Temperament and Children: Profiles of Individual Differences [1st ed.] 9783030622077, 9783030622084

Citation preview

Roy P. Martin A. Michele Lease Helena R. Slobodskaya

Temperament and Children Profiles of Individual Differences

Temperament and Children

Roy P. Martin • A. Michele Lease Helena R. Slobodskaya

Temperament and Children Profiles of Individual Differences

Roy P. Martin Educational Psychology University of Georgia Athens, GA, USA

A. Michele Lease Educational Psychology University of Georgia Athens, GA, USA

Helena R. Slobodskaya Child Development and Individual Differences Institute of Physiology and Basic Medicine Novosibirsk, Russia

ISBN 978-3-030-62207-7    ISBN 978-3-030-62208-4 (eBook) https://doi.org/10.1007/978-3-030-62208-4 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

This book presents a new way to think about individual differences in the behavior of children. We argue that to be a successful parent, teacher, child psychologist, or pediatric medical professional, it is important to understand why one child tends to be shy when meeting new people while another is talkative, initiates conversation, and seems generally excited in this situation. But the tendency to be shy is only one of the behavioral tendencies that plays a central role in child development. Characteristics like the energy level of the child, their tendency to experience joy, the tendency to express anger in the face of frustration, as well as the tendencies to be argumentative, distractible, and disorganized play a role in how the child develops. These individual differences are viewed by contemporary researchers in child development as temperamental differences. This term is used to describe individual differences in behavior that can be observed in infancy and early childhood, which are substantially influenced by genetic mechanisms and have some degree of stability throughout the life of the individual. One goal of this book is to provide a review of some of the most influential thoughts about temperamental differences in children. However, the primary goal is to describe the development of a taxonomy of temperamental profile types. Psychologists have been systematically studying individual differences in children for more than 100 years. However, the vast majority of this research has been designed to identify and measure one or a few individual characteristics (e.g., intelligence, social inhibition, activity level) and to study the effects of these characteristics on the developmental progress of children. A problem occurs, however, in that these behavioral tendencies are considered individually. A child’s interaction with parents, teachers, and peers is determined not by any one of these behavioral tendencies, but by the pattern or profile of all these characteristics. Thus, the central feature of this book is the description of a system of commonly occurring temperament profiles. A further goal was to shed light on the educational, mental health, and social outcomes for children who exhibit each profile. The data reported in this book is primarily focused on children in the age range of 8 to 12 years. This age range was selected because it represents a period of intense social learning. Our research was based on the assumption that all this social v

vi

Preface

learning is heavily influenced by the child’s temperamental characteristics expressed as a profile. It is also based on the hypothesis that parents, teachers, and peers by middle childhood are beginning to come to a consensus on the individual difference characteristics of children. The following pages will show that by middle childhood, parents, teachers, and peers have a pretty good fix on what each child is like as a social being. It is the hope of the authors that the normally occurring (non-pathological) behavioral profiles described in the volume will help parents and teachers better understand the children in their care. It is also our hope that these profiles will be useful to child psychologists and other mental health professionals as they design their assessments, make diagnostic decisions, and help parents and others optimize their interactions with their children. Further, we hope that our approach will point researchers in education and psychology toward some of the ways that studying children with similar behavioral profiles augments other methods of studying individual differences in children. In order to meet our goals, the book can be thought of as two books in one, with the two parts designed to meet the needs of different types of users. The main body of the text is a summary of our research and is designed for those primarily interested in the descriptions of the profile types and their effects on parenting, teaching, and peer interactions. The text is written with limited discussion of statistical procedures and other issues in data analyses; no statistical training is assumed or necessary to follow the discussion. The text is augmented by an extended number of appendices designed for readers with a particular interest in the methods used to develop and validate the profile types. The appendices are provided for research professionals, mental health professionals, and students in child development who may wish to examine in a critical fashion the methodological procedures used. The appendices assume at least intermediate levels of statistical training. Athens, GA, USA Athens, GA, USA  Novosibirsk, Russia

Roy P. Martin A. Michele Lease Helena R. Slobodskaya

Contents

Part I Introduction 1 We Need a New Model of Normal Behavior Differences in Children������������������������������������������������������������������������������������������������    3 1.1 Introduction��������������������������������������������������������������������������������������    3 1.2 Confusion About Normal Behavior Differences������������������������������    4 1.3 Toward a New Model������������������������������������������������������������������������    7 1.4 Cultural Differences��������������������������������������������������������������������������    9 1.5 Purpose����������������������������������������������������������������������������������������������   11 References��������������������������������������������������������������������������������������������������   12 2 The Most Important Behavioral Traits of Children ����������������������������   13 2.1 What Are the Most Important Behavioral Traits?����������������������������   13 2.2 Descriptions of Fifteen Behavior Tendencies ����������������������������������   14 2.3 Combinations of Related Tendencies������������������������������������������������   16 2.4 Definitions: Behavior, Temperament, Personality, Types, and Taxonomies��������������������������������������������������������������������   18 2.5 Parent and Teacher Perceptions of Child Behavior��������������������������   21 References��������������������������������������������������������������������������������������������������   24 3 Development of Behavioral Profiles ������������������������������������������������������   27 3.1 Parent, Teacher, and Peer Samples����������������������������������������������������   27 3.2 Focus on Middle Childhood��������������������������������������������������������������   28 3.3 Development of Profiles��������������������������������������������������������������������   30 3.4 The Eight-Profile Model ������������������������������������������������������������������   31 References��������������������������������������������������������������������������������������������������   34 Part II The Seven Behavioral Profiles 4 Exceptionally Well-Adjusted High Achievers����������������������������������������   37 4.1 Introduction��������������������������������������������������������������������������������������   37 4.2 Parent-Teacher Perceptions of Behavioral Characteristics ��������������   37 4.3 Demographic Characteristics������������������������������������������������������������   39 vii

viii

Contents

4.4 Academic Ability and Motivation����������������������������������������������������   40 4.5 Prosocial Behaviors, Compliance, and Likeability��������������������������   41 4.6 Social Status��������������������������������������������������������������������������������������   44 4.7 Influence on Peers ����������������������������������������������������������������������������   44 4.8 Behavior Problems����������������������������������������������������������������������������   45 References��������������������������������������������������������������������������������������������������   49 5 Fostering the Development of Well-­Adjusted Children������������������������   51 5.1 General Principles����������������������������������������������������������������������������   51 5.2 Behavioral Assets������������������������������������������������������������������������������   52 5.3 Behavioral Risks ������������������������������������������������������������������������������   54 5.4 The Gifted-Creative Child Issue ������������������������������������������������������   58 5.5 A Personal Postscript������������������������������������������������������������������������   61 References��������������������������������������������������������������������������������������������������   62 6 Average Children: Two Temperamental Profiles����������������������������������   63 6.1 Introduction��������������������������������������������������������������������������������������   63 6.2 Parent/Teacher Perceptions of Behavioral Characteristics ��������������   64 6.3 Demographic Characteristics������������������������������������������������������������   67 6.4 Academic Ability and Motivation����������������������������������������������������   68 6.5 Prosocial Behavior, Compliance, and Likeability����������������������������   71 6.6 Social Status��������������������������������������������������������������������������������������   74 6.7 Influence on Peers ����������������������������������������������������������������������������   75 6.8 Behavior Problems����������������������������������������������������������������������������   77 References��������������������������������������������������������������������������������������������������   79 7 Fostering the Development of Average Children����������������������������������   81 7.1 Introduction��������������������������������������������������������������������������������������   81 7.2 Effects of Being Perceived as Average���������������������������������������������   82 7.3 Developmental Assets and Risks������������������������������������������������������   85 7.4 Interventions for Average Children��������������������������������������������������   87 References��������������������������������������������������������������������������������������������������   88 8 Shy and Socially Withdrawn Children: Two Temperament Profiles����������������������������������������������������������������������   89 8.1 Introduction��������������������������������������������������������������������������������������   89 8.2 Parent/Teacher Perception of Temperamental Characteristics����������   90 8.3 Demographic Characteristics������������������������������������������������������������   94 8.4 Academic Ability and Motivation����������������������������������������������������   95 8.5 Prosocial Behavior, Compliance, and Likeability����������������������������   97 8.6 Social Status��������������������������������������������������������������������������������������   99 8.7 Influence on Peers ����������������������������������������������������������������������������  101 8.8 Behavior Problems����������������������������������������������������������������������������  102 References��������������������������������������������������������������������������������������������������  106

Contents

ix

9 Fostering the Development of Shy, Withdrawn Children��������������������  107 9.1 Introduction��������������������������������������������������������������������������������������  107 9.2 Developmental Assets ����������������������������������������������������������������������  108 9.3 Developmental Risks������������������������������������������������������������������������  109 9.4 Differences Between Two Types of Withdrawn Children����������������  110 9.5 General Guidelines for Parenting/Teaching Withdrawn Children��������������������������������������������������������������������������  114 References��������������������������������������������������������������������������������������������������  116 10 Poorly Self-Regulated Children: Two Temperament Profiles��������������  119 10.1 Introduction������������������������������������������������������������������������������������  119 10.2 Different Profile Types��������������������������������������������������������������������  120 10.3 Demographic Characteristics����������������������������������������������������������  124 10.4 Academic Ability and Motivation��������������������������������������������������  124 10.5 Compliance ������������������������������������������������������������������������������������  127 10.6 Prosocial Behaviors and Likeability ����������������������������������������������  127 10.7 Social Status������������������������������������������������������������������������������������  129 10.8 Influence on Peers ��������������������������������������������������������������������������  130 10.9 Behavior Problems��������������������������������������������������������������������������  131 10.10 The Highly Achieving, Highly Emotional Profile��������������������������  134 References��������������������������������������������������������������������������������������������������  135 11 Fostering the Development of Poorly Self-­Regulated Children����������  137 11.1 Summary of Temperamental Characteristics����������������������������������  137 11.2 The Meaning of Poor Self-Regulation��������������������������������������������  138 11.3 Impulsive Extraverts ����������������������������������������������������������������������  140 11.4 Differences Between Profiles����������������������������������������������������������  141 11.5 Parental Influence on Self-Regulation��������������������������������������������  142 11.6 General Guidelines for Parenting and Teaching ����������������������������  144 References��������������������������������������������������������������������������������������������������  147 Part III Stability, Causes, and Implications for Diagnosis 12 Stability of Temperament Traits and Profiles����������������������������������������  151 12.1 Introduction������������������������������������������������������������������������������������  151 12.2 Long-Term Prediction��������������������������������������������������������������������  152 12.3 Mean Age Differences in Temperament-Related Behaviors����������  152 12.4 Stability of Individual Differences in Temperamental Traits����������  153 12.5 Stability of Profiles ������������������������������������������������������������������������  155 References��������������������������������������������������������������������������������������������������  157 13 Why Children Exhibit Different Behavioral Patterns��������������������������  159 13.1 Nature and Nurture: A Little History����������������������������������������������  159 13.2 Estimates of Genetic Effects: Behavior Genetics ��������������������������  160 13.3 Estimates of Genetic Effects: Molecular Genetic ��������������������������  161 13.4 Epigenesis ��������������������������������������������������������������������������������������  162

x

Contents

13.5 The Differential Susceptibility Hypothesis������������������������������������  163 13.6 Genetic and Environmental Effects Are Correlated������������������������  164 13.7 Other Considerations����������������������������������������������������������������������  166 13.8 Evolution and Individual Differences in Behavior ������������������������  166 13.9 Evolution and the Seven-Profile Model������������������������������������������  168 References��������������������������������������������������������������������������������������������������  170 14 Diagnostic Implications of Temperament-­Based Profiles��������������������  171 14.1 The Narrowing Conception of Normal ������������������������������������������  171 14.2 Increasing Rates of Diagnoses of Mental Disorders and Special Education Placement ��������������������������������������������������  172 14.3 Prescriptions for Psychoactive Medication������������������������������������  174 14.4 A Problem or Progress?������������������������������������������������������������������  174 14.5 Have We Lost the Concept of “Normal”?��������������������������������������  176 14.6 The Medical Model and Common Problems of Children: A Mismatch������������������������������������������������������������������������������������  177 14.7 The Seven-Profile Model and Problematic Behavior����������������������  179 References��������������������������������������������������������������������������������������������������  184 15 Limitations, Major Findings, and Implications������������������������������������  185 15.1 Limitations��������������������������������������������������������������������������������������  185 15.1.1 All Models Are Wrong to Some Extent����������������������������  185 15.1.2 More or Less than Seven Profiles May Prove Useful��������������������������������������������������������������������������������  186 15.1.3 Longitudinal Changes in Profile Stability Were Not Studied ����������������������������������������������������������������������  187 15.1.4 Latent Profile Analysis Currently Has Limitations����������  188 15.1.5 Models Are Only as Good as the Measurements Made ��������������������������������������������������������  188 15.2 Summary of Findings����������������������������������������������������������������������  189 15.3 Implications������������������������������������������������������������������������������������  190 15.3.1 Similarity in Profile Structure Across Raters and Cultural Environments Reinforces the Ideas that Temperament-Related Individual Differences Are a Natural Part of the Human Condition����������������������������  190 15.3.2 The Social Environment Stabilizes Behaviors������������������  191 15.3.3 Research on Individual Differences in Temperament-­Related Behavioral Tendencies Strongly Implies that Relying on One Parenting or Educational Approach Toward Behavior Management Will Be Ineffective if Not Harmful ������������  192

Contents

xi



15.3.4 Appropriate Clinical Diagnosis of Behavior Problems Requires an Understanding of the Normal (Non-­pathological) Behavioral Tendencies of the Child����������������������������������������������������������������������������  194 15.3.5 Allowing Children to Be Who They Are Is Likely to Have the Most Positive Effect on Their Developmental Path����������������������������������������������������������  194 References��������������������������������������������������������������������������������������������������  195 Appendix A: Study Participants ��������������������������������������������������������������������  197 Appendix B: Measurement Methods��������������������������������������������������������������  203 Appendix C: Statistical Issues������������������������������������������������������������������������  209 Appendix D: Profiles of Child Behavior: A Brief Review����������������������������  217 Appendix E: Results of the Latent Profile Analyses ������������������������������������  225 Appendix F: Academic Ability and Achievement Motivation����������������������  235 Appendix G: Compliance/Obedience ������������������������������������������������������������  241 Appendix H: Peer Perceptions of Positive Emotionality, Empathy, and Likeability��������������������������������������������������������������������������������������������������  245 Appendix I: Influence on Peers����������������������������������������������������������������������  249 Appendix J: Social Status��������������������������������������������������������������������������������  255 Appendix K: Parental Perceptions of Child Behavior Problems����������������  261 Appendix L: Peer and Teacher Perception of Aggressive Behavior������������  267 Appendix M: Child’s View of Their Behavior Problems������������������������������  273 References ��������������������������������������������������������������������������������������������������������  275 Index������������������������������������������������������������������������������������������������������������������  279

About the Authors

Roy P. Martin, PhD  is an emeritus professor at the University of Georgia, where he has served as a trainer of school psychologists in the Department of Educational Psychology. He has written widely on temperament and personality assessment, with particular application to educational settings. A. Michele Lease, PhD  is a professor of educational psychology at the University of Georgia, a licensed psychologist, and has trained school psychologists for over 20 years. Her primary areas of expertise are child psychopathology and social interaction patterns of children. In particular, she focuses on the characteristics of children who have acquired high social status within the peer social network and the ways in which they influence peers to conform to academic, behavioral, and social norms. Helena R. Slobodskaya, MD, PhD, DSc  is the principal research scientist at the Institute of Physiology and Basic Medicine and professor at the Novosibirsk State University, Russia. Her research focuses on child temperament and personality, and the association of these behavioral characteristics with child well-being and mental health problems in cross-cultural perspective.

xiii

List of Figures

Fig. 4.1 Comparison of cluster mean scores for Exceptionally Well-Adjusted High Achieving Children based on US and Russian parent, and teacher ratings (data presented in percentile form) ������������������������������������������������������ 38 Fig. 4.2 Peer perceptions of prosocial behaviors of the Well-Adjusted cluster (Data are in percentile form)����������������������������������������������������� 42 Fig. 4.3 Peer perceptions of four indicators of social status by Well-Adjusted children (data in percentiles form)�������������������������� 43 Fig. 4.4 Peer perceptions of five areas of influence by Well-Adjusted children (data in percentile form) �������������������������������������������������������� 45 Fig. 6.1 Two average behavioral profiles (in percentile form) �������������������������� 64 Fig. 6.2 Comparison of profiles of High Average Self-Regulators across three samples (data in percentile form)�������������������������������������� 65 Fig. 6.3 Comparison of profiles of Low Average Self-Regulators across three samples (data in percentile form)�������������������������������������� 66 Fig. 6.4 Grade point average of children in the High Average and Low Average Self-Regulator clusters compared to all other students in the Russian sample ������������������������������������������ 68 Fig. 6.5 Comparison of peer perceptions of academic ability and motivation for High Average and Low Average Self-Regulators (data are in percentile form)��������������������������������������� 69 Fig. 6.6 Comparison of self-rated academic ability and motivation of High Average and Low Average Self-Regulators by child gender ��������������� 70 Fig. 6.7 Comparison of compliance scores for High Average and Low Average Self-Regulators: parent, teacher, and self-ratings (data are in percentile form)��������������������������������������� 72 Fig. 6.8 Comparison of High Average and Low Average Self-Regulators by gender for the propensity to express positive emotion and empathy�������������������������������������������������������������� 73

xv

xvi

List of Figures

Fig. 6.9 Peer-rated influence of High Average and Low Average Self-Regulators in five areas of school-based behavior (Georgia Students, Cohort A, data in percentile form) ������������������������ 75 Fig. 6.10 Peer-rated influence of High Average and Low Average Self-Regulators in five areas of school-based behavior (Georgia Students, Cohort B, data in percentile form) ������������������������ 76 Fig. 6.11 Comparison of parent-rated behavior problems scores of High Average and Low Average Self-Regulators (Russian sample, data in Percentile Form) ������������������������������������������ 76 Fig. 8.1 Two types of socially withdrawn profiles: universal sample (data in percentile form)����������������������������������������������������������������������� 91 Fig. 8.2 Comparison of the Withdrawn High Achiever profile obtained from three independent samples (data in percentile form) ������������������ 93 Fig. 8.3 Comparison of the Withdrawn Low Achiever profile obtained from three samples (data in percentile form) ������������������������ 93 Fig. 8.4 Comparison of prosocial behavioral characteristics of Withdrawn High Achievers and Withdrawn Low Achievers (data in percentile form)����������������������������������������������������������������������� 98 Fig. 8.5 Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on four measures of social status (data in percentile form)��������������������������������������������� 100 Fig. 8.6 Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on five areas of social influence (data in percentile form)������������������������������������������������������ 102 Fig. 8.7 Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on four types of behavior problems as rated by Russian Parents���������������������������������������������������������������� 103 Fig. 8.8 Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on measures of social, verbal, and physical aggression as perceived by peers (data in percentile form) ��������������� 104 Fig. 10.1 Behavior profiles of two types of poorly self-regulated children: higher academic ability and lower academic ability (data in percentile form)��������������������������������������������������������������������� 121 Fig. 10.2 Comparison of the poorly self-regulated higher ability profiles obtained from US parents and teachers and Russian parents (data in percentile form)��������������������������������������������������������� 122 Fig. 10.3 Comparison of the poor self-regulated lower ability profile obtained from US parents and teachers (data in percentile form)������ 123 Fig. 10.4 Comparison of parental (US and Russian), teacher, peer, and self-ratings of academic ability and motivation for children exhibiting the poorly self-regulated higher ability profile (data in percentile form)���������������������������������������������� 126

List of Figures

xvii

Fig. 10.5 Comparison of parent/teacher, peer, and self-perceptions of academic ability and motivation for the poorly self-regulated lower ability cluster (data in percentile form)������������� 126 Fig. 10.6 Peer perceptions of prosocial behavior and likeability for the poorly self-regulated children in two student cohorts ������������ 128 Fig. 10.7 Comparison of social status scores of two clusters of poorly self-regulated children in two student cohorts ������������������� 130 Fig. 10.8 Comparison of social influence of two clusters of poorly self-regulated children in two student cohorts (data in percentile form)��������������������������������������������������������������������� 131 Fig. 10.9 Comparison of two clusters of poorly self-regulated children in two student cohorts on peer perceptions of three types of aggression (data in percentile form)������������������������������������� 133 Fig. 10.10 Comparison of US parent and US teacher profiles for the poorly regulated highly emotional cluster (data in percentile form)��������������������������������������������������������������������� 134 Fig. 13.1 Hypothetical outcome of aggression scores for highly sensitive (orchids) and less sensitive children (dandelions) and for children who have high and low genetic predispositions toward anger and are experiencing three levels of environmental support ��������������������������������������������������������� 164 Fig. E.1 Graph of change in information criteria for increasingly complex models: US parent ratings ��������������������������������������������������� 226 Fig. E.2 Graph of change in information criteria for increasingly complex models: Russian parent ratings��������������������������������������������� 229 Fig. E.3 Graph of change in information criteria for increasingly complex models: US teacher ratings��������������������������������������������������� 231 Fig. F.1 Grade point average (raw score form) by profile and child gender: Russian sample. Profile labels: 1a—Well-Adjusted High Achievers; 2a—High Average Self-Regulators; 2b—Low Average Self-Regulators; 3a—Withdrawn High Achievers; 3b—Withdrawn Lower Achiever; 4a—Poorly Regulated Higher Ability Under-Achievers; 4b—Poorly Self-Regulated Lower Achievers ������ 236

List of Tables

Table 2.1 Higher-order behavioral tendencies��������������������������������������������������� 18 Table 3.1 The seven most common behavioral profiles ������������������������������������ 32 Table 4.1 Percentage of children in the Exceptionally -Adjusted Higher Achiever cluster who are perceived to have significant behavior problemsa����������������������������������������������� 47 Table 6.1 Percentage of children in the two average clusters who are perceived to have significant behavior problems����������������� 78 Table 8.1 Percentage of children in the Withdrawn High Achiever and Withdrawn Low Achiever clusters who are perceived to have significant behavior problems ��������������������������������������������� 103 Table 8.2 Percentage of children in the Withdrawn High Achiever and Withdrawn Low Achiever clusters who have significant behavior problemsa based on peer and self-report��������������������������� 105 Table 10.1 Percentage of children in the two poorly self-regulated clusters who have significant behavior problemsa based on peer and self-report��������������������������������������������������������������������� 133 Table A.1 The number of children studied by sample ������������������������������������� 198 Table A.2 Comparison of characteristics of US parent-rated sample to the US population������������������������������������������������������������������������ 198 Table A.3 Geographic distribution of US parent-­rated samples: percentage of respondents by region������������������������������������������������ 199 Table A.4 Demographic characteristic of the two cohorts of Georgia students��������������������������������������������������������������������������� 200 Table A.5 Characteristics of the Russian parent-rated sample������������������������� 201

xix

xx

List of Tables

Table B.1 Correlations of scales chosen to create aggregate measures������������ 205 Table B.2 Reliabilitya,b of aggregate and single scales used as indicator variables for US parents, Russian parents, and US teachers ������������������������������������������������������������������������������� 205 Table B.4 Correlations among temperament-related indicators: US teachers��������������������������������������������������������������������������������������� 206 Table B.3 Correlations among temperament-related indicators: US parentsa��������������������������������������������������������������������������������������� 206 Table B.5 Correlations among temperament-related indicators: Russian parents��������������������������������������������������������������������������������� 207 Table E.1 Change in latent profile indices from three- to nine-cluster models: US parent samplea ������������������������������������������ 226 Table E.2 Indicator mean scores for the eight-cluster modela: US parent sample������������������������������������������������������������������������������ 227 Table E.3 Change in latent profile indices from threeto nine-cluster models: Russian parent-­rated samplea ��������������������� 228 Table E.4 Indicator mean scores for the eight-cluster model: Russian parent samplea��������������������������������������������������������������������� 229 Table E.5 Change in latent profile indices from threeto nine-cluster models: US teacher samplea ������������������������������������ 230 Table E.6 Indicator mean scores for the eight-cluster model: US teacher samplea��������������������������������������������������������������������������� 232 Table E.7 Fit results for the eight-cluster model: US parents, teachers, and Russian parent combined (universal) sample ������������ 233 Table E.8 Indicator mean scores for the eight-cluster model: universal samplea������������������������������������������������������������������������������ 233 Table E.9 Standard deviations of profile means for the eight-cluster model: all samples ��������������������������������������������������������������������������� 234 Table F.1 Correlation among peer and self-assessment measures of academic ability and achievement motivation (Cohort B) ���������� 235 Table F.2 Academic grade point average: significantly different homogeneous subsets by profiles ���������������������������������������������������� 236 Table F.3 Peer perception of academic ability: significantly different homogeneous subsets by profiles��������������������������������������� 237 Table F.4 Peer perception of achievement motivation: significantly different homogeneous subsets by profiles��������������������������������������� 238 Table F.5 Self-perception of academic ability (Cohort B): significantly different homogeneous subsets by profile ������������������ 239 Table F.6 Self-perception of academic motivation (Cohort B): significantly different homogeneous subsets by profiles������������������ 239 Table G.1 Russian parents’ ratings of compliance: significantly different homogeneous subsets by profile������������������������������������������������������ 242

List of Tables

xxi

Table G.2 US parental ratings of compliance: significantly different homogeneous subsets by profile ��������������������������������������� 243 Table G.3 US teacher ratings of compliance: significantly different homogeneous subsets by profile ��������������������������������������� 243 Table H.1 Peer perception of positive emotionality (Cohort A): significantly different homogeneous subsets by profiles������������������ 246 Table H.2 Peer perception of empathy: significantly different homogeneous subsets by profile������������������������������������������������������ 246 Table H.3 Peer perception of likeability: significantly different homogeneous subsets by profile������������������������������������������������������ 247 Table I.1 Table I.2 Table I.3 Table I.4 Table I.5

Table J.1 Table J.2 Table J.3 Table J.4 Table J.5

Peer perception of academic influence: significantly different homogeneous subsets by profile ��������������������������������������� 252 Peer perception of sports influence: significantly different homogeneous subsets of profiles��������������������������������������� 252 Peer perception of youth cultural trends: significantly different homogeneous subsets by profile ������������������ 253 Peer perception of influence on fantasy games: significantly different homogeneous subsets of behavioral profile ������������������������������������������������������������������������ 253 Peer perception of influence on inappropriate behavior of others: significantly different homogeneous subsets by profile������������������������������������������������������������������������������ 253 Peer perception of popularity: significantly different homogeneous subsets of profile���������������������������������������� 258 Peer perception of leadership: significantly different homogeneous subsets of profile���������������������������������������� 258 Peer perception (Cohort A) of admiration: significantly different homogeneous subsets of profile ������������������� 259 Peer perception (Cohort A) of social prominence: significantly different homogeneous subsets by profile ������������������ 259 Peer perception (Cohort B) of social prominence: significantly different homogeneous subsets by profile ������������������ 259

Table K.1 Russian parental ratings of hyperactivity: significantly different homogeneous subsets by profile ������������������ 263 Table K.2 Russian parental ratings of emotional problems: significantly different homogeneous subsets by profile ������������������ 264 Table K.3 Russian parental ratings of conduct problems: significantly different homogeneous subsets by profile ������������������ 264 Table K.4 Russian parental ratings of peer problems: significantly different homogeneous subsets by profile ������������������ 264 Table K.5 Russian parental ratings of all types of behavior problems: significantly different homogeneous subsets of profile ������������������� 265

xxii

List of Tables

Table K.6 Percentage of children exhibiting each profile type who have significant behavior problemsa ���������������������������������������� 265 Table L.1 Peer perception of physical aggression: significantly different homogeneous subsets by profile (Cohort A)��������������������� 270 Table L.2 Peer perception of physical aggression: significantly different homogeneous subsets by profile (Cohort B)��������������������� 271 Table L.3 Teacher perception of physical aggression (Cohort A): significantly different homogeneous subsets by profile ������������������ 271 Table L.4 Peer perception of verbal/social aggression: significantly different homogeneous subsets by profile (Cohort A)��������������������� 271 Table L.5 Peer perception of verbal/social aggression: significantly different homogeneous subsets by profile (Cohort B)������������������������������������������������������������������������ 272 Table L.6 Teacher perception of verbal/social aggression (Cohort A): significantly different homogeneous subsets by profile ������������������ 272

Part I

Introduction

The purpose of this book is to describe a set of seven behavioral profiles that can be used by parents, teachers, policy makers, and developmental researchers to better understand the social and emotional characteristics of children. Part I provides an overview of the research on which the seven-profile model of normal behavior was based. This part contains three chapters. Chapter 1 discusses the need for a new way to look at normal individual differences in children. Chapter 2 discusses the most important behavioral characteristics of children as viewed by parents and teacher. These behavioral traits, interpreted in the context of current scientific knowledge, served as the foundational descriptors of the individual differences that were measured in the 2300 children that were studied. Chapter 3 briefly describes the samples that were studied and the seven-profile model that was developed. The three chapters in Part I can be thought of as a brief summary of the research rationale and methods on which our model of child behavior is based. It was designed to be read as an overview, with details on many aspects of the research provided in appendices. Some readers may wish to initially skip this part, and begin with Part II which describes the seven behavioral profiles and their implications for parenting and teaching. However, as questions arise about how and why these profiles were developed, the material in Part I will prove useful.

