a 1 or 2 on a 7-point party identification scale. When Republicans are consistent in their framing of abortion but Democrats are inconsistent, Democratic identifiers (both weak and strong) drop from a total of 52 percent in the high matching condition (row 1 of Table 8.4) to 42 percent (row 3). That said, Republican identifiers (both weak and strong) drop from 38 percent in row 1 to 34 percent in row 3. When both presidents and members of Congress from both parties provide highly variant (inconsistent) messages on abortion, Republicans have their best performance in Table 8.4 while Democrats have their worst.

Great Communicators? 151 Table 8.4 Effects of Presidential and Congressional Framing of Abortion on Party Identification (%) Consistency High DP/CD High RP/CR High DP/RP Low CD/CR Low DP/CD High RP/CR High DP/CD Low RP/CR Low DP/CD Low RP/CR

Strong Weak Democrats Democrats

Independents Weak Strong Republicans Republicans

37

15

10

18

20

36

14

10

18

22

30

12

16

16

18

32

17

11

19

21

29

17

13

22

19

Note High = the highest matching variable score for the framing source, Low = the lowest matching variable score for the framing source. DP = Democratic Presidents, CD = Congressional Democrats, RP = Republican Presidents, CR = Congressional Republicans.

Comparing the framing effects on taxes and abortion, the claim that the more substantive effects should be present for abortion rather than taxes is confirmed. While the difference in total Republicans and Democrats is never greater than seven on the tax issue, partisan abortion issue frames can influence up to a predicted 14-point advantage for Democrats.

Discussion This chapter presents the first systematic evidence that variability in presidential and congressional issue framing helps to explain changes in party identification. The results presented here suggest that partisan elites—and especially presidents—can nudge long-term trends in party identification on cultural issues of high import to the public by virtue of presenting consistent, competing, interparty frames over time. Thus, presidents, and other elite partisans (though to a lesser degree), may be a missing link to understanding how conflict extension, and changes in partisanship more generally, actually occur. Interestingly, the presence of consistent, competing frames does not simply benefit both parties equally. Republicans do better when engaged in a clear framing battle with Democrats on taxes while Democrats enjoy the advantage on the abortion issue. Sometimes, the biggest advantages come when presidents provide a clear message, but members of Congress from the same party are less dependable in their framing of the issue. The implication may be that parties need to frame issues consistently enough to develop a brand name, but not so dogmatically as to be perceived as too rigid by moderates. There is much more work to be done: including adding the examinations of argument quality in competitive elite issue framing over time (Chong and Druckman 2007), elite framing and television news coverage, how other issues

152 Impact have been framed (Lindaman and Haider-Markel 2002), and how framing worked in other political eras. The results also point to the importance of political debate. When the parties offer consistent, competing positions on issues, the prospects for legitimate democratic choice exist. When one side is overrepresented in media coverage, as presidents often are (especially on foreign policy issues), the possibility for political manipulation is more likely (Page and Shapiro 1992; Zaller 1992). Also left out of this chapter is a careful treatment of why the particular framing strategies of party elites vary so much. Clearly, the fact that presidents framed the abortion and tax issues increasingly consistently from 1975 to 2000 suggests that these elite politicians believed they had generally found frames that worked; or, at least, that they had found frames that worked on a targeted portion of the population. How frames develop, why they change, and what kinds of frames are most successful on what kinds of people remain open, crucially important questions (but see Steensland 2008 and the other chapters in this volume). Previous analyses of framing effects questioned citizen competence because citizens seemed to shift their preferences quite easily, depending on the frame most recently presented to them. The analyses presented here suggest that, when presidents and Congress members present consistent and competing frames on important issues, people become aware of partisan differences and, under some conditions, identify with a political party because of what they have learned.

Great Communicators? 153

Appendix A Table 8.A1 Tax Frames and Changes in Party Identification, 1975–2000 Independent variables

Party identification Party identification Party identification MIP taxes = 1 MIP taxes = 0

Number of stories

0.0000 (0.001) Democratic 0.0469 Congressional match (0.08) Republican 0.0053 Congressional match (0.015) Democratic –0.1332 Presidential match (0.16) Republican 0.1301 Presidential match (0.118) Religiosity –0.029+ (0.019) Age –0.0669*** (0.012) Gender 0.0491 (0.04) Race 0.2095*** (0.02) Education 0.2086*** (0.024) South –0.2193*** (0.044) Income 0.1667*** (0.019) Ideology 7pt 0.6677*** (0.015) Party strength –0.1402*** (0.02) MIP tax 0.1047** (0.043) Political knowledge 0.3452*** (0.077) Constant 1.269*** (0.168) Wald Chi2 2954.84(16)*** N 8922 R2 0.69

–0.003 (0.003) 0.1543 (0.14) –0.0113 (0.03) 0.2152 (0.307) 0.4787** (0.218) –0.0195 (0.032) –0.0829*** (0.02) 0.1478** (0.218) 0.1782*** (0.038) 0.1935*** (0.04) –0.2944*** (0.078) 0.1073*** (0.035) 0.7425*** (0.027) –0.058+ (0.036) NA

0.0008 (0.002) –0.0097 (0.098) 0.0102 (0.018) –0.1909 (0.191) 0.0102 (0.143) –0.0347+ (0.023) –0.0591*** (0.015) 0.002 (0.049) 0.2206*** (0.023) 0.213*** (0.029) –0.1764*** (0.054) 0.1932*** (0.023) 0.6348*** (0.018) –0.1852*** (0.025) NA

0.335** (0.089) 0.9812*** (0.299) 1048.97(15)*** 2877 0.96

0.3418*** (0.81) 1.500*** (0.206) 1908.12(15)*** 6045 0.57

+ p < 0.15, *p < 0.1, **p < 0.05, ***p < 0.01. Sources: American National Election Studies, author

154 Impact

Appendix B Table 8.B1 Abortion Frames and Changes in Party Identification, 1975–2000 Independent variables Number of stories Democratic Congressional match Republican Congressional match Democratic Presidential match Republican Presidential match Religiosity Age Gender Race Education South Income Ideology 7pt Party strength Political knowledge MIP abortion Constant Wald Chi2 N R2

Party identification Party identification Party identification MIP abortion = 1 MIP abortion = 0 0.0035** (0.001) –1.2621 (0.959) 1.6214+ (1.12) –0.2086** (0.092) 0.3913** (0.165) –0.0503** (0.021) –0.0682*** (0.014) 0.0447 (0.044) 0.226*** (0.021) 0.2141*** (0.027) –0.1974*** (0.049) 0.1562*** (0.021) 0.6737*** (0.017) –0.1474*** (0.027) 0.1977* (0.105) 0.1443*** (0.056) 1.244*** 2636.81(18) 7361 0.86

–0.0019 (0.004) –1.976** (0.959) 2.619** (1.12) –2.409** (1.18) 2.805** (1.31) –0.0069 (0.05) –0.0411 (0.032) 0.1146 (0.102) 0.2676*** (0.047) 0.2036*** (0.06) –0.2082** (0.106) 0.1888*** (0.049) 0.672*** (0.039) 0.0676 (0.053) 0.1267* (0.066) NA 0.8207 (0.597) 627.27(18)*** 1372 0.92

+ p < 0.15, *p < 0.1, **p < 0.05, ***p < 0.01. Sources: American National Election Studies, author

0.0046*** (0.002) 0.0075 (0.522) –0.2443 (0.595) –0.1665 (0.3726) 0.3651* (0.203) –0.0578** (0.023) –0.0677*** (0.015) 0.0273 (0.05) 0.2127*** (0.024) 0.2155*** (0.03) –0.1817*** (0.055) 0.1476*** (0.024) 0.666*** (0.018) –0.1944*** (0.025) 0.072*** (0.032) NA 1.28*** (0.274) 2005.8(18)*** 5989 0.97

Great Communicators? 155

Notes 1. Of course, there is considerable variation in the abortion preferences of both Republican and Democratic voters (Fiorina 2005) 2. The coding universe runs from 1975 to 2000, but some of my statistical analyses occur from 1980 to 2000 owing to public opinion data limitations. 3. Layman and Carsey (2002) argue that three dimensions exist: one for racial issues, one for cultural issues, and one for social-welfare issues. This is a legitimate perspective, though, as Kellstedt (2003) carefully demonstrates (analyzing media coverage of elite issue frames regarding racial issues), the public’s racial issue dimension “fused onto” the social-welfare dimension over time. This is also true of partisan elites (Poole and Rosenthal 1997) 4. This, to the dismay of some media critics, occurs even when an issue has more than two sides. Typically, only two sides of an issue are reported; just as typically, these sides are Republican and Democrat. 5. See also Carmines et al. (2005) for a detailed factor analysis demonstrating the placement of these and other major political issues on a social-welfare issues factor and a cultural issues factor. 6. From this point on, I will use “articles” as a global representation of articles and editorials. 7. I did not code stories that mentioned abortion in the following contexts: (1) letters to the editor; (2) the character of a book, play, movie, or poem does or does not have an abortion; (3) stories about other nations’ abortion policies that were not related to the United States. 8. I did not code letters to the editor, articles about other nations’ tax policies, and the like. Even after narrowing the search terms to the most relevant words and phrases that also provided a wide range of articles, 409 stories were excluded from the analysis because of a lack of applicability. 9. See Wagner (2006) for a comprehensive explanation of the content analysis. Intercoder reliability was always above 88 percent, regardless of the coding category. 10. This choice allowed me to continue the time-series coding in years when a party was out of the White House. For example, Walter Mondale’s media coverage provides the data from 1981 to 1984 for the Democratic Presidential Match variable. 11. Coders referenced copies of the relevant sections of each party’s platform statements on abortion and taxes during the coding of each article to determine whether the frame was consistent with the party platform, inconsistent with the party platform, or indeterminate. Frames consistent with the party’s platform were coded 1 for “match” while frames inconsistent or indeterminate were recoded to both equal 0 for “non-match.” 12. While I have content analyzed media data from 1975 to 2000, the ANES only asks consistent questions about each issue from 1980 to 2000, so I restrict the more formal analyses to that period. 13. For example, the 1986 matching scores for Democratic and Republican presidents and Congress members are all calculated by adding together the scores for 1985 and 1986. 14. Some may bristle at controlling for partisan extremity in a model seeking to explain partisan change. See Hetherington’s (2001) analysis examining the increasing polarization in the electorate which includes an independent variable measuring party strength. 15. Recall that the importance measure here is a rather blunt instrument that cobbles together social-welfare issues that include taxes.

156 Impact

Works Cited Adams, Greg D. 1997. “Abortion: Evidence of an Issue Evolution.” American Journal of Political Science 41: 718–737. Alford, John R., Carolyn L. Funk, and John R. Hibbing. 2005. “Are Political Orientations Genetically Transmitted?” American Political Science Review 99: 153–167. Alpern, David M. 1975. “Abortion and the Law” Newsweek (March 3): 18. Brians, Craig Leonard and Martin P. Wattenberg. 1996. “Campaign Issue Knowledge and Salience: Comparing Reception from TV Commercials, TV News, and Newspapers.” American Journal of Political Science 40: 172–193. Brody, Richard A. and Benjamin I. Page. 1972. “The Assessment of Policy Voting.” American Political Science Review 66 (June): 450–458. Brooks, David. 1999. “Clintonizing the GOP.” Newsweek (February 8): 35. Campbell, Angus, Phillip Converse, Warren Miller, and Donald Stokes. 1960. The American Voter. New York: Wiley. Carmines, Edward G. and James A. Stimson. 1989. Issue Evolution: Race and the Transformation of American Politics. Princeton, NJ: Princeton University Press. Carmines, Edward G. and Michael W. Wagner. 2006. “Political Issues and Party Alignments: Assessing the Issue Evolution Perspective.” Annual Review of Political Science 9: 67–81. Carmines, Edward G., Michael J. Ensley, and Michael W. Wagner. 2005. “Party Polarization and Mass Partisan Identification in the United States, 1972–2000.” Presented at the Annual Meeting of the Midwest Political Science Association, Chicago, IL. Carsey, Thomas M. and Geoffrey C. Layman. 2006. “Changing Sides or Changing Minds? Party Identification and Policy Preferences in the American Electorate.” American Journal of Political Science 50(2): 464–477. Chong, Dennis and James N. Druckman. 2007. “Framing Theory.” Annual Review of Political Science 10: 103–126. Cook, Timothy E. 1998. Governing with the News: The News Media as a Political Institution. Chicago, IL: The University of Chicago Press. Druckman, James N. 2001. “On the Limits of Framing Effects.” The Journal of Politics 63(4): 1041–1066. Druckman, James N. 2004. “Political Preference Formation: Competition, Deliberation, and the (Ir)relevance of Framing Effects.” American Political Science Review 98(4): 671–686. Edelman, Murray J. 1988. Constructing the Political Spectacle. Chicago, IL: The University of Chicago Press. Entman, Robert M. 1989. “How the Media Affect What People Think: An Information Processing Approach.” Journal of Politics 51: 347–370. Fiorina, Morris P. 1981. Retrospective Voting in American National Elections. New Haven, CT: Yale University Press. Fiorina, Morris P. (with Samuel J. Abrams and Jeremy C. Pope). 2005. Culture War? The Myth of a Polarized America. New York: Pearson Longman. Gerrity, Jessica C., Michael W. Wagner, and Edward G. Carmines. n.d. “The Partisan Dynamics of Political Issues: The Role of Interest Groups and the Mass Media in the Evolution of Abortion and Crime.” Paper presented at the 2005 Annual Meeting of the Midwest Political Science Association.

Great Communicators? 157 Graber, Doris A. 2006. Mass Media and American Politics, 7th edition. Washington DC: CQ Press. Green, Donald, Bradley Palmquist, and Eric Schickler. 2002. Partisan Hearts and Minds: Political Parties and the Social Identities of Voters. New Haven, CT: Yale University Press. Hetherington. Marc J. 2001. “Resurgent Mass Partisanship: The Role of Political Elites.” American Political Science Review 95(3). Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago, IL: The University of Chicago Press. Iyengar, Shanto and Donald R. Kinder. 1987. News That Matters. Chicago, IL: The University of Chicago Press. Jones, Bryan D. and Frank R. Baumgartner. 2005. The Politics of Attention. Chicago, IL: The University of Chicago Press. Kellstedt, Paul M. 2003. The Mass Media and the Dynamics of American Racial Attitudes. Cambridge: Cambridge University Press. Layman, Geoffrey. 2001. The Great Divide: Religious and Cultural Conflict in American Party Politics. New York: Columbia University Press. Layman, Geoffrey C. and Thomas M. Carsey. 2002. “Party Polarization and Conflict Extension in the American Electorate.” American Journal of Political Science 46(4): 786–802. Lindaman, Kara and Donald P. Haider-Markel. 2002. “Issue Evolution, Political Parties, and the Culture Wars.” Political Research Quarterly 55: 91–110. Nelson, Thomas E. and Kinder, Donald R. 1996. “Issue Frames and Group-Centrism in American Public Opinion.” The Journal of Politics 58(4): 1055–1078. Newsweek. 1982. “Abortion: The Debate Begins” Newsweek (August 30): 29. Page, Benjamin I. and Shapiro, Robert Y. 1992. The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago, IL: The University of Chicago Press. Petrocik, John R. 1996. “Issue Ownership in Presidential Elections, with a 1980 Case Study,” American Journal of Political Science 40(3): 825–850. Poole, Keith and Howard Rosenthal. 1997. Congress: A Political-Economic History of Roll-Call Voting. New York: Oxford University Press. Riker, William. 1990. “Heresthetic and Rhetoric in the Spatial Model.” In James Enelow and Melvin Hinich, editors, Advances in the Spatial Theory of Voting. Cambridge: Cambridge University Press. Schattschneider, E.E. 1942. Party Government. New York: Rinehart. Schattschneider, E.E. 1960. The Semisovereign People: A Realist’s View of Democracy in America. New York: Holt, Rinehart & Winston. Shafer, Byron E. and William J.M. Claggett. 1995. The Two Majorities: The Issue Context of Modern American Politics. Baltimore, MD: Johns Hopkins University Press. Sniderman, Paul 2000. “Taking Sides: A Fixed Choice Theory of Political Reasoning.” In Arthur Lupia, Mathew D. McCubbins, and Samuel L. Popkin, editors, Elements of Reason: Cognition, Choice, and the Bounds of Rationality. Cambridge: Cambridge University Press. Sniderman, Paul M. and Sean M. Theriault. 2004. “The Structure of Political Argument and the Logic of Issue Framing.” In Willem E. Saris and Paul M. Sniderman, editors, Studies in Public Opinion. Princeton, NJ: Princeton University Press. Steensland, Brian. 2008. “Why Do Policy Frames Change? Actor-Idea Coevolution in Debates over Welfare Reform.” Social Forces 86: 1027–1054.

158 Impact Stimson, James A. 2004. Tides of Consent: How Public Opinion Shapes American Politics. New York: Cambridge University Press. Sundquist, James L. 1983. Dynamics of the Party System: Alignment and Realignment of Political Parties in the United States. Washington, DC: The Brookings Institution. Tormala, Zakary L. and Richard E. Petty. 2002. “What Doesn’t Kill Me Makes Me Stronger: The Effects of Resisting Persuasion on Attitude Certainty.” Journal of Personality and Social Psychology 83: 1298–1313. Tormala, Zakary L. and Richard E. Petty. 2004. “Resistance to Persuasion and Attitude Certainty: The Moderating Role of Elaboration.” Personality and Social Psychology Bulletin 30: 1446–1457. Wagner, Michael W. 2006. “Think of It This Way: The Conditional Influence of Competitive Partisan Issue Framing on Public Opinion and Party Identification.” Dissertation, Indiana University. Wagner, Michael W. 2007. “The Utility of Staying on Message: Competing Partisan Frames and Public Awareness of Elite Differences on Political Issues.” The Forum 5(3), Article 7. Wright, Gerald C. and Brian F. Schaffner. 2002. “The Influence of Party: Evidence from the State Legislatures.” American Political Science Review 96(2). Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge University Press.

9

The Decline of the Death Penalty How Media Framing Changed Capital Punishment in America Frank R. Baumgartner, Suzanna Linn, and Amber E. Boydstun1

Introduction A majority of Americans has long supported the death penalty and, on moral grounds, most still do. However, like any issue under the right circumstances, even capital punishment is subject to framing effects. In this chapter we draw from a larger book project (Baumgartner et al. 2008) to explore the impact of framing on public policy—specifically, on the number of human beings sentenced to death each year. We document the dramatic rise of a new way of thinking about the death penalty in America—the “innocence frame.” This frame rests on the notion that the American criminal justice system, dealing as it does with thousands of cases, is not immune from mistakes. Attention to the possibility of error in the particular case of the death penalty raises concerns among even those who support capital punishment in the abstract, as most Americans do. As the innocence frame has become more prominent in public discussion, the moralitybased and constitutionality-based frames that have traditionally dominated the death-penalty debate have declined in relative importance. The result of this shift in framing has been a marked shift in aggregate public opinion and, even more strikingly, a dramatic decline in the willingness of juries to impose death sentences across the country. Yet while the importance of framing has been well-documented (in the previous eight chapters of this book, for instance!) and, as we will show, the rise of the innocence frame in the death-penalty debate has been especially dramatic, we must consider the null hypothesis that capital punishment policy is driven not by media coverage but by underlying phenomena in the criminal justice system alone. Our analyses refute this null hypothesis. We demonstrate the powerful effects of framing on death sentences even while controlling for the crime rate and the number of individuals exonerated from death row. Media framing in the case of the death penalty has a profound influence on public policy above and beyond these real-world cues. The link between framing and public policy is well established. Shifts in the framing of nuclear power, pesticides, tobacco, and, we demonstrate here, the

160 Impact death penalty, have dramatic effects on policy (Baumgartner and Jones 1993; Schneider and Ingram 1993; Riker 1986). Framing drives policy making through a number of different channels, and key among these is public opinion. As many authors have shown, the way an issue is defined can powerfully affect the audience’s perception of the issue (see Chapters 6 and 7; see also Druckman 2001; Jacoby 2000; Nelson et al. 1997; Nelson and Oxley 1999; Pollock 1994; Terkildsen and Schnell 1997; Tversky and Kahneman 1986). Sometimes the effects that framing has on public opinion and policy making serve to augment events, sometimes they trump events, and sometimes framing effects can set in motion a social cascade—the rapid spread of information, like wildfire, throughout and between social networks. Such is the case in the recent history of the death penalty. Arguments that the system is broken—that the process, run by bureaucrats, is inevitably prone to errors—have resulted in a dramatic reduction in the number of death sentences handed down by juries in the United States, beyond that explained by the empirical evidence that mistakes are made. Further, as we show in greater detail in the larger project from which this chapter is drawn (Baumgartner et al. 2008), initial success in drawing attention to the new innocence frame contributed to the system of positive feedback, promoting the mobilization of student-led “innocence projects” in many journalism and law schools throughout the country. These innocence projects have uncovered more examples of innocents on death row, which in turn have led to greater attention to the problem. The positive-feedback mechanism we note here is typical of many social cascades and in this case explains a dramatic explosion in attention to the concept of innocence since the late 1990s. Here, we expand on our previous analyses, tracing framing of the deathpenalty debate and its impact on public policy and building a statistical model that predicts an important measure of policy outcomes and shows strong framing effects even while controlling for relevant real-world indicators. Our focus is on the rise of the appeal of innocence-based arguments against the death penalty and their effect on the annual number of death sentences in the United States. Our emphasis here is on aggregate, national-level framing effects, public opinion, and policy change. We begin by documenting and measuring the rise of the innocence frame, presenting evidence that public discussion of the death penalty has been altered by a new and unprecedented focus on the possibility of errors in the system, an eventuality with which no one is comfortable. The “innocence frame” has resonated more than previous arguments, bringing together a cluster of previously existing, but distinct, arguments into a single frame likely to have greater effect on public discourse than the same arguments considered separately. More broadly, we show attention has focused on different dimensions of the debate during different historical periods. Through a comprehensive analysis of the content of nearly four thousand stories relating to the death penalty in the past 45 years, we trace these substantive shifts in the nature of the debate. And against the backdrop of this shifting debate we demonstrate the unprece-

The Decline of the Death Penalty 161 dented power of the new “innocence frame” in influencing public policy, measured here by the annual number of death sentences issued in the United States over this same time period.

A Theory of Dynamic Framing and a Method of Measurement Many have noted the importance of issue-framing, or issue-definition, in public policy making. From Schattschneider (1960) to Cobb and Elder (1972) to Kingdon (1984) to Baumgartner and Jones (1993), such a focus has been a staple in studies of agenda-setting. Baumgartner and Jones were arguably the most systematic in tracing the level and tone of news coverage to particular issues over time. They showed for example that news coverage toward the pesticides and nuclear industries was sometimes overwhelmingly positive, but later switched to be overwhelmingly negative. Public policy, they argued, was closely related to these shifting foci of public discussion—where there were “waves of enthusiasm,” government actions supported the development of the industry; where “waves of criticism” were apparent, officials attacked and regulated those same industries (Baumgartner and Jones 1993, p. 84). “Good news” and “bad news” have policy consequences. Policy issues are much more complex than only positive and negative, good and bad, or any other single dimension; more recently the focus in the empirical study of issue-definition has moved to the analysis of how multidimensional issues are simplified in public debate so that only a few dimensions are the object of significant discussion. The multidimensional nature of policy debates has been the object of attention from dozens of scholars ranging from Riker (1986) to Schneider and Ingram (1993) to Stone (1989, 2002) to Poole and Rosenthal (1991), who note that, no matter how complex the underlying set of issues, congressional responses can be arrayed along one or two dimensions. Bryan Jones (2001) discussed these issues in some detail in developing a model of decision making based on multidimensional choice. In their most recent work, Jones and Baumgartner (2005) have developed a model focusing on the implications of bounded rationality and the implications of “attention-shifting,” which is when individuals or organizations move their focus from one dimension of a debate to another (we could also call this process “frame-shifting”). They argue that attention-shifts are inevitable in any complex decision-making environment and that these shifts explain the punctuated-equilibrium nature of public policy response. Most policies, most of the time, follow a strongly inertial, status-quo-oriented track, but are occasionally up-ended through fundamental reconsiderations of how to approach the issue: attention-shifts. We believe that the death penalty has undergone such a transformation in recent years, and hence we study it in considerable detail here. The death penalty, of course, in many ways represents a “most difficult case” for framing effects, since for most people it has been so strongly associated with deeply felt religious or moral views. These views do not change lightly. Thus, the shift that we

162 Impact document is all the more remarkable because it reorients the debate away from the morality dimension to an entirely different way of thinking of the issue. If it can happen on such a morally charged issue as capital punishment, it can happen to virtually any issue. The death penalty itself is the object of a voluminous literature and we do not propose a comprehensive analysis of the topic (see Banner 2002 for an excellent recent overview; see also Bedau 1997; Haines 1996; Jost 2001). We focus here on demonstrating quantitatively the important shifts in attention that have occurred in media coverage of the death penalty, the impact of these shifts on the overall tone of coverage, and finally the impact of the tone of discussion on public policy outcomes. We begin with a simple content analysis of New York Times coverage of the topic since 1960 using abstracts archived in the New York Times Index.2 Our comprehensive coding system allows us to go into much greater detail than others have done before to show the shifting topics of public attention, and in doing so we show how feasible it is to incorporate multidimensional coding into content analysis.

Media Coverage of the Death Penalty The death penalty has long been understood and discussed in the media in terms of constitutionality and morality. Since the mid-1990s, however, a new issue definition has arisen, and dramatically so. The innocence frame rests on the idea that, no matter what one thinks about the morality of capital punishment, we should ask whether the justice system is capable of administering the penalty across thousands of individual judicial proceedings with no errors. The idea of flaws in the system, of innocent people being on death row, of the wrong people possibly being executed, has transformed the debate. In 1996, thirty stories appeared in the New York Times concerning capital punishment; the bulk of these reported opinions, news, or events leading toward the application of the death penalty. In 2000, 235 stories appeared, and the overwhelming majority of them were highly critical of capital punishment. In just a few short years, the issue was reframed to focus on errors, mistakes, and the possibility of executing the wrong person. In Chapter 6, Druckman explains how the effectiveness of a frame can be altered not only by the quality of the frame but also by the amount of competition among distinct frames in the debate. His discussion helps to explain the profound effect of the innocence frame on capital punishment policy that we will demonstrate in this chapter. Not only is the innocence frame highly compelling— even the staunchest death-penalty supporters have to agree that the state should not execute innocent people—but it has virtually drowned out all competing definitions of the issue. As we document here, the death-penalty debate of the late 1990s and early 2000s was dominated, saturated in fact, by this single frame. Figure 9.1 shows the number of stories in the New York Times relating to capital punishment from 1960 to 2005.3 A total of 3,939 stories appeared during this time, with substantial peaks of coverage in 1976–77, just after the Supreme

The Decline of the Death Penalty 163 250

Number of stories

200

150

100

50

0 1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

Figure 9.1 The Number of Stories on Capital Punishment in the New York Times Index, 1960–2005

Court’s Gregg v. Georgia ruling reinstating the death penalty (following the 1972 Furman v. Georgia decision invalidating state capital punishment laws), and then again in 2000. During these two peaks, the newspaper carried over 150 stories per year: more than three stories per week. The figure makes clear that the death penalty emerged onto the media agenda in the 1970s; there was little coverage, less than one story per week, before 1972. Attention to the death penalty remained high through the 1980s and 1990s, with the exception of a brief dip in 1996–98. Most notably, since 1999 coverage has grown substantially, even though there has been no monumental Supreme Court decision such as those of 1972 and 1976. Rather, more recent coverage, especially the unprecedented level of coverage in 2000, has related to various challenges to the system based on juvenile offenders, the mentally handicapped, and the concept of innocence. The number of front-page stories has grown as well: from just one in 1960, there were two in 1970, four in 1980, eight in 1990, and 19 in 2000. Capital punishment contains many different dimensions of debate, from constitutionality to morality to efficacy and others. We coded each New York Times abstract for the presence of a comprehensive list of 65 different arguments, or frames, clustered into seven main themes, or frame dimensions: efficacy, morality, fairness, constitutionality, cost, mode of execution, and international concerns. We show in this section how the component arguments have waxed and waned over time, how the frame dimension, or topic, of discussion is systematically related to the tone of the debate, and specifically how the tone of the debate has shifted in response to the shifting dimensions of discussion. We begin in Figure 9.2 with a simple presentation of the number of stories using each frame dimension.