Chapter 1

We Need a New Model of Normal Behavior Differences in Children …every snowflake is unique, and no category will do complete justice to everyone … But intelligence depends on lumping together things that share properties, so that we are not flabbergasted by every new thing we encounter. – (Steven Pinker, The Blank Slate, p. 203)

1.1  Introduction From the moment of conception, the infrastructure has been put in place that fosters individual differences in the behavior of children. Roughly 99% of the 21,0001 genes and 3 billion base pairs that constitute our heredity are shared with all other human beings [1]. A few of these genes, however, contribute to our differences. Some of these genes influence height, weight, eye color, and other somatic structural differences, but others influence brain structure and neurological processes. A range of environmental factors, some occurring in the uterine environment, influence genetic activity and neurological processes. These genetic and environmental influences create individual differences in behavior. These behavior differences are the subject of this book. Even before a baby is born, mothers of more than one child notice that some of their babies were more physically active than others. Some rotate their bodies, poke and kick their long-suffering mothers during the latter stages of pregnancy, while others seldom move. Within the first few months of life, parents began to observe further differences. As they observe their child and the children of friends and acquaintances, they realize that some babies are happy and content most of the time while others are fussy and difficult to console. Some are particularly responsive to the smiles of their parents while others are not. By the time the child enters the preschool years and progresses into middle childhood, parents may have noticed other differences in children. Some children adapt quickly to new environments; others take longer. Some may show signs of fear when meeting a stranger while others seem energized and excited. Some seem to spend most of their time running, jumping, and climbing while others sit quietly. Some children play with one object for a considerable period of time while others  Geneticists continue to update this figure. At this date, 21,000 is a good approximation.

1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_1

3

4

1  We Need a New Model of Normal Behavior Differences in Children

move from object to object, never engaging with anything for very long. Later in childhood, parents, teachers, and peers recognize differences in expressions of happiness, caring for the feelings of others, school achievement, intellectual interests, ability to plan ahead, social involvements, general fearfulness, argumentativeness, expressions of frustration, and desire to conform to societal expectations, just to mention some of the variation. Are these behavioral tendencies observed in childhood important? Are they predictive of future outcomes in adolescences and adulthood? Are some of these tendencies a sign of pathology? How stable are they? Until the 1960s, such questions were not given much research attention. Since that time, stimulated by the seminal New York Longitudinal Temperament Study of Alexander Thomas and Stella Chess, there has been an increasing focus on individual differences in behavior of young child. The most striking findings from the research is that differences in behavior, even those that are observed very early in life, have a relationship to important aspects of life experienced throughout childhood and adolescence, and even in adulthood. Individual differences in behavior observed at age 3, for example, has been linked to adult mental illness, number of years of educational attainment, number of adjudicated criminal offenses, marital discord and divorce, and income levels [2–8]. Does this research indicate that life patterns are set at birth? This is obviously not the case. Common observations of the lives of children as they mature, as well as decades of research, clearly show that our behavior is not hard wired. While associations between behavioral tendencies in childhood and adult behavior have often been found, they tend to be modest. But if we dig into the data a little deeper, some children are found to exhibit the same general behavior pattern over much of their childhood and adolescences, and some are more variable. For a group of 100 children, some children will be perceived to be quite consistent in their behavior pattern when assessed in the 1st grade and again in the 5th, but a few will show dramatic changes. This raises the issue of how basic behavioral tendencies might be sustained by the social environment that the child experiences and how these tendencies may be altered by that environment. It also raises questions about how genetic factors influence environments (more about this later), and how genetic factors may influence which children are most susceptible to environmental influences.

1.2  Confusion About Normal Behavior Differences The most important environmental circumstances experienced by young children (infancy through age 7) are determined by parents and by parental surrogates (day care workers, preschool and early elementary school teachers). They provide not only the physical and much of the social environment for the child, but they also create the learning opportunities the child experiences. As development progresses through middle childhood, the peer group plays an increasing role but parents and teachers continue to have influence throughout development.

1.2  Confusion About Normal Behavior Differences

5

How parents and teachers think about the behavior they observe determines to a great extent how they react to their children. Stated another way, how parents and teachers conceptualize what they observe influences the environment they create for the child. If a parent feels that their son is easy going and “sweet,” but their daughter is emotionally difficult, this perception alters their reactions to both children. Interactions with the easy-going son are comfortable and relaxed, while parental interactions with the daughter are made in anticipation of an emotional outburst with the resulting parental behavior having an undercurrent of anxiety or irritability. Further, how the parent (or teacher) thinks about the causes or reasons for these two types of behavior patterns affect their interactions with the child. If the mother believes that the emotionally difficult daughter is behaving in this way on purpose (i.e., is willful), then parental responses are more likely to be centered on punishment. If the father views his daughter as emotionally fragile, he may often give in to her unreasonable demands. The research community has made important progress in understanding individual differences in the behavior of children. However, much important information has not been communicated appropriately to the public at large. As a result, most adults who observe the behavior of children are confused about the variation they observe, particularly about the factors that influence this variation. Some are not confused; they simply have mistaken ideas about why children are so different. As some wise soul has said (maybe it was Yogi Berra), What you don’t know can be a problem, but often it is what you know for certain (that is mistaken) that causes the biggest problems. Part of the confusion occurs because many factors have been described by psychologists and other professionals as important influences on child behavior. A large number of genetic, physiological, and environmental factors have been shown to influence behavior. For example, some children are more irritable than others. They respond to the normal frustrations of life with expressions of negative emotion (crying, screaming, verbal aggression directed at others) more often than other children. There is evidence that this is a genetically linked tendency. Studies of twins and adopted children indicated that more than 50% of the variation among children in expression of anger is related to genetic variability [1]. Further, some specific genes have been found to be associated with the tendency to express anger such as the dopamine D4 receptor (DRD4) [9, 10]. However, irritability and expressions of anger can be modeled from parental responses to frustration (e.g., she observes her mother behaving this way), can result from harsh parenting practices (i.e., poor parental skill in teaching self-regulation of emotion), and can be linked to a persistent lack of sleep, physical health problems, feelings of rejection by peers, or other factors [11–13]. In any one case, it is difficult for parents to understand the factors that affect their child’s behavior. Further, all these factors may be related to each other in some complex way. For example, children who frequently express negative emotion have more social problems because their negative emotion makes them more difficult as social partners. They are more frequently avoided or rejected by peers. This response of peers fosters a view on the part of the child that the social world is a place filled with

6

1  We Need a New Model of Normal Behavior Differences in Children

rejection and hostility. This perception, in turn, decreases expressions of positive emotions and empathy for others. Further, the tendency to express negative emotion is related to other broad behavioral characteristics such as the ability to self-regulate attention. Because understanding all these influences and their interaction in still far from complete, there is a lack of consensus even among experts on the influence of genetic, physiological, and social influences on individual differences in behavior. As a result, parents, teachers, and policy makers are faced with differing opinions from experts some of which are contradictory. Another contributor to parental and teacher confusion about individual differences in child behavior can be attributed to poor communication to the general public from mental health professionals regarding how normal individual differences relate to diagnosed psychopathology. Psychological and psychiatric researchers have made major strides in the past 40 years in the understanding of pathological patterns of behavior in children. Behavioral characteristics of children who are having significant social, emotional, and educational problems have been described with increasing sophistication and the genetic and environmental factors associated with the behavior patterns are becoming better understood. Further, this understanding has been broadly transmitted to the public through mass media. Thus, many parents have at least a fuzzy notion of the characteristics of children diagnosed with learning disabilities, attention deficit hyperactivity disorder, autism spectrum disorders, bipolar disorder, and major depression. Diagnostic categories have been so widely disseminated and assimilated by the public that they have strongly influenced the ways in which parents and teachers think about and describe child behavior. Two of the authors have served as the director of a training clinic at The University of Georgia and have encountered parents who describe their child’s behavior in terms they had borrowed from their understanding of psychopathology. “My child seems normal, but he is awkward and has few friends. I worry he may be borderline autistic.” “My child is bright, but has ADHD-like behaviors; he just can’t focus.” “My child has a very dark emotional life; she has many worries and sometimes seems depressed.” The parents that came to the clinic would often see behavior of young children (one clinic focused on children ages 3 through 7) that was well within the normal range of individual differences as a sign of pathology. The psychopathology lens through which parents and teachers view child behavior has created confusion. In particular, it has often resulted in normal occurring individual differences becoming defined as abnormal. The training of psychologists suffers from the same bias. The training of the first author (RM) predominantly focused on abnormality including diagnostic criteria for psychiatric disorders as well as educational disabilities. These difficulties affect from 3% to 20% of the child and adolescent population depending on the severity of the disorders, the type of disability, and the method of diagnosis. This training did little to address the remaining 80–97% of the population. All this training in pathology did not help the understanding of normal individual differences in behavior. It also did not help understand the connection between normal variation and diagnosable conditions.

1.3  Toward a New Model

7

In addition to the complexity of understanding the factors that affect behavior, and the emphasis on psychopathology, an additional factor that creates confusion is the way that social and developmental science has approached individual differences in child behavior. Work in this field has been almost exclusively done in what has been referred to as the variable-centered approach. This approach to research typically focuses on one or a few behavioral traits (i.e., relatively stable patterns of behavior such as the tendency to be anxious). Then the relationship of these traits with some developmental outcome of interest is studied (e.g., grade point average, reading skills, symptoms of anxiety). There have been tens of thousands of such studies investigating hundreds of behavioral tendencies. Because so many behavioral traits have been studied and the relationship between traits is complex, it is difficult to summarize in a way that can be communicated to the general public. This research is even difficult for experts in the field to assimilate.

1.3  Toward a New Model Communication and scientific progress might be enhanced by a different way of thinking about individual differences in children. The approach is based on the logic that in order to understand child behavior, it is necessary to look at the full complement of cognitive, social, and emotional characteristics of each child. Some selection of attributes is necessary because there are many ways that children differ from one another. But in childhood, perhaps 20 characteristics have been most studied and have been found to affect development in profound ways. If measures were obtained on each of these characteristics, then each child could be described by a profile of scores. For example, Joseph might be high on the tendency to be shy, low on the tendency to be aggressive, average on compassion for others, high on motivation to do well in school, and above average on distractibility (ability to control his attentional focus). In other words, child characteristics would be summarized by a profile. This is often referred to as the child-centered or person-centered approach to developmental research. This approach, at first, sounds similar to what is the predominant method of studying child development (the variable-centered approach), and perhaps sounds more complicated. The child-centered approach begins with the delineation of the child’s location on a wide domain of trait scores. But there is a known characteristic of child behavior that leads to a surprising outcome. Many of the behavioral tendencies of children are moderately correlated with one another. For example, if we know that Joseph is shy, it is likely that he is less physically active than his same-age peers. Not all shy children are low on physically activity, but shy children have a high probability that this is the case. Such relationships among behavioral tendencies are complex when 10 or 20 characteristics are studied. But the surprising outcome is that in a large group of children, a small number of profiles or patterns account for most individual differences among those children. In large samples of children less than 10 profiles account for most of the variation in child behavior.

8

1  We Need a New Model of Normal Behavior Differences in Children

The investigation of behavioral profiles across a wide range of characteristics has been advocated in psychological research by a number of scholars [14–18], yet it is still seldom done. Many advantages of the approach have been described but two will be discussed here. The first advantage is a logical one. Optimal performance in any environment requires a number of characteristics and skills. For example, choosing the athletes that will most likely be successful on a high school basketball team might include measures of (a) physical strength, (b) height, (c) weight, (d) foot speed, (e) hand-eye coordination, (f) competitive desire, (g) responsiveness to coaching, (h) commitment to training, (i) past experience in sports, (j) ability to get along with teammates, and (k) academic ability. While a few of these characteristics might be considered most important, the best procedure to select athletes would involve consideration of the entire profile. As any coach will tell you, many gifted athletes have found themselves eliminated from the team because they could not meet academic requirement for participating in high school sports or could not get along with their teammates. The second advantage is that the child-centered approach is consistent with the way that human beings form concepts. What do human beings do when they are faced with the problem of understanding a new phenomenon that is important to their wellbeing? New parents and teachers are faced with this problem as they try to understand the varied ways that children respond to their social environment. Cognitive psychologists have observed that through the processes of induction, human beings assimilate their observations into categories based on similarities in the phenomena observed [19]. The process is much like that of the botanist who wants to try to understand the different types of trees in the forest. It is important to understand the types of trees because some flourish in one set of environmental conditions and others flourish in a different set of conditions. In a similar fashion, the parent with two or more children and the teacher with 25 students need to conceptualize individual differences in behavior so they can adjust their own behavior to fit the needs of the child. So how do parents or teachers come to understand the multifaceted displays of behavior they observe? Returning to the analogy of the botanist, the first task is to observe how the trees are different. Thus, the botanist might observe that trees vary in height, differ in the size and shape of their leaves, the bark of the tree varies in the degree of smoothness, some are evergreens, and some lose their leaves during the autumn season. In order to develop a meaningful and useful categorization of trees, the botanist must decide on the most important characteristics that differentiate one tree from another. The same process is necessary for parents. Eye color is one way that children differ, but it may not be the most critical for understanding their developmental progress. Emotional irritability, the speed with which the child learns to read, expressions of happiness and joy, attention span, and the empathy they show for others may be more important characteristics. Once the behavioral tendencies of the child are understood by a parent, their observations lead to the formation of a categorical summative description of these observations (e.g., a concept). As the eminent cognitive psychologist Steven Pinker has stated (see quotation at the beginning of this chapter), “intelligence depends on

1.4  Cultural Differences

9

lumping together things that share properties so that we are not flabbergasted by every new thing we encounter” [19]. Thus, human cognitive processes naturally lead to category formation. Parents create for themselves a category of child behavior based on the behavioral tendencies they observe. Because this process of induction and concept formation is so ubiquitous, parents and teachers almost always have a categorical model in mind when thinking about child behavior. We have already seen that parents may inappropriately use their understanding of psychiatric diagnostic categories to form their concept of why their child behaves as he does. Listening to parents and teachers for many years, I have often heard the following two concepts: “That child is spoiled.” “My child is lazy.” These categories are made from observations of behavior mixed with thoughts about the causes of the behavior. Describing Mary as “spoiled” indicates that the teacher believes certain behavioral tendencies (a profile) occur when children are given too many privileges. Describing Steven as “lazy” indicates that the teacher believes he does not do his homework and does not pay attention in class (as well as other tendencies) because he does not have motivation to achieve. These examples indicate that parents and teachers already have conceptual categories that they believe help them understand the behavior of children. However, many such categories are not helpful in fostering optimal development. In some cases, such categorical descriptors have been passed down by generations of parents and have come to be culturally based truisms. Such erroneous culturally shared conceptions of child behavior cause real harm. For example, they may simply be based on prejudicial thinking about a family, an ethnic/racial group, or gender. This brings us to one of the primary goals of this book. Based on recently developed statistical techniques, we have developed a model of behavioral profiles that are common across children in middle childhood. We believe that this model provides important advantages in capturing individual differences in children and will enhance understanding of commonly occurring types of behavior tendencies. The research on which this model was developed was based on the hypothesis that when parents and teachers are asked to describe the cognitive, social, and emotional behaviors of their children, the resulting measurements can be categorized based on a relatively small number of profiles. The emphasis of the research was on the most common profiles describing child behavior. The research was not directed toward relatively rare psychopathological behavior patterns. Based on prior research, we anticipated that we would find between 4 and 10 such profiles. We found seven. Most of this book is devoted to describing the characteristics of the children who exhibit these seven behavior patterns.

1.4  Cultural Differences We wondered if parents from different cultural groups would produce similar types of profiles. We read every day in professional and mass media about how socioeconomic differences in family, subcultural and ethnic differences even within the same

10

1  We Need a New Model of Normal Behavior Differences in Children

broad culture affect many aspects of family life and schooling. Thus, it was possible that the seven-cluster model that was found in ratings by US parents and teachers may be a cultural limited phenomenon. There is, however, reason to believe that analyses of parent ratings in various cultural settings will produce a similar set of behavioral profiles. Thousands of studies of temperamental and personality differences (these terms are described below) across cultures has supported the idea that the same broad traits are observed in all cultures and subcultural groups. There are shy children and socially outgoing children in every cultural setting, in every socioeconomic circumstance, in every ethnic or racial group. Likewise, there are children in every setting who have difficulty self-regulating their emotional expression in response to frustration. The culture may have somewhat different values about these behaviors and parents in these cultures may respond to the child differently, but the variation is present everywhere. In addition, in modern life, there are a relative common set of responses to some child behavior whether the child is born into a Brazilian, Chinese, French, Russian, or North American family. For example, some children find minor frustration particularly distressful and as a result exhibit more anger and general emotional upset than others. Parents, siblings, and peers who are in the presence of such distressed children typically find these emotional expressions upsetting, particularly if they are prolonged and intense. The behavioral variation in frustration tolerance (sometime referred to as irritability) will be noticed in almost all families, regardless of culture or socioeconomic circumstances, and will generally be experienced as aversive by parents and siblings. Another type of variability in child behavior is related to how quickly children learn. Children vary greatly in the ability to learn abstract concepts and in the speed of learning new skills. In the twenty-first century, the pace of social change has placed a heavy burden on the ability to learn many skills including the ability to utilize abstract symbols (e.g., reading, coding, and mathematics). These pressures have fostered an acute awareness of individual differences in the ability to learn abstract material (one form of intelligence), and the characteristic behaviors that foster learning (e.g., ability to control attention). Learning in school requires the ability to coexist with other learners, to cooperate with them, and to subordinate one’s desires to that of the teacher and the learning group. There is increased awareness by parents and teachers of individual differences in social intelligence and social skills that foster learning. Given that genetic predispositions in behavior vary and this variation is likely to occur in groups of children everywhere, and given that there are common pressures related to individual differences of children, we hypothesized that parents and teacher from different cultural settings will have observed a similar range of behavioral traits and that these traits will be correlated with one another in a similar fashion. This, in turn, leads to the hypothesis that the characteristics of behavioral profiles in different cultural settings will be similar. This hypothesis was generally supported in our research with regard to the two cultural groups studied. An eight-­ cluster model fit data well in all three of our samples (US parents, Russian Parents, and US teachers). However, when analyzed separately, the Russian parents

1.5 Purpose

11

produced one different profiled than was found for US parents and teachers. When the samples were combined, this difference disappeared (i.e., all eight profiles found in US children were again found), but one of the profiles defined a very small number of children. For this reason, seven of the profiles are described in detail in this book; these are the seven profiles that under the most rigorous analyses were found to replicate across all three samples.

1.5  Purpose The purpose of this book is to make some headway toward minimizing confusion about commonly occurring patterns of individual differences among children. Child behavior is complex. The concepts outlined in this book provide one way to help assimilate what is known about the variability of behavior among children. Thus, the first goal is to present a model of the most common behavior profiles of children. In an effort to help understand normal individual differences, our research is based on the notion that it is useful and scientifically justified to classify behavioral tendencies of children into a small number of profiles. These profiles are derived statistically from parental and teacher descriptions of children’s behavior. These profiles are described as latent, meaning that they were unknown to the researcher prior to doing the analysis and are unknown to the observers of the child’s behavior who are providing the measurements (parents, teachers). In other words, we do not ask parents to fit their child into an existing system of behavior profiles based on theory or expert opinion. The research is designed to find the behavior profiles that are implicit in the behavioral descriptions parents and teachers provide. As children mature from the preschool years through latter childhood, the peer group begins to have a greater and greater impact on their life. It becomes then important to determine if these profiles developed from adult perspectives are related to the ways that children think about each other and interactions they have with one another. From a scientific point of view, there is evidence that temperamental differences in children in early childhood are related to behaviors in adulthood including predictions of divorce or mental illness. What is less understood are the mechanisms through which early appearing behavior differences are maintained or weakened by the social group. Peer perceptions and interactions is one potential mechanism. In order to study peer interactions, the research we report here is based on children in the age range of 8 through 12  years of age. This age was chosen because this is that period of development in which peer interactions and relationships begin to take on a life of their own, outside the purview of adults, adult supervision, and adult constraints. This is an age range in which interactions with peers play a central role in the lives of children. It is also the age at which children can report on their attitudes and behaviors toward other children.

12

1  We Need a New Model of Normal Behavior Differences in Children

References 1. Plomin, R. (2018). Blueprint: How DNA makes us who we are. Cambridge, MA: MIT Press. 2. Caspi, A., Begg, D., Dickson, N., Harrington, H., Langley, J., Moffitt, T.  E., & Silva, P.  A. (1997). Personality differences predict health-risk behaviors in young adulthood: Evidence from a longitudinal study. Journal of Personality and Social Psychology, 73, 1052–1063. 3. Caspi, A., Henry, B., McGee, R. O., Moffitt, T. E., & Silva, P. A. (1995). Developmental origins of child and adolescent behavior problems: From age three to age fifteen. Child Development, 66, 55–68. 4. Caspi, A., Moffitt, T. E., Newman, D. L., & Silva, P. A. (1996). Behavioral observations at age 3 years predict adult psychiatric disorders. Archives of General Psychiatry, 53, 1033–1039. 5. Caspi, A., & Silva, P. A. (1995). Temperamental qualities at age 3 predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486–498. 6. Henry, B., Caspi, A., Moffitt, T. E., & Silva, P. A. (1996). Temperamental and familial predictors of violent and nonviolent criminal convictions: Age 3 to age 18. Developmental Psychology, 32, 614–623. 7. Chess, S., & Thomas, A. (1987). Origins and evolution of behavior disorders. Cambridge, MA: Harvard University Press. 8. Kagan, J. (1994). Galen’s prophecy. New York: Basic Books. 9. Saudino, K. J. (2005). Behavioral genetics and child temperament. Journal of Developmental and Behavioral Pediatrics, 26, 214–223. 10. Saudino, K.  J., & Wang, M. (2012). Quantitative and molecular genetic studies of temperament. In M.  Zentner & R.  L. Shiner (Eds.), Handbook of temperament (pp.  315–446). New York: Guilford. 11. Deater-Deckard, K., & Wang, Z. (2012). Anger and irritability. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 124–144). New York: Guilford. 12. Deater-Deckard, K., Beekman, C., Wang, Z., Kim, J., Petrill, S.  A., Thompson, L.  A., & DeThorne, L. (2010). Approach/positive anticipation, frustration/anger, and overt aggression in childhood. Journal of Personality, 78, 991–1010. 13. Dodge, K., Coie, J. D., & Lynam, D. (2006). Aggression and antisocial behavior in youth. In W. Damon, R. Lerner, & N. Eisenberg (Eds.), Handbook of child psychology (pp. 719–788). New York: Wiley. 14. Bergman, L. R. (2001). A person approach in research on adolescence: Some methodological challenges. Journal of Adolescent Research, 16, 28–53. 15. Bergman, L. R., & Magnusson, D. (1997). A person-oriented approach in research on developmental psychopathology. Development and Psychopathology, 9, 291–319. 16. Kagan, J. (1997). Conceptualizing psychopathology: The importance of developmental profiles. Development and Psychopathology, 9, 321–334. 17. Lanza, S., & Cooper, B. R. (2016). Latent class analysis for developmental research. Child Development Perspectives, 10, 59–64. 18. Magnusson, D. (1988). Individual development from an interactional perspective. Hillsdale, NJ: Erlbaum. 19. Pinker, S. (2000). The blank slate. New York: Viking.