Fairness

Constitutionality

140

140

120

120

100

100

80

80

60

60

40

40

20

20

0 1960

1970

1980

1990

2000

0 1960

1970

Morality 140

120

120

100

100

80

80

60

60

40

40

20

20 1970

1980

1990

2000

0 1960

1970

Cost 140

120

120

100

100

80

80

60

60

40

40

20

20 1970

1980

2000

1980

1990

2000

1990

2000

Efficacy

140

0 1960

1990

Mode

140

0 1960

1980

1990

2000

0 1960

1970

1980

International 140 120 100 80 60 40 20 0 1960

1970

1980

1990

2000

Figure 9.2 The Number of Stories on Each Topic, or Dimension, of Attention in the New York Times Index, 1960–2005

The Decline of the Death Penalty 165 Issues of constitutionality are the single most common theme in New York Times coverage over most of this period; 1,467 stories mentioned discussions of this type, with peaks coming in 1972, 1976 and the years following that, in the mid- to late 1980s, and finally in the early twenty-first century as the constitutionality of capital punishment for juveniles and the mentally handicapped became important controversies. Morality issues have been less prominent over time, with a total of 622 stories focusing on these. Discussion of morality has been prominent since 1972, especially from 1972 to 1978 when the constitutionality of the entire death penalty was hotly debated. Since the 1970s, moral issues have never completely disappeared from the media agenda, but they have been significantly less prevalent. The fairness dimension, of which the innocence frame is one part, was not prominent before the 1980s, but grew rapidly beginning in 1983. It reached a peak in 2000 with 134 stories in that year alone; over the entire period there were 1,099 stories with fairness arguments. Many stories fit into other categories focusing on international comparisons, efficacy (whether the penalty serves as a deterrent or not, for example), cost, mode of execution, or “other” topics. Yet none of these categories was used in more than 241 stories across the entire period. In general, we can see from Figure 9.2 that constitutionality is a perennial theme; that morality has been an important theme as well; and that the fairness dimension (driven primarily by the innocence frame) has shown a dramatic increase from virtually no coverage before the 1980s to constituting more than half of the entire amount of coverage annually in many recent years. Some topics, such as the heinousness of a crime, are almost always associated with an overall pro-death-penalty theme of coverage. Others, such as international comparisons, lead almost always to an anti-death-penalty story. In fact, knowing the topic, or frame dimension, of the story allows us to make a good prediction of the tone, as Figure 9.3 shows. Stories mentioning morality issues may mention arguments on either side of the death-penalty debate, of course. These stories in fact are quite evenly split between pro- and anti-death-penalty tones. Similarly, stories mentioning constitutionality questions may focus on issues favoring the defendant or those favoring the state; here too the tone is quite split (though with a net anti-deathpenalty tendency of 60 to 40 percent). When the topic shifts to fairness, however, the vast majority of the coded stories are anti-death penalty. International comparisons, while less common, similarly have a powerful anti-death-penalty bias. The various tendencies present in Figure 9.3 make clear that the shifting attention to various topics that we demonstrated in Figure 9.2 can be expected to relate to powerful changes in the overall tone of media coverage over time. In coding New York Times coverage of the death penalty, we also noted whether the abstract mentioned anything about the victim(s) or the defendant. Figure 9.4 shows that this simple question goes far in determining the overall tone of the story. Most stories mention neither a victim nor a defendant, discussing instead such things as the general implications of Supreme Court decisions or state legislative

166 Impact

80%

81%

84%

20%

19%

16%

Cost (15)

Fairness (920)

International (91)

40%

69%

31%

Efficacy (176)

48%

60%

Constitutional (1,200)

Moral (525)

52%

52%

Mode (195)

48%

Pro-death penalty

Anti-death penalty

Figure 9.3 The Topic Determines the Tone Note: The figure shows the percentage of stories on each topic that were coded pro- or anti-death penalty, 1960–2005. Number of stories (with pro or anti tone) given in parentheses. Stories that were neutral or uncodeable by tone are not included.

debates about revising capital punishment laws and procedures. In fact, more than two-thirds of the stories (2,817 abstracts) mentioned neither a victim nor a defendant, and, of the 2,159 stories in this group that were given either a proor anti-death-penalty tone, about 40 percent were coded pro-death penalty, very similar to the overall total. Over 570 stories mentioned something about the victim with no discussion of the defendant; of the 505 stories among this group that were toned pro- or anti- approximately 71 percent had a pro-death-penalty tone. Similarly, 392 stories mentioned something about the defendant without mentioning the victim; of the 308 toned stories in this group only 21 percent were coded pro-death penalty. Those stories that mentioned both the victim and the defendant were similar to the stories mentioning neither in their overall tone. Clearly, something about discussing the defendant either humanizes him or her, raises questions about imperfections or problems in the trial, or both. Similarly, a focus on the victims of crime typically corresponds with pro-death-penalty stories. Whereas stories that mention the defendant tend to focus on the time period after the crime (the trial, the defendant’s rehabilitation, etc.), stories that mention the victim tend to focus on the crime itself, usually highlighting the heinousness of the incident.4 Combined, the data in Figures 9.3 and 9.4 make clear why proponents and opponents of the death penalty have different foci. While certainly not all of those

The Decline of the Death Penalty 167

36%

73%

64%

27%

The victim (640) Pro-death penalty

The defendant (443) Anti-death penalty

Figure 9.4 Tone and Mention of Victim and Defendant Note: The figure shows the percentage of stories mentioning either the victim or the defendant that were coded pro- or anti-death penalty, 1960–2005. The number of stories (with pro or anti tone) is shown in parentheses. Stories that were neutral or uncodeable by tone are not included.

close to the victims of violent crime support the death penalty, attention to victims is clearly associated with more stories favorable to the application of the death penalty. Thus, the subject matter of the story goes far in determining the tone. Since, as shown in Figure 9.2, the frame dimensions that dominate the debate change over time, we should expect the tone of the debate to shift as well. In Figure 9.5 we trace relative attention to the victim and the defendant since 1960. Figure 9.5 shows the increased discussion of the victims of capital crimes during the 1970s (associated with significantly more pro-death-penalty discussion in general during that time), some decline in these numbers after the reinstatement of the death penalty in 1976, a steady rise in attention to the victims of crime in discussions of death-penalty issues until 1993, and then a dramatic and sustained reversal after that date. Since 1993, attention has increasingly focused on questions relating to the defendants in criminal trials rather than to victims. This change, of course, is strongly associated with the increased concern with innocence, fairness, and the accuracy of judicial proceedings.

168 Impact

60

50

40

30

20

10

0 1960

1965

1970

1975

1980 Victim

1985

1990

1995

2000

2005

Defendant

Figure 9.5 The Number of Stories Mentioning the Victim and the Defendant, 1960–2005

We can see the dramatic rise of the innocence frame in particular by examining the number of stories each year presenting any of the following: (1) claims of innocence; (2) problems relating to evidence used in trial; (3) problems or imperfections in the justice system; or (4) characteristics of the defendant. This cluster of considerations, ranging from simple humanization of the defendant to demonstrations of actual innocence through exonerations, has always been present in the death-penalty debate, as Figure 9.6 shows. However, none of these issues was a prominent aspect of media coverage of the death penalty until they collectively surged to unprecedented levels of coverage in 2000. From 1960 to the mid-1980s there was trivial coverage of these questions, typically fewer than ten stories mentioning them each year. Coverage grew from the 1980s to the 1990s, then rocketed to new levels in 2000. We have shown how the nature of this debate has been transformed over the last 45 years. The shifting foci of attention have caused systematic variations in the overall tone of attention over time. Figure 9.7 shows these trends, summarized in the “Net Tone” of New York Times coverage: the number of pro- minus the number of anti-death-penalty stories. The data in Figure 9.7 capture the net results of all the shifting foci of debate we have examined. As attention has moved from the victims to the defendants, from morality and constitutionality to the possible imperfections in the justice system itself, the net result has been a movement towards greater anti-deathpenalty reporting. Of course, the data also show the rise of the pro-death-penalty

The Decline of the Death Penalty 169

160

100

80

60

40

20

0 1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

Figure 9.6 The Number of Stories Mentioning Innocence, Evidence, Flaws in the System, or Defendant Characteristics, 1960–2005 40

Pro-death-penalty stories minus anti-death-penalty stories

20 0 –20 –40 –60 –80 –100 –120 1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

Figure 9.7 The Net Tone of New York Times Index Coverage, 1960–2005

discussion during the 1970s and 1980s. In sum, our combined analysis of the topics and the tone of coverage allows us to address the multidimensional nature of the debate as well as the simpler question of its tone. We have developed a single measure valid for tone across the entire historical period, and we have shown how Net Tone is explained not by the same arguments over time, but

170 Impact by different arguments and shifting attention during different historical periods. The topics of discussion explain the tone, and the tone explains the public policy response (as we show in the next section). Shifting attention from one topic to another therefore can be linked clearly to important policy shifts, reversing longstanding policy equilibria.

The Rise of the Innocence Frame Why the sudden surge in attention to innocence around the year 2000? Was it because the notion that there could be innocent individuals on death row had not previously occurred to anyone? Certainly not. As Figure 9.6 shows, stories featuring questions of innocence appeared in the New York Times as early as 1960, and probably decades earlier. Publicized instances of wrongful conviction and exonerations in the United States date back at least to 1819 (Banner 2002, p. 122). And in modern times especially, Americans are no strangers to the concept of wrongful conviction. While the rate of exonerations from death row has increased in recent years (to an average of 6.7 each year between 1999 and 2006), 76 of the 123 people exonerated from death row as of 2006 were exonerated before the surge in attention to innocence began in the year 1999, at an average rate of about three each year between 1973 and 1998 (Death Penalty Information Center 2006). In other words, whatever flaws exist in the capital punishment system, they are not new. The underlying phenomena of capital punishment policy have not changed. Attention has. Indeed, even controlling for the increase in the exoneration rate, individuals exonerated between 1999 and 2005 received an average of 40 news stories in major US newspapers, as opposed to an average of three stories devoted to individuals exonerated between 1973 and 1991 (Baumgartner et al. 2008). So why all the recent fuss about innocence? There are, of course, many reasons. Some would point to individual factors, such as the actions of George Ryan— previously a staunch supporter of the death penalty—who as Governor of Illinois in 2000 placed a moratorium on that state’s death penalty and then in 2003 commuted the sentences of all death row inmates to life in prison; to the presidential campaign between Al Gore and George W. Bush in 2000, which brought many national journalists to Texas and led to significant scrutiny of the candidates’ views on the death penalty; to the decline in violent crime starting in the 1990s; or to other factors. Some would argue that the innocence frame took hold only because of new scientific technologies, chief among these forensic DNA profiling, developed in 1984 and first used in a criminal conviction in 1988, which provides overwhelming evidence of innocence in particular criminal cases. Yet, again, the facts of the death penalty have changed very little over time. Of the 130 individuals exonerated from death row between 1973 and 2008 (and of the 82 exonerations since the first exoneration based on DNA evidence in 1993), DNA evidence played a substantial factor in proving the individual’s innocence in only 19 cases (Death Penalty Information Center 2006).