Chapter 2

The Most Important Behavioral Traits of Children

2.1  What Are the Most Important Behavioral Traits? Our research was designed to measure the most important ways in which children differ in their behavior. Important behavioral tendencies were defined as those characteristics that are relatively stable over time and situation and which have the most impact on the child’s current adjustment. Stated another way, we wanted to find those characteristic behavior patterns that play a meaningful role in the way the child adapts to the family, school, and peer environment. One child can be differentiated from another on a wide range of characteristics. There are differences in observable aspects of their bodies (skin color, eye color, hair color, height, weight, etc.), in their abilities (scholastic, athletic, musical, etc.), their cultural/ethnic heritage, their socioeconomic status, or their gender identity. They also vary in their academic cognitive ability, emotional tendencies (e.g., tendencies toward anger or fear) and their typical social behaviors (e.g., tendencies toward leadership, shyness, friendliness, empathy). What behavioral characteristics from this long list are most important? We chose to base our research efforts on those characteristics that parents and teachers believe are most important. This approach is contrary to the approach taken by most researchers who rely on expert judgements (other researchers). However, what parents think are important characteristics of their children determine in large part how they will interact with their child, the learning opportunities they will provide, and the attempts they will make to change the child’s behavior. Thus, how parents conceptualize the most salient behaviors of their children matters. In understanding parental perceptions, we relied heavily on the research of Charles Halverson, Gedolph Kohnstamm, Ivan Mervielde, and their collaborators. They initiated a project some 30 years ago in which they collected simple parental descriptions of children. In most settings, parents were asked to provide a brief description of their child. The rationale for this approach was that human beings are

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_2

13

14

2  The Most Important Behavioral Traits of Children

thought to develop and use words and concepts that have an importance to them (sometimes referred to as the natural language hypothesis). For example, the native people of Alaska have many words describing different kinds of snow because the varieties of snow matter in their lives. The native peoples of the tropics have no need for such a precise differentiation of snow, so their language does not include these fine-grained distinctions. Parental descriptions, usually a few phrases or sentences, were collected in eight countries [1–3]. Approximately 40,000 descriptions were obtained. From the list of adjectives and phrases, groups of experts in child behavior sorted the descriptors hierarchically into categories based on similarity of meaning. Fifteen categories of behavior were determined to capture the vast majority of the adjectives and descriptive phrases used by parents (see Appendix B for details). One remarkable aspect of the descriptions provided by parents was that they seldom mentioned issues of stature (height, weight) or other physical characteristics. Also, health status or specific talents were seldom mentioned. The vast majority of the adjectives and descriptive phrases were related to behavioral characteristics of their child. Parents observe and remember most clearly those temperamental or personality characteristics (definitions provided below) of their children that have the most effect on their daily life in the family, at school, and with peers. From the 15 categories of cognitive, emotional, and social characteristics, Halverson and colleagues developed a scale designed to measure each of the 15 characteristic traits. This initial scale (referred to as The Inventory of Child Individual Differences— ICID) was then given to samples of parents in four countries (the United States, China, Greece, and the Netherlands) and further refined based on these data. Several different forms of this measure were developed, including a short form of the questionnaire (ICID-S [4]). Recently, Martin and Halverson have developed a second-­ generation instrument called the Survey of Individual Differences of Children and Adolescents (SIDCA), which was designed to be a further refinement of this set of instruments [5]. All of the different forms of the instrument measure the same 15 characteristics. The following is a description of each of the behavioral characteristics measured.

2.2  Descriptions of Fifteen Behavior Tendencies Intelligence  This scale is a measure of the speed of learning new skills and the ability to think through and solve problems. It is also a measure of the parent’s (or teacher’s) perception of the verbal ability of the child. Intelligence was included on these measures of behavioral traits because parents very often describe children in terms of their speed of learning new concepts and skills. Also, intelligence has been shown to have many associations with the temperamental and personality characteristics of children and is central to school adjustment. “Intelligence,” as measured here, should be conceptualized as the parent’s (or teacher’s) perception of their child’s ability to learn academic concepts and skills (e.g., reading).

2.2  Descriptions of Fifteen Behavior Tendencies

15

Openness to Experience/Curiosity  This scale is a measure of the child’s interest in the world. In particular, do they have an interest in many different things? By the time the child is in middle childhood, it is also a measure of openness to new ideas. Children with high scores are perceived to have many interests, and to enjoy the consideration of new ideas. Achievement Motivation  Children who score high on this scale are perceived to be motivated to improve their academic skills and knowledge, and to meet the demands of the culture for school achievement. Achievement motivation is believed to be an indication of the desire to improve one’s social status in the eyes of others as well as a measure of intrinsic motivation. By the time the child reaches middle childhood, parents and teachers focus intensely on academic motivation. Positive Emotionality  This scale was designed to be a measure of the extent to which the child expresses contentment, joy, and happiness. Children with high scores on this scale are easy to be with and often enhance the mood of those with whom they interact. Thus, children with high levels of this characteristic are typically perceived as friendly. Consideration-of-Others  Children with high scores on this scale are perceived to be helpful, thoughtful, and loving toward others. Scores are interpreted as a measure of altruism and empathy. Activity Level  This is a measure of the energy level and physical vigor of the child. Children who receive high scores on this scale often run rather than walk, like to play outside, like to participate in sports, and generally are engaged in high levels of gross motor movement. Negative Emotionality  This scale is a measure of irritability, or the tendency to become emotionally upset. Events that may cause this reaction are frustration when not obtaining a goal or desired outcome, as well as situations that are stressful and induce fear. Negative emotionality is often expressed as anger or verbal aggression (e.g., criticism) directed toward others, but also includes crying and other responses related to fear. Antagonism  This scale measures the tendency to engage in antisocial behaviors such as being rude, sneaky, aggressive, or disrespectful to adults. It is a measure of the tendency to break rules set by adults. Strong-Willed  Children who obtain a high score on this scale are perceived to be argumentative, stubborn, and generally want to impose their wants and attitudes on others. They want their own way and often seek dominance in social interactions. Distractibility  Children who obtain a high score on this scale have a low ability to self-regulate attention toward obtaining long-range goals. They generally have a short attention span and are more forgetful than their peers.

16

2  The Most Important Behavioral Traits of Children

Disorganized  This is a measure of the general tendency to be messy; that is, to have a low need for order of one’s belongings. It also is a measure of carelessness. Inhibition-to-the-Unfamiliar  Children with high scores on this scale have a tendency to avoid new situations and people. If they cannot avoid the new situation, they may look away or sit passively observing people they do not know well. They may warm up to the new situation, given time, and become progressively less apprehensive. However, they take longer to integrate into new social situations than their peers. Insecure/Fearful  This scale captures the tendency to feel insecure about one’s abilities or social standing. It also measures the tendency of the child to be generally fearful. Fearfulness and social insecurity are substantially related in middle childhood. Social Withdrawal  Children with high scores on this scale are often described as introverted. They have fewer friends than other children and may have difficulty making friends. Some of these children simply prefer being by themselves; they prefer the world of things versus the world of people. This scale measures a variety of reasons for social isolation and is a broader concept than inhibition-to-the-unfamiliar. Compliance  The compliance scale is a measure of the extent to which the child adapts quickly to adult social rules and expectations. Highly compliant children seem to have a desire to please others. We do not argue that the 15 traits measured by the instruments used in this research assess all the important individual differences of children. For example, they do not measure talents or abilities in the athletic, musical, or artistic realms. They do not measure physical attractiveness. Likewise, physical health issues are not included. All these individual differences as well as many others (e.g., physical stature, gender identification) can have an effect on the life of the child. The characteristics measured in our research are those behavioral tendencies considered most important by parents in many cultures and are the characteristics that are relevant to all children regardless of social class, ethnicity/race, or cultural background.

2.3  Combinations of Related Tendencies As we examined the results in large samples of child measurements completed by US parents and teachers, it was clear that some of the 15 traits were substantially related to one another; that is, they are correlated. This situation indicated some traits of the 15 scale could be combined resulting in a smaller set of measures of more general behavioral tendencies. There are statistical and measurement reasons to combine substantially correlated scales. For example, combining measurements

2.3  Combinations of Related Tendencies

17

that are highly correlated increases the reliability of the resulting measurement. But such combinations should be based on a theoretical foundation as well as statistical considerations. We have conceptualized the broad behavior tendencies measured in our research as temperamental traits. Temperamental traits are those personality characteristics that can be observed early in childhood. These traits are observed throughout childhood and adolescence. Further, these broad behavior tendencies are likely to generate more specific behavioral tendencies as the child develops, although these more specific tendencies will be moderately to highly related (correlated) to one another. Most temperament researchers have identified six broad traits: positive emotionality, prosocial behavior, negative emotionality/irritability, distractibility (ability to control attention), activity level, and inhibition to the unfamiliar (shyness) [6]. In addition to these six temperamental constructs, the measurements obtained included parental and teacher perceptions of the academic ability and motivation of the child. Historically, cognition and emotion (most temperamental characteristics are thought of as grounded in emotional responding) have been treated as two distinct sets of processes. In determining what is most important in adaptation over the life span, most theoreticians have given primary consideration to cognition. However, the close connection of affect and cognition has been addressed by Block. “Rather than awarding preeminence to cognition […] I believe that cognition […] should be viewed as only one way of responding to affective imbalances. Affect is not in the service of cognition; instead I look at cognition […] as a way […] of reacting and responding to affect (Block, 2002, p. xiii [7]). In addition to Block’s theoretical perspective, a number of empirical studies have shown that specific cognitive abilities in children are related to temperamental characteristics. For example, in one study, self-regulation of motor behavior and attention were found to have a strong relationship to expressive vocabulary development in a sample of preschoolers [8]. Influenced by Bloch’s theoretical position, by empirical research on the relationship of temperamental characteristic to cognitive skill development, and the importance placed on academic ability by parents and teachers, cognitive ability has been included as a trait that is central to our understanding of individual differences in children. Since some of the 15 scales we originally measured are substantially correlated, some scales were combined to form a broader measure. There are a number of statistical methods for creating these combined scales, but we chose to aggregate substantially correlated measures by summing the scale scores giving equal weight to each of the constituent scales. This resulted in six scales that were aligned with existing research on fundamental temperament traits. When the two cognitive scales were added, this resulted in eight broad-spectrum scales. Table  2.1 provides the name for these measures as well as the scales that were combined. (For more details on measurement issues, see Appendix B). The names given to the eight broad behavioral tendencies in Table 2.1 are brief descriptors and are, therefore, incomplete. For example, irritability/antagonism includes emotional responses related to frustration but also includes argumentative behavior as well as antisocial behaviors. The Poor Attention Regulation scale

18

2  The Most Important Behavioral Traits of Children

Table 2.1  Higher-order behavioral tendencies Higher-order behavioral tendency Academic Ability Achievement Motivation Prosocial Behavior Activity Level Irritability/Antagonism

Poor Attention Regulation Social Withdrawal/Shy Insecure/Fearful

Scales measuring these constructs Intelligence Openness-To-Experience Achievement Motivation Positive Emotionality Consideration-of-Others Activity Level Negative Emotionality Antagonism Strong-Willed Behavior Distractibility Disorganization Social Withdrawal Inhibition to the Unfamiliar Insecurity/Fearfulness

includes elements of distractibility as well as disorganization. These are lumped together because they are theoretically related and empirically linked in the parental and teacher ratings of children we studied. (Much will be said about the meaning of these broad behavioral tendencies in the subsequent chapters.) Achievement Motivation, Activity Level, and Insecurity/Fearfulness are not aggregated with other constructs because they did not correlate substantially with other measures. The Compliance scale was not included in the creation of behavioral profiles because it correlated highly with many other scales. (The role of compliance will be discussed in descriptions of each profile.)

2.4  D  efinitions: Behavior, Temperament, Personality, Types, and Taxonomies The behavioral traits measured in our research are often considered by psychologists to be temperamental or personality traits. Understanding what is meant by these terms requires a brief sojourn into the realm of personality psychology. Behavior can be described at a number of levels of abstraction from very broad tendencies to specific behaviors that occur in response to a specific type of environmental stimulation. Psychologists conceptualize behavior as hierarchically organized [9]. For example, negative emotionality describes a general tendency to be irritable, to respond to even low levels of stress with hostility, anger, or fear. Negative emotionality is considered a trait, because it is a generalized tendency and because it is relatively stable. The child exhibiting a high level of this trait cries and gets angry more often than others at home, at school, and at grandma’s house. Many

2.4  Definitions: Behavior, Temperament, Personality, Types, and Taxonomies

19

specific behaviors are indicators of this trait. For example, the child becomes emotionally upset when they are having difficulty learning a new skill, or becomes angry with playmates when a game does not work out well for them. The tendency to become upset with difficult learning tasks and becoming angry with playmates are related to one another (are correlated) in the sense that if a child tends to exhibit one of these sets of behavior at a high frequency, they tend to exhibit the other at a high frequency. In the example above, the trait of negative emotionality is inferred based on many different instances of related behaviors, and two examples are given. However, each of these two sets of behaviors is complex; each set is made up of many more specific behaviors. For example, we could assess the frequency in which the child becomes upset when learning a new cognitive skill (e.g., reading), or when learning a new physical skill (learning to ride a bicycle). Thus, behaviors can be progressively differentiated into finer gradations. This brief discussion leads to a few definitions that will be used throughout this book. Behavior is defined as a relatively specific motor, verbal, or emotional expression that can be observed in an environmental context (e.g., child resists when asked to go to bed). Specific behaviors are known to be affected by the specific environmental circumstances that the child is exposed to at the time they exhibit the response. A trait is a more general tendency to respond to differing environmental circumstances in a similar manner. More precisely, a personality trait is a characteristic pattern of thinking, feeling, or behaving that tends to be consistent over time and across relevant situations. Traits are an aggregate of correlated behavioral tendencies. An important difference between specific behaviors and higher-level traits is that traits are much more stable across situations and time than are specific behaviors. They are defined and measured in a manner to increase their stability. Thus, the number of phone calls you make to friends on one day is an indication of a tendency to be social, but daily phone calls are highly variable. They are affected by the circumstances surrounding your day. Adding up the number of phone calls made to friends over a month, the number of text messages sent, and the number of times during the day an individual is observed talking to others for a month is a much more stable measure of the tendency to be social. Better trait measurement of children occurs when parents and teachers provide ratings of many specific behaviors, as observed in multiple settings, and complete ratings many times over months or years. In the pages that follow, some traits that will be discussed are labeled as temperamental traits and some as personality traits. While there is some disagreement among experts about how these concepts should be differentiated, there is general agreement that temperamental traits are a subclass of personality traits. Specially, they are those traits that can be readily observed in very young children (toddlers and infants). Some can also be observed in higher animals (e.g., chimpanzees, dogs) [10–12]. Social inhibition in the presence of unfamiliar others (strangers) is one example. This trait is commonly referred to as shyness. Social inhibition can be observed within the first year of life and is easily observed as the child reaches the

20

2  The Most Important Behavioral Traits of Children

preschool period. Temperamental traits are not only observed in early childhood, they can be observed throughout the life span. Temperamental traits have been shown to have a genetic origin, but they are influenced by a variety of environmental factors occurring from the time of conception (e.g., they can be affected by the inter-­ uterine environment) and throughout life. For example, with sensitive and gradual exposure, children who are afraid of dogs can come to enjoy interactions with dogs and other animals. In addition, with similar gradual exposure to feared social interactions (e.g., fear of strangers) children can become progressively less fearful in a broad range of social interactions. All these cumulative experiences can lessen the initial fearfulness of the child. Other traits are not observable in the first years of life. These include traits like Openness to New Ideas and Experiences, Achievement Motivation, and Disorganization. These traits are labeled as personality traits, indicating that the behaviors are observable later in development (some by age 4 or 5, others somewhat later). Personality traits are thought by most experts to develop from temperament traits as they interact with environmental circumstances. That is, temperamental traits differentiate with time, experience, and the increasing capabilities of the person. The terms “temperament” and “personality” are used in our discussion, because there is a very long history in psychology for both terms. However, the distinction between temperament and personality is of little practical importance other than at what stage in development the behaviors can be observed. Both temperament and personality are influenced by genetic and environmental factors. By middle childhood, both classic temperament traits and personality traits are clearly seen. All of the traits measured by the eight broad scales described in Table 2.1, except one (Achievement Motivation), contain a behavioral characteristic that can be observed in early childhood. Even aspects of cognitive ability are observable very early in development. Thus, the traits that are the focus of our research are conceptualized as temperament-based. However, because of theoretical issues regarding temperamental and personality characteristics, we often will sometimes describe these traits as behavioral traits or behavioral tendencies. These terms do not contain some of the surplice meaning that is often placed on the words “temperamental” or “personality.” In the minds of some, these terms imply a highly stable tendency to behave that have a strong genetic basis. We take a more nuanced position. Yes, there is considerable stability in the behavioral traits of children 1 (see Chap. 12), and there is a clear genetic influence on these behavioral traits (both temperamental and personality traits, see Chap. 13). But stability and genetic influences on behavior depend on the behavioral characteristic being considered. Further, both the stability and heritability of behaviors can be influenced by a variety of environmental factors. While relying heavily on temperament research for the theoretical foundation for the traits we have measured, we will often describe the measured traits simply as behavioral traits. Temperament-based traits have been used in our research to create behavior profiles. Based on statistical modeling, the data in this book demonstrate that it is useful to think of a small number of profiles of child behavior traits. A behavior type is defined as a profile of scores across traits. The traits are said to be indicators of the

2.5  Parent and Teacher Perceptions of Child Behavior

21

type, because the type is created using scores on measures of these traits. For example, one type we have found is a pattern in which the child receives above-average scores on Academic Ability, Academic Motivation, Prosocial Behavior, Social-­ Withdrawal, and Insecurity/Fearfulness. This group of children has low scores on Activity Level, Attention Problems, and Irritability/Antagonism. This behavior profile is characteristic of a cluster of children (the group of children exhibiting the profile) or a type (when referring to the profile). A cluster describes a group of children who are relatively homogeneous in their pattern of scores across all traits measured. Finally, the word taxonomy is used to indicate a group of behavior types or profiles. The word also has the connotation that the behavior patterns that are described cover the full range of observed behavior patterns in a population. There are two caveats. Our taxonomy does not consider very low frequency rare profiles of behavior, many of which would be considered as indicative of psychopathology. Also, the taxonomy developed is dependent on the traits used as indicators. If, for example, physical attractiveness was added to the set of indicators under investigation the taxonomy would be altered.

2.5  Parent and Teacher Perceptions of Child Behavior Most research on behavioral traits in adolescents and adults is based on self-­ perceptions. In this kind of measurement, the individual describes their own behavioral tendencies. However, young children do not have the ability to view their own behavior as objectively as parents and teachers. Further, some children simply have not thought about how they behave in different environments. Many do not have the language ability to communicate these differences even if they are aware of them. Thus, the measurement of behavioral characteristics in children relies primarily on observations by others (e.g., parents and teachers). Some researchers also utilize laboratory observations in specific experimental circumstances. Each measurement method has advantages and disadvantages. When trying to determine the behavior tendencies of individual children, parents have some advantages over other observers. The primary advantage is that they have watched their child grow and develop from the time of birth. This long history cannot be duplicated by teachers, field researchers (e.g., researchers who go into homes or schools to observe child behavior), or laboratory researchers. A second advantage is that parents have watched their child behave in a number of situations, a much wider range than teachers or researchers. Thus, they have seen how the child responds to strangers, to each parent individually, to extended family members, and to siblings and friends. They have also observed the child respond to a range of disappointments and situations, such as when punished for a transgression. This range of settings and circumstances across time is unique to parents. Third, most children tend to express their emotions more fully in the safe confines of the home than in public. Some children, for example, are viewed by their teachers as

22

2  The Most Important Behavioral Traits of Children

exhibiting strong self-regulation of emotion in the classroom. But the parent may observe more intense emotional expressions because the family is perceived by the child as an environment in which displays of strong emotion is tolerated or even encouraged. The first author (RM) had a vivid experience like this when his younger son was 5 years old. When he picked him up after school one day, his son was very irritable and argumentative. No matter what was said his son would disagree. After a while, RM asked if something was bothering him. After a pregnant pause, his son said, “I have been good all day, and I can’t be good anymore.” The perceptions of parents regarding their child’s behavior have been shown to play a significant role in how they interact with the child. Parental behaviors based on those perceptions in turn become the most important environmental experience for that child from birth to middle childhood. Thus, perceptions have a direct effect on the quality of the interactions between parents and children. Further, parental reactions begin to be internalized by the child, and thus, play an important role in how the child views themselves. Thus, parental perception of their children is an important determinant of the child’s view of the social world and of their place in that world. However, parental perceptions of their children are influenced by several factors that limit their reliability and validity. First, parents have limited information on what is typical behavior for a child at any age. Measurement of anything (e.g., temperature, shoe size) involves comparisons to some standard. In the realm of child behavior, the comparison is most often to how the typical child of a given age behave. Parents often do not have access to this kind of information, these developmental norms. They simply do not observe a wide range of children, so they must rely on behavior of siblings (in families with multiple children), on the views of other adults in their social world (extended family, friends), and on the views of teachers for this information. Some parents read books on child development or seek the advice of medical or psychological professionals. However, it remains difficult for a parent to say whether the behavior their child exhibits is typical or not. Further, the personality of the parent affects how they view the behaviors of their child. Quiet, introspective parents will view the behavior of the typical 8-year-old as more troubling than more active, extroverted parents will. The typical 8-year-old is active, loud, messy, and somewhat disorganized as viewed from adult standards— more active, loud parents will not find this behavior so aversive. Parents who are experiencing various forms of stress (e.g., economic pressures, very demanding work environments, mental health problems) find common behaviors of children more distressing than parents in less stressful circumstances. Also, parents who are physically or mentally ill (e.g., depressed) tend to view their children’s behaviors as more extreme [13]. Our approach was to study the perceptions of large numbers of parents. With large numbers (500–1000 in each sample), many of these biasing effects are controlled or at least minimized. The assumption of such research is that if the sample is large, infrequent individual circumstances of families or individual parental personalities will not strongly affect the average perception of all families and thus the

2.5  Parent and Teacher Perceptions of Child Behavior

23

average measurement is a good representation of the population of parents and children. Whereas parental report has many advantages as a source of objective measurement, teachers have a unique perspective on child behavior due to their professional role. The primary advantage of behavioral measurements obtained from teachers is that they have in front of them every day a group of children of similar age; thus, they have a comparison group against which to judge the behavior of any individual child. Teachers who have been in the profession for more than a few years have observed the behavior patterns of hundreds of children. This opportunity places the observations made by teachers in a privileged and pivotal position in understanding individual differences among children. Teachers also have the advantage of seeing the behavior of children in a context where the child must perform tasks that are often difficult. This includes academic tasks as well as social interactions with peers that require complex social skills. Temperamental characteristics are often most easily observed in situations where the child must cope with new challenges and stressful circumstances. The challenges faced by children in classrooms include the necessity to attend to the instruction of the teacher, to keep ones belongings organized, perform in a public setting (e.g., answer questions posed by the teacher in front of peers), to meet demands for rapid changes in activities, to respond to expectations and social pressures from peers, and to manage their anxiety about test performance. In these situations, how the child responds provides an index of their typical coping mechanisms and emotional responses. Further, many temperamental and personality characteristics involve aspects of the social responses of children. Teachers have the opportunity to observe child behavior in the context of interactions with their peer group. Thus, the child’s responses to peers who exhibit different types of behavior can be observed. Finally, unlike trained observers viewing the child for the first time (in a research context), teachers have the opportunity to view child behavior daily for months (typically 9–10 months per school year). This allows the teacher to see changes in behavior over time as well as observe consistent patterns. Observations of teachers, like those of parents, are subject to a variety of factors that limit their reliability and validity. Teachers differ in their personalities. These differences are likely to have effects on their perceptions of their student’s behavior. Being responsible for the behavior of 15–30 children in one room places specific demands on teachers, which effect how tolerant they are of some types of child behavior. Highly active or aggressive behaviors do not fit this environment well, so even small transgressions are attended to and affect the teacher’s overall impression of the child. Finally, teachers do not observe the behavior of children in as wide a range of environments as parents. The discussion of the different opportunities of parents and teachers to observe child behavior indicates that a high level of agreement between a given parent and teacher regarding an individual child is unlikely. Indeed, years of research has indicated that parent-teacher agreement is low when individual traits or levels of behavior problems are measured [14–17].

24

2  The Most Important Behavioral Traits of Children

In the current analysis, we eliminated one source of potential disagreement. Ratings of child behavior by teachers were centered at the mean of all children in his/her class (i.e., the average score of all children studied was set at 00.0). Thus, in each classroom, a child’s score on any characteristic was calculated in reference to the average in that class. This helped control for teacher biasing effect. The average score on each measure obtained from parents in different samples was also set to the mean of that sample. The average score in the US parent sample for each trait was set to zero as was the average score in the Russian parent sample. Thus, our research did not consider the question of whether Russian parents thought their children were more active than did US parents. Our research was focused on individual differences—not average cultural differences. This procedure was justified because our focus was on how the behavioral characteristics of children differ from the average child within that environment as perceived by a specific type of observer.

References 1. Halverson, C. F., Havill, V. L., Deal, J., Baker, S. R., Victor, J. B., Pavlopoulos, V., Besevegis, E., & Wen, L. (2003). Personality structure as derived from parental ratings of free descriptions of children: The inventory of child individual differences. Journal of Personality, 71, 995–1026. 2. Havill, V.  L., Allen, K., Halverson, C.  F., & Kohnstamm, G.  A. (1994). Parents’ use of Big Five categories in their natural language descriptions of children. In C.  F. Halverson, G. A. Kohnstamm, & R. P. Martin (Eds.), The developing structure of temperament and personality from infancy to adulthood (pp. 371–386). Hillsdale, NJ: Erlbaum. 3. Kohnstamm, G.  A., Halverson, C.  F., Jr., Mervielde, I., & Havill, V.  L. (1998). Parental descriptions of child personality: Developmental antecedents of the Big Five? Mahway, NJ: Erlbaum. 4. Deal, J. E., Halverson, C. F., Martin, R. P., Victor, J., & Baker, S. (2007). The inventory of children’s individual differences: Development and validation of a short version. Journal of Personality Assessment, 89, 162–166. 5. Martin, R. P., & Halverson, C. F., Jr. (2015). Survey of individual differences of children and adolescents. Athens, GA: Author. 6. Zentner, M., & Shiner, R. L. (2012). Handbook of temperament. (Part II: Basic temperament traits) (pp. 69–168). New York: Guilford. 7. Block, J. (2002). Personality as an affect-processing system: Toward an integrative theory. Hillsdale, NJ: Erlbaum. 8. Bohlmann, N. L., Maier, M. F., & Palacios, N. (2015). Bidirectionality in self-regulation andexpressive vocabulary: Comparison between monolingual and dual language learners inpreschool. Child Development, 86, 1094–1111. 9. Tachett, J.  L., Slobodskaya, H.  R., Mar, R.  A., Deal, J., Halverson, C.  F., Jr., Baker, S.  R., Pavlopoulos, V., & Besevegis, E. (2012). The hierarchical structure of childhood personality in five countries: Continuity from early childhood to early adolescence. Journal of Personality, 80, 847–879. 10. Barr, C. S. (2012). Temperament in animals. In M. Zentner & R. Shiner (Eds.), Handbook of temperament (pp. 251–272). New York: Guilford. 11. Belyaev, D. K. (1969). Domestication of animals. Science Journal (UK), 5, 47–52. 12. Diamond, S. (1957). Personality and temperament. New York: Harper.

References

25

13. Richters, E. (1992). Depressed mothers as informants about their children. A critical review of the evidence for distortion. Psychological Bulletin, 112, 485–499. 14. Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). Child/adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychological Bulletin, 101, 213–232. 15. Lee, S. W., Elliot, J., & Barbour, J. D. (1994). A comparison of cross-informant behavior ratings in school-based diagnosis. Behavior Disorders, 19, 87–97. 16. Los Reyes, A., & Kazdin, A. E. (2005). Informant discrepancies in the assessment of childhood psychopathology: A critical review, theoretical framework, and recommendations for further study. Psychological Bulletin, 131, 483–509. 17. Major, S. O., Seabra-Santos, M. J., & Martin, R. P. (2015). Are we talking about the same child? Parent-teacher ratings of preschoolers’ social-emotional behaviors. Psychology in the Schools, 52, 789–799.

Chapter 3

Development of Behavioral Profiles

3.1  Parent, Teacher, and Peer Samples The research reported in this volume is based primarily on the analysis of information obtained from three different samples of research participants. Two separate groups of children from the United States were studied. One group was a sample of children whose behavior characteristics were described by their parents. A nationally representative sample of parents was selected from all regions of the United States who had children in the age range of 3–18 years of age. Parents responded to the Survey of Individual Differences of Children and Adolescence (SIDCA) administered via computer. From this sample, 1150 parental descriptions were selected of children who were in middle childhood (ages 8 through 12 years). (See Appendix A for details.) A sample of children ages 8 through 12 had been recently studied by Michele Lease and her students at the University of Georgia. As part of this research, teachers responded to a brief form of the Inventory of Child Individual Differences (ICID), which assessed all 15 of the same characteristics that were measured in the US parent sample. Data were available on 912 students in the third through fifth grades in Georgia, and only children ages 8 through 12 were included in the analysis. (See Appendix A for details.) Lease and colleagues also collected extensive information from these students regarding their perceptions of their own behavior, and their attitudes about other students in their classroom. This information allowed for the comparison of behavior profiles developed from parents and teachers of children in the same age range, and also allowed for the determination of how these parent and teacher profiles related to the self-perceptions of children and to peers’ perceptions of them. Finally, data had been collected in Russia on parental descriptions of child behavior using the ICID, a form of the same measure used the collect data from teachers in the United States. The research team headed by Helena Slobodskaya

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_3

27

28

3  Development of Behavioral Profiles

graciously joined our project and provided substantial information on the academic performance and behavior problems of these students. From their larger sample, 538 students were selected who were in the same age range (8 through 12) as the US parent and teacher samples. (See Appendix A for details.) Having data describing the same behavioral characteristics from these three large samples allowed us to determine if similar profiles would be obtained from parents and teachers in one culture as well as from parents across different cultural settings. Because data had been collected on the characteristics of the parents and teachers, this allowed for analysis of how the behavior profiles that were created were associated with these characteristics (e.g., comparison of profiles of mothers and fathers). Finally, the associations of profiles of behavioral tendencies as perceived by parents and teachers to behaviors observed by peers could be assessed.

3.2  Focus on Middle Childhood Five major categories of childhood can be defined as follows: infancy (birth to 1 year), toddlerhood (1–2 years), early childhood (3 through 7 years), middle childhood (8 through 12 years), and adolescence (13 through 19 years). These typical categories used in child development research are somewhat arbitrary. However, they are useful in describing many aspects of development at different ages. One such consideration is the delineation of the social influences that are associated with each age. Human beings are social beings. Thus, clarifying the major social influences on child behavior at any age can be useful. In infancy, the child is almost exclusively under the protection of parents, or immediate family members. In toddlerhood, children typically have a somewhat widened interaction with others, including extended family, and for some child care workers and other children. By early childhood, the principal influences are still family members but very often teachers or other trained child care workers become involved. There is also increasing involvement with same-age peers. Movement into middle childhood represents a major transition in the social world of the child for two central reasons. First, this is the period in the child’s life that one’s peer group begins to become a dominant force. Through interactions with the peer group, the child begins to understand how they are perceived by peers. In turn, this influences the child’s self-perception in a way that does not occur at earlier ages. These processes occur at this age because children have developed the cognitive ability to observe and conceptualize how they fit into the world of their peers. They begin to have an appreciation for their place in some general status hierarchy, as well as their relative abilities in academics, sports, and other activities. Their friendships grow in complexity across this developmental period and, as they enter late childhood, they increasingly explore cross-gender interactions and budding romantic relationships.

3.2  Focus on Middle Childhood

29

Why is it that we remember the names and the behavioral characteristics of many of the children we knew in middle childhood, but not one of our playmates or classmates in kindergarten, first or second grade? For most of us, these middle childhood friends were our first real friends. Coping with this new social world also required a lot of time and emotional energy. For example, I (RM) remember Mary Jane. She was bright, seemed to always have the right answer when the teacher called on her, played the violin in a performance for the entire school with no apparent anxiety, and was friendly to almost everyone. I liked Mary Jane, but was too shy to approach her. Joe was great at basketball, but was very competitive and often blamed others when his team did not win. I played with him, but it was not always fun when he got angry. I remember specific boys who I thought were dangerous to be around, others who were very smart but not very friendly, and others who were just average in almost every way. I still remember these children and their names, because at this age, these social relationships were crucial. They left a lasting impression and began the process of learning who I was. The second major transition for most children at about age 8 is that the adult world now expects performances that meet cultural expectations. Families impose more task on the child (household chores) and teachers make increasing demands for organizing materials, paying attention to instruction, planning, and executing larger projects. Continuing the personal reflections of the first author, I learned from the feedback from teachers that I was an above average student, but my performance was hampered by “excessive daydreaming,” and a lack of organization. (These were comments made on several report cards over the years.) I was also under-­performing in the area of reading particularly in the third and fourth grades. I did not really learn to read anything until the beginning of third grade. I remember really concentrating on these skills for the first time at the end of second grade because all my friends were reading comic books and I wanted to be able to do this also. I can remember how I performed academically in third and fourth grades, but not in the first and second grades. This difference illustrates the point that my social world (parents, teachers, peers) was beginning to communicate specific expectations much more directly at this age than previously. As one index of this transition and its demands, most initial referrals to school psychologists by parents and teachers in the United States occur between the ages of 7 and 9 years old. Such referrals represent concerns on the part of parents and teachers that the child is not meeting social, emotional, or educational expectations. This is another manifestation of the increased cultural demands placed on children at this age in the realms of academic performance as well as social and emotional functioning. For all these reasons, individual differences in the behavior of children become particularly apparent in middle childhood. They become apparent to parents, teachers, the peer group, and to the children themselves. Further, those behavioral tendencies that were observed in early stages of development have become more stable. Is my child’s interest in books at age 3 really a sign of intelligence and academic potential, or is this an isolated transient behavior? Is my child’s fearfulness and

30

3  Development of Behavioral Profiles

social inhibition something they will “get over,” or is this the way they are like to interact with the world in the future? The answers to these questions begin to be clarified in middle childhood to a greater extent than at early points in development. Specific to the issues of interest in our research, this is also a period in which most children have the ability to respond to questions about themselves and about others in their environment. This makes it possible to determine the effects of child behavior profiles on how children view themselves and how they view their peers. This is the first time in a child’s life that these kinds of questions can be reliably addressed.

3.3  Development of Profiles Based on parent and teacher responses, children were given a score for each of eight behavioral characteristics (the six temperamental characteristic and two cognitive characteristics described in Chap. 2) measured in all three samples. Children varied widely in their score on each characteristic. All behavior traits measured were very nearly normally distributed; that is, a few children had very high or very low scores but most were in the middle range of each behavioral trait. The goal of the research was to identify groups of children (referred to as clusters) who exhibited commonly occurring behavioral profiles. Such profiles can be determined by a consensus of experts (as was historically done in psychiatry and abnormal psychology in the delineation of childhood psychopathology) or through statistical methods. The statistical approach was used in the current context. The specific technique is referred to as latent profile analysis. A latent profile is a description of a group (cluster) of individuals that share a pattern of behavior (a behavioral profile). The profiles are latent or unobserved in the sense that they are not known at the time of data collection or analysis. If a researcher knows the gender of the children, then gender is a known category, and data for males and females can be used to place children in separate groups based on gender. This is not a latent category. In the current research, the different clusters of children were not known prior to the research by parents, teachers, or the researchers. The profiles were extracted based on the pattern of behavioral traits measured in each sample. Latent profile analysis is a technique in which the goal is to determine the smallest number of latent clusters that is sufficient to account for the associations observed among the measured variables. Each cluster that is isolated will contain individuals that are relatively homogeneous, or similar, with regard to their scores on the eight behavioral scales. The cluster of individuals with a given profile is typically identified by their average score (the mean) on each of the eight scales. All individuals in the group do not have the same score on each scale, but the scores of the individuals in the group are more similar to the other children in the group than to children in any other group. A central question in latent profile research is how many clusters defined by behavioral profiles best fit the data. It is customary to test a wide range of

3.4  The Eight-Profile Model

31

models—models with varying numbers of clusters—in order to find the best one according to a number of statistical criteria. In the current case, we usually tested models with 3 through 10 profiles then determined which one fit the data best. Previous research on children’s behavioral profiles had found somewhere between 3 and 9 profiles provided the best fit. (See Appendix D for a brief review of some of the most influential studies of temperamental or personality profiles in children.) Our research was aimed at identifying the most common behavioral profiles recognized by parents and teachers not including low incidence profiles (e.g., the profile associated with intellectual disability and psychopathology). We set a size limit on the smallest cluster of children that we would consider keeping. Clusters could not contain less than 2% of the sample. Because such small cluster may occur due to randomly occurring measurement issues, this limit also helps to ensure that the clusters of children we obtained were meaningful and were more likely to be found in different samples of children. The research was also guided by the idea that these profiles are the result of biologically driven individual differences in children shaped through interaction with the environment. Since we know of no reason to suspect large cultural differences in the biology of individual differences among children, and because industrialized countries place similar demand on parents and children, we expected to find a set of behavioral profiles that parents and teachers implicitly share across cultural settings. By “implicitly,” we mean that individual parents and teachers are likely to be unaware of the profiles. We believe latent profile analysis of large samples is currently the best method available to isolate these behavioral profiles. (For an extended discussion of the methods used to latent profiles, see Appendices C and E.)

3.4  The Eight-Profile Model To determine how many profiles were latent in the parental (US and Russian) and teacher (US) descriptions of child behavioral characteristics, we initially tested each sample separately. For each sample we determined which model (from 3 to 10 profiles) best fit the data using a wide range of statistical indices. We then tested the proposition that the profiles from the separate analyses could be considered to be statistically similar using recently developed techniques [1]. This was done for each combination of samples (US and Russian parents, US parents and US teachers, Russian parents and US teachers). (See Appendix E for details.) The results of all these analyses point toward the eight-profile model as being the best fit to the data. However, science is never easy and there are always complications. First, researchers look at a large number of indices to determine what model fits the data best. These different indices seldom all agree. Thus, the researcher must rely, in addition to statistical indices, on theory and past research (as well as experience with children) to guide the selection of the most useful and meaningful profile model. Stated another way, sometimes the statistical model that seems the best does not make much sense from what is known about the behavior patterns of children.

32

3  Development of Behavioral Profiles

In summarizing the known work on behavioral profiles in children (see Appendix D), most researchers studying the temperamental characteristics of young children (e.g., preschoolers and toddlers) obtain four broad clusters. These clusters have been given various names but can be described as: (a) Adaptable/Happy/Social; (b) Average/Typical; (c) Withdrawn/Insecure; (d) Poorly Self-Regulated. However, we also included measures of academic ability and achievement motivation to the list of temperamental characteristics that we studied, and these characteristics were included in only two other studies of which we are aware. These two investigations focused on behavior problems in middle childhood and included some measure of academic problems [2, 3]. Those investigations isolated seven profiles based on teacher ratings of behavior, and nine profiles based on parental ratings. We expected to obtain between five and nine profiles. The profiles were also expected to describe clusters of children that were subdivisions of the four general clusters of temperamental characteristics that had been obtained from analysis of younger children. For example, we might obtain two or more clusters of children in the Poorly Self-­ Regulated cluster. Based on our statistical analyses, the extent to which different models replicated across samples, and how well the results fit with prior research, we determine that an eight-cluster model fit the data best. However, there is a caveat. One of the clusters obtained from the 8-cluster model included less than 2% of our sample. Thus, this cluster was eliminated from most of our discussion because we are primarily interested in the most common profiles. This reduced our model to seven common profiles. Table 3.1 summarizes the results of the analyses. One profile was isolated from all analyses, which described an Adaptable/Happy/Social group of children. These children also were perceived to be among the most academically talented and achievement oriented. The label given this profile was Exceptionally Well-Adjusted High Achiever. Parents and teachers thought these children had strong intellectual ability and as well as high achievement motivation. It is recognized that Table 3.1  The seven most common behavioral profiles Broad categories Adaptable/Happy/ Social Average/Typical Withdrawn/Shy/ Insecure Poorly Self-Regulated

Specific profiles Exceptionally Well-Adjusted High Achieversa High Average Self-Regulators Low Average Self-Regulators Withdrawn High (Academic) Achievers Withdrawn Lower (Academic) Achievers Poorly Self-Regulated Higher (Academic) Achievers Poorly Self-Regulated Lower (Academic) Achievers

Abbreviated label 1a 2a 2b 3a 3b 4a 4b

In order to shorten the label applied to this profile, we will often refer to the cluster as “Well-­ Adjusted”

a

3.4  The Eight-Profile Model

33

achievement can take place in a variety of settings even for children in middle childhood (e.g., in sports activities, helping with adult tasks), but the measures obtained were most directly related to academic ability and achievement. Two clusters of children were isolated who had nearly average scores on all behavioral measures. The two average profiles differed primarily in the extent to which the children were able to regulate their emotions, attention, and motor behavior. Thus, one profile was labeled High Average Self-Regulators, while the other was labeled Low Average Self-Regulators. It must be emphasized that both of these groups had scores in the average range on all characteristics, but subsequent analyses of achievement and peer perceptions indicated that the two average groups had very different outcomes in middle childhood. Two behavioral profiles were obtained which described children who were withdrawn, shy, and insecure. One group was characterized by very high self-regulation and high achievement. The other group was characterized by low self-regulation, low achievement, and social withdrawal. As we proceeded further in our understanding of this latter group, this cluster had much in common with the broad category of poorly self-regulated children. However, in all subsequent discussions of this cluster of children, they will be described in conjunction with the withdrawn cluster as they shared the unique social response of being socially inhibited, which, in turn, had consequences for peer group interactions. Two additional profiles were found which described children who had difficulty controlling their attention, expressions of negative emotions (anger, frustration, etc.), and their motor behavior. The two clusters of children exhibiting these profiles were differentiated from one another primarily by parental and teacher perceptions of their academic ability. Thus, one profile was labeled Poorly Self-Regulated Higher Ability and the other Poorly Self-Regulated Lower Ability. The brief verbal labels given to each profile do not do justice to all the characteristics that make up the behavioral pattern of the children in each cluster. These labels are used simply to identify the profile. Further, there is variation in the behavioral characteristics associated with any profile. Thus, not all children exhibiting the profile have all the characteristics at the same level as indicted by the label. Even the shortened verbal label is awkward and wordy, so as the reader become familiar with each profile, sometimes we will refer to each cluster by a number and letter (this is limited to the appendices primarily). For example, for profile type 2a, the number “2” indicates that this child is in the broad average profile type, and the letter “a” indicates the specific profile within the broader category. We make no assumption that all the children identified by a given profile will continue to exhibit that pattern of characteristics as their development progresses. There is a good deal of evidence that there is moderate stability in temperamental and personality characteristics. Thus, some children will move from one profile type to another as they get older. Further, even the characteristics of the profiles might change somewhat as these children enter adolescence. The profiles isolated in this research constitute a snapshot at one point in development. (See Chap. 12  for a discussion of profile stability.)

34

3  Development of Behavioral Profiles

The following eight chapters describe the profiles in each of the four broad categories. The characteristics of the children in each cluster are described in Chaps. 4, 6, 8, and 10. Further, the characteristics of the children and their families in each cluster are described, as well as their academic achievement and peer relationships. The even numbered chapters are a summary of all of our research on each of the four broad profile categories. Chapters 5, 7, 9 and 11, respectively, describe issues that arise in providing optimal guidance of children in each profile by parents and teachers. The implications for the future development of children in each cluster are also discussed with an emphasis on both the advantages and disadvantages of each behavior profile. For details regarding analyses of all the data studied in this project, see appendices F through N. These appendices cover analyses of academic ability, motivation, and achievement (F); compliance with adult rules and expectations (G); prosocial behaviors such as consideration for others (H); peer-perceived social influence (I); peer-perceived social status of profile types (J); parental perceptions of behavior problems (K); peer- and teacher-perceived tendencies toward aggressive behavior (L); and self-perceptions of behavior problems (M).

References 1. Morin, A. J. S., Meyer, J. P., Creusier, J., & Bietry, F. (2016). Multiple-group analysis of similarity in latent profile solutions. Organizational Research Methods, 19, 231–254. 2. Kamphaus, R.  W., Huberty, C.  J., DiStefano, C., & Petoskey, M.  D. (1997). A typology of teacher-rated child behavior for a national U.S. sample. Journal of Abnormal Child Psychology, 25, 453–463. 3. Kamphaus, R.  W., Petosky, M.  D., Cody, A.  H., Rowe, E.  W., Huberty, C.  J., & Reynolds, C. R. (1999). A typology of parent rated child behavior for a national U.S. sample. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 40, 1–10.

Part II

The Seven Behavioral Profiles

Part II is devoted to a detailed description of each of the seven profiles that were isolated from parent and teacher descriptions of the behavior of children in middle childhood. The findings of our research regarding the academic achievement of children exhibiting each profile is described. Further, self-perceptions and peer perceptions of academic talent, achievement motivation, social status, likeability, influence on other children, and behavior problems of children exhibiting each behavioral profile are described. These data provide validation that the profiles we have isolated have a real-world impact on the lives of children. For purposes of convenience, the seven profiles are described in four chapters. Chapter 4 summarizes the characteristics of children we labeled Exceptionally Well-­ Adjusted High Academic Achievers. Chapter 5 describes some of the behavioral assets and risks of children who exhibit this profile in middle childhood, as well as a few broad guidelines for parents and teachers about maximizing the developmental potential of this cluster of children. Chapter 6 describes two profile types describing behavior patterns of average children, as well as how these children perceive themselves and how they are perceived by their peers. These two groups are perceived by parents and teachers as having average academic talents and motivation, as well as typical behavioral tendencies. One of the groups of average children is described as High Average Self-Regulators and the other as Low Average Self-­Regulators. The two groups are differentiated on the basis of academic achievement and their ability to self-regulate attention, expression of negative emotion, and activity level. Chapter 7 describes specific issues affecting the development of average children. Chapter 8 summarize our findings regarding the characteristics of two profile types described as Withdrawn High Academic Achievers and Withdrawn Lower Academic Achievers. Issues that are known to affect the development of shy and socially withdrawn children are described as well as a few broad guidelines for parents and teachers are presented in Chapter 9. Two groups of children who have difficulty self-regulating their behavior are described in Chapter 10. Assets and developmental risks of these groups are summarized in Chapter 11 as well as broad guidelines regarding the role of parents and teachers in foster the develop of these children.

Chapter 4

Exceptionally Well-Adjusted High Achievers

4.1  Introduction Do you remember those children in your elementary classroom who almost always got the answer right and who seemed to know what the teacher was talking about when the rest of the class was lost? In addition, these children also seemed happy, were kind to others, and were popular. I (RM) remember Mary Jane in the third grade. She played the violin at an assembly of the entire school and seemed to have no trace of anxiety. She got good grades and was admired by many of her peers. Mary Jane, at least in the perception of one of her classmates (me), was a prototype of the well-adjusted child in middle childhood. The purpose of this chapter is to summarize the findings of our research regarding the behavioral characteristics of that cluster of children who were defined by the profile we have designated as “Exceptionally Well-Adjusted High Achievers.” (For convenience, the profile will be referred to as Well-Adjusted throughout this discussion. In the appendices, this group of children is labeled cluster 1a.)

4.2  Parent-Teacher Perceptions of Behavioral Characteristics In middle childhood, there is a group of children that in the perception of US and Russian parents as well as US teachers express more joy and general happiness, are more empathetic, and are more considerate than other children. In addition, they are physically vigorous often exhibiting high energy levels. They exhibit less negative emotionality than their peers. For example, when inevitable disappointments, conflicts, or frustrations occur, this group of children is less likely to become intensely emotional or to express anger. They also engage in less rule-breaking behavior in

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_4

37

38

4  Exceptionally Well-Adjusted High Achievers

the home or at school. They are perceived to be less argumentative and stubborn than their peers. They have good control of their attention (i.e., less distractible), and have the ability to organize and keep track of their belongings. They are socially outgoing (i.e., low scores on shyness and social withdrawal) and are not insecure about their abilities. Finally, they are perceived to be more academically talented and achievement motivated. There is remarkable agreement between US and Russian Parents, as well as US teachers about the behavioral characteristics of this cluster of children. Figure 4.1 presents the profile means (average score on each behavioral characteristic in percentile form) for each sample. Children in the Well-Adjusted cluster were perceived as being at or near the 90th percentile in academic ability, achievement motivation, prosocial behavior, and activity level. They were also perceived to be in the 10th to 20th percentile in irritability/antagonism, attention problems, inhibition/social withdrawal, and insecurity/fearfulness. This indicates that some of these children occasionally exhibited these more antisocial behaviors, but this occurred infrequently. There was also general agreement among parents and teachers regarding the percentage of children in each sample that are perceived to exhibit this profile. For US parents, 10% of all children studied exhibited the Well-Adjusted High Achiever profile; for Russian parents, it was 11.3%. US teacher were somewhat more generous including 15.7% of their students in this category. When all samples were combined (2359 children), the percentage was 12.1%. Other researchers who have studied children at various ages have found a similar cluster, although the label provided varies. Prior to the 1960s, children with similar characteristics to the Exceptionally Well-Adjusted High Achieving cluster were referred to in terms such as “nice children” or “obedient children.” As more scientific approaches to grouping children according to their temperamental or 100 90 80 70 60 50 40 30 20 10 0

U.S. Parents Russian Parents U.S. Teachers

Fig. 4.1  Comparison of cluster mean scores for Exceptionally Well-Adjusted High Achieving Children based on US and Russian parent, and teacher ratings (data presented in percentile form)

4.3  Demographic Characteristics

39

personality characteristics began to appear in the 1970s, these children were referred to as “Easy” [1, 2], “Confident” [3, 4], “Resilient” [5–7], “Well-Adapted” [8–10], “Expressive” [11], and “Well-Regulated, Positive Reactive” [12]. Each of these labels reflected some of the salient characteristics of the cluster. “Easy” and “Resilient,” for example, emphasized that these children are adaptable to a variety of situation and life circumstances. “Confident” emphasized their vigor and zeal in approaching new tasks. “Well-Regulated, Positive Reactive” emphasized the ability of children in this group to regulate their emotions and behavior in accordance with the demands of their environment (e.g., the classroom) and with their frequent expression of positive emotions (e.g., smiling, laughter).