The Decline of the Death Penalty 171 Perhaps more important than actual advancements in DNA testing is the newfound (or renewed) public fascination with scientific evidence, and DNA evidence in particular, as captured and fueled by popular television shows like CSI and popular events like the O.J. Simpson trial. A growing popular cultural awareness of the complexities of evidence collection and testing in criminal cases has certainly led to public awareness of the myriad human errors (or intentional falsifications) that can happen anywhere along the chain of evidence. Yet a broad cultural awareness of the possibility of wrongful conviction is not enough by itself to explain the pronounced surge in news attention to innocence we have chronicled. In our view, it is impossible to pinpoint any single cause of the rise of the innocence frame; indeed it is unimportant to do so. As we discuss in more detail in the book from which this study is drawn, the sudden and dramatic rise of the innocence frame represents a collective attention-shift; a cascade of attention produced by multiple independent yet mutually reinforcing factors operating in a positive-feedback system. Legal scholars, judges, and journalists began focusing their attention on issues of innocence in the 1970s and 1980s, not coincidentally in the wake of Watergate and the resulting wave of government distrust. This academic and legal work led to the creation of the first university-based “innocence projects”—pro-bono legal aid clinics, wherein law and journalism students and their professors work to uncover evidence of wrongful convictions. Yet despite the beginnings of the innocence projects in the years shortly after the reinstatement of the death penalty, it would be more than a decade before the innocence frame would grab hold. Attention, after all, is not proportional to urgency. Public focus shifted to the question of innocence only after a nexus of political forces, including media coverage, pushed it in that direction. Before the innocence frame would penetrate social consciousness, it would take the establishment of another eighty legal aid clinics under the Innocence Project model, the exonerations of another two dozen innocent defendants from death row, the concerted efforts of DC lobbyists, several more academic publications, and many, many national news stories. At some point in the mid-1990s, however, these elements coalesced. And once past a critical threshold, the innocence frame took public attention by storm, exploding onto the agenda and dominating the debate. In short, events, organizational efforts, governmental actions, and media framing have moved in unintended tandem to break the death-penalty debate wide open, exposing a new dimension of innocence and redefining the policy issue. In a positive-feedback system, each event makes the next event more likely. No single event can be said to be the one that caused the others; they all reciprocally have influenced one another, each reinforcing the trend toward greater attention to questions of innocence. The fascination with glamorous TV portrayals of DNA evidence surely contributed to the rise of the innocence frame, but so did the unglamorous academic publications of the 1970s and 1980s; for that matter, so did Watergate.

172 Impact Most importantly, this pattern of attention-shifting—whereby a series of independent forces leads, via a long-winding system of positive feedback, to an abrupt reframing of an issue and a sharp change in policy—is not unique to the death penalty. Observing how a morally charged issue like capital punishment can be redefined through the forces of positive feedback tells us that virtually every policy debate is susceptible to these same forces; virtually every debate can be reframed. And given what we know about the give and take between positive- and negative-feedback dynamics in the political system, when “successful” reframing occurs it is likely to be both sudden and dramatic, just as we have documented in the case of capital punishment.

The Impact of Framing on Capital Sentencing The number of Americans sentenced to death has varied quite substantially over time. From 1961 until the 1972 moratorium approximately 100 Americans were sentenced to death annually. A substantial drop in numbers followed the moratorium, but death sentences quickly began to rise in anticipation of the reinstatement of the death penalty, hitting nearly 300 in 1975. The number of yearly sentences dropped after this reinstatement surge, but to levels higher than the pre-moratorium years. And for two decades the number of sentences continued to rise, growing from 137 in 1977 to 317 in 1996. Then, in 1997, the trend began a dramatic reversal that continues today; only 128 Americans were sentenced to death in 2005, less than one-half the number of nine years before and a lower number than in any year since the moratorium. Figure 9.8 shows the annual numbers of capital sentences, executions, death row inmates, and homicides over time. The number of executions has always been substantially lower than the number of death sentences, of course. In fact, a substantial majority of death sentences are overturned on appeal, often after many years of the inmate living on death row. While overturned convictions are not exonerations, they do indicate serious flaws in the original trial or sentencing phase: serious enough that an appeals court orders the initial decision vacated and sent back for a new trial. Executions peaked in the US in 1935 and declined regularly from that date until the late 1960s, when they reached zero, where the series stayed until 1977 when Gary Graham became the first modern executionee. Executions grew substantially in the 1990s until peaking at nearly 100 in 1999; since then they have declined by almost half. Figure 9.8 also shows the size of the various death rows across the country; cumulatively over 3,500 individuals were on death row in the late 1990s, a number which had grown steadily from fewer than 500 in the 1950s, 1960s, and 1970s. Both the number of executions and the number of people on death row are driven largely by the number of individuals sentenced to death, and this number as we have mentioned has fluctuated widely over the past decades. There was particularly wild variation in the period immediately surrounding the constitutional ban and subsequent reinstatement of the death penalty in 1972

400

4000

350

3500

300

3000

250

2500

200

2000

150

1500

100

1000

50

500

0 1930

0 1940

1950

1960

1970

1980

1990

2000

Sentences (left axis)

Executions (left axis)

Death Row (right axis)

Homicides (right axis)

Homicides (in 10s) and Death Row population

Sentences and executions

The Decline of the Death Penalty 173

Figure 9.8 Capital Sentences, Executions, and the Death Row Population Primary sources: Snell (2005) and Fox and Zawitz (2006)

and 1976, respectively. As discussed above, the number of death sentences increased steadily each year from 1977 until a peak of 317 in 1996 before declining to less than one-half that number just a few years later. What explains the variation in the number of death sentences issued by juries? Certainly media coverage plays a large part. In particular, attention to the innocence argument and the attendant anti-death-penalty tone should reduce support for the death penalty and prime jurors to resist meting it out. We have shown that different topics lead to different tone in media coverage and expect that the tone of media coverage directly influences the number of death sentences juries hand down. But there is more to this story. We assess the responsiveness of death sentences to media framing, but also to homicides (shown in Figure 9.8), public opinion (i.e., the percentage of US survey respondents in favor of the death penalty in the case of murder), and the objective evidence that the system is prone to error (i.e., the number of individuals exonerated of the crimes for which they were sentenced to death). We also control for the effects of the constitutional ban on executions. We test the hypothesis that media framing influences death sentences using a simple multivariate time-series regression. Drawing on previous work, we hypothesize that, as violent crime rates increase, the number of death sentences should increase. Not only do the numbers of potential capital cases before juries increase with the number of murders, but also concern about crime

174 Impact increases with crime rates, leading to the view that more should be done to be tough on crime—namely sentence more people to death. We use the annual number of homicides (measured in thousands, but shown in Figure 9.8 in tens for purposes of scaling) as reported in the FBI Uniform Crime Reports as our indicator of violent crime rates. Public opinion has been linked to political behavior and public policy in a variety of settings (Erikson et al. 2002; Page and Shapiro 1983). Public sentiment with regard to the death penalty is historically both supportive and stable. Yet recent years have seen a small but steady and sustained drop in support. The hypothesized effect on the number of death sentences takes two forms. The first is simple and direct. As more Americans come to oppose the death penalty, the number of death sentences is expected to drop; displeasure with the policy among citizens selected for jury duty should result in fewer death sentences. The second effect is indirect. As opinion becomes less supportive, prosecutors become less inclined to pursue the death penalty—the costs, in both dollars and political capital become too high—resulting in fewer capital trials and, thus, fewer death sentences. Our measure of public opinion is an annual time-series created by combining information from all available survey data pertaining to the death penalty using Stimson’s (1999) dimensional analysis algorithm. The constitutional ban on the death penalty also had a predictable and dramatic effect on the annual number of death sentences, of course. In 1973, following the 1972 Furman v. Georgia decision to abolish the death penalty, the number of death sentences fell. Prior to Gregg v. Georgia in 1976, strong anticipation of the death penalty’s reinstatement explains the jump in death sentences in 1975 as pent-up demand for the policy was realized. We thus account for the beginning and the end of the moratorium period with variables measured 1 in 1973, the first full year of the moratorium, and 1 in 1975, the year that by all historical accounts the reversal was socially anticipated, and 0 in all other time periods. Much of the variation in the number of death sentences has come in the last decade, the period in which the innocence frame has risen and grown. It is reasonable to ask whether mounting evidence of the flaws in the system, rather than media coverage, explains the drop in death sentences. We test whether the number of exonerations alone or in conjunction with media coverage can explain the drop. We perform this test by including a measure of the number of individuals exonerated in each year, beginning with the reinstatement of capital punishment in 1976. By including in our model a measure of exonerations alongside the Net Tone measure of media framing from Figure 9.7, we can determine empirically whether it is the direct evidence of innocence itself (the prevalence of exonerations) or the media coverage of this concept of innocence that explains the shifts in death sentences that we observe. In fact, as mentioned earlier and demonstrated in more detail elsewhere (Baumgartner et al. 2008), individual exonerations generate widely divergent amounts of news coverage, with much more coverage of recent exonerations than those occurring in previous years. There are marginally more exonerations

The Decline of the Death Penalty 175 today than 30 years ago, but on average much more news coverage of those exonerations that have occurred since the innocence frame came to dominate media discussion of the death penalty. The average number of stories an individual exonerated from death row today is likely to get is more than 13 times the number that someone exonerated in the pre-innocence-frame era could expect. Something changed, and it was not the facts. Exonerees are simply more newsworthy today than before the innocence movement began. Our analysis covers the time period from 1961 to 2005, giving us 45 years of data with which to work in testing our hypotheses. For each year, we want to know whether we can predict the number of death sentences and the relative role of each of the variables in explaining the number of death sentences. The model we estimate is a simple regression analysis of the number of death sentences as a function of the number in the previous year—recognizing the inertial nature of the number of death sentences—and the previous year’s media framing (Net Tone), homicide level, and opinion, accounting as well for both the beginning and end of the moratorium period.5 Analyses that include exonerations of necessity begin in 1976. Analysis Our results are presented in Table 9.1. We focus our attention on the influence of media framing, comparing the magnitude of the estimated effects with those of the number of homicides. The entries in the table show the expected change in annual death sentences for a one-unit change in each of the row variables. The results show strong and robust framing effects. Our model of death sentences shows significant inertia, here directly assessed by the size of the coefficient of the number of sentences in the previous year. Looking first at Model 1, the first coefficient tells us that when large numbers of people were sentenced to death last year, it is likely that a large number will be sentenced to death this year. The number of sentences in the previous year propagates forward at the rate of 0.316, meaning that every death sentence in a given year is by itself statistically responsible for one-third of a death sentence in the following year. Obviously, death sentences cannot occur in thirds, but the point is that, even holding all other factors in the system constant, death sentences beget more death sentences at a rate of 0.316. This inertia has two important implications. First, it makes clear that we should not expect dramatic shifts in death sentences in any given year; the series has a significant bias toward the status quo. The value in one year is a major predictor of the value of the series in the subsequent year. Second, and obviously related to the first, it takes several years for the effect of new information to enter the system and reach its full impact. To calculate the full effect of any variable in the system, consider that it will have an impact in the first year equal to the coefficient listed, but in each subsequent year it will continue to affect the system at the rate of nearly 32 percent of its initial impact, then 32 percent of that number in the following year, and so on. This dynamic means that the long-term impact

176 Impact Table 9.1 Explaining the Number of Annual Death Sentences Model 1 (1961–2005) Sentencest–1 Net Tonet–1 Opiniont–1 Homicidest–1 (thousands)

0.316+ (0.097) 0.453+ (0.137) 5.059* (1.069) 0.817 (1.437)

Exonerationst–1

Model 2 (1973–2005) 0.339+ (0.109) 0.361# (0.210) 5.007+ (0.137) 2.423 (2.708) –0.966 (1.721)

Cumulative Exonerationst 1973 1975 Constant N R2 RMSE Std. dev.

Model 3 (1973–2005) 0.327+ (0.108) 0.432# (0.240) 4.906+ (1.423) 2.747 (2.727)

0.069 (0.197) –67.80+ (25.80) 129.49* (25.34) 22.92# (19.20) 44 0.930 23.97 83.70

129.68* (28.21) –10.44 (56.78) 32 0.836 25.27 62.32

131.29* (28.24) –17.97 (56.49) 32 0.834 25.37 62.32

Note Entries are regression coefficients; standard errors are in parentheses. * p < 0.001, + p < 0.05, # p < 0.10, one-tailed.

of shifts in any of the independent variables is about 1.5 times as great as its immediate impact, and that it takes about four years for effects to reach their full impact. So we have modeled a slowly evolving system; the first coefficient in the system shows the degree of inertia, in Model 1 about 32 percent (with very similar numbers for Models 2 and 3). Why might this inertia occur? As we noted above, death sentences vary substantially in number over time, and yet these changes evolve slowly from year to year. We can think of at least two reasons this might be the case. First, the prosecutors making decisions about whether to pursue the death penalty tend to be the same from one year to another; only slowly are prosecutors replaced. Second, standard operating procedures in the justice system mean that the process is sticky and, thus, change is very slow to occur. These two facts contribute to the likelihood that the number of death sentences handed down in a given year will look something like that in the previous year. Finally, to the degree that the processes we describe affect public opinion and juries across the country, it is clear that all people would not become aware of changes instantaneously; it takes time for new information to filter through the system. Media coverage shifting from positive to negative in tone is filtered through conversations and from one media outlet to another; nothing in this process is necessarily expected to occur instantaneously or to affect public opinion and individual thinking immediately.

The Decline of the Death Penalty 177 Despite the slowly evolving nature of the change that we observe, it is also clear that inertia is far from the full story here; change does occur. Looking at the results of Net Tone in Table 9.1, we see that the effects of media framing are substantively and statistically significant. For every ten more pro-death-penalty stories in a given year, we expect nearly five more death sentences. But, again, about 32 percent of this effect is carried forward in the next year and 32 percent of the remaining effect carried forward into the next year after that, and so on. Thus, a pro-death-penalty swing of ten points in media coverage is expected to produce about 6.5 more death sentences ((10 × .453) × 1.46 = 6.61) over four years. Pro-death-penalty coverage in the New York Times reached its pinnacle in 1973 and again in 1992 with a Net Tone of 36 more pro-death-penalty stories than anti-death-penalty stories. The situation was reversed with a 105-story antideath-penalty advantage in 2000. This 141-point swing in Net Tone translates through Model 1, then, to explain a decline of 93 death sentences after the variable reaches its full effect (64 in the short term). Compare this effect to that of homicides.6 For much of this period, homicides increased by an average of about 1,000 per year. The predicted effect on death sentences is just one additional sentence in the long run. Over the period of our study, homicides in the US have varied from 8,530 (in 1962) to 24,703 (in 1991)—a shift of 16,173. Translating this number into its effect on death sentences shows a long-term impact of 20 additional death sentences. The effect is certainly important, but considerably smaller than that of Net Tone. To further draw comparisons between the effects of Net Tone and homicides, we compare the effects of a standard deviation change in each on the number of death sentences. A standard deviation change in media framing produces an expected change in sentences of over 20, while the comparable change in homicides produces only an expected change of five death sentences.7 We see that Net Tone has an effect almost four times as large as that of homicides, using equivalent measures of each. Public opinion also plays a role in determining the number of death sentences in a given year.8 The estimated short-term effect of a one-point shift in net public opinion is to change the number of death sentences by over five (5.06). The long-term effect of this change is over seven additional death sentences (5.06 × 1.46 = 7.39). When public opinion becomes more pro-death penalty, on average we experience more death sentences. When it becomes less pro-death penalty, we experience fewer death sentences on average, as hypothesized. In fact, the evidence we present provides the upper bounds on the effects of opinion on sentences, and we do not draw out the dynamic implications of these shocks because of the mutual effects of sentences on opinions. The positive and significant finding provides further evidence for the social cascade theory: that is, for the theory that “successful” shifts in issue-definition do not occur incrementally, but rather in surges that lead to dramatic change throughout a policy system. Media framing effects build momentum in the death-penalty system, influencing sentencing directly as well as indirectly through public opinion.