4.3  Demographic Characteristics The Well-Adjusted profile describes a pattern of behavior that appears more mature and adult-like than would be expected of most children in middle childhood. Gender differences might be expected in the percentage of children who exhibit this profile because there is ample evidence that girls are more biologically and psychologically mature (e.g., more self-regulated) than boys at this age [13, 14]. Based on descriptions by US parents, 43.9% of the children exhibiting the Well-Adjusted profile were girls. However, for Russian parents, 55.1% of the children who exhibited this behavioral pattern were female, and a similar percentage (57.6%) was found for US teachers. Thus, for Russian parents and US teachers, girls were viewed as exhibiting the Well-Adjusted profile at a somewhat higher rate than boys, but this was not true for US parents. Other researchers who have studied behavioral profiles in middle childhood have tended to find an excess of females in this type of profile. In one study, parental ratings resulted in 57% of the Well-Adjusted cluster being female [8]. In another study, using teacher ratings, 61% of this group was female [9]. Thus, we conclude that there is a modestly higher percentage of females in middle childhood who both parents and teachers perceive to be exceptionally well-adjusted and high achieving. Temperament/personality has seldom been found to relate to ethnicity or race. However, much of this past research is based on self-perceptions. Perceptions of parents and teachers might be somewhat different. For the US parent sample, we asked parents to identify the best descriptor of their child’s ethnicity/race and these were placed into five categories: African American (9.9%), Asian American (5.0%), European American (54.1%), Hispanic American (8.2%), and Other (22.8%; included a variety of small ethnic groups plus a larger group referred to as “mixed”). Because the number of children in most minority ethnic groups was small, we could not make meaningful comparisons for specific groups. We did, however, do an analysis of European Americans compared to various combinations of minority groups (e.g., all other groups considered as minority, then all other groups but Asian

40

4  Exceptionally Well-Adjusted High Achievers

American’s considered as minority), and found that in no analysis were there significant differences in the proportion of children in the majority and minority categories who exhibited the Well-Adjusted profile. For the teacher-rated sample, which included only teachers in the State of Georgia, the only minority group of any size was African Americans. There were no significant differences in the percentage of African American and European American children who were viewed by teachers as exhibiting the Well-Adjusted profile. Thus, being well-adjusted, happy, and socially integrated, as well as achievement oriented was not related in any meaningful way to race/ethnicity among the middle childhood children that were studied. Maternal education level is often found to contribute to the prediction of child behavior and other developmental outcomes. In the US parent sample, more children exhibiting the Well-Adjusted profile had parents who had a college education or advanced training (57.3%) than had lower levels of education. For the Russian sample, a similar pattern emerged. We conclude that there is a modest overrepresentation of children in the Well-Adjusted cluster whose parents had higher levels of education. In summary, there is very little evidence that the Well-Adjusted behavioral profile is related in meaningful ways to the ethnicity/race, but there is a modest association with socioeconomic circumstances of the children and their families. There is some indication from our research and the broader literature in child development that girls in middle childhood exhibit better self-regulation than boys, and thus a higher percentage of girls exhibit the Well-Adjusted profile than boys. However, even these differences were not large.

4.4  Academic Ability and Motivation Two of the central characteristics of the Well-Adjusted profile are that parents and teachers view children in this cluster as having substantial academic talent (e.g., intelligence and openness to new ideas) and as having higher levels of motivation to achieve than their peers. One question that our research addressed is the extent to which teacher and parental perceptions relate to (a) the actual grades the children obtain, (b) the perceptions of peers about the academic motivations and talents of this group, and (c) to the perceptions of the children themselves about their own academic ability. Academic grade point averages were available for the Russian sample that was studied by Slobodskaya and colleagues. Parents reported their children’s grade point average (GPA) based on the most recent report card. Children in the Well-­ Adjusted cluster obtained the highest GPA of all profile types. Also, there was a tendency for girls to have higher GPAs than boys. How do the classmates of students in middle childhood view their peers’ academic talent and motivation? This question addresses the awareness of children in this age range of the abilities of their peers. Data relevant to this question were obtained from two groups of students in Georgia by Lease and colleagues. The two

4.5  Prosocial Behaviors, Compliance, and Likeability

41

groups of students differed somewhat in socioeconomic and ethnic/race characteristic. In one cohort of students in the fourth and fifth grades (Cohort A), children nominated from their classroom those who were academic talented and motivated to achieve. (Cohort A students were from middle-class families, but they were of moderately lower socioeconomic circumstances and served a higher proportion of African American students than Cohort B.) The specific questions were as follows: 1 . This person makes good grades, is smart, and usually knows the right answer. 2. This is a person who tries hard to do good school work. Children in the Well-Adjusted cluster had the highest average (i.e., mean) score of all clusters for both questions 1 and 2. Thus, the peer group perceives children in this cluster in much the same way as teachers and parents. They believe that these children have more academic ability and are more motivated to do academic work. Students in Cohort B of the Georgia student sample were asked to rate their own academic ability and motivation. The specific questions were as follows: 1. If you were to list all of the students in your grade from worst to best in school work, where would you put yourself. (Rated on a five-point scale from “one of the worst” to “five of the best.”) 2. For me, being good at my schoolwork is … (“not at all important” to “very important”). Again, children in the Well-Adjusted cluster had the highest mean score both for self-rated academic ability and motivation. Thus, parents, teachers, peers, and the children themselves all view the children who exhibit the Well-Adjusted profile as more academically talented and more motivated to do well in school compared to same-age peers. The grades given by Russian students support this view. (See Appendix F for more details on the relationship of temperament profiles to academic achievement and motivation.)

4.5  Prosocial Behaviors, Compliance, and Likeability One characteristic of the group of children who exhibit the Well-Adjusted profile is that they exhibit behavior patterns that contribute to positive relationships with others. For example, they are viewed by parents and teachers as displaying a positive mood and they seem to behave toward peers in an empathic way more often than do other children. They also engage in less social antagonism such as criticism of peers. This temperament profile was hypothesized to foster the social attractiveness of this group of children both for adults and their peers. One attribute of the children who exhibit the Well-Adjusted profile is that they are compliant; that is, they are sensitive to the desires and expectations of adults and peers in their environment and have a desire to meet these expectations. They want to please others. This line of reasoning is based on the observation that children exhibiting this pattern of behaviors actually do meet the expectations of most

42

4  Exceptionally Well-Adjusted High Achievers

parents and teachers. They do well in school, are perceived by parents and teachers as seldom exhibiting negative emotions in a prolonged and intense fashion, and are seldom argumentative and antagonistic to authority, or toward peers. Data collected from all three samples in our project (Russian parents, US parents, US teachers) directly addressed this issue using a measure of compliance (see Appendix G for details). Children in the Well-Adjusted cluster, both in the United States and in Russia, obtained compliance scores from parents that were in the 91st percentile for their respective samples, meaning that these children were more compliant than 91 of 100 of their peers. The teacher score was at the 84th percentile. There is no doubt that this temperamental profile defined children who are compliant with the expectations of the adults in their world. Some children seem to be happy and joyful most of the time even in the face of setbacks and frustrations. This tendency is labeled “positive emotionality” by many psychologists and is a fundamental temperamental characteristic of human beings. It can be observed in infancy and the preschool years as well as later in development. Positive mood is a social lubricant; it attracts other people (adults as well as same-age peers) and makes social interactions more comfortable. Children exhibiting the Well-Adjusted profile were defined in part by parent and teacher perceptions that they were often in a positive mood. We wondered if the peer group viewed the children in this cluster in the same way. Figure 4.2 presents the peer nomination scores for positive emotionality for US students (likeability was assessed an analyzed separately for Cohort A and B) . (For details, see Appendix H). Their scores were near the 70th percentile. Clearly, Well-Adjusted children were perceived by their peers as exhibiting more positive emotionality than other children. 85 80 75 70 65 60 55 50 Posive Emoonality

Empathy

Likeability-A

Likeability-B

Fig. 4.2  Peer perceptions of prosocial behaviors of the Well-Adjusted cluster (Data are in percentile form)

4.5  Prosocial Behaviors, Compliance, and Likeability

43

90 85 80 75 70 65 60 55 50 45 40

Fig. 4.3  Peer perceptions of four indicators of social status by Well-Adjusted children (data in percentiles form)

Empathy toward others is another aspect of what we have labeled as prosocial behavior. Figure 4.3 presents peer nominations scores for behaviors such as cheering up peers when they are sad or upset about something, or showing sympathy to a peer who is sad, hurt, or upset (see Appendix H for more details). Children in the Well-Adjusted cluster again were near the 70th percentile in empathy in the perception of peers. Georgia students were asked to indicate those children who they would like to play with most and also to indicate those children they would like to play with least. A score was then calculated in which the number of “like least” nominations was subtracted from the number of “like most” nominations. This is a common practice in research on children’s peer relationships to determine how well-liked or social preferred a child is among their peers. The scores for children in the Well-Adjusted cluster are presented in Fig.  4.2. This nomination procedure indicated that Well-­ Adjusted children are well-liked. However, there was a difference in the likeability scores of this cluster of children for the two student samples (Cohorts A and B). Well-Adjusted children were viewed as much more attractive companions in cohort A than in cohort B. (As a reminder, the Cohort A sample contained a higher percentage of African American children than Cohort B, and while both cohorts were lower middle class, Cohort A children were somewhat less advantaged.) In summary, children who exhibit the temperament profile we have labeled Well-­ Adjusted are perceived to be compliant by parents and teachers, and are perceived by their peers as unusually happy and empathetic toward others. They are also viewed as the type of child that others want to spend time with in a social situation (i.e., they were liked).

44

4  Exceptionally Well-Adjusted High Achievers

4.6  Social Status Being perceived by adults and your peers as bright, academically capable and motivated, caring, and likeable are all adaptive characteristics associated with the Well-­ Adjusted personality profile. As socially positive as these behaviors are, they might not translate into other indicators of social status, such as popularity, in the eyes of one’s peer group. Social status was evaluated in our research in four ways. First, it was directly measured based on peer nominations. Perceived popularity was the number of nominations each child received as most popular minus the number of nominations they received as least popular. Second, in one cohort of students in Georgia (Cohort A), children were asked to nominate those peers that were most admired and the children who other children want to be like. Third, there was a measure of social prominence; that is, students nominated those children who are perceived to be well known by almost everyone and thought to be really “cool.” Fourth, one aspect of social status is how often a person is likely to be chosen for leadership roles. Scores for leadership were based on the number of nominations by the peer group in response to questions in the form of: “This person gets chosen by other children as the leader.” In Cohort A, children could nominate peers from their classroom, whereas in Cohort B children could nominate peers from their grade level. Figure 4.3 presents the social status characteristics of the children in the Well-Adjusted cluster. Children who exhibited the Well-Adjusted behavioral profile (based on parent and teacher ratings) were perceived by their peers as the most popular group of children in their class. They were often given leadership roles by their peers and were generally admired. They were also socially prominent in the sense that they were thought of as “cool,” and were well known by many children. Note, however, that some forms of social status are affected by the composition of the peer group. Well-Adjusted children in Cohort A were not viewed as being as “cool” as Well-­ Adjusted children in Cohort B. However, all indicators of social status show that children with the behavioral profile that was labeled Well-Adjusted have much higher social status than the average child in their peer group.

4.7  Influence on Peers Do popularity, leadership, admiration, and social prominence translate into social influence on others? Influence was measured based on peer nominations of children who were particularly influential in five areas: academics, sports, youth cultural trends (clothing, music, slang), fantasy/make-believe games, and inappropriate behavior in the classroom (e.g., talking back to the teacher, breaking rules if the teacher is out of the room). In both student cohorts, children exhibiting the Well-­ Adjusted personality profile were nominated far more frequently than their peers as being able to influence others in the area of academics (see Fig. 4.4.). Further, they

4.8  Behavior Problems

45

90 80 70 60 50 40 30

Fig. 4.4  Peer perceptions of five areas of influence by Well-Adjusted children (data in percentile form)

were far below average in influencing others to initiate inappropriate behavior in the classroom. As has been discussed previously, this cluster of children is unusually compliant to adult rules and the desire to please adults seems to have kept them from instigating these rule-breaking behaviors. Again, there was a marked difference in the influence of Well-Adjusted children depending on the student cohort being examined. Students in Cohort B viewed their peers who exhibited the Well-Adjusted profile as more influential in all areas than did children in Cohort A. This is consistent with the previously reported finding that children in the Well-Adjusted cluster were not considered to be as “cool” or well known in Cohort A as they were in Cohort B. Students in Cohort B came from families with somewhat higher socioeconomic status, and from more European American families. The children in Cohort A who were most influential had different behavioral characteristics than the academically oriented, self-controlled, and compliant children in the Well-Adjusted cluster as we will see latter. Thus, characteristics of the peer group can have an important effect on what characteristics that are associated with influence.

4.8  Behavior Problems Another important aspect of the way that the social environment views children is the extent to which they exhibit behavior that is viewed as problematic by parents and teachers. This type of behavior in child development research is often referred to as “behavior problems.” In an extreme form, behavior problems can be an indication of psychopathology although most children exhibit some level of these behaviors. Russian parents completed a commonly used measure of such problems (for

46

4  Exceptionally Well-Adjusted High Achievers

details see Appendix K) and scores were obtained for four types of problems (Hyperactivity, Emotional Problems, Conduct Problems, and Problems with Peers). The Well-Adjusted children were far below average in all areas, with scores at the 32nd percentile. However, this does not mean that none of the children who exhibit this temperamental profile exhibit problematic behavior. One aspect of our research was designed to determine the number of children in this group who were perceived to have significant behavior problems by their parents. A significant behavior problem was defined as being at least one standard deviation above the mean of all children studied (i.e., having a score greater than the 84th percentile). This level of problems does not necessarily indicate that they would receive a clinical diagnosis, but it does indicate that they exhibited problematic behaviors at a much higher frequency than their peers. Using this definition of significant problems, 8.0% of the Well-Adjusted children had significant problems with Hyperactivity, and 9.5% had Conduct Problems (did not follow rules) based on Russian parental perceptions. Much smaller percentages had Emotional Problems including problems with anxiety and depression (1.6%) and Peer Problems (4.8%). Given our definition of significant problems (being one standard deviation above the mean), 16% of all children in this sample have significant problems. Thus, if significant problems are present to the same degree within each cluster, then we would expect to see close to 16% within clusters having a significant problem. However, a smaller percentage of the children in the Well-­ Adjusted cluster had significant problems than was typical of their peers; that is, the percentages in this cluster were well below 16%. In order to examine the issue of behavior problems more closely, we examined measurements of aggressive behaviors that were available for the sample of US teachers, children’s same-age peers at school, and the children themselves. One measure that was obtained from US teachers (children in Cohort A) was a 10-item questionnaire that asked about each child’s social aggression (e.g., attempts to exclude other children from their social group), verbal aggression (e.g., saying mean or unkind things to another child), and physical aggression (e.g., pushing or hitting another child). The score was primarily focused on social and verbal aggression (see Appendix L). It is interesting that teachers’ ratings isolated about as many students in the Well-Adjusted cluster as having significant levels of social and verbal aggression (18%) as other children in the sample (16% would be the expected value for the sample as a whole). In the US teacher sample, children’s same-age peers from both cohorts were asked to nominate children from the classroom (Cohort A) or from their grade (Cohort B) who exhibited social, verbal, or impulsive physical aggression. Fewer children were in the significantly aggressive range based on peer nominations than was typical of all children in the sample. Of all types of aggression, the highest percentage was for verbal aggression for these Well-Adjusted children. With regard to self-ratings of aggression, about an average percentage of children in the Well-Adjusted group labeled themselves as significantly aggressive. This self-perception is consistent with the view held by teachers (see Table 4.1).

4.8  Behavior Problems

47

Table 4.1  Percentage of children in the Exceptionally Well-Adjusted Higher Achiever cluster who are perceived to have significant behavior problemsa Rater group Measure US teachers  Cohort A Rating scaleb Student peers  Cohort A Nominationsc  Cohort A  Cohort B  Cohort B  Cohort A Self-ratings  Cohort A Rating scaled  Cohort B Rating itemf Self-ratings  Cohort A BASCe  Cohort A BASC  Cohort A BASC  Cohort A BASC

Problem type

Percent with significant problems (%)

General aggression

18.1

Impulsive/aggressive Socially aggressive Verbally aggressive Bullying others Shy/socially inhibited

5.6 8.5 10.0 3.3 11.3

Aggression 12.9 Motivation for rule breaking 11.9 Anxiety Depression Interpersonal problems School maladaptation

18.6 31.4 10.0 38.6

Significant behavior problems were defined as scores at or greater than the 84th percentile for the sample of children studied (>1 s.d.) b A 10-item rating scale addressing social, verbal, and physical aggression c A score based on number of peer nominations for each child in the sample, standardized for each classroom in the cohort. d A 4-item rating scale addressing social, verbal, and physical aggression e Behavior Assessment System for Children-Self-Report Form (for details, see Appendix M) f One item measuring motivation to engage in inappropriate behavior in the classroom (for details, see Appendix L) a

About half of the public-school sample of children in Georgia responded to a standardized measure of behavior problems. (The Behavior Assessment System for Children [15], see Appendix M for details.) One of the most interesting findings from these analyses was that significant self-ratings of depression were found for a higher percentage of students in the Well-Adjusted temperament cluster than occurred for the sample as a whole. These well-adjusted students also reported somewhat higher levels of negative attitudes toward school than their peers. So despite excellent performance, a relatively high percentage of children in the Well-­ Adjusted cluster reported that they disliked school (see Table 4.1) and felt significant depression. The level of self-perceived depression, as well as negative attitudes toward school, is at odds with what would be expected from parent and teacher behavioral ratings. This cluster of children was perceived by the adults in their environment as unusually happy and as seldom appearing insecure. Perhaps the constant pressure to perform academically at a high level and to adhere to adult expectations contributed to feelings of generalized anxiety and depression. The direction of causation runs in

48

4  Exceptionally Well-Adjusted High Achievers

both directions, however. It is possible that a heightened level of anxiety may increase adherence to adult expectation for performance and behavior. In general, then, children who exhibit the Well-Adjusted profile (as perceived by parents and teachers) have lower levels of behavioral problems than peers. However, some of these “model” children are aggressive (primarily socially and verbally), enjoy being noncompliant with authority, and experience high levels of depression. Many of these children (more than one-third) also report they dislike school. Thus, while this group has a low probability of exhibiting significant behavior problems, this does not mean that all children defined by this behavioral profile experience low levels of problematic behavior. Clearly, some do. Behavior problems and psychopathology have often been studied in clusters of children defined by temperamental characteristics. In the classic study of temperament in young children, Chess and Thomas [1, 2] created three profiles primarily based on clinical judgment. These groups were labeled as follows: Easy, Slow-To-­ Warm-Up, and Difficult. The “Easy” group was most similar to the Well-Adjusted profile just described and was defined by being highly adaptability to change as well as being of predominantly of positive mood. When upset, these children expressed their distress with mild to moderate intensity. These children were assessed repeatedly from early childhood through adulthood. The primary finding of this research effort was that the “Easy” group of children (defined in the preschool years) required less support from their family, from school personnel and from mental health professionals than the other groups throughout childhood and early adulthood. However, their research indicated, as does ours, that some of these children need mental health support from time to time. DiStefano and colleagues [16] obtained seven profiles based on teacher ratings of large samples of public school children using a standardized behavior problem measure. One of the profiles was labeled “Well-Adjusted,” and this profile had much in common with the profile of the same name from our research. This research team examined a number of school-based outcomes indicative of school maladjustment including suspensions, physical aggression offenses, verbal aggression offenses, sexual offenses, and referrals for psychological testing. They found that the children defined as Well-Adjusted had the lowest number of infractions of any kind and the lowest number of special education referrals of any cluster in their study. In one of the foundational studies of behavioral profile types, Caspi and colleagues [3, 17, 18] obtained ratings of behavior of a large sample (over 900) of children in New Zealand at age 3. The ratings of 22 behaviors were made by examiners immediately after a 90-minute session of psychological testing. Using statistical techniques, all children in the sample were placed into five homogeneous groups based on their profiles on the 22 behaviors measured. The groups were labelled Well-Adjusted, Under-Controlled, Confident, Reserved, and Inhibited. These children have been followed and reassessed repeatedly into adulthood. In one study, Slutske et  al. [18] investigated the self-reported gambling problems of this birth cohort at age 32. Of all five temperamental categories, the children who were defined as Well-Adjusted children at age 3 had the lowest level of gambling problems at age 32. Specifically, while approximately 12% of the Well-Adjusted group

References

49

had gambling problems (primarily males), this was a lower percentage than all other groups in the sample. These effects could not be explained by differences in childhood IQ or family socioeconomic status. In two similar studies by members of this research group [3, 17], the level of psychopathology and the quality of interpersonal functioning in social networks in the home, in romantic relationships, and in work relationships at age 21 were investigated. These researchers found that the Well-­ Adjusted group defined by their temperament profile at age 3 had the lowest levels of interpersonal problems and psychopathology at age 21. These studies strongly support not only the behavioral resilience of children similar to our Well-Adjusted profile, but show that temperamental profiles obtained in childhood are not limited in their impact to the age group in which they were assessed. Temperament has a lasting influence.

References 1. Chess, S., & Thomas, A. (1987). Origins and evolution of behavior disorders. Cambridge, MA: Harvard University Press. 2. Thomas, A., Chess, S., & Birch, H. G. (1970). The origins of personality. Scientific American, 223, 102–109. 3. Caspi, A., & Silva, P. A. (1995). Temperamental qualities at age 3 predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486–498. 4. Janson, H., & Mathiesen, K. S. (2008). Temperament profiles from infancy to middle childhood: Development and associations with behavior problems. Developmental Psychology, 44, 1314–1328. 5. Block, J. (1971). Lives through time. Berkeley, CA: Bancroft Books. 6. Asendorph, J. B., Borkenau, P., Ostendorf, f., & Van Aken, M. A. G. (2001). Carving personality description at its joints: Confirmation of three replicable personality prototypes for both children and adults. European Journal of Personality, 15, 169–198. 7. Robins, R. W., John, O. P., Caspi, A., Moffitt, T. E., & Stouthamer-Loeber, M. (1996). Resilient, overcontrolled, and undercontrolled boys: Three replicable personality types. Journal of Personality and Social Psychology, 70, 157–171. 8. Kamphaus, R.  W., Huberty, C.  J., DiStefano, C., & Petoskey, M.  D. (1997). A typology of teacher-rated child behavior for a national U.S. sample. Journal of Abnormal Child Psychology, 25, 453–463. 9. Kamphaus, R.  W., Petosky, M.  D., Cody, A.  H., Rowe, E.  W., Huberty, C.  J., & Reynolds, C. R. (1999). A typology of parent rated child behavior for a national U.S. sample. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 40, 1–10. 10. Orpinas, P., Raczynski, K., Bandalos, D., Peters, J. W., & Colman, L. (2014). Latent profile analysis of sixth graders based on teacher ratings: Association with school dropout. School Psychology Quarterly, 30, 577–592. 11. Van den Akker, A. L., Dekovic, M., Prinzie, P., & Asscher, J. J. (2010). Toddler’s temperament profiles: Stability and relations to negative and positive parenting. Journal of Abnormal Child Psychology, 38, 485–495. 12. Scott, B. G., Lemery-Chalfant, K., Clifford, S., Tein, J. Y., Stoll, R., & Goldsmith, H. H. (2016). A twin factor mixture modeling approach to childhood temperament: Differential heritability. Child Development, 87, 1940–1955. 13. Else-Quest, N. M. (2012). Gender differences in temperament. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 479–496). New York: Guilford.

50

4  Exceptionally Well-Adjusted High Achievers

14. Else-Quest, N. M., Hyde, J. S., Goldsmith, H. H., & Van Hulle, C. A. (2006). Gender differences in temperament: A meta-analysis. Psychological Bulletin, 132, 33–72. 15. Reynolds, C. R., & Kamphaus, R. W. (2004). Behavior assessment system for children-selfreport form – 2. New York: Pearson. 16. DiStefano, C., Kamphaus, R. W., Horne, A. M., & Winsor, A. P. (2005). Behavioral adjustment in the U.S. Elementary school: Cross-validation of a person-oriented typology of risk. Journal of Psychoeducational Assessment, 21, 338–357. 17. Newman, D. L., Caspi, A., Moffitt, T. E., & Silva, P. A. (1997). Antecedents of adult interpersonal functioning: Effects of individual differences in age 3 temperament. Developmental Psychology, 33, 206–217. 18. Slutske, W. S., Moffitt, T. E., Poulton, R., & Caspi, A. (2012). Under-controlled temperament age 3 predicts disordered gambling at age 32: A longitudinal study of a complete cohort. Psychological Science, 23, 510–516.

Chapter 5

Fostering the Development of Well-­Adjusted Children

5.1  General Principles For children whose behavior fits any of the seven behavior profiles, there are a few general principles that serve as the foundations for all interventions by parents or teachers. These principles are based on temperament theory and research on the stability of temperament in childhood (see Chap. 12). These principles have also been fundamental in the major intervention programs based on temperament theory [1–4] and related behavior change interventions. 1. Profiles are relatively stable particularly from middle childhood onward. Interventions are not aimed at changing the basic temperamental structure of the child. 2. Parents and teachers should understand that children come into the world with a huge range of individual differences, and should accept and appreciate the characteristic behavior patterns of each child in their care. 3. The behavior of children will change from time to time during development as their understanding of themselves and their skills develop. Also, environmental changes will alter their behavior. However, there is a tendency for children to come back to their characteristic pattern of behavior after short-term perturbations. 4. If a child is persistently engaged in a behavior that is troubling to those around them or to the child themselves, intervention efforts by parents and teachers should be focused on helping the child make adjustments of specific behaviors in a given environmental circumstance. 5. Most successful interventions are designed to alter the environmental circumstance that fosters a behavior that is troubling to the child or to those around him/her.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_5

51

52

5  Fostering the Development of Well-Adjusted Children

Every parent wants to know how their child “will turn out.” Will they be happy? Will they find meaningful social relationships and stable intimate relationship with a life partner? Will they accomplish great things? Psychological theory and research have addressed these questions and is fairly good at making predictions for groups of children. These predictions are much like the predictions made of weather. They are stated in terms of probabilities. Based on the information developed from the current study and a great deal of other research in child development, children who exhibit the Well-Adjusted behavioral profile in middle childhood have a high probability of having a smooth adjustment to adolescence and the demands of adult life. However, many factors can derail this developmental trajectory. One way to conceptualize these predictions is to think about the assets and risks that a particular behavior pattern (like the Well-Adjusted profile) has for the future of the child. The discussion of risks also leads directly toward a discussion of what parents and educators can do to help mitigate the risks of a particular behavior pattern.

5.2  Behavioral Assets Children who are in the Well-Adjusted cluster in middle childhood are blessed with many of the characteristics that enhance cognitive, social, and emotional development. One of these assets is that they have the ability to learn quickly. In particular, by middle childhood, they have demonstrated the ability to learn abstract symbols and concepts more readily than their peers. Almost all adults have learned the basic concepts and skills taught in elementary schools. These include the ability to read (interpret abstract written symbols) and the ability to do simple arithmetic calculations (e.g., addition, long division, multiplication). However, some children have the ability to learn these concepts and skills after only a few exposures while others take longer. Children who can learn these skills quickly have the ability to progress to higher levels of conceptual learning and find enjoyment in learning these skills and concepts. In the realm of athletics and music, the ability to learn quickly is often referred to as “talent.” Well-Adjusted children are talented learners. The importance of being a talented learner cannot be overemphasized. Academic learning (symbol manipulation) is only one form of learning, and there are clearly other forms that affect the development of any individual [5]. One example is the learning of motor skills. Another is the ability to understand one’s position in space (e.g., knowing ones orientation in relationship to the cardinal directions). There is, also, the critical ability to understand the social world. Although there are debates about the various form of learning (sometimes conceptualized as different kinds of intelligence), one thing is clear. The ability to learn to manipulate and conceptualize verbal and mathematical symbols plays a major role in coping with the demands of the twenty-first century. It affects not only formal learning and progress in schooling, but it applies to all the learning that takes place outside of classrooms. For example, the skills involved in programming a computer are now beginning to be

5.2  Behavioral Assets

53

systematically taught in some elementary school. A more basic example is the ability to utilize technological machines. Most of what children know about their cell phones is learned on their own or from peers. As these kinds of skills become progressively more important, the speed with which one learns to use these tools makes a real difference in how well one copes this technological environment. In addition to a talent for learning, Well-Adjusted children have a talent for creating a supportive social environment. Their parents, teachers, and peers find them socially attractive. They are perceived by their peers as likeable; however, at least some of these children are socially and verbally aggressive to peers, which is a behavior that can help establish and maintain a popular status among peers. They are happy and energetic. Most adults and peers find that Well-Adjusted children are socially attractive companions who foster positive emotions in those with whom they interact. This means that these children have created a positive social environment for themselves. This supportive environment further enhances their development progress. Children who are well-liked are also those are most likely to develop and sustain friendships in their peer groups. There is a strong literature in developmental psychology on the benefits of friendship for children and adolescents. High-quality relationships with friends have been shown to be related to many aspects of wellbeing [6]. For example, there is considerable evidence that friendships provide security and social support, and help children develop social problem-solving skills [7]. Sharing feelings with friends helps children progress from being egocentric to being able to understand the perspective of others [8]. Every human being has difficult days, difficult decisions, and problems that must be solved. All these problems are more likely to be coped with in an effective way if the individual has a social support system in which people want to reach out and provide help, emotionally or instrumentally. Children who exhibit the Well-Adjusted personality profile have enormous advantages that enhance their social skills and ability to cope with adversity. At one point in his career, the first author worked with families who had a preschool child who had some type of physical or intellectual disability. Many of these children had cerebral palsy, a birth-related defect that caused brain damage that resulted in poor motor control. Interviews with family members revealed that they understood the physical problems their children faced, but some of the parents were concerned about the behavioral characteristics of their children. Children with cerebral palsy, regardless of the degree of their disability, vary greatly in their social and emotional behavior. Some are temperamentally sunny and happy, while others are irritable, fearful, or aggressive. Children with disabilities have a particular need for help from others. It is important that the helper (therapists, teachers) find the interaction with the child rewarding. Observations of the interactions of the children with cerebral palsy with their physical and occupational therapists indicated that the quality of supportive training the therapists provided depended to some extent on the temperamental qualities of child. If the child was happy, smiling, and joyful, they received more sensitive therapy and a greater quantity of therapy than if they were irritable and difficult to manage. Therapists simply enjoyed interactions with

54

5  Fostering the Development of Well-Adjusted Children

the children who had a sunny disposition, and they found their work as a helper meaningful and rewarding. Probably few of the children defined here as Well-Adjusted have a readily observed disability. On the contrary, they are perceived to have many intellectual and behavioral advantages. But like all of us, they need social support because they will inevitably experience failures and frustrations. Their cognitive ability to learn how to navigate complex social relationships and the emotional talents which enhance social engagement greatly facilitate social interaction and the establishment of a supportive social environment. The Well-Adjusted cluster of children has another advantage not shared by many of their peers. They have unusual talent for self-regulation of their own behavior, including regulation of attention and emotions. For example, these children are perceived by parents, teachers, and their peers as very active and energetic. However, this energy is displayed in appropriate ways. Well-Adjusted children have learned to regulate when and where they exhibit their energetic behavior. On average, they also have very few problems with distractibility, aggression, antisocial behavior, insecurity, and fearfulness.

5.3  Behavioral Risks It is one of the major tenants of this book that all temperament-based profiles have advantages and all have risks. Some temperament patterns have fewer risks than others, and children in the Well-Adjusted cluster in middle childhood are likely to have lower risks for adverse developmental outcomes than peers. But all behavioral profiles contain risks simply because different life tasks, different problems, and different environments require different behaviors to maximize the chance of having a productive, happy, and socially integrated life. Our data indicate that these Well-Adjusted children exhibit fewer behavior problems than the majority of their peers. However, at least some of the children with this profile have social and emotional issues that require adult monitoring and at times may require professional support. The data reported in the previous chapter point out, for example, that approximately 10% of Russian parents perceive that their Well-Adjusted children exhibit a significant level of conduct problems (e.g., breaking of rules, are antagonistic to social norms). Ratings of aggressive tendencies, particularly social aggression (i.e., attempting to exclude others from their social group, verbally demeaning statements) by US teachers indicated that 18% of this sample engaged in significant social and verbal aggression. Student peer nominations of a cohort of Georgia students indicated that about 10% of the Well-­ Adjusted students were socially and verbally aggressive. While all these rates of antagonistic behaviors are lower than the rates for any other cluster in our research, the data indicate that some of the children in this group engage in significant levels of socially antagonistic behavior. Aggression and rule-breaking behaviors are referred by psychologists as “externalizing” problems, and clearly children in the

5.3  Behavioral Risks

55

Well-Adjusted group are not immune from this type of problematic behavior. In fact, these types of aggressive and antagonistic tendencies are exhibited by some popular children when trying to influence their peers [9]. Data from the children’s peers indicate that about 11% of the Well-Adjusted children were significantly socially inhibited and shy in new social situations, and 31% of the Well-Adjusted children in one cohort of students self-reported that they experienced depression. Excessive social inhibition, social anxiety, and depression are types of “internalizing” problems, and Well-Adjusted children experience this category of psychological problems at a rate that is near average for their peer group, and perhaps even higher for feelings related to depression. Thus, some children who may appear well-adjusted to their parents and teachers are reporting significant levels of internalizing problems. A recent report of the prevalence of mental disorders in children was summarized from five major longitudinal studies carried out in the United States, Switzerland, and New Zealand [10]. Each study followed the children from childhood or adolescence into adulthood. On average, 60–85% of the participants developed diagnosable psychiatric disorders over a 12- to 30-year span. For most, this was a transient episode much like having a broken bone or kidney stones. Sufferers experience impaired functioning, sought some type of mental health intervention, and recovered. But this type of data indicate that the majority of people experience mental health problems, and these will occur in all likelihood for those who are in our Well-Adjusted cluster in middle childhood. Children exhibiting the Well-Adjusted profile are characterized by strong levels of self-regulation. The source of this self-regulation is rooted in many factors. But from a psychological point of view, it has been shown to be, in part, the result of greater apprehension about transgressing [11, 12]. The behavior profile based on parent and teacher reports does not indicate that parents are aware of the apprehensive nature of this group of children. However, when the children in our study completed self-reports of the feelings of anxiety and depression, a higher percentage of children in the Well-Adjusted cluster reported significant symptoms of depression than their peers. These data should not be interpreted as depression that would reach diagnosable levels, but rather a tendency to experience anxious and depressed feelings. These data support the notion that the self-regulation of at least some of the children in this cluster is based on enhanced sensitivity to parent and teacher disapproval. Thus, one risk for children in the Well-Adjusted cluster is that they have a need to meet the expectations of the adults and peers in their social world. In many ways, if it is not excessive, this stress to perform at a high level and to meet the expectations of others has a positive social effect. It enhances the learning of self-­ regulatory processes, thus increasing academic performance and social attractiveness. In addition to the need to meet expectations of others, Well-Adjusted children run several other risks that might impair their development and create significant behavioral and mental health issues. The primary risks occur when there is a major change in their environment. There are many ways in which this can occur. Consider the example of increased competition. It is inevitable that the child who is the

56

5  Fostering the Development of Well-Adjusted Children

brightest and most highly achieving in their elementary school class will encounter a situation in which they are not the academic star. This may occur if they are placed in classes for gifted and creative children. It may occur if they are in advanced placement classes in high school. It may occur when they are admitted to a highly selective college, or to a graduate or professional degree program. It may occur when they get a position in a highly competitive industry. What happens when they discover that they are not the brightest person in the room? Some Well-Adjusted children are at particular risk when this situation occurs because learning has come so easily they have been able to perform at high levels with very little effort throughout their prior schooling. When the learning environment becomes more demanding and achievement depends on prolonged effort, some of these children have not learned the required study skills and time management skills required by the new situation. They can no longer write the paper the night before it is due. The paper may take days of prolonged research and repeated editing, or in the case of a Ph.D. dissertation, it may take years. This scenario will happen to all children at some point in their development. It is inevitable. What happens, then, when hastily thrown together papers begin to receive poor evaluations from teachers? Does the child start to believe that the work in the advanced class is too hard and they just cannot do it? Do they stop trying to compete with their peers because they believe they are not smart enough? This situation may cause the child to withdraw from some social interactions and to be more emotional and irritable with parents, teachers, and peers. In addition to a perceived loss of social status, this situation frustrates the need of the Well-Adjusted child to meet the expectations of others. Some will learn new skills as they cope with the increased competition, while others will struggle and begin to perform at a level that is below their capabilities; in the language of educators, they will be under-achievers. Let us consider the example of Tom Brokaw, the famous television journalist who related in his autobiography (A Long Way from Home: Growing up in the American Heartland) a common story of under-achievement. Brokaw enjoyed great success in elementary and secondary school. In high school he was a leader of his debate club, the starting guard on the varsity basketball team, and president of the student council. He was selected by his teachers to attend Boys State, a summer experience designed to train future leaders in state government. While there, he was elected Governor (the highest office) by his peers. He clearly had the characteristics of what we have labeled the Well-Adjusted behavioral profile. After high school graduation, he enrolled in The University of Iowa, one of the most rigorous institutions of learning in the upper-mid west. However, Tom came from a small town (Yankton, South Dakota). He was unprepared for the competence of his peers and the academic challenges he faced at The University of Iowa. Further, his parents had not gone to college so they were not able to prepare him for this environment. He reports meeting students from Chicago who seemed much more self-confident, assertive, and competent than he had ever encountered before. His response to this situation was less than optimal. “Instead of hitting the books with dedication, I cruised the student union, bedazzled by the rich population of fetching coeds, especially those from the moneyed suburbs of North Chicago.… I drank beer

5.3  Behavioral Risks

57

and played pool and seldom studied for an exam until the night before, if then” (Brokaw, 2002, pp. 215–216). After experiencing very poor academic success, he was transferred to the University of South Dakota, but continued the same behavior pattern in this new environment. One of his professors finally sat him down and said, “I think you should drop out of school. Get all of this wine, women, and song out of your system and come back when you’ll do yourself some good, your parents, and this school” (p.  218). (I am thankful to Tom Hebert for supplying this example) [13]. Tom Brokaw did begin to get his academic life in order, and engaged in a professional life that was exemplary. The new behavior pattern of consistent hard work and achievement was probably brought about by a range of factors, including a supportive wife, the realization of new responsibilities, and the accumulation of life experiences. He exemplified one of the principles set forth at the beginning of this chapter; that is, often individuals bounce back to their characteristic behavior profile after it has been altered by some environmental perturbation. Thus, some part of Brokaw’s success can also be attributed to his innate talents for learning, social interactions, and leadership that were first seen in middle childhood and adolescence. If conditions had not been positive and supportive in his young adulthood, he may have taken a different path. In a recent interview, the singer Michael Buble attributed his success to 10% talent, 40% hard work, and 50% luck. While it is impossible in hindsight for anyone to say why their life has progressed as it has, those who come to social prominence have done so based on a combination of talents (including a variety of behavioral propensities), social supports, hard work, and random effects we think of a luck. In addition to environmental changes, such as increased task demands, many Well-Adjusted children also run a risk that is a consequence of their empathy for others and their desire to please others. One such risk occurs when an empathetic child feels they must hide their ability so that they do not make others feel uncomfortable. A recent television interview of a child (about aged 10) was revealing. He told his parents he did not want to play competitive sports anymore because winning made him uncomfortable. When asked why, he said he felt bad for the children who lost. This particular child who seemed to have many characteristics typical of the of the Well-Adjusted cluster came from a family of collegiate and professional football players, was large for his age, and was considered an outstanding athlete by his coaches. But dominating others physically is a necessary part of football and consistently winning as his team tended to do, simply was at odds with his social sensitivity and empathy for other children. The first author of this book had a personal experience involving one of his most promising doctoral students who met all the characteristics of the Well-Adjusted profile. She had been a leader in high school, achieved at the highest level in her classes, and was universally liked by her peers and by the faculty. After completing most of her work on a dissertation, she came into my office one day to notify me that she did not plan to finish her dissertation and was withdrawing from the program. I was astonished as her progress had been consistent and rapid up to that point. When asked what led to this decision, after some hesitation she said that her husband had

58

5  Fostering the Development of Well-Adjusted Children

not finished his doctoral work at a nearby university and she worried that if she completed her dissertation it would make him feel inadequate and insecure. She viewed this decision as placing her marriage ahead of her career. A third kind of problem that some Well-Adjusted children are particularly prone to is also related to their social skill and desire to please others. Being the center of attention is wonderful and many Well-Adjusted children have experienced the “spot light” of peer attention. They are often admired. Their desire to please others can lead to self-limiting behaviors through a tendency to do what the other popular students do. This can include alcohol, drugs, petty theft, driving too fast, or risky sexual exploration. While these are likely to be short-term episodic behaviors, some can lead to problems that create further long-term complications. For example, illegal drug use, if it results in addiction or a police record can create serious impediments to adjustment and achievement. For some children exhibiting the Well-Adjusted profile, there is a different kind of risk based on the desire to please. Adolescence is a time of identity formation which necessarily requires breaking away in some way from family and community expectations. The strong tendency toward compliance of children in the Well-­ Adjusted cluster can lead to being overly constrained by family concerns and expectations, or by spending a great deal of energy in an attempt to meet family and community expectations. Finding one’s own way in life is a necessary part of an optimal developmental path. This necessarily involves not doing all the things that the family, school, or community expect of you. To be overly limited by these expectations stunts the psychological and social growth of any individual.

5.4  The Gifted-Creative Child Issue There is a vast research literature on the characteristics of gifted children. Gifted children are typically defined as being above the 90th percentile (in some locations the 96th percentile) on several measures of general intellectual ability and school achievement, as well as being above the 90th percentile on measures of intrinsic motivation to learn. A sizeable proportion of the children with the Well-Adjusted profile meet the criteria for specialized instruction in programs for gifted children. The behavioral characteristics of gifted children have often been researched. Hebert has summarized some of this literature and describes the social and emotional behaviors most often evidenced in gifted students as having (a) a perfectionistic tendency, (b) an internal motivation to learn, (c) unusual emotional sensitivity, (d) empathy for others, (e) a higher level of moral maturity than many peers of the same age, (f) a highly developed sense of humor, and (g) being resilient to adverse environmental events [13]. Almost all of these characteristics were directly measured in our research and were found to be characteristics of the Well-Adjusted child profile. What is noteworthy is that this profile of characteristics is recognized by parents in Russia and the United States, most of whom know nothing of the professional literature on gifted children. This profile was so coherent and distinct that,

5.4  The Gifted-Creative Child Issue

59

in the behavioral ratings of all children, it could be mathematically isolated and was strikingly similar across two cultural groups and across observers from different settings (US parents and teachers). Programs for gifted children range from simple offerings of more advanced or accelerated curricula, to integrated programs that focus on the psychological, social, and intellectual needs of this population. Regardless of the quality of the program, having a child placed in the “gifted program” often takes on additional meaning. For some parents, having your child classified as “gifted” is a symbol of their own social status. Others may feel that they must foster “unusual intellectual gifts” by enrolling them in specialized advanced tutoring, summer program, or even moving the family to a location where the schools are thought to have better programs for advanced students. The child may find that a gifted placement results in even more homework than they already had. It may also increase the demands on the child, because now the child has to compete for recognition against a highly selected and talented group of children. Thus, the placement in gifted education programs can place stressful demands on families and on the child. The other side effect of “gifted placements” is that many bright and psychological competent, self-regulated children who are assessed for placement in the program are not selected. Their intelligence test scores may simply be a few points below state- or school district-specified cut-off scores. This perceived “failure” to be recognized as gifted can create adverse effects on the child’s sense of self and may be particularly difficult for parents who are highly invested in the schooling success of their child. This can result in the parents seeking independent verification from other psychological examiners (sometimes referred to as “shopping” for the qualifying examination scores) to validate the parental view that the child is truly exceptional. The pressures on  parents who perceive their child as academically talented is sometimes based on the idea that the life course of the child depends on being officially recognized as gifted at an early age. This is an unjustified assumption. These children have talents that will, in all probability, play a lasting role in their development whether or not they are officially recognized as being gifted. Many studies of placements in highly selective schools at any level of education have shown that the performance of students when compared to the performance of similarly talented students in less selective schools is not different [14]. In the 1920s, Lewis Terman at Stanford University began a longitudinal study of gifted children which remains the longest running longitudinal study in the world. The sample was large (approximately 1500 children in 1928); the children were selected from the public schools of California primarily based on scores on the Stanford-Binet intelligence test. The sample was 90% white and the majority came from middle- or upper-class families. These children were not selected based on a broad assessment of behavior, but they did have high intelligence test scores. Terman and colleagues found in early follow-up studies that the children were generally well-integrated in their social groups in school, had few behavior problems, and

60

5  Fostering the Development of Well-Adjusted Children

achieved at a higher level in school than their peers. As was the case in our research, these gifted children were even physically taller than their peers on average. One of the most notable findings from Terman’s research was that his subjects did not in general become the movers and shakers of American society. Long-term follow-up research found that a few of the children became eminent scholars or successful entrepreneurs; however, the majority lived more typical, and some would say, mundane lives. Many had jobs that would be considered working-class by today’s standards (e.g., policeman, secretaries). Some studies found that Terman’s children in adulthood did about as well as would be expected from a random sampling of children with similar middle- and upper-middle class backgrounds. What is the lesson for our Well-Adjusted children? We, as a society, may place expectations on these children that are unjustified. Children like those who exhibit the Well-Adjusted profile early in life are the subject of achievement desires of their parents, school, and communities. We all want to bask in the glory that these children seem bound to achieve. However, life-time achievement depends on many factors, including good fortune (luck), good health, strong family and community support, supportive life partners, an economy that fosters opportunity, access to opportunity that is not limited by racial/ethnic prejudice, and many others. Perhaps other factors are at work as well, particularly those that foster a consistent desire to achieve societally sanctioned goals. Achievement that is recognized by the broad community depends on having high levels of motivation or drive. This drive can come from a desire to escape poverty, a desire to prove you are as good as the best in a given field, or the motive to overcome a personal difficulty. A recent book by Weiner [15] on the nature of creativity and genius points out that many of the most creative individuals seem to be motivated by a difficult childhood (loss of a parent, abusive parent), a psychological disability (a learning disability), or being a member of a social group that has experienced being marginalized. Highly creative persons are also typically perceived by their peers as “a little different.” “Eccentric, barefoot, and endearingly stubborn, Socrates occupied that precarious position that all geniuses do—perched between insider and outsider. Far enough outside the mainstream to see the world through fresh eyes yet close enough so that those fresh insights resonate with others” (Weiner, 2016, p. 22). Weiner points out that Edison was partially deaf; Beethoven was socially awkward, slovenly, and hard of hearing; Alexander Graham Bell and Picasso were dyslexic; and Michelangelo was afflicted with several illnesses and was in great pain when he painted the Sistine Chapel. Weiner postulates that socially comfortable people seldom become eminent through making a new discovery or creating a new product. Could it be, then, that most children who exhibit our Well-Adjusted profiles are so socially comfortable, having been admired for their behavior and intellect since childhood, and having a strong desire to please others, that these gifts lead to a comfortable life, but inhibit creativity at the highest level? The jury is still out on this proposition and far more long-term longitudinal studies of development are needed to clearly answer such questions.

5.5  A Personal Postscript

61

5.5  A Personal Postscript To illustrate some of the points made about well-adjusted children, I (RM) will provide a brief summary of my own developmental history, of course with the caveat that it is based on self-reflections with all the possible errors that such self-­reflections are prone to. My temperamental and academic behavior during middle childhood had some of the characteristics of the Well-Adjusted Profile, although not all. In particular, I performed academically at a somewhat lower level than the average child in the Well-Adjusted cluster, and had more difficulty focusing attention than many of my peers. My teachers reported to my parents that I was under-­achieving. Like Tom Brokaw (the famous TV broadcaster mentioned above), I was a product of small-town life and working-class parents, neither of whom had attended college. During my elementary school years, my parents moved from a rural, tiny farming community in New Mexico to Los Alamos, the home of the research laboratory that produced the first atomic bomb. My mother worked in the local hospital as a nurse and my father, initially, as an ambulance driver. Many children in my new school were the sons and daughters of renowned scientists, and I soon learned I was not the brightest child in the class. Nevertheless, I was an A or B student in most classes. In high school, I continued this level of performance although in more demanding classes that required persistent effort and prolonged attention (mathematics, foreign languages), I performed at a lower level. I found some success in music and by my junior year in high school was selected as class president and was nominated for Boys State by my teachers (a summer program designed for student leaders to teach them about state and local government). Remember, that Tom Brokaw, was also given this honor. While there, a central activity was the “election of state and county officers.” I was elected mailman–a low status position–while Tom Brokow was elected Governor of his State, the highest position. Clearly, he had more political skill at an early age that I did. However, I was asked to perform (a violin solo) at the culminating event, so my peers and the adult organizers saw me as “the musician” I guess. I then attended a music conservatory upon graduation. I was completely unprepared for college life and for the dedication required to compete in a music conservatory. One of my roommates was the son of cellist of the Los Angeles Philharmonic and he had soloed with this orchestra while a teenager. Several had already had important “gigs” as jazz players in New  York clubs. Everyone seemed more skilled and more sophisticated than me. On the academic level, I floated along in the gentleman “B/C” range. My life in the conservatory was filled with self-perceptions of failure and the experience resulted in profound performance anxiety. I transferred after 1  year to my home state university leaving music to study psychology. My academic performance slowly improved, but it was not until I reached graduate school that I had learned the skills necessary to succeed in academic life, had the motivation to apply my skills, and had a supportive learning environment. My personal example may illustrate several points. First, many children who are perceived by peers and adults in middle childhood as “well adjusted” (in the manner

62

5  Fostering the Development of Well-Adjusted Children

described in this book) have difficulties when the demands of their environment change. My basic temperament and personality had not changed, but my behavior was affected by a changed social environment. Many observers would not have placed me in the Well-Adjusted cluster during my adolescence. Finally, I seemed to have returned to some equilibrium after experiencing success in graduate school. This positive adjustment was made possible under the helpful and supportive guidance of many fine teachers and colleagues. I have had a meaningful life as an academic; I find it a good fit for my temperamental profile. However, like many others who have been lucky enough to find “their niche,” it was due to some mix of social, emotional, and cognitive behavioral tendencies, effort, social support, and a generous portion of random events (luck).

References 1. McClowry, S. G. (2003). Your child’s unique temperament: Insights and strategies for responsive parenting. Champaign, IL: Research Press. 2. McClowry, S. G. (2014). Temperament-based elementary classroom management. Lanham, MD: Rowman & Littlefield. 3. McClowry, S.  G. (2016). Using what works: Elementary school classroom management. Lanham, MD: Rowman & Littlefield. 4. O’Connor, E. E., Cappella, E., McCormick, M. P., & McClowry, S. G. (2014). An examination of the efficacy of INSIGHTS in enhancing the academic and behavioral development of children in early grades. Journal of Educational Psychology, 106, 1156–1169. 5. Gardner, H. (1983; 2011). Frames of mind. New York: Basic Books. 6. Van Aken, M. A. G., & Asendorpf, J. B. (1997). Support by parents, classmates, friends, and siblings in preadolescence: Covariation and compensation across relationships. Journal of Social and Personal Relationships, 14, 79–93. 7. Shaffer, D.  R. (2000). Social and personality development (4th ed.). Belmont, CA: Wadsworth Pub. 8. Fonzi, A., Schneider, B.  H., Tani, F., & TomadAa, G. (1997). Predicting children friendship status from their dyadic interaction in structured situations of potential conflict. Child Development, 68, 496–506. 9. Lease, A.  M., Kwon, K., Lovelace, M., & Huang, H. (2020). Peer influence in elementary school: The importance of assessing the likeability of popular children. Journal of Genetic Psychology: Research and Theory on Human Development, 181(2–3), 95–110. 10. Schaefer, J. D., Caspi, A., Belsky, D. W., Harrington, H., Houts, R., Horwood, L. J., Hussong, A., Ramrakha, S., Poulton, R., & Moffitt, T. E. (2017). Enduring mental health: Prevalence and prediction. Journal of Abnormal Psychology, 126, 212–224. 11. Kochanska, G. (1997). Multiple pathways to conscience for children with different temperaments: From toddlerhood to age 5. Developmental Psychology, 33, 228–240. 12. Kochanska, G., Tjebkes, T.  L., & Forman, D.  R. (1998). Children’s emerging regulation of conduct: Restraint, compliance, and internalization from infancy to second year. Child Development, 69, 1378–1389. 13. Hebert, T. (2011). Understanding the social and emotional lives of gifted students. Waco, TX: Prufrock Press. 14. Plomin, R. (2018). Blueprint: How DNA makes us who we are. Cambridge, MA: Massachusetts Institute of Technology. 15. Weiner, E. (2016). The geography of genius. New York: Simon & Schuster.

Chapter 6

Average Children: Two Temperamental Profiles

6.1  Introduction The majority of my friends (RM) in middle childhood were “typical kids.” They did well enough in school, did not need special education assistance, got along with most of their peers, and did not cause parents and teachers any more distress than was typical of children in middle childhood. I had a friend, George (name has been changed), who I continued to be friendly with even into my college years. Throughout schooling, he achieved at the B/C level, with an occasional A. He was not always organized well enough or motivated to turn in his homework in a timely fashion, but his performance would not have been viewed as problematic by his teachers. They had far more difficult students to deal with. He seemed content with his academic performance. His goal seemed to be to do the required work but with no more effort than was necessary. He was not withdrawn or shy, and he had a number of good friends. In fact, he was a little more social than many of his peers, including me. His primary focus seemed to be playing card games and marbles (a fad back in the 1950s in my school). He was not an outstanding athlete but played schoolyard sports with some skill. He was generally a happy person. He went to college—I remember him as a good bridge player—but never finished his degree. He found work for a while doing manual labor for a large corporate entity and became a supervisor after a few years. Unfortunately, we lost track of one another after this time. He was a good friend, a comfortable companion, a nice guy. As I look back on the personality characteristics of my friend, George, in his middle childhood years he would have had many characteristics in common with the broad category of children that we identified as “Average/Typical.”