178 Impact In addition to the effects of media framing, homicides, and opinion, Model 1 also controls for the effects of the constitutional ban on death sentences. We find that the estimated effects of the onset and end of the constitutional ban on the death penalty were to drop death sentences by about 68 (estimated as 1 × –67.80) and to increase death sentences by 129, respectively, controlling for the levels of homicides, Net Tone, and opinion that existed at that time (1 × 129.49). These are contemporaneous effects and so again underestimate the total effect. The total effect is the now familiar 1.46 times the contemporaneous effect; for the onset the effect is –98.99, and for the end of the moratorium the effect is 189.06. That structural changes to the death-penalty system exert such big effects should come as no surprise. One reason to include these effects in the model is so that we know that the other effects we measure are in addition to these effects, not simply reflecting them. Consider the effect of the death-penalty moratorium in 1972, which was responsible, statistically speaking, for 99 fewer people being sentenced to death in the long term, as compared to the impact of the 141-point swing in the Net Tone of media coverage between 1992 and 2000, which was statistically responsible in the long term for 93 fewer death sentences. The shift in media framing we have documented produced nearly as large an effect on capital sentencing as did the complete system shut down ordered by the Supreme Court. Without any direct checks and balances on the judicial system, the media nevertheless wield tremendous influence. We turn now to testing the alternative hypothesis that exonerations rather than media framing explain the annual number of death sentences. Both Model 2 and Model 3, also presented in Table 9.1, are run using data from 1973 through 2005. Model 2 includes the number of exonerations in the previous year. Model 3 presents an alternative operationalization of the same concept, this time including the cumulative number of exonerations from 1976 to the year in question rather than only the number in each previous year. This alternate specification allows us to test for the possibility that it took the momentum of increasing numbers of exonerations over several years to affect public policy. The main finding from these results is that exonerations themselves have no independent effect on death sentences, a finding that persists when Net Tone is omitted from the models. Further, while the effect of the number of exonerations in the last period is negative, as we would expect, the cumulative number of exonerations has an estimated effect of zero. Importantly, the inclusion of exonerations does not change the conclusion that media coverage has a substantively important impact on the annual number of death sentences. The estimated effects of all the variables in the model are highly stable.

Conclusions The number of death sentences is an important indicator of public policy, one that has changed a lot over the last four decades, ranging from a low of 42 in 1973 (when executions were rendered unconstitutional) to a high of 317 in

The Decline of the Death Penalty 179 1996, just before the rise of the innocence frame. We have shown that this variation is highly predictable. The number of death sentences responds to media attention, to public opinion, and to homicide levels. It is especially responsive to the tenor of media framing. In particular, the sentencing rate is almost four times more responsive to Net Tone than to homicides. And the recent shift in media coverage from pro-death-penalty coverage focused on morality-based and constitutionality-based frames to anti-death-penalty coverage focused on the innocence frame was responsible for nearly as big a drop in death sentences as was the moratorium placed on capital punishment in 1972. In contrast, exonerations from death row have no significant effect on the number of people sentenced to death. Taken with our knowledge of actual media coverage, we can say a lot about the nature of policy change with respect to the death penalty. The paths that homicides, media framing, and public opinion have followed in recent time periods have produced a dramatic decrease in the number of people sentenced to death. Declines in the number of homicides and a consistent anti-deathpenalty tenor in media framing—rooted in the innocence frame—have pulled opinion in a distinctly anti-death-penalty direction and, both directly and indirectly via public opinion, have led to a reduction in the number of death sentences. Together, these variables help us to understand the dramatic decline in the number of death sentences handed down by juries in the last ten years, from 317 in 1996 to four-tenths as many (128) in 2005. This conclusion is meaningful and, moreover, it is sensible. The innocence movement we have documented appears to have led to a shift in the focus of the death-penalty debate. This redefinition has had a substantial effect on public policy. The story we tell draws first on real events and then on media’s presentation of them. The consequence of these events and media framing, as we have shown, is fundamentally, but slowly, to alter public policy. This is a straightforward story, told with many tables and figures. Framing matters.

Appendix: Description of Data This Appendix offers summary descriptions of the data sources and collection procedures for the six major annual data series presented in this chapter: death sentences, death row inmate population, executions, exonerations, homicides, and net public opinion support. Death Sentences: The number of death sentences handed down each year, as provided by the following sources: 1961–1972 data from United States Department of Justice 1972; 1973–2005 data from Snell (2005: Appendix Table 2: Prisoners Sentenced to Death and the Outcome Sentence, by Year of Sentencing, 1973–2005). Death Row Inmates: The total number of individuals under sentence of death in the United States in each year. We follow the Death Penalty Information Center in supplementing the primary Bureau of Statistics data

180 Impact source with the NAACP estimate for the most recent year, as reported by Snell (2005: Figure 9.1: Persons under Sentence of Death 1953–2005). Executions: The number of individuals executed in the United States in each year, as reported by Snell (2005: Figure 9.4: Persons Executed, 1930–2005). Exonerations: The number of individuals exonerated from death row in the United States in each year, as calculated from DPIC (2006). The Death Penalty Information Center makes the following statement regarding the criteria for identifying exonerations: “The DPIC uses the traditional objective criteria that have determined innocence since the founding of this country. In order to be included on the list, defendants must have been convicted and sentenced to death, and subsequently either: a) their conviction was overturned and they were acquitted at a re-trial, or all charges were dismissed; or b) they were given an absolute pardon by the governor based on new evidence of innocence. The list includes cases in which the release occurred in 1973 or later.” Homicides: The number of homicides committed in the United States each year, as provided by the following sources: 1961–1985 data from Fox and Zawitz (2006: Table of Homicide Victimization, 1950–2004); 1986–2005 data from Uniform Crime Reporting Program (2006: Table 1: Crime in the United States by Volume and Rate per 100,000 Inhabitants, 1986–2005). Net Public Support for the Death Penalty: A measure of public support for the death penalty as calculated from data collected from the following source: Gallup Organization, public opinion surveys conducted November 11, 1953–May 5, 2006. In this chapter, we use a variable we call “net support.” This variable is calculated simply from subtracting, at each point in time, our measure of death-penalty opposition from our measure of death-penalty support (i.e., % pro – % anti). To gather these measures of death-penalty support and opposition, we began by searching on iPoll for all public opinion polls on “capital punishment” or “death penalty.” Our search yielded 780 survey items taken between December, 1936 and May, 2006. These surveys were conducted by a wide range of different survey organizations (Gallup, CBS, Roper, etc.) and employed an even wider range of question types (“Do you believe in the death penalty?,” “Are you in favor of the death penalty for murder?,” “Do you think the death penalty prevents crime?,” “Are you in favor of the death penalty for persons convicted of rape?,” etc.). Although there is significant variance across these surveys in terms of survey house and question wording, we are able to incorporate different questions asked by different organizations by using the WCalc algorithm, created by James Stimson, which we describe below. Through use of this algorithm, questions with important differences in question wording but with a common root subject—including those questions listed above and many more—can each contribute information to our overall measure of public opinion on the death penalty. Of the full set of 780 question items we downloaded from iPoll, 292 surveys meet our criteria for inclusion: directly relevant to the death-penalty debate, given to a random sample of national adults, containing responses that can be categorized into pro- and anti-death-penalty valences, and asked two or more times by the same survey

The Decline of the Death Penalty 181 organization. The surveys we used represented 19 distinct survey organizations and 35 distinct question wordings. In all, the data hold a total of 65 survey organization and question wording combinations. In order to utilize this data set of 292 surveys from multiple survey houses and with multiple question types, we employed the WCalc Public Opinion Dimensional Extraction Algorithm created by James Stimson (software retrieved March 23, 2007, from http://www.unc.edu/~jstimson/resource.html). In a manner similar to dynamic factor analysis, this software program calculates how the survey marginals (i.e., percentage values) for each survey question asked by each organization change over time. Having calculated relative change scores for each individual survey question series, the algorithm extracts the latent dimension underlying the shared patterns of variance across these changes, producing a single series of public opinion data. The algorithm is also equipped with an optional smoothing function, which we choose to employ in order to minimize the “noise” inherent in this kind of survey data. The result: two smoothed time-series, one representing aggregate support for the death penalty and the other representing aggregate opposition. Subtracting the opposition values from the support values, we obtain the final series of “Net Support,” which we employ in our analysis.

Notes 1. We gratefully acknowledge the support of the Department of Political Science at Penn State University, the Miller-LaVigne Graduate Fellowship, the Miller-LaVigne Professorship, and the College of the Liberal Arts. Baumgartner also gratefully acknowledges the support of the Camargo Foundation where he was a Fellow during the time when parts of this chapter were drafted. 2. We use the New York Times Index because the Times represents the best single national source for public policy stories. Our main interest is in how coverage shifts over time, so we are less concerned with how the Times may differ from other newspapers in its editorial stance on the death penalty or coverage of the issue; the key question is how coverage changes over time. To test the robustness of our findings using the Times, we also reviewed coverage in the Readers’ Guide to Periodical Literature. Although there is only a partial correlation of 0.53 between the two series overall (1960–2003), both show the same dramatic rise in attention in recent years coinciding with the rise in the innocence frame. In the year 2000, both the Times and the Readers’ Guide printed a higher number of stories—235 and 106, respectively—than in any previous year, and both surges were dominated by arguments against the death penalty. On the use of the New York Times as an indicator of media coverage, see Althaus et al. (2001), McCombs and Reynolds (2002), Soroka (2002), and Van Belle (2003). We also compare the Times to nine other newspapers and show high correlations in coverage over time (0.70) and, in particular, in coverage of the innocence frame (0.90); see Baumgartner et al. (2008). 3. We coded every abstract listed under the heading “capital punishment” in the New York Times Index, noting whether or not the abstract mentioned any of an exhaustive list of 65 different arguments, with abstracts being allowed to receive multiple codes as appropriate. Thanks to Cheryl Feeley for doing the bulk of this work for her Senior Thesis and for allowing us to use and update the data she collected. See Baumgartner and Boydstun (2005) for a description of the coding process.

182 Impact 4. Contrary to some common expectations, what the story mentioned about the victims, including whether they were police officers, women, children, or if there were multiple victims, had no significant impact on the tone of the story overall; any discussion of the victim was related to an overall pro-death-penalty tone. Similarly, there were few differences across types of defendants: be they female, of various racial categories, etc., any discussion of the defendant correlated highly with an overall anti-death-penalty tone. There is one notable exception to this finding: if the defendant was characterized as a terrorist, the tone was more likely to be pro-death penalty. There were few such cases, however, as a proportion of the total. 5. The number of death sentences is large enough that the variable is approximately normally distributed so that OLS regression, rather than models designed especially for the unique problems associated with count data, is appropriate for the analysis. 6. We estimated statistical tests—Granger causality tests—that allow us to test the null hypothesis that each of the processes we care about predicts the others. We find that homicide levels are predicted by the number of death sentences that occur in a given year. This finding means that the estimated effect of homicides on death sentences is, strictly speaking, biased. We find, however, that in estimating the full system of equations and interpreting effects in the context of a vector autoregression (VAR) we draw almost identical inferences to those produced by the simple regression analysis that we report here. Of particular note, Net Tone is weakly exogenous and, thus, unbiased. Given the complexities of the VAR, we choose to present the single equation in Table 9.1. 7. The standard deviation of Net Tone is just over 30 stories; that of homicides 4,508. 8. As with homicides, we find that we cannot rule out the possibility that the number of death sentences influences opinion when we test that hypothesis using Granger causality tests. This finding means that our estimated coefficient on public opinion is biased. Using more statistically sophisticated techniques that handle this possibility— again the VAR—we find that the dynamic effects we report overestimate the effect of opinion somewhat.

Works Cited Althaus, Scott L., Jill A. Edy, and Patricia F. Phalen. 2001. “Using Substitutes for Fulltext News Stories in Content Analysis: Which Text Is Best?” American Journal of Political Science 45(3): 707–723. Banner, Stuart. 2002. The Death Penalty: An American History. Cambridge, MA: Harvard University Press. Baumgartner, Frank R. and Bryan D. Jones. 1993. Agendas and Instability in American Politics. Chicago, IL: The University of Chicago Press. Baumgartner, Frank R. and Amber E. Boydstun. 2005. New York Times Capital Punishment Coverage, 1960–2005. Codebook available online at: http://www. personal.psu.edu/frb1/Innocence/DP_NYT_Codebook.pdf. Baumgartner, Frank R., Suzanna L. De Boef, and Amber E. Boydstun. 2008. The Decline of the Death Penalty and the Discovery of Innocence. New York: Cambridge University Press. Bedau, Hugo Adam, editor. 1997. The Death Penalty in America: Current Controversies. New York: Oxford University Press. Cobb, Roger W. and Charles D. Elder. 1972. Participation in American Politics: The Dynamics of Agenda-Building. Baltimore, MD: The Johns Hopkins University Press.