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_6

63

64

6  Average Children: Two Temperamental Profiles

6.2  Parent/Teacher Perceptions of Behavioral Characteristics We found many children in middle childhood who exhibited behavioral tendencies that were perceived by their parents and teachers as being average. Children who have this profile do not stand out as being unusually talented academically or unusually empathetic toward others. Their activity level seems about average for children their age, as is their self-regulation of negative emotions and their ability to focus attention. For research purposes, a profile was considered average if the mean scores of children exhibiting the profile were between one-half of a standard deviation below the mean and one-half of a standard deviation above the mean (from the 31st to the 69th percentile) on all characteristics. This broad category included 64% of all children in our large combined sample. (The combined sample included children whose behavior was rated by US parents, a sample rated by US teachers, and a sample rated by Russian parents. In total, it contained 2349 children.) However, our research indicated that there were two average profiles. Children in one profile are labeled “High Average Self-Regulators,” while the cluster of children exhibiting the other profile was labeled “Low Average Self-Regulators.” These brief labels do not capture all the characteristics of these two groups, but these labels emphasize one of the major factors in their profile. The High Average Self-­ Regulators have a greater ability to regulate their activity level, emotions, and attention than the Low Average Self-Regulators, although both sets of scores fell within the average range. The characteristics of these two groups are graphically presented in Fig. 6.1. The two profile types are mirror images of one another; that is, the High Average Self-Regulators are perceived by parents and teachers to have high average academic talent and achievement motivation. They also exhibit high average prosocial characteristics including expressions of happiness and empathy for others. The Low 80 70 60 50 40 30 20 10 0

Fig. 6.1  Two average behavioral profiles (in percentile form)

High Average SelfRegulators Low Average SelfRegulators

6.2  Parent/Teacher Perceptions of Behavioral Characteristics

65

Average Self-Regulators are perceived by their parents and teachers as having low average academic ability and as being less achievement motivated than their peers. Their prosocial behaviors are also marginally below average. With regard to expressions of irritable and antagonistic behavior, again the two typical types are mirror images of one another, with the High Average Self-Regulator cluster of children having less of a tendency to express anger or to be aggressive than their Low Average peers. One of the most distinctive characteristics of the Low Average Self-Regulator group is that they are viewed by parents and teachers as having higher levels of shyness (social inhibition and social withdrawal) than most of their peers; in contrast, the High Average Self-Regulators are perceived to have lower levels of shyness and social withdrawal. Self-regulation of attention and feelings of insecurity follow in a similar pattern. One goal of our research was to determine if both types of average profiles were identified from all three of our samples when analyzed separately. Figure 6.2 presents the mean scores on the eight behavioral measures for each sample (US parents, U.S. teachers, Russian parents) for the High Average Self-Regulator cluster. Figure  6.3 presents the profile for each rater group for the Low Average Self-­ Regulator cluster. Both graphs illustrate that parents from two cultures and teachers in the United States viewed these children as average on every measure; that is, the range of scores was roughly from the 30th to the 70th percentile. Further, these graphs illustrate that each rater group produced very similar profiles for each of these clusters of children. Analyses of ratings of US parents and teachers indicate similar scores for both profiles, whereas Russian parents had modestly lower ratings of prosocial behavior and activity level for the High Average Self-Regulators. Overall, these differences were not of practical importance, in fact, statistical analyses revealed that 80 70 60 50 40 30

U.S. Parents

20

Russian Parents U.S. Teachers

Fig. 6.2  Comparison of profiles of High Average Self-Regulators across three samples (data in percentile form)

66

6  Average Children: Two Temperamental Profiles

80 70 60 50 40 30

U.S. Parents

20

Russian Parents U.S. Teachers

Fig. 6.3  Comparison of profiles of Low Average Self-Regulators across three samples (data in percentile form)

these profiles were not significantly different so were combined to create universal sample profiles. While all three rater groups identified similar profiles for both clusters of average children, the percentage of children exhibiting these profiles varied. For Russian parents, the High Average Self-Regulator profile described 22.3% of that sample, which was similar to that of US teachers (20.6%). However, analysis of data from US parental ratings produced a High Average Self-Regulator profile that included 32.6% of the sample. For the Low Average Self-Regulator profile, the percentages were 40.1% (Russian parents), 37.1% (US parents), and 32.2% (US teachers). When all these groups were combined (universal sample), the percentage of the High Average group contained 23.4% of the sample and the Low Average cluster included 41.0%. It is noteworthy that in all samples, the percentage of children in the Low Average Self-Regulators was higher than for High Average Self-Regulators. Average groups have been isolated by other researchers; however, most other studies have been of preschool-age children. Two research groups have studied children in middle childhood and identified latent profiles similar to the average profiles found in our research. Kamphaus and colleagues [1–3] studied large samples (N > 2000) of children age 6 through 11 and found two groups of children that had some similarities to our High Average and Low Average clusters. The first, labeled Adapted (11% of their sample), had good social skills, average scores on most clinical scales, and low scores on attention problems. This group had some characteristics that were similar to the High Average Self-Regulated cluster. A second group isolated by Kamphaus and colleagues was labeled “Average” (17% of their sample) because they had near average scores on all personality and behavior problem measures. Scott et al. [4] studied approximately 800 twin pairs (1574 individuals) with a mean age of 7.4 years using a measure of temperament. They found four profile

6.3  Demographic Characteristics

67

types one of which was labeled “Regulated Typical Reactive” describing 33.9% of the sample. These children had near average scores on all 12 temperamental variables used to identify the profiles. This group looked much like the combination of our average groups.

6.3  Demographic Characteristics Analyses on all three samples (US parents, US teachers, Russian parents), separately and in combination, indicated that there were no differences in the percentage of boys and girls in the High Average Self-Regulator cluster or the Low Average Self-Regulator cluster. Further, within the US parent sample, there were no differences between mothers and fathers as respondents. In other words, mothers and fathers had similar views of the behavioral characteristics of children in this age group. However, the children exhibiting the two average profiles did differ for the US parent sample in the education level of the parents. Fifty-five percent of the High Average Self-Regulators had parents who had college degrees, whereas 43% of the Low Average Self-Regulators had parents with college degrees. There was no statistically significant difference between the two average profiles in the number of years of education of mothers for the Russian sample: however, the High Average Self-Regulator profile contained children whose mothers had approximately one more year of education than did mothers of children in the Low Average Self-­ Regulator cluster. Studies of children temperament and/or personality have seldom found ethnic or race differences. However, we were interested to see if the race/ethnicity composition of the two average clusters might be different when profiles were based on parent and teacher ratings. We studied this question in several ways. For the US parent sample, we asked parents to identify the best descriptor of their child’s race/ ethnicity and these were placed into five categories: African American (9.9%), Asian American (5.0%), European American (54.1%), Hispanic American (8.2%), and Other (22.8%; included a variety of smaller ethnic groups plus a larger group referred to as “mixed race” or “multiracial”). First, we compared the percentages in each average group to those in all other groups and found no significant differences in either of the average clusters. Second, we compared the ethnic/race composition of the two average clusters and again found no significant differences. Thus, in the view of US parents, there is no association with ethnicity/race for either of the two average profile types. For the US teacher sample from the State of Georgia, 27.6% were African American, 66.0% European American, and 6.4% were all other minority groups. The High Average Self-Regulator cluster was found to contain 20.4% of the African American children and 74.4% European American children. Thus, in the view of these teachers, a higher percentage of children of European American ethnicity were found to be in the High Average Self-Regulated cluster of children than would be expected. This difference was statistically significant. Interestingly, there was no

68

6  Average Children: Two Temperamental Profiles

effect of ethnicity/race on the percentage of children in the Low Average Self-­ Regulator cluster. Thus, teacher perceptions indicated that there was a small excess of European American children in the High Average Self-Regulator cluster, but the ethnicity/race of the Low Average Self-Regulator cluster reflected the percentages in the sample.

6.4  Academic Ability and Motivation Parents and teachers rated the academic ability and achievement motivation of children and students. As noted previously, High Average Self-Regulators were rated as somewhat above average in ability and motivation, whereas those exhibiting the Low Average Self-Regulator profile were perceived to be somewhat below average. These perceptions became part of the profile characteristics defining the two average clusters. The question addressed in this section is, “To what extent are the perceptions of parents and teachers related to the actual grades the children in these two clusters obtained?” Academic grade point averages were available for the Russian sample. Parents reported the grade point average (GPA) from their child’s most recent report card. Grades were assigned to Russian students on a 5-point scale with “5” indicating the highest level of achievement. The GPA of Russian students in the High and Low Average Self-Regulator clusters are presented in Fig. 6.4 as well as the GPA of all remaining students in the sample. Analyses indicated the achievement of males and females was statistically different.

4.6 4.4 4.2 4 Males

3.8

Females

3.6 3.4 3.2 High Average SelfRegulators

Low Average SelfRegulators

All Other Students

Fig. 6.4  Grade point average of children in the High Average and Low Average Self-Regulator clusters compared to all other students in the Russian sample

6.4  Academic Ability and Motivation

69

The average GPA of all children in the Russian sample was 3.97. Consistent with parental ratings of academic ability and motivation, the children in the High Average Self-Regulated cluster had a mean GPA that was moderately above the mean of the sample (4.25), and the Low Average Self-Regulated cluster had a mean GPA that was moderately below the mean of the sample (3.81). However, there were significant differences between boys and girls with boys achieving at a lower level for both groups and for all the remaining students in the sample. This kind of sex difference in academic performance has been found in most countries [5]. In sum, findings for GPA were closely related to the ratings of academic ability and achievement motivation provided by parents and teachers. Peer Ratings of Academic Ability  Students in the third through fifth grades in Georgia in Cohort A (451 students) nominated those classmates who they thought were academically talented and motivated to achieve. The specific questions were as follows: 1 . This person tries hard to do good school work (measure of motivation). 2. This person makes good grades, is smart, and usually knows the right answer. (measure of ability) The score given to each child was the number of nominations they received from all of their classmates taking into account the number of nominations possible (the number of classmates making nominations). In response to the question related to perceptions of ability, boys and girls described by the High Average Self-Regulator profile received very similar scores; they were both at the 56th percentile of all children in the sample. For the nominations related to achievement motivation, children in the High Average cluster received scores at the 57th percentile again with differences between boys and girls not being significant (see Fig. 6.5). It is clear that the perceptions of the peer groups did not reflect the sex differences in actual school grades. 60 55 50 45 High Average Self Regulators

40

Low Average Self Regulators

35 30 25 20

Academic Ability

Academic Movaon

Fig. 6.5  Comparison of peer perceptions of academic ability and motivation for High Average and Low Average Self-Regulators (data are in percentile form)

70

6  Average Children: Two Temperamental Profiles

How well did peers understand the academic ability and motivation level of their fellow students who had been determined to be in the Low Average Self-Regulation cluster? The short answer is that they had a good understanding of these characteristics. Academic ability scores based on peer nominations were at the 44th percentile and academic motivation scores were at the 37th percentile. Thus, for children within the average range, peers perceived a large difference between High Average and Low Average Self-Regulators in the areas of academic ability and motivation. These peer perceptions coincided very closely with expected values based on temperament profiles derived from parent and teacher perceptions. It is worth emphasizing that the profile assignment of each child was not known by peers. Self-Rated Academic Ability and Motivation  Ratings by students of their own academic ability and motivation were available for one cohort of Georgia students (Cohort B, 446 students). Children were asked, “If you were to list all of the students in your grade from worst to best in school work, where would you put yourself?” Also, children were asked, “How good would you be at learning something new in school?” Responses were “1” = one of the worst, “5” = one of the best. The responses to the two scores were summed to obtain the score for each individual. Further, since Russian data indicated that girls had stronger academic performance than boys, self-ratings were broken down by the sex of the child. These data are presented in Fig. 6.6. Comparing the graphs for the High Average (four bars to the left) and Low Average Self-Regulators (four bars to the right), it is clear that both for academic ability and achievement motivation, the children in the High Average Self-Regulator cluster had more positive self-perceptions than did children in the Low Average 70 65 60 55 50 45

Boys

40

Girls

35 30 25 20 High Average SR- High Average SR- Low Average SR- Low Average SR-Ability -Movaon -Ability -Movaon

Fig. 6.6  Comparison of self-rated academic ability and motivation of High Average and Low Average Self-Regulators by child gender

6.5  Prosocial Behavior, Compliance, and Likeability

71

cluster. Both boys and girls in the High Average Self-Regulator cluster described their academic ability as above average, with boys giving themselves slightly higher scores than girls. They also described their academic motivation as above average with girls perceiving that their motivation was higher than boys. Boys in the Low Average Self-Regulator cluster described their ability and motivation as considerably below average, while both ability and motivation scores for girls were near average. In summary, these children in the third through fifth grades had an understanding of their academic ability and achievement motivation that was highly consistent with the perception of their peers, as well as the perceptions of their parents and teachers. All rater sources (parents, teachers, peers, children themselves) recognized that these are typical children with respect to their abilities and motivations (i.e., scores are near average). Further, all raters had a similar understanding of the differentiation between the two average clusters on these characteristics. Self-­ perceptions of academic ability and motivation are related in complex ways to the sex of the child. Boys in the High Average Self-Regulation cluster overestimate their ability (when compared to teachers and peers), and girls in the Low Average Self-Regulated cluster overestimate their ability and motivation. (Details regarding these analyses can be found in Appendix F.)

6.5  Prosocial Behavior, Compliance, and Likeability Prosocial behaviors, as they were defined in our research, include a wide range of positive emotional reactions to other people, including caring and empathetic behavior. Thus, the highly prosocial child is viewed as often happy and joyful, as well as empathic about the frustrations and difficulties faced by others. Finally, prosocial children tend to be agreeable. They are cooperative with their peers and try to meet the expectations of their parents and teachers. A number of researchers [6, 7] have argued that a range of prosocial behaviors (including all of those studied in our research) can be thought of as part of a broad constellation of behaviors they label as “kindness.” This broad constellation includes positive attitudes, feelings, and behaviors toward others including empathy, generosity, and altruism. All these behaviors have characteristics that benefit other human beings. This constellation of behavior tendencies meets all criteria to be called a temperament. As reviewed by Knafo and Israel [7], there are large individual differences in this constellation that can be observed from childhood through adulthood. It is moderately stable across situations, and aspects of the constellation have been shown to be relatively stable across childhood. Prosocial behavior has been found to be related to biological mechanisms such as brain function indices [7]. Parents from the United States and Russia as well as teachers from the United States rated children on their tendency to express a positive mood and their tendency to be empathetic and caring of others. These ratings became part of the data that resulted in temperament profiles. With regard to children in the average

72

6  Average Children: Two Temperamental Profiles

category, both clusters (High and Low Average Self-Regulators) had prosocial ratings that were in the average range. However, children in the High Average Self-­ Regulator cluster had higher prosocial scores than those did those in the Low Average cluster. The question addressed in the following section is, “To what extent do these prosocial characteristics as reflected in the two average temperament profiles relate to other measures of prosocial behavior completed by parents, teachers, and same-­ age peers?” First, we consider compliance with adult expectations and rules. Parent, Teacher, and Self-Ratings of Compliance  US parents, US teachers, and Russian parents all provided ratings of how compliant children were with rules and expectations set down in the home and school. This measure can be thought of as somewhat broader than simple rule following. It is also a measure of social sensitivity, thoughtfulness, and a desire to please others. The compliance score obtained from parents and teachers was not included in the analyses used to identify temperament clusters. A measure assessing students’ self-perception of their own compliance with adult rules and expectations was also obtained. Mean compliance scores of children from both average clusters are summarized in Fig. 6.7. The data in this figure indicate that compliance scores differed substantially between the two average temperament types. High Average Self-Regulators had much higher scores based on parent and teacher perception than did Low Average Self-Regulators. The difference was smaller for student self-perception. Interestingly, compliance scores of boys and girls were not significantly different regardless of who provided the ratings. Overall, the ratings from parents, teachers, and the students themselves for compliance nicely paralleled the prosocial ratings embedded in the latent temperament profile model. 80 70 60 50

High Average SelfRegulators

40

Low Average SelfRegulators

30 20 U.S. Parents

Russian Parents

U.S. Teachers Student-Self Rang

Fig. 6.7  Comparison of compliance scores for High Average and Low Average Self-Regulators: parent, teacher, and self-ratings (data are in percentile form)

6.5  Prosocial Behavior, Compliance, and Likeability

73

Peer Perception of Positive Emotion and Empathy  Positive mood, one aspect of prosocial behavior, is often referred to as positive emotionality. It is measured by assessing the amount of joy, happiness, and contentment that is expressed by the child. It is an important component of the prosocial abilities of children in that positive emotion creates a positive social atmosphere. It is contagious. It tends to make others feel more positive emotion and, thus, makes the happy person more socially attractive. In our research, parents and teachers viewed children who exhibited high levels of positive emotion as also exhibiting high levels of empathy toward others. Thus, we expected that peer ratings of these two characteristics would also be similar. Peer-based positive emotion scores were higher for children in the High Average Self-Regulator cluster than for those in the Low Average Self-Regulator cluster. However, same-aged peers perceived girls in both clusters to express more positive emotion and empathy than boys. Gender differences for High Average Self-­ Regulators were particularly large (Fig. 6.8). Peer Perception of Likeability  Children in both cohorts of Georgia students were asked to nominate peers in their classroom (Cohort A) or in their grade (Cohort B) who they would like most to play with and children they would least like to play with. Each child’s score created by likeability (or “social preference”) was created by subtracting the number of “least liked” nominations from the number of “most liked” nominations. Likeability represents one concrete social manifestation of the prosocial characteristics of the child. Higher levels of social sensitivity, positive emotionality, and empathy should enhance peers’ desire to interact and play with them. The act of

80 70 60 50

Positive Emotion Empathy

40 30 20 High Average High Average Low Average Low Average SR--Boys SR--Girls SR--Boys SR--Girls

Fig. 6.8  Comparison of High Average and Low Average Self-Regulators by gender for the propensity to express positive emotion and empathy

74

6  Average Children: Two Temperamental Profiles

choosing a peer to spend time with is where the “rubber hit the road” when it comes to the social effects of child temperament. Consistent with expectations, the High Average cluster had higher peer nomination scores on likeability (mean score was the 58th percentile) than children in the Low Average Self-Regulator cluster (mean score was 44th percentile). Note that both mean scores were well within the average range, but clearly different from one another. In summary, those children who exhibited the High Average and Low Average Self-Regulator profiles were generally perceived to exhibit near average levels of prosocial behavior by their peers. This was true whether prosocial behavior was measured as compliance (desire to please others), positive emotionality, empathy, or likeability. Further, in all cases the parent and teacher ratings of compliance, and peer group measures of positive mood, empathy, and likeability indicated that the two types of average children have a different impact on the perceptions and behaviors of their peers. High Average Self-Regulators were more socially attractive to their peers than Low Average Self-Regulators.

6.6  Social Status Likeability is considered to be one aspect of social status by researchers—the affective regard that peers hold for the child. However, it is a common observation that individuals who engender high affective regard from others are not necessarily popular, which is an aspect of social status that taps into a person’s reputation. History is replete with strong leaders who were admired but not necessarily well-liked. In our research, the reputational aspects of children’s social status in middle childhood, as viewed by peers, was measured in three different ways. Students were asked to nominate children from their class (procedure used in cohort A) or from their grade (procedure used in cohort B) who had the following characteristics. 1. Which of your classmates are the most popular at school? Which are the least popular? (Least popular scores were subtracted from most popular scores.) 2. This person gets chosen by the others as the leader. Other people like to have this person in charge. 3. This person is really cool. Just about everybody in school knows this person. Children in the High Average Self-Regulation cluster were rated by their peers as being in the high average range in all three indices of social status (popularity, leadership ability, and standing out from the crowd—cool), whereas children in the Low Average Self-Regulation cluster are rated as having low average reputational status. The High Average cluster tended to have scores around the 58th percentile across all measures of reputational status, while scores for the Low Average cluster were near the 40th percentile. These results indicate that reputational status was closely related to parent/teacher temperament profile models.

6.7  Influence on Peers

75

6.7  Influence on Peers Children in middle childhood have considerable influence on the behaviors and attitudes of one another. While influence is related to social status, it is not the same construct [8]. Further, influence can be broken down into the specific arenas in which children influence one another. The question we addressed was: To what extent to do children in the High Average and Low Average Self-Regulation clusters have influence on their peers, and to what extent is the influence on peers different for these two clusters? Influence was measured based on nominations by each student in a classroom of those of their peers who were particularly influential in each of five areas of life: academics, sports, youth cultural trends (clothing, music, slang), make-believe games, and inappropriate behavior in the classroom (talking back to the teacher, breaking rules if the teacher is out of the room). The influence that children in one cohort of Georgia had on their peers was quite different from the influence in the other cohort (see Figs. 6.9 and 6.10 below). In Cohort A, children in the High Average and Low Average Self-Regulation clusters had less influence than most children in the sample; that is, most of the influence scores were below the 50th percentile. The only exception was that children in the Low Average Self-Regulator cluster had above average influence on the inappropriate behavior of peers. Their influence was in the direction of increasing rule-­breaking behavior by others. However, inspection of Fig. 6.11 indicates that in Cohort B, High Average Self-­ Regulators were considerably more influential in most areas than the typical child in the cohort (i.e., had scores greater than the 50th percentile), and specifically were more influential than the Low Average Self-Regulator peers. The exception was in

70 65 60 55 50 45 40 35 30 25 20

High Average SelfRegulators Low Average SelfRegulators

Fig. 6.9  Peer-rated influence of High Average and Low Average Self-Regulators in five areas of school-based behavior (Georgia Students, Cohort A, data in percentile form)

76

6  Average Children: Two Temperamental Profiles

70 65 60 55 50 45 40 35 30 25 20

High Average SelfRegulators Low Average SelfRegulators

Fig. 6.10  Peer-rated influence of High Average and Low Average Self-Regulators in five areas of school-based behavior (Georgia Students, Cohort B, data in percentile form)

60 55 50 45 40 35 30 25 20

High Average SelfRegulators Low Average SelfRegulators

Fig. 6.11  Comparison of parent-rated behavior problems scores of High Average and Low Average Self-Regulators (Russian sample, data in Percentile Form)

the area of inappropriate behavior where both High Average and Low Average Self-­ Regulators had less influence than most children. The differences in the peer influence for these two cohorts of students indicate that influence on one’s peers is strongly affected by the nature of the social group. Schools attended by children in Cohort A served a somewhat less privileged population than schools in Cohort B.  Cohort B schools also had a larger majority of European American children. In these Cohort B schools, High Average Self-­ Regulators had more influence on their peers than they did in Cohort A.

6.8  Behavior Problems

77

6.8  Behavior Problems We expected that parental ratings of the mental health and behavior problems exhibited by these two groups of average students would be nearly average with the High Average Self-Regulators having fewer problems than their Low Average Self-­ Regulators. On a commonly used measure of mental health and behavior problems (for details see Appendix K), scores were obtained on the Russian sample for four types of problems (Hyperactivity, Emotional, Conduct, and Problems with Peers). Emotional problems are an index of depressive and anxiety symptoms, and conduct problems are an index of rule-breaking, antagonistic, and aggressive behaviors. Analyses of ratings of behavior problems by Russian parents were consistent with expectations (see Fig.  6.11). There was a sizeable difference between the behavior problem symptoms exhibited by High Average and Low Average Self-­ Regulators. High Average children exhibited very low scores in all areas, whereas Low Average Self-Regulators obtained scores marginally above average. Another aspect of our research was to determine the extent to which children in these two groups had significant mental health and behavioral problems. A significant behavior problem was defined as being at least one standard deviation above the mean of all children studied (i.e., having a score greater than the 84th percentile). This level of problems does not necessarily indicate that the child would receive a clinical diagnosis, but it does indicate that the child exhibited problematic behaviors at a much higher frequency than their peers. Table 6.1 compares the percentage of children who had significant mental health and behavior problems for the two average temperament profiles. These data illustrate that the High Average Self-Regulator cluster included a much smaller percentage of members with significant problematic behaviors than the Low Average Self-Regulator cluster. The High Average group had about the same percentage of problems as children in the Well-Adjusted cluster (see Chap. 4). The percentage of Low Average Self-Regulated children exhibiting significant behavior problems was about at the level that was typical of all children in the sample. In order to examine this issue of mental health problems more closely, measurements of behavior problems were available from a sample of US teachers, student peers, and the children themselves. One measure that was obtained from US teachers (for children in Cohort A) was a 10-item questionnaire that asked about each child’s aggressive behavior including social aggression (e.g., attempts to exclude other children from their social group), verbal aggression (e.g., saying unkind things to another child), and physical aggression (e.g., pushing or hitting another child). Although all types of aggression were included, the measure was primarily focused on social and verbal aggression. Also, student peers in both cohorts of Georgia students were asked to nominate children from the classroom (Cohort A) or from their grade (Cohort B) who exhibited social, verbal, or physical aggression. Finally, self-­ ratings were available on two aspects of conduct problems. One was the tendency to engage in aggression, and the other was the motivation to engage in behavior that broke classroom rules. In addition, self-ratings on a standardized measure of

78

6  Average Children: Two Temperamental Profiles

Table 6.1  Percentage of children in the two average clusters who are perceived to have significant behavior problems Raters Georgia teachers  Cohort A Student peers  Cohort A  Cohort B  Cohort B  Cohort A Self-ratings  Cohort A  Cohort B  Cohort A  Cohort A  Cohort A  Cohort A

Problem type

Percent with significant problemsa High averageb Low averageb

General aggressionc

16.4

19.5

Socially aggressived Verbally aggressived Bullying othersd Shy/socially inhibitedd

4.5 11.9 4.0 13.6

20.6 12.7 9.6 16.4

General aggressione Motivation for rule breakingg Anxietyf Depressionf Interpersonal problemsf School problemsf

14.5 13.9 21.0 26.9 3.9 26.4

20.6 22.6 17.7 22.0 9.8 29.7

Significant behavior problems were defined as scores at or greater than the 84th percentile for the sample of children studied (>1 s.d.) b Children exhibiting the High Average Self-Regulation and Low Average Self-Regulation profiles based on parent and teacher assessments of child temperament and personality c A 10-item rating scale addressing social, verbal, and physical aggression (see Appendix L) d A score based on number of peer nominations for each child in the sample (see Appendix L) e A self-rating score addressing social, verbal, and physical aggression (see Appendix M) f Behavior Assessment System for Children-Self-Report Form (for details see Appendix M) g One item measuring motivation to engage in inappropriate behavior in the classroom (for details see Appendix N) a

anxiety, depression, interpersonal problems, and school problems was obtained (for details see Appendix M). Teacher-, peer-, and self-rated behavior problem data are summarized in Table 6.1. To interpret this table, it is necessary to remember that 16% of all students in our Georgia samples would be expected to have significant problems in any of these areas. (This percentage is simply the result of using the 84th percentile—one standard deviation above the mean) as the definition of a significant problem.) Data in Table  6.1 reveal that, for many problems, the percentages for both the High Average Self-Regulators and the Low Average Self-Regulators were quite close to this 16% figure. This indicated that the number of children in these clusters who exhibited significant problems was typical of their age group. However, there was a general tendency for the Low Average Self-Regulators to have higher levels of significant problems than the High Average Self-regulators. In particular, they had higher rates of social aggression (e.g., excluding others from their social group). Several other trends are noticeable. First, High Average Self-Regulators consistently had lower aggression scores as perceived by peers than was typical of the whole sample of students. However, the self-ratings of aggression were close to

References

79

typical levels (16%). Second, both the High Average and Low Average Self-­ Regulators indicated that they experienced higher rates of depression and school-­ related problems than their peers. Finally, both groups had below average rates of self-rated interpersonal problems. This indicates that these children felt they had good relationships with their peers.

References 1. DiStefano, C., Kamphaus, R. W., Horne, A. M., & Winsor, A. P. (2005). Behavioral adjustment in the U.S. Elementary school: Cross-validation of a person-oriented typology of risk. Journal of Psychoeducational Assessment, 21, 338–357. 2. Kamphaus, R.  W., Huberty, C.  J., DiStefano, C., & Petoskey, M.  D. (1997). A typology of teacher-rated child behavior for a national U.S. sample. Journal of Abnormal Child Psychology, 25, 453–463. 3. Kamphaus, R.  W., Petosky, M.  D., Cody, A.  H., Rowe, E.  W., Huberty, C.  J., & Reynolds, C. R. (1999). A typology of parent rated child behavior for a national U.S. sample. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 40, 1–10. 4. Scott, B. G., Lemery-Chalfant, K., Clifford, S., Tein, J. Y., Stoll, R., & Goldsmith, H. H. (2016). A twin factor mixture modeling approach to childhood temperament: Differential heritability. Child Development, 87, 1940–1955. 5. Spinath, B., Eckert, C., & Steinmayr, R. (2014). Gender differences in school success: What are the roles of students’ intelligence, personality and motivation. Educational Research, 56, 230–243. 6. Eisenberg, N., Guthrie, I.  K., Murphy, B.  C., Shepard, S.  A., Cumberland, A., & Carlo, G. (1999). Consistency and development of prosocial dispositions: A longitudinal study. Child Development, 70, 1360–1374. 7. Knafo, A., & Israel, S. (2012). Empathy, prosocial behavior, and other aspects of kindness. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 168–182). New York: Guilford. 8. Lease, A.  M., Kwon, K., Lovelace, M., & Huang, H. (2020). Peer influence in elementary school: The importance of assessing the likeability of popular children. Journal of Genetic Psychology: Research and Theory on Human Development, 181(2–3), 95–110.

Chapter 7

Fostering the Development of Average Children

7.1  Introduction We should not be surprised that a large percentage of children in any age group have average academic ability, average achievement motivation, average levels of empathy, positive emotion, and social inhibition (shyness) and average capacity for self-­ regulation of emotion, motor activity, and attentional focusing. In large samples, measurements of most characteristics result in a bell-shaped or normal curve with few children at the extremes and most in the middle ranges. There are strong biological and societal mechanisms that tend to produce large numbers of typical children. Since all the behaviors we measured in our research have been shown to be influenced by genetic factors, and because these predispositions are influenced by 100s (in some cases 1000s) of genes, the behaviors will be normally distributed. By definition, this means that most children will be in the average range (see Chap. 13 for more detail.) Many environmental factors also tend to influence children to exhibit behaviors in the average range. For example, most of us are average or “good enough” parents [1]. We learn how to parent based on broad cultural norms as well as more specific norms from subcultures based on ethnicity/race or socioeconomic status. Thus, physical abuse of children is statistically rare (although not rare enough), and optimal parenting that maximizes the cognitive, social, and emotion potential of the child is also rare. Most parents make mistakes out of ignorance of what is optimal practice because their understanding of children and how to manage their child’s behavior are simply mistaken. Further, their own temperamental and personality characteristics get in the way of being the parent that they want to be, should be, or could be under ideal circumstances. Their tendencies to impulsive expressions of anger, for example, limit their ability to think in the moment about what is best for the child. Depression, alcoholism, drug abuse, and even normal personality differences of parents can have detrimental effects on parenting behavior. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_7

81

82

7  Fostering the Development of Average Children

A similar analysis can be made of teaching. In any given decade, there is a national norm for what constitutes “best practices” for teaching, and most teachers approximate this cultural norm. Some are exceptional, with innate talent for connecting to the lives of children and through diligent study of teaching methods. Some perform at a less than professional level due to poor training, being professionally “burned out,” health problems, or their own psychological makeup. Learning at school is about more than reading, writing, and arithmetic. It is a place where children learn how to behave based on observing how their peers behave [2]. Further, in middle childhood, there is strong social pressure to conform. This creates a situation where most children attempt to exhibit similar patterns of behavior. All these social processes affect the way children learn to cope with the academic and social world and tend to produce a modal or typical behavior pattern. There is another reason that children are placed in the average range on many behavioral characteristics. Parents vary in sensitivity to the social and emotional characteristics of their children. Some parents are tuned into the behavioral tendencies and moods of their children as well as their typical behavioral transgressions (e.g., aggression toward peers), for example, while others parents are not. Alternatively, they may notice emotional or behavioral tendencies but write these off with statements such as “that’s just the way children are.” These complex processes related to how parents and teachers make judgments about the behavior of children are not well understood. However, it seems logical that some children end up with average scores on behavior rating scales because parents and teachers are simply not aware of their behavioral tendencies and/or are not accurately evaluating the behaviors they observe. Both processes could lead parents to choose the average response option (the middle number on a rating scale) to most questions about the child’s behavior.

7.2  Effects of Being Perceived as Average While there are strong socialization pressures to conform to some modal behavior pattern, some children stand out from the crowd. They may have outstanding talents and, if these talents are supported by their parents and/or the community, they become positively affirmed outliers. Children gifted in athletics, the arts, or in academic pursuits are commonly observed examples. Some children are socially gifted, with sophisticated social skills, and become “social stars” in their peer group. On the other end of the continuum, those children who have social and emotional difficulties, or difficulties meeting the demands of schooling, also standout from the norm. Both of these groups of children in the context of the family, community, or the classroom are evaluated against a backdrop of what is perceived to be typical for the children their age. Children who are perceived as average are the background against which exceptional, statistically rare, children are the figure. Exceptional children stand out in the classroom and in the home, and, as a result, receive unusual levels of attention. It follows that average children get less attention

7.2  Effects of Being Perceived as Average

83

than do their more exceptional peers. Average children, as a group, are thought to require less specialized and intensive instructional or behavioral supports from parents and teachers. This is particularly evident in schools where children who have been officially designated as exceptional (either on the gifted/talented end of the spectrum or as disabled) receive more expensive individual attention in special classes or in specialized remedial activities. But even within a typical classroom, teacher attention is allocated differentially. Early in his career, the first author (RM) specifically addressed the question of what kind of student in a typical public school classroom gets the most attention from teachers. The focus of the study were in individual interactions that first-grade teachers had with students in their classroom. A large number of classrooms were studied (typically for one school day) and every interaction the teacher had with a student was coded. The research revealed that teachers focus their interactions disproportionately on two groups, children who were academically skilled, and children who the teacher worried were likely to cause disruptive behaviors in the classroom (most often boys). Children whose behavior and academic skill were in the average range simply obtained fewer individual interactions from their teachers than these exceptional children received [3]. Further, teachers were almost completely unaware that they were providing differential rates of interactions with children [4]. The majority of teachers thought they were treating each child the same and were giving similar attention to all children. This kind of difference in the focus of attention operates at the peer level as well. Average children are not the most admired, most well-known, or most popular children as has been documented in the research reported here and elsewhere. They are less likely to be thought of as influential in affecting the behavior of their peers in the areas of academics, sports, cultural trends, or make-believe games. They also are not the most socially isolated. They are the perceptual background in their peer group. What effect occurs when a child is perceived as the perceptual background. The psychological literature is very thin with regard to research on the developmental implications of being average in the eyes of one’s parents, teachers, and peers. The vast majority of research efforts have been on those that are distinct from average; that is, the gifted and talented, as well as the socially difficult and troubled. Society needs to know how to foster the intellectual gifts of those who are quick learners and are highly motivated by abstract concepts (e.g., mathematics). These leaders may provide advances in science, industry, and the arts that result in a better life for others. Also, all cultures have a vested interest in understanding the developmental pathways that lead to psychiatric difficulties, antisocial behavior, or an impaired ability to make a living. So a disproportionate amount of psychological research is devoted to these issues. Thus, the typical child is the perceptual background for researchers, and their funding agencies, too. Despite this relative lack of attention by researchers on average children, there is informative developmental theory and some research which provides guidelines for the understanding of children who are typical or average in their academic abilities and temperamental profile. One theoretical model seems particularly important. We

84

7  Fostering the Development of Average Children

have emphasized that all human beings come into the world with individual differences in emotion and the capacity to demonstrate talent as part of their genetic potential (DNA). But this genetic predisposition shapes and is shaped by a complex array of environmental factors, particularly in the social world. It may be that the role of the environment is different for the average child than for children at the extremes. Several theorists have posited that the role of the social environment differs depending on whether one is at the extremes of personality or around the middle of the distribution [5, 6]. Specifically, it is theorized that children at the extremes of the distribution of a behavior in any characteristic tend to create an environment that fosters that characteristic behavior. For example, in the study reported above regarding interactions between teachers and individual students, exceptional children (both talented and disruptive) tended to demand attention from the teacher. Thus, these children co-created their own environment related to teacher attention. For the talented students, they often elicit an unusually supportive environment in which they were praised or encouraged to show their talents. For the disruptive children, an environment is created that is replete with high levels of adult disapproval and behavioral corrections. Exceptionally talented child in the areas of the arts or sports also play a role in how their own environment is structured. People single them out for praise and put them in positions to exercise and build on those talents. This can occur for temperamental differences as well. The unusually caring and empathetic child in middle childhood might be offered more chances to baby sit the neighbors’ children or pets, or take on other similar positions. In this way, the empathic tendency leads to opportunities to learn further caring-related behavior. As is the case in economics, the rich (talented children) get richer. On the other hand, children who are often of negative mood, are belligerent, and argumentative, are often socially isolated or are the object of social or even physical aggression. These children begin to see the world as a threatening place. In fact, it is more threatening for them. Their behavior has helped to create an environment in which other people cannot be trusted to be caring, understanding, and supportive. The theory posits that the average or more typical child does not have the same kind of social impact, so they do not “make” an environment that fosters the development of their skills. This point of view leads to the conclusion that, for average children, the environment created by parents and teachers play an outsized role in creating the environment that average children live in. They are more influenced by the parenting and schooling environment than are their exceptional peers. They are more dependent on the instructor, their friends, and their family to foster their developmental path, both academically and socially. If the family environment is particularly conducive to learning to read and to discussing books, the child in the middle range of talent in the areas of language and reading will be particularly helped by this environment. A teacher who models self-regulation of negative emotion (e.g., anger, frustration) as she deals with difficult behaviors in the classroom helps the average child develop emotional self-regulation skills perhaps more so than peers with high or low levels of self-regulation. Those who already possess excellent emotional self-regulation skills will learn less from this modeling experience, and those who are highly prone to anger may also gain less from this teacher.

7.3  Developmental Assets and Risks

85

7.3  Developmental Assets and Risks Up to this point, we have been describing the children who exhibit average behavior tendencies and average academic talent as if all these children fit one temperamental profile. However, as we have seen, within the average range there is considerable variation in temperamental tendencies and talent. In all three samples that we have studied (children assessed by US parents, Russian parents, US teachers), two distinct groups of average children were consistently found. One group was perceived to have moderately above average academic performance and motivation, and moderately below average tendencies toward irritability, aggression, attention problems, as well as problems with shyness and social isolation (i.e., the High Average Self-­Regulator cluster). A second group (Low Average Self-Regulators) was identified that had low average academic performance and motivation, and had moderately more difficulties regulating their attention, irritability, and antagonistic tendencies. They also had higher levels of social isolation. While all these tendencies were within the average range, these two groups in the middle school years lived in two distinct social worlds. High Average Self-Regulators, for the most part, were considered psychologically and socially healthy, typical children. They are more well-liked by their peers, were perceived as happier, more empathetic than children in the Low Average Self-­ Regulated cluster; furthermore, they also have higher reputational regard as shown on multiple indicators of social status (e.g., popularity, leadership). Consistent with what has been said above regarding the influence of the environment on children in the middle range of personality characteristics, peer perceptions of children in the High Average group were substantially influenced by the characteristics of their peer group. In more middle-class schooling environments, the High Average Self-­Regulators enjoyed more social status and perceptions of likeability than in schooling environments in which there with a higher percentage of lower-middle class children. Their self-restraint and high-average academic ability might not be as valued and supported by their peers in the less academically oriented social world of the schools serving children from less advantaged homes. But all in all, High Average Self-Regulators had average to moderately above average peer relationships and status. Children who exhibit the High Average Self-Regulator profile have much in common with children exhibiting the Well-Adjusted High Achieving profile, although their behavior proclivities are perceived by parents, teachers, and their peers as less extreme. Therefore, they have many of the assets and liabilities associated with this more extreme cluster. On the positive side, they are generally well-­ liked and have many of the prosocial skills characteristic of Well-Adjusted High Achievers. This helps them obtain the parental, teacher, and peer supports that enhance their developmental progress. It is important to be aware that some children who are perceived to exhibit the High Average Self-Regulator profile actually have talents and skills that have not yet been fully developed or recognized. Children develop social, emotional, and academic skills at different rates. Some of these children who are perceived in middle childhood as average will blossom with maturity into Well-Adjusted High Achievers. This may be particularly true of children in lower socioeconomic

86

7  Fostering the Development of Average Children

circumstances who have not had the learning opportunities that have been given to more advantaged children. Further, all members of the child’s social group (adults as well as peers) are subject to prejudices related to ethnicity/race and the socioeconomic background of the child. Some teachers, for example, may have a biased view of the behavior of children who come to school in dirty or ill-fitting clothes due to preconceptions about children who live in poverty. In general, the developmental outlook for children who exhibit the High Average Self-Regulator profile is positive. It is helpful for parents and teachers to understand that these average children need extra support or special opportunities to fully develop their social, emotional, and learning capacities. Some suggestions for all average children are described below. Children in the Low Average Self-Regulation group have more difficulties in school than their High Average peers. In our research, they were perceived to have academic ability and motivation that was around the 40th percentile and this perception was consistent with the actual grades given in the Russian sample. Further, in the US sample, these children were perceived by their peers to have below average academic talent and motivation. They were, also, less likely to be chosen as play partners by peers and had a lower frequency of being nominated as popular or influential by other students. A careful look at the temperament profile of the children in the Low Average Self-Regulated cluster indicates that they are less energetic and more socially reticent than their peers. They also have attention problems. This profile has similar characteristics to what some prominent psychologists have labeled “sluggish cognitive tempo” [7]. While this label describes a more extreme form of these behaviors, a listing of the characteristics of children who fit the sluggish cognitive tempo (SCT) label is instructive for its similarity to the Low Average Self-Regulators. SCT children are described as having difficulty with (a) selective or focused attention, (b) are hypoactive, (c) daydreaming excessively, (d) having a slow working speed, (e) often staring into space, and (f) seeming to be in a mental fog. SCT has been the subject of controversy for decades. It is sometimes thought of as a type of attention deficit hyperactivity disorder (ADHD), but these children do not respond positively to stimulant medications as most children with ADHD do. Stimulants simply have little or no effect. Many criticize the diagnostic category (it was not included in current editions of the Diagnostic and Statistical Manual of the American Psychiatric Association or the International Classification of Disease) and rightly point out that there could be many causes for inattention, slow working speed, and low activity level. Whether characteristics of sluggish cognitive tempo apply to the children in the Low Average Self-Regulation group or not, schooling for this group is difficult. Since low average academic ability is combined with a low average ability to self-­ regulate  negative emotions and attention, children in the Low Average Self-­ Regulation cluster seems likely to find both the academic and social aspects of schooling to be challenging. For this group, school is something you do because you have to; it is not often an environment that creates excitement or joy. This seems likely to result in higher levels of dropping out of school and in lower probabilities of attending advanced academic training.

7.4  Interventions for Average Children

87

It is important to emphasize that in all three samples this was the largest single cluster of children, comprising 37.1 percent of all children in the universal sample (when all of our samples were combined). Thus, these children create a special challenge for school administrators and teachers. Their behavior and academic problems are not extreme enough to warrant special education interventions, yet they struggle more than other average children in an academic environment. Many of the children in this cluster might be those referred to by educators in the United States as “slow learners.” Slow learners are children who simply need more repetitions to learn new material. This can be caused by a somewhat slower speed of cognitive maturation. All children do not mature at the same rate. In some cases, a particular skill, such as learning to read, that is particularly difficult at one age will be relatively easy at a later age, enabling the child to catch up with peers. In other cases, children will continue to progress in their academic skills at a somewhat slower rate than peers. Academic ability is normally distributed, so some children will simply continue to take more repetitions and need more support to learn some material throughout the lifespan. This places more pressure on parents and teachers to provide skilled support of learning. For some children in this cluster, they might have a specific processing deficit that contributes to a learning disability, even though they have average or above average academic potential. These children may have a specific learning disability and will need professional help and perhaps specialized supports in school.

7.4  Interventions for Average Children Every child deserves the opportunity to feel competent and valued. Average children, particularly Low Average Self-Regulators, may have difficulty getting this opportunity in academic settings. For both profile types, many average children can find particularly meaningful growth opportunities in environments that are outside of school. Opportunities provided by faith-based institutions, by outdoor experiences and sports, as well as experiences in the arts (e.g., playing a musical instrument) can be very important for children in both average groups. By middle childhood, many children are capable of working alongside their parents and learning skills related to the parent’s work or hobbies. This type of experience is more difficult in the twenty-first century than previously, because so many parents work in environments that make observational learning and participation by children impossible. Further, the tasks would not be meaningful to children in middle childhood. However, there are family activities that are appropriate to children (e.g., hiking, gardening, carpentry, hunting) that can be shared and that will be meaningful for some in this age group, allowing them to experience a sense of accomplishment. Activities at school or home emphasizing group-oriented cooperative activities can also be helpful. Project-oriented activities in groups at school can be particularly useful not only to help with learning academic material—but to also help

88

7  Fostering the Development of Average Children

develop the full range of social and emotional skills required to succeed in life. The emphasis in all these activities should be on finding meaningful and supportive activities that allow average children to have their “moment in the sun.” The art of teaching any type of skills (social, emotional, cognitive, or motor skills) is for the parent or teacher to monitor the ability of the child to attend to the task as well as the speed at which the child is learning. For average children, the ability to remain attentive to learning materials is shorter and speed of learning is slower than is the case for children in the Well-Adapted Higher Achiever profile. Shorter exposures and breaking tasks down into smaller steps are required. However, within the average range there is a great deal of variability. Low Average Self-­ Regulators will require significantly longer exposures and smaller steps to learn the same skills than their peers who exhibit the High Average Self-Regulator profile. For teachers in most settings, one of the most difficult challenges is to individualize instruction enough that children at all levels of ability can profit from the instruction. In fact, given the factory system of education that predominates in the United States, it is difficult if not impossible. But simply understanding the profiles we have described can go a long way toward helping provide a cognitive structure for knowing how to individualize instruction. This is particularly important in the realm of emotional and behavioral regulation. Most beginning teachers have been responsibly well-trained in curriculum and cognitive instructional methods. Their training is much less adequate for understanding the emotional and behavior tendencies of their students. Like cognitive ability, the range of ability to inhibit impulsive expression of negative emotion, for example, in a typical fourth grade classroom is enormous. If the teacher can understand the contributors to these individual differences—and approach the learning of self-regulation like he would the learning of cognitive skills (small step for those that need it)—the outlook for more productive schooling experiences is improved.

References 1. Scarr, S. (1992). Developmental theories for the 1990s: Development and individual differences. Child Development, 63(1), 19. 2. Bukowski, W. M., Laursen, B., & Rubin, K. H. (Eds.). (2018). Handbook of peer interactions, relationships, and groups (2nd ed.). New York: Guilford. 3. Martin, R. P. (1972). Student sex and behavior as determinants of the type and frequency of student-teacher contacts. Journal of School Psychology, 10, 339–347. 4. Martin, R. P., & Keller, A. (1976). Teacher awareness of classroom dyadic interaction. Journal of School Psychology, 14, 47–55. 5. Bell, R.  Q. (1979). Parent, child, and reciprocal influences. American Psychologist, 34, 821–826. 6. Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype-environment effects. Child Development, 54, 424–435. 7. Barkley, R. A. (2014). Sluggish cognitive tempo (concentration deficit disorder?): Current status, future directions, and a plea to change the name. Journal of Abnormal Child Psychology, 42, 117–125.

Chapter 8

Shy and Socially Withdrawn Children: Two Temperament Profiles

8.1  Introduction Most people can remember a friend in middle childhood who was shy. I (RM) remember two children from my middle childhood years (grades 3 through 5) who were particularly socially withdrawn. Many of us in this age range were shy, but these two individuals stood out. One, I will call him Craig, was very “bookish,” what we might now refer to as a “geek.” He was brilliant, always obtaining grades near the top of the class, and particularly knowledgeable in the areas of mathematics and science. I think he was bored with much of his school work. He was socially awkward, and, although he had few friends, he seemed not to care about the social world. He seemed to spend most of his time reading. Karen was shy, insecure, and also did very well in school. However, she seemed to desire social interaction and had a wider range of friends than Craig. I guess, in retrospect, her most distinguishing characteristic was an insecure type of social anxiety. A relatively small percentage of children in middle childhood are as shy or socially withdrawn as Craig and Karen. There are multiple reasons for social withdrawal as exemplified by these two examples. One type of social withdrawal seems to be related to what some psychologists refer to as inhibition-to-the-unfamiliar. This type of inhibition describes a reluctance to engage strangers [1]. This form of social inhibition is often referred to as shyness. Shy children at a birthday party will often watch others from the periphery of the group for a while before gradually participating in activities and conversation. The degree to which this type is inhibited varies greatly. Some children, after a brief period of withdrawal, begin to initiate conversation and participate in activities. Others stay withdrawn from the group for a longer time, initiating interactions cautiously and slowly, but some never become comfortable in a situation like a birthday party.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_8

89

90

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

Shyness may involve social withdrawal due to inhibition in novel situations, but often may also be the result of the perception of being evaluated negatively. Thus, some children in middle childhood are reluctant to enter a particular group activity because they fear they will be rejected. Some might not join the pick-up basketball game because they fear their skills are not as strong as others in the group. Other children may be particularly aware of what clothes others wear and realize they do not fit in. This type of social withdrawal due to feelings of being negatively evaluated becomes more prevalent in middle childhood, as social comparison processes become more common and as the child becomes more aware of the social processes operating in their peer group. Inhibition-to-the-unfamiliar and shyness based on perceived social evaluation have much in common in that they are associated with anxiety. It has been suggested by some psychologists that shyness is a manifestation of an “approach-avoidance” conflict [2]. The approach motivation is the desire to interact with one’s peers, whereas the avoidance motivation is based on the fear that interacting will result in some form of rejection. My description of Karen described above best fits this type of social withdrawal. There is another group of children who seem not to be shy but are socially withdrawn. These children seem to be less motivated to be social than others simply because they are less interested in the social world in general. When observing children at play, some children are more interested in interacting with a particular electronic game than with a group playing a board game. They are not fearful or anxious about acceptance—they just prefer not to play with others and enjoy solitary activity. They seem unaware of being socially isolated and/or simply do not have a strong need to be accepted. My memory of my friend Craig in the fourth grade seemed to be an example of this type of socially withdrawn child. In our research one scale was constructed from items related to shyness and inhibition-to-the-unfamiliar, and another to the tendency to be simply socially withdrawn. The same latent profile structure was obtained when these scales were combined or were analyzed separately. This implies that parents and teachers are not discriminating between differing types of motivations for social withdrawal. Most parents and teachers appear to attribute all social withdrawal to temperamental shyness. For these reasons, scores for the two scales were combined in the analysis used to identify the temperamental profiles.

8.2  Parent/Teacher Perception of Temperamental Characteristics Based on parent and teacher assessments of the all eight behavioral tendencies of children in middle childhood, two profiles describing children who were withdrawn were found. Figure  8.1 presents the two profiles in percentile form (the average score of all children in the sample is at the 50th percentile). The data presented are based on the combined sample of ratings of child temperament from US and Russian

91

8.2  Parent/Teacher Perception of Temperamental Characteristics 100 90 80 70 60 50 40 30 20 10 0

Withdrawn High Achievers Withdrawn Low Achievers

Fig. 8.1  Two types of socially withdrawn profiles: universal sample (data in percentile form)

parents as well as US teachers (combined total of 2359 children). The two profiles have little in common other than the children exhibiting these profiles