The Decline of the Death Penalty 183 Death Penalty Information Center (DPIC). 2006. “Innocence: List of Those Freed From Death Row.” Retrieved March 13, 2007, from http://www.deathpenalty info.org/article.php?scid=6&did=110. Druckman, James N. 2001. “On the Limits of Framing Effects: Who Can Frame?” Journal of Politics 63(4): 1041–1066. Erikson, Robert S., James A. Stimson, and Michael MacKuen. 2002. The Macro Polity. New York: Cambridge University Press. Fox, James Alan and Marianne W. Zawitz. 2006. “Homicide Trends in the United States.” Washington, DC: United States Department of Justice, Bureau of Justice Statistics. Retrieved March 16, 2007, from http://www.ojp.usdoj.gov/bjs/ homicide/tables/totalstab.htm. Furman v. Georgia, 408 U.S. 238 (1972). Gallup Organization. Public opinion surveys conducted November 11, 1953–May 5, 2006. Retrieved March 18, 2007, from the iPOLL Databank, The Roper Center for Public Opinion Research, University of Connecticut, http://www.ropercenter. uconn.edu/ipoll.html. Gregg v. Georgia, 428 U.S. 153 (1976). Haines, Herbert H. 1996. Against Capital Punishment: The Anti-Death Penalty Movement in America, 1972–1994. New York: Oxford University Press. Jacoby, William G. 2000. “Issue Framing and Public Opinion on Government Spending.” American Journal of Political Science 44(4): 750–767. Jones, Bryan D. 2001. Politics and the Architecture of Choice: Bounded Rationality and Governance. Chicago, IL: The University of Chicago Press. Jones, Bryan D. and Frank R. Baumgartner. 2005. The Politics of Attention. Chicago, IL: The University of Chicago Press. Jost, Kenneth. 2001. “Rethinking the Death Penalty.” CQ Researcher 11 (November 16): 945–968. Kingdon, John W. 1984. Agendas, Alternatives, and Public Policies. Boston, MA: Little, Brown. McCombs, Maxwell and Amy Reynolds. 2002. “News Influence on Our Pictures in the World.” In Jennings Bryant and Dolf Zillman, editors, Media Effects: Advances in Theory and Research. Hillsdale, NJ: Erlbaum. Nelson, Thomas E. and Zoe M. Oxley. 1999. “Issue Framing Effects on Belief Importance and Opinion.” The Journal of Politics 61: 1040–1067. Nelson, Thomas E., Rosalee A. Clawson, and Zoe M. Oxley. 1997. “Media Framing of a Civil Liberties Conflict and Its Effect on Tolerance.” American Political Science Review 91(3): 567–583. Page, Benjamin I. and Robert Y. Shapiro. 1983. “Effects of Public Opinion on Public Policy.” The American Political Science Review 77(1): 175–190. Pollock, Philip H. III. 1994. “Issues, Values, and Critical Moments: Did ‘Magic’ Johnson Transform Public Opinion on AIDS?” American Journal of Political Science 38(2): 426–446. Poole, Keith T. and Howard Rosenthal. 1991. “Patterns of Congressional Voting.” American Journal of Political Science 35(1): 228–278. Riker, William H. 1986. The Art of Political Manipulation. New Haven, CT: Yale University Press. Schattschneider, E.E. 1960. The Semisovereign People. New York: Holt, Rinehart & Winston.

184 Impact Schneider, Anne and Helen Ingram. 1993. “Social Construction of Target Populations: Implications for Politics and Policy.” The American Political Science Review 87(2): 334–347. Snell, Tracy L. 2005. Capital Punishment, 2005 (NCJ-215083). Washington, DC: United States Department of Justice, Bureau of Justice Statistics. Retrieved March 13, 2007, from http://www.ojp.usdoj.gov/bjs/pub/pdf/cp05.pdf. Soroka, Stuart N. 2002. Agenda-Setting Dynamics in Canada. Vancouver, Canada: UBC Press. Stimson, James A. 1999. Public Opinion in America: Moods, Cycles, and Swings, 2nd edition. Boulder, CO: Westview Press. Stone, Deborah A. 1989. “Causal Stories and the Formation of Policy Agendas.” Political Science Quarterly 104(2): 281–300. Stone, Deborah A. 2002. Policy Paradox: The Art of Political Decision Making, revised edition. New York: Norton & Company. Terkildsen, Nayda and Frauke Schnell. 1997. “How Media Frames Move Public Opinion: An Analysis of the Women’s Movement.” Political Research Quarterly 50(4): 879–900. Tversky, Amos and Daniel Kahneman. 1986. “Rational Choice and the Framing of Decisions.” Journal of Business 59: 251–284. Uniform Crime Reporting Program. 2006. Crime in the United States 2005. Washington, DC: United States Department of Justice, Federal Bureau of Investigation. Retrieved March 16, 2007, from http://www.fbi.gov/ucr/05cius/ data/table_01.html. Van Belle, Douglas A. 2003. “Bureaucratic Responsiveness to the News Media: Comparing the Influence of the New York Times and Network Television News Coverage on US Foreign Aid Allocations.” Political Communication 20(3): 263–285.

10 Framing Research The Next Steps Shanto Iyengar

As the chapters in this volume so ably demonstrate, the concept of framing provides political scientists with considerable explanatory leverage over both elite and mass political behavior. The chapters in Part I indicate that at the elite level strategic framing or “crafted talk” has emerged as an essential ingredient of political leadership. In the era of media-based politics, politicians have good reasons to be hypersensitive to the power of both imagery and language. Public personas and performances are orchestrated accordingly. In this sense, the framing phenomenon is simply an extension of William Riker’s dominance principle, i.e., that politicians strategically select or define the issues for public debate (Riker 1993). Applied to the mass public, framing represents a context-sensitive explanation for shifts in attitudes or opinions. In contrast to the dominant “dispositional” paradigm, which posits that individuals derive their opinions on specific issues or events from longstanding political predispositions including political party affiliation or liberal–conservative ideology, framing implies a more dynamic, circumstantially bound process of opinion formation in which the prevailing frames in elite rhetoric and news media coverage shape mass opinion. In this concluding chapter, I attempt to catalogue the vast literature on framing according to the alternative definitions of the key concept. I then offer two recommendations for future research. On the one hand, I argue in favor of a narrower and more precise definition that distinguishes framing effects from persuasion (or frames from messages). On the other hand, I recommend that the sample of potential frames be expanded to include non-verbal, visual cues.

Varieties of Framing The power of the framing paradigm derives, in part, from its overarching generality. To frame is to present information in a particular manner. In this information age, individuals are subject to continuous framing. For questions of politics, however, the two principal “presenters” are the news media and public officials. Scholars have therefore sought to identify the relevant media frames—presentations associated with particular news sources or genres of journalism—as well as frames associated with the topical or subject matter emphases in elite rhetoric.

186 Impact In the case of broadcast news, Iyengar (1991) identified two distinctive frames within national newscasts. The dominant—in terms of sheer frequency— “episodic” news frame takes the form of an event- or person-oriented report that illustrates public policy debates in terms of particular instances. Poverty might be depicted episodically by a news report describing the plight of a homeless family, or terrorism might be framed in terms of the latest suicide bombing in Baghdad. The thematic frame, in contrast, depicts issues at a higher level of abstraction and generality. A thematic report on poverty might focus on changes in government welfare spending or on congressional debates over the future of the food stamps program. Thus, the essential difference between episodic and thematic framing is the level of analysis: episodic framing considers issues at the level of concrete events or persons while thematic framing presents issues at the level of the collectivity. In visual terms, episodic news is characterized by “good pictures,” while thematic reports invariably feature “talking heads.” Given the commercial constraints facing broadcasters, it is hardly surprising that episodic news predominates. Anecdotal accounts of complex issues are more riveting and emotion-laden than “deep backgrounders,” whereas thematic reports require relatively in-depth and interpretive reporting which are not only more expensive to produce but also likely to drive away the median viewer. The use of episodic over thematic framing in broadcast news programs has important consequences for the audience’s attitudes. Exposure to episodic framing breeds individualistic as opposed to societal attributions of responsibility; national issues such as crime, terrorism, or poverty are traced to private actions and motives rather than deep-seated societal or structural forces (see Iyengar 1991). Over the long run, therefore, the effect of television news programming is to trivialize political discourse and weaken the ability of the viewing public to hold elected officials accountable for policy failures. Building on the idea of episodic framing, Gilliam and Iyengar (1999) introduced the related concept of a news “script” for covering public affairs issues. As originally formulated by cognitive psychologists (among others, see Abelson 1976), a script refers to “a coherent sequence of events expected by the individual, involving him either as a participant or as an observer” (Abelson 1976, p. 33). In addition to capturing mundane everyday experiences (e.g., the “restaurant script”), scripts represent narrative or text-based knowledge conveyed by fiction, drama, advertising, or journalism. Gilliam and Iyengar established that local television news coverage of crime systematically features one particular script or pattern of reporting. This script focuses on particular acts of violent crime and includes the appearance of an alleged suspect or perpetrator. In urban areas, the role of perpetrator is typically filled by young, non-white males. Given the large audience for local newscasts and the pervasiveness of crime coverage (see Iyengar and McGrady 2007), the crime script has become common knowledge for the audience. Thus, when viewers encounter incomplete versions of the script, they “fill in” the missing information and make appropriate (that is, script-based) inferences. In Gilliam and Iyengar’s study, a significant proportion of respondents who were exposed to news reports that excluded all references

Framing Research: The Next Steps 187 to a perpetrator recalled the presence of a non-white suspect (Gilliam and Iyengar 1999). Unlike the distinction between episodic and thematic framing, which corresponds to generic forms of news coverage that cut across subject matter, topical frames derive from the substantive content of media or elite presentations. News coverage of an election might dwell on the policy positions separating the candidates (the “issues” frame), the personal attributes of the candidates (the “personality” frame) or, as is more typically the case, the latest poll reports and analysis of the candidates’ prospects (the “horse race” frame). In this sense, framing is less a matter of presentation per se and more a question of sampling content. The number of available frames for any given political event depends simply on the number of dimensions or sets of considerations for considering the target event. An alternative scheme for classifying frames is based not on genre of presentation or the specific subject matter content of particular frames but rather, the level of semantic or audiovisual similarity across frames. As developed by Druckman (2001a, 2001b), operational definitions of the framing concept can be arrayed along a continuum ranging from presentations that differ only minimally in substantive content (“equivalence” framing) to presentations accompanied by numerous content differences (“emphasis” framing). The great majority of framing studies produced by political science and mass communications scholars embody the emphasis-oriented, less precise definition of framing. A final basis for cataloguing the framing literature corresponds to the distinction between one-sided and two-sided messages in persuasion research. As originally developed in psychology (by Tversky and Kahneman), the concept of framing was defined in terms of a “gain” versus “loss” presentation. As the concept began to travel across disciplines, however, it came to be defined more loosely in terms of emphasis or issue framing. Empirical tests of framing typically focused on the effects of exposure to some particular perspective on a policy debate (e.g., freedom of speech versus public safety consequences in the case of the rights of dissenting groups). More recently, scholars have begun to incorporate more elaborate emphasis framing designs in which study participants are exposed simultaneously to not just one but a pair of competing emphases on contentious issues. Under conditions of synchronous exposure (exposure at one point in time) to two-sided framing, the immediate effects on issue opinions are muted as participants tend to fall back on general predispositions as opinion cues (see, for instance, the chapters by Druckman and Schaffner and Atkinson in this volume; also Sniderman and Theriault 2004; Chong and Druckman 2007, 2008). Under conditions of repeated exposure, susceptibility to framing becomes contingent on both time of exposure and receivers’ information-processing strategies (Chong and Druckman 2008). In essence, sizable framing effects emerge among people classified as memory-based processors, but these effects dissipate rapidly. Among people assumed to be online processors, initial or instantaneous framing effects tend to be more modest, but the effects persist over time.1

188 Impact As this brief snapshot of the literature suggests, the concept of framing is applied quite differently across psychology, sociology, political science, and communications. In operational terms, however, there is less diversity; researchers have converged on a relatively loose definition of framing as information that conveys differing perspectives on some event or issue. As such, the majority of observed framing effects represent differences in opinion that cannot be attributed exclusively to differences in presentation. Emphasis-based frames not only vary the perspective or underlying dimension for considering an event (e.g., freedom of speech in the case of some particular dissenting group) but also differ in several other respects. In one of the more precise emphasis framing studies, for instance (Druckman and Nelson 2003), approximately 50 words were unique to each of the competing “free speech” and “special interest” frames. While it is admittedly difficult to anticipate differential responses to specific words, the fact remains that semantic variability (especially noun variability) represents multiple and hidden confounded variables. At some point, as the content gap between emphasis frames grows, framing becomes indistinguishable from persuasion— changes in opinion prompted by exposure to one message rather than another. In short, the widespread adoption of the emphasis over the equivalence mode of framing makes it much more difficult to observe framing effects per se. Frames have morphed into messages.

A Return to Equivalence Framing One of the reasons for the popularity of the emphasis approach is the ease with which researchers can construct framing manipulations. It is a relatively simple matter to string together three or four paragraphs of text that provide different considerations concerning some underlying attitude target. But, as noted above, different words may convey more than differences in perspective and different individuals may “read” the same words quite differently. One solution to the problem of semantic ambiguity is to utilize non-verbal cues as the vehicle for assessing framing effects. A picture may be worth a thousand words, but for framing research pictorial manipulations have the added value of precision. It is possible to create alternative versions of a picture that differ along a specific dimension, but which remain identical on all other observable dimensions so that any variation in the audience response can be attributable only to the manipulated dimension. In a recent series of experiments, for instance, Bailenson and his collaborators (Bailenson et al. 2009) used face morphing techniques to alter a candidate’s facial resemblance to individual voters. (Their methodology requires access to a photographic database of voters’ faces and the ability to match individual faces to survey responses.) The researchers morphed a target candidate’s face with either a particular respondent’s face or with the face of some other respondent. For any given respondent, therefore, the candidate appeared either similar or dissimilar (see Figure 10.1 for an example of the similarity manipulation).

Framing Research: The Next Steps 189

Subject

“George Bush”

60:40 blend

Subject

“John Kerry”

60:40 blend

Figure 10.1 Example of Similarity Manipulation

Given the wealth of evidence suggesting that similarity is a compelling basis for attraction, Bailenson and his collaborators were able to compare the effects of facial similarity with similarity based on partisanship or policy preferences, or similarity according to group affiliation (i.e., gender or race). As expected, the effects of facial similarity were much weaker than the effects of partisan or ideological proximity, but they were significant nonetheless. The morphing of photographs can be generalized to incorporate dimensions other than an individual’s facial structure. It is also possible to manipulate the prominence of group-based attributes associated with race or ethnicity. Social psychologists have shown that relatively prototypic Afrocentric facial features elicit more stereotypic evaluations of African-Americans than relatively ambiguous or racially neutral features (see Blair et al. 2002; Pizzi et al. 2005). In the case of skin complexion, for instance, there is ample evidence that the darker the target non-white individual, the more hostile the evaluations expressed by whites (for a review of the evidence, see Hochschild and Weaver 2007). Applying these findings concerning non-verbal racial cues to the political arena, we might expect that darkening a non-white candidate’s complexion would have the effect of weakening that candidate’s standing among white voters. To test this proposition, researchers presented a national sample of registered voters with relatively lighter and darker images of Senator Obama at two different stages of

190 Impact the 2008 campaign—early February and late October. The results showed that whites’ evaluations of Obama were significantly affected by skin complexion in February, but that, as the campaign progressed and voters learned more about Obama, their evaluations came to be dominated by traditional cues (e.g., party affiliation and issue opinions) and were thus unaffected by the complexion manipulation (Iyengar et al. 2009). As the facial similarity and complexion studies suggest, visual rather than semantic frames provide researchers with tight control over particular attributes of candidates. A candidate might be framed as an in-group member on the basis of race, gender, age, or other physical attributes. Not only does the visual medium provide greater precision and a return to the tradition of equivalence framing, visual stimuli are also essential ingredients of the daily stream of political information. In terms of sheer size, the broadcast news audience dominates the audience for print media; non-verbal cues thus represent an ecologically valid test of framing effects on mass opinion.

Summing Up The papers published in this volume provide compelling testimony to the explanatory power of the framing paradigm. No matter how it is defined, the concept of framing has had a major impact in Political Science and Communications scholarship. But as the number of papers on framing has increased, the boundaries between framing and other forms of media or elite influence on public opinion have become blurred. Framing is often indistinguishable from priming or persuasion. I have tried to suggest in this closing chapter that researchers should devote more attention—both at the conceptual and operational levels—to distinguishing between framing and other contextual, short-term influences on public opinion. The first step toward this goal is the adoption of a more precise definition of frames. Eventually, I anticipate that the revival of equivalence framing will contribute to the development of a more powerful theoretical framework for understanding the conditions under which framing effects do or do not occur.

Note 1. Memory-based processors store incoming information about any given attitude target and retrieve this information when formulating their attitudes. Online processors, on the other hand, update their attitudes continuously and then discard the information on which the updating occurred (see Lodge et al. 1989).

Works Cited Abelson, Robert. 1976. “Script Processing in Attitude Formation and Decision Making.” In J.S. Carroll and J.W. Payne, editors, Cognition and Social Behavior. Hillsdale, NJ: Lawrence Erlbaum Associates. Bailenson, Jeremy, Shanto Iyengar, and Nick Yee. 2009. “Facial Similarity between Voters and Candidates Causes Influence.” Public Opinion Quarterly, in press.

Framing Research: The Next Steps 191 Blair, I.V., C.M. Judd, M.S. Sadler, and C. Jenkins 2002. “The Role of Afrocentric Features in Person Perception: Judging by Features and Categories.” Journal of Personality and Social Psychology 83: 5–25. Chong, Dennis and James Druckman. 2007. “Framing Public Opinion in Competitive Democracies.” American Political Science Review 101: 637–655. Chong, Dennis and James Druckman. 2008. “Dynamic Public Opinion: Framing Effects over Time.” Unpublished manuscript, Department of Political Science, Northwestern University. Druckman, James. 2001a. “On the Limits of Framing Effects.” Journal of Politics 63: 1041–1066. Druckman, James. 2001b. “Evaluating Framing Effects.” Journal of Economic Psychology 22: 91–101. Druckman, James and Kjersten Nelson. 2003. “Framing and Deliberation.” American Journal of Political Science 47(4): 729–745. Gilliam, Frank and Shanto Iyengar. 1999. “Prime Suspects: The Influence of Local Television News on the Viewing Public.” American Journal of Political Science 44: 560–573. Hochschild, Jennifer and V. Weaver. 2007. “The Skin Color Paradox and the American Racial Order.” Social Forces 83: 643–670. Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago, IL: The University of Chicago Press. Iyengar, Shanto and Jennifer McGrady. 2007. Media Politics: A Citizen’s Guide. New York: W.W. Norton. Iyengar, Shanto, Kyu Hahn, Christopher Dial, and Mahzarin Banaji. 2009. “Racial Cues in Electoral Choice: Some Experimental Evidence.” Manuscript in preparation. Lodge, Milton, Kathleen McGraw, and Patrick Stroh. 1989. “An Impression-driven Model of Candidate Evaluation.” American Political Science Review 83: 399–419. Pizzi, W.J., I.V. Blair, and C.M. Judd. 2005. “Discrimination in Sentencing on the Basis of Afrocentric Features.” Michigan Journal of Race & Law 10: 327–355. Riker, William. 1993. “Rhetorical Interactions in the Ratification Campaign.” In W. Riker, editor, Agenda Formation. Ann Arbor: University of Michigan Press. Sniderman, Paul and Sean Theriault. 2004. “The Structure of Political Argument and the Logic of Issue Framing.” In W.E. Saris and P.M. Sniderman, editors, Studies in Public Opinion. Princeton, NJ: Princeton University Press.

Contributors

Taylor Ansley graduated from Davidson College, North Carolina, in May 2007, cum laude, with high honors in Political Science. He received the J. Harris Proctor Award, presented annually to an outstanding political science major, and was selected as a William Holt Terry Fellow, a Bank of America Kemp Scholar, and a Presidential Fellow with the Center for the Study of the Presidency. His research at Davidson focused on political communications and media coverage of political debates, with a particular eye to new media and online communications. Mary Layton Atkinson is a doctoral candidate in political science at the University of North Carolina, Chapel Hill. Her research interests include media bias, media effects, and public opinion on social and morality policies. Frank R. Baumgartner is the Richard Richardson Distinguished Professor of Political Science at The University of North Carolina, Chapel Hill. He is co-author of The Decline of the Death Penalty and the Discovery of Innocence, winner of the 2008 Gladys M. Kammerer Award for the best political science publication in the field of US national policy. His other publications include Comparative Studies of Policy Agendas (2007); The Politics of Attention: How Government Prioritizes Problems (with Bryan D. Jones, 2005); Policy Dynamics (with Bryan D. Jones, 2002); and Agendas and Instability in American Politics (with Bryan D. Jones, 1993), winner of the 2001 Aaron Wildavsky Award for a work that continues to influence the study of public policy at least ten years after its publication. Amber E. Boydstun is an Assistant Professor of Political Science at the University of California, Davis. She is co-author of The Decline of the Death Penalty and the Discovery of Innocence, winner of the 2008 Gladys M. Kammerer Award for the best political science publication in the field of US national policy. Her work examines the relationship between media framing and agenda control and has also appeared in Mass Communication and Society. James N. Druckman is a Professor of Political Science and a fellow at the Institute for Policy Research at Northwestern University, Illinois. His most recent work examines how citizens make political, economic, and social

Contributors 193 decisions in various contexts. He has published articles in several journals including American Political Science Review, American Journal of Political Science, and The Journal of Politics and he currently edits the journal Public Opinion Quarterly. Jessica C. Gerrity is Assistant Professor of Political Science at Washington College, where she teaches courses on American politics. She is co-author of Congress in the Public Mind. Her research focuses on interest groups, Congress, women and politics and public opinion. She worked for Sen. Jack Reed in the Senate as an American Political Science Association Congressional Fellow in 2006–2007. Douglas B. Harris is Associate Professor of Political Science at Loyola University, Maryland. His research on Congress, political parties, and media politics has been published in several scholarly journals and he is co-author of The Austin/Boston Connection: Five Decades of House Democratic Leadership, 1937–1989. Shanto Iyengar is the Harry and Norman Chandler Professor in Communication and Professor of Political Science at Stanford University, California. Iyengar’s books include Media Politics, Going Negative: How Political Advertisements Shrink and Polarize the Electorate (co-authored with Stephen Ansolabehere), Do the Media Govern? (co-edited with Richard Reeves), and Is Anyone Responsible? How Television Frames Political Issues. His research has also been published in several journals including American Political Science Review, Communication Research, Journal of Personality and Social Psychology, and Public Opinion Quarterly. Suzanna Linn is Professor of Political Science at The Pennsylvania State University. She is co-author of The Decline of the Death Penalty and the Discovery of Innocence, winner of the 2008 Gladys M. Kammerer Award for the best political science publication in the field of US national policy. Her research examines the dynamics of public opinion, elections, and public policy and the statistical methods used to analyze them and has appeared in journals such as American Political Science Review, American Journal of Politics, Political Analysis, and Statistics in Medicine. Thomas E. Nelson is Associate Professor of Political Science at the Ohio State University. His research interests include the effect of political communication on attitude formation and change, social stereotypes and intergroup relations, and social-cognitive approaches to public opinion. His work has appeared in the American Political Science Review, The Journal of Personality and Social Psychology, and The Journal of Politics and he was the winner of the 2002 Erik Erikson award for early career achievement from the International Society for Political Psychology. Brian F. Schaffner is an Associate Professor of Political Science at the University of Massachusetts, Amherst. His research focusing on public opinion,

194 Contributors campaigns and elections, political parties, and legislative politics has appeared in several journals including American Political Science Review, American Journal of Political Science, Public Opinion Quarterly, and Political Communication. Schaffner edits the journal Congress & the Presidency, is co-author of Parties, Politics and Elections in America (6th edition) and is a regular contributor to the award-winning political site Pollster.com. Patrick J. Sellers is Professor of Political Science at Davidson College, North Carolina. His current research explores the interaction among journalists and members of Congress. Sellers is author of Cycles of Spin: Strategic Communication in the U.S. Congress and he has published his work in the American Political Science Review, American Journal of Political Science, The Journal of Politics, and other political science journals. Allyson F. Shortle earned her BA at Union College in Schenectady, New York and is currently working on her Ph.D at The Ohio State University. Her research interests include public opinion, race, gender, and political psychology. Michael W. Wagner is Assistant Professor of Political Science at the University of Nebraska, Lincoln. He is co-author of Congress in the Public Mind. His current research focuses on partisan issue framing, campaign coverage, and biology and politics. His research appears in Political Research Quarterly, American Politics Research, The Annual Review of Political Science, and several other journals and book chapters. Dana E. Wittmer earned her BA from The University of California at Santa Barbara and is currently working on her Ph.D. at The Ohio State University. Her research interests include gender, race, and political psychology.

Index

60 Plus Association 123 abortion 2–5, 36, 48, 60–78, 103, 123, 136–7, 139–52, 154 pro-choice 60, 62–3, 66, 70–4, 136–7, 166 pro-life 36, 62, 66–74, 136–7 Abu Ghraib 102–3 academic freedom see values ACLU (American Civil Liberties Union) 23, 102 African-Americans 189 Alan Guttmacher Institute, The 61 alienation 81 Amended Substitute Senate Bill 1 15–16 American Recovery and Reinvestment Act 1 analogy 27–8, 36 ANES (American National Election Study) 137, 146–8, 150 Ansley, Taylor 4, 36, 74, 78 Apple 94 Archer, Bill 123 Armstrong, Jerome 85 Atkinson, Mary Layton 3, 4, 27, 144, 186 attention-shifting 161 Atwater, Frank 136 Baghdad 186 Bailenson, Jeremy 188–9 Baltimore, Maryland 41 Baumgartner, Frank 5, 49, 159, 161 Begala, Paul 45 Behe, Michael 29 Benchmark H, Life Sciences, Grade 10 16–17

Bible 22 Big Tent 136 Birney, Mayling 125, 132 Blagojevich, Rod 104, 105, 109 Blethen, Frank 123 blog 83–4, 90–2 Boydstun, Amber E. 5, 49, 158 Bright, Jean Hay 88 Brooks, David 136 Brown, Sherrod 79, 91–3 Brown University 16 Brownback, Sam 12, 73 Bush, George H. W. 145–6 Bush, George W. 11, 41, 71, 145, 170 administration 49, 50 legislation signed 61 presidential campaign 79 Byrd, Robert 89 California 45 Calvert, John 15 Canady, Charles 62 capital punishment 163 Capitol Hill 53 Cardin, Ben 88–9 Carmines, Edward G. 139 Carsey, Thomas M. 147 Carter, Jimmy 136, 143, 146 Cartern, Jack 88 Case Western Reserve University 16 Casey, Bob 87 casino 3, 102, 104–9, 112–13, 115–21 CBS 51 Cheney, Richard 51, 54 Chicago, Illinois 104–5 Chicago Tribune 105

196 Index Chong, Dennis 101–4, 106–7, 109, 114, 127, 187 Christian 22, 28 Christian fundamentalism 30 Church, Frank 136 Church-state separation 21 Cincinnati, Ohio 62 Claggett, William 139 Clawson, Rosalee A. 2, 49, 66, 78, 103, 119 Cleveland Plain Dealer, The 92 Clinton, Bill 46, 143, 145–6 administration 50 health care reform initiative 14 second term 125 Clinton, Hillary 93 Cobb, Roger W. 161 Cochran, Michael 17 Collins, Francis 22 Columbus Dispatch (Ohio), The 92 Columbus, Ohio 16 Confederacy 147 conflict extension perspective 138 conflict frames 48, 51–2 Congress 2, 45, 47, 49–51, 53, 55, 60, 89, 103, 123, 137, 142–4, 151–2 and abortion 66, 68, 71, 73, 149, 151 members of 45, 62, 64, 72, 74, 78, 125 Congressional Record 72 Connecticut 88 contested value 14, 35 Contract with America 45 Cook County, Georgia 108–9 Copernicus 20 COS (Conservative Opportunity Society) 52 creationism 11–12, 33–4 young-Earth creationism 20 CRS (Congressional Research Service) 125 CSI: Crime Scene Investigation 171 cultural issues 137, 139, 141, 147, 151, 171 CWA (Concerned Women of America) 60, 62, 67–71, 73 Daley, Richard 105 Dallas, Texas 62

Darwin, Charles 11 Darwinism 22, 28 Dawkins, Richard 22 de Vellis, Phil 93 Dean, Howard 86 death penalty 5, 159, 161–2, 175, 177 debate about 160, 168 moratorium 178 opposition to 178 support for 178 death sentences 175–6, 178 death tax see inheritance tax Deaver, Mike 54 De Boef, Suzanna L. 5, 159 Democratic Message Board 49 Democrats 2, 27, 41, 44–6, 48–50, 52, 79, 88, 90, 92–3, 104, 109, 127–8, 130–1, 136–7, 142–3, 146, 151 arguments and strategy 43, 122 caucus 45–6,49 DCCC (Democratic Congressional Campaign Committee) 47 DNC (Democratic National Committee) 47 in the House 45, 51, 54 issue credibility 126 leaders 47, 49, 50, 54 Presidents 144 public perception 49, 149–50 DeWine, Mike 91–3 Discovery Institute 16–17 DMC (Democratic Message Group) 45 DNA 170–1 Dogmatic Darwinsim 27–8, 30, 32–4 Dole, Bob 145 dominance principle 185 Dover, Pennsylvania 12 Druckman, James 3, 4, 6, 36, 53, 64, 78, 101–4, 106–7, 109, 114, 121, 126, 127, 162, 186, 187 Dukakis, Michael 145 Edelman, Murray 140 Edwards, Don 45 egalitarianism 13 Eisenhower, Dwight D. 141 Elder, Charles D. 161 Election Day 79–80, 104–6, 108 emphasis framing see framing

Index 197 Enlightenment 19 Entman, Robert 2, 35 environmentalism 13 episodic frames see framing equivalence framing see framing estate tax see inheritance tax evolution 11, 17, 21, 29, 32 excludability 82 exclusive value` 14, 35 Facebook 79 FBI 174 federal estate tax see inheritance tax First Amendment 21 Florida 62 Ford, Gerald 144–5 Foreign Intelligence Surveillance Act 94 Foster Grandparents Programs 54 Fox, Michael J. 88 frame dimension see framing 165 frame-shifting 161 framing 13, 18, 30, 35 causal frames 48 emphasis framing 187 episodic frames 43, 48, 51, 186, 187 equivalence framing 187–8 fairness 13, 15, 20, 28, 30, 32 frame dimension 13, 18, 30, 36 framing tactics 23, 27 group-centric frames 43, 48 innocence frame 5, 36, 159–61, 165, 168, 174, 177 issue framing 161 linkage frames 44 message frames 42, 44 strong frames 121 thematic frames 43, 186–7 tragic and rare frame 63, 66 FRC (Family Research Council) 60, 62, 67–9, 72 free riding 81–2 free speech 13, 20, 28, 30, 32, 188 Freedman, Paul 63 Frost, Graeme 41, 43 Furman v. Georgia 163, 174 Gallup Organization 63 Georgia 44 Gephardt, Richard 45, 49

Gerrity, Jessica C. 3, 5, 36, 60, 78, 103, 123 Gilliam, Frank 186 Gingrich, Newt 4, 45–6, 47, 49, 52–4, 123 God 22 Gonzales v. Carhart 61 GOP see Republicans Gore, Al 170 Gorney, Cynthia 62 Graham, Gary 172 Grassley, Charles 44 Gray, Barbara 27 Green Party 104 Greenbriar Resort 54 Gregg v. Georgia 163, 174 group-centric frames see framing Harris, Doug 4–5, 41, 103 Hart, Peter 46 Haskell, Martin 62 Hatch, Orrin 44 Helms, Jesse 136 Helmsley, Leona 49 heresthetics 63 Holyoak, Keith James 29 House Republican Conference 52 HPV vaccine 70 Human Subjects pool 30 humanitarianism 13 Hutcheson, Ron 11 Icons of Evolution 17 ICR (International Communications Research) 127 Idaho 136 Illinois 104–6, 108, 170 indexing 5 individualism 13 information costs 81 Ingram, Helen 161 inheritance tax 36, 121, 123–4, 126, 129–32 death tax 3, 4, 27, 36, 64, 121–31, 133 estate tax 3, 27, 122–30, 132 innocence projects 171 innocence frame see framing institutional role assignment 23

198 Index intellectual integrity 20, 35 intelligent design 11–13, 15–23, 27–30, 32–4 Intelligent Design Network 15 Internet 74, 83–4, 86, 90, 93 Iowa 44–5 iPhone 94 Iraq 44, 105 Iraq War 44, 78 issue categories and labels 27, 30, 36 issue evolution 139 issue ownership 142 issue-definition 161 issue-framing see framing iTouch 94 Iyengar, Shanto 3, 35, 50, 185, 186 Johnson, Douglas 62 Jones, Bryan 161 Jones III, John E. 12 Kahneman, Daniel 4, 37 Kaiser Family Foundation 127 Kansas 12, 19 State Board of Education 16 Keenan, Nancy 71–2 Kellstedt, Paul 141–2 Kennedy, Edward 12, 88 Kennedy School of Government 127 Kids First Act 44 Kinder, Donald R. 43, 55 King, Steven 45 Kingdon, John 161 Kitzmiller, Tammy 12 Kitzmiller v. Dover Area School District 12, 17 Klansmen 101 Knight-Ridder Newspapers 11 Krauss, Lawrence 16 Ku Klux Klan 101–3 Kuklinski, James H. 133 Lattimer, Robert 16 Layman, Geoffrey 148 Leadership Management Model 46 Lexis-Nexis 141 Liberal values Life Sciences, Grade 10 Indicator 23 16–17

Linder, John 44 linkage frames see framing Lorge, Robert 88 Luntz, Frank 45, 123–5 Lupia, A. 126 McCain, John 78–9, 86, 122 McCaskill, Claire 86, 88 McConnell, Mitch 44 McCubbins, Matthew D. 126 McDermott, Jim 44 Maine 88 Marburger III, John H. 11 marketplace of ideas 3, 28, 30, 32–4, 103, 121, 137 Martin, Jim 123 Maryland 88–9 Massachusetts 88 Medicare 47, 84 Menendez, Robert 87 message framing see framing 42, 45 Meyer, Stephen 16 Michelman, Kate 66, 69 Miller, Kenneth 16, 22 Minneapolis Star-Tribune 62 Minnesota 86 Missouri 145 Mondale, Walter 86 moral values see values morality policy 64 Moran, Jim 71 Moyers, Bill 51 Murray, Charles 49 MyBarackObama.com 79 NAACP 85 NAF (National Abortion Federation) 62 NARAL (National Abortion Rights Action League) 5, 60, 62–3, 65–72, 102 National Center for Science Education 12 National Enquirer 102, 126 Nebraska 61 Nelson, Thomas E. 2, 43, 49, 54, 66, 78, 103, 121 Net Tone 168, 174, 177–8 Nevada 88 New Deal 139

Index 199 New Jersey 87 New York Times 5, 125–6, 137, 141–2, 162–3, 165, 168, 170, 177 New York Times Index 162 New Yorker 123 Newsweek 51, 136–7, 141–2, 144–5 non-excludable benefit 81 NPR (National Public Radio) 127 NRLC (National Right to Life Committee) 60–2, 65–6, 69–70, 72 National Right to Life News 60, 70 O.J. Simpson trial 171 Obama, Barack 1, 41, 78–9, 93–4, 122, 189–90 obligatory conservative 70 Of Pandas and People 12 Ohio 12–13, 15–16, 18–21, 27, 91, 92 2007 Senate Election in 79–80 General Assembly 15 State Board of Education 15–17 Ohio State University, The 30 O’Neill, Tip 4, 45–6, 51 open-mindedness see values Owens Fink, Deborah 17 Oxley, Zoe M. 2, 49, 66, 78, 103, 118 Packwood, Bob 136 Partial-birth Abortion Ban Act 60–1, 62–3, 65, 67–8, 71, 73–4, 123 party identification 137–8 Pennsylvania 12, 73, 87 People Like US 51 Petrocik, John 142 Pitts, Joe 73 Planning Authority Team 46 Poole, Keith 161 Presidency 46, 53 Princehouse, Patricia 16 public policy 160–1 qualification tactics 23 Reagan, Ronald 46, 49, 51, 145–6 administration 54 and abortion 136, 143 Reaganomics 50 religious sensibilities 19

Republicans 2, 13, 27, 41, 44, 46, 49, 51–2, 62, 86, 89–90, 104, 121–4, 127–30, 136–7, 142–3, 146, 148, 150–1 arguments and strategy 53, 122–3 in the House 45–7, 51, 53–4, 132 issue credibility 126 leaders 45, 49, 51 Presidents 144 Press Secretaries 51 public perception 49, 149–50 RPC (Republican Policy Committee) 51 strategies of 122 views on abortion 145 voter identification 149–50 rhetorical devices 18 Riker, William 63, 140, 161, 185 Roe v. Wade 68, 136 Rosenthal, Howard 161 RU-486 69 Ryan, George 105, 170 Santorum, Rick 12 Saunders, William 69, 72 Schaffner, James 3, 4, 27, 144, 187 Schapiro, Ian 122, 125, 131 Schattschneider, E.E. 140, 161 SCHIP (State Children’s Health Insurance Program) 41, 43–4, 49 Schlesinger, Alan 88 Schneider, Anne 161 Science Academic Content Standards 15–16 Science Excellence for all Ohioans 15–16 Science Model Curriculum Set A 17 Science Standards Writing Committee 16 scientific discovery 19, 28, 35 Seattle Times 123 Seattle, Washington 16 secular religion 21 Sellers, Patrick J. 4, 36, 74, 78 Senate 4, 44, 63, 72–3, 86 2007 campaigns for 80, 85–6, 91 Sessions, Pete 44 Shafer, Byron 139 Sniderman, Paul 53, 101, 103, 112, 187 social cueing tactics 23 Social Security 123

200 Index social-welfare issues 137, 139, 141, 147–8 Soldano, Pat 123, 125 State of Tennessee v. Scopes 23 Steele, Michael 88 Stenberg v. Carhart 61 Stimson, James A. 139 Stimson’s dimensional analysis algorithm 174 Stone, Deborah 161 Strategy Whip Organization 46 strong frames see framing Supreme Court 21, 61, 66, 162–3, 165 taxes 2–4, 27, 36, 45–6, 49–51, 64, 104–6, 109, 121–49, 151 Team Obama 79 Tennessee 102 Tester, John 90 Texas 44, 62, 170 Thagard, Paul 29 Theistic Evolutionists 22 thematic frames see framing Theriault, Sean 53, 101, 103, 112, 187 TiVo 80 Topinka, Judy 104–5, 109, 115 tragic and rare frame see framing Trojan Horse 3, 27–8, 30, 32–3 Tversky, Amos 4, 36 US House of Representatives 4, 41, 44–5, 47, 61–2, 68, 72–3 Utah 44 values 18, 37 liberal values 19 moral values 13 open-mindedness 12–13, 28, 32

value association 14 value conflict 23 value definition and delimitation 23 value diminishment 14 value disassociation 14 value enhancement 14 value expansion 23 value ranking 23 value recruitment 3, 13, 23, 66 VAT (Value Actions Team) 73 Ventura, Jesse 86 Virginia 71 Wagner, Michael 4, 36, 69 Wallace, Alfred Russell 11 Washington Post, The 49 Washington, DC 1, 35, 109, 122, 171 Washington, State of 44 Watergate 171 waves of criticism 161 waves of enthusiasm 161 Ways and Means Committee 123 Webster v. Reproductive Health Services 66 Weicker, Lowell 136 Wells, Jonathan 16–17 West Virginia 89 White House 46, 53, 54, 125, 136, 143, 146 Whitney, Richard 104–5, 109 Who Decides Campaign 66–7 Wikipedia 88 Winston, David 41 Wisconsin 88 Wright, James 54 Wright, Wendy 67–71, 73 Wyoming 90 YouTube 79, 84, 88, 91, 93–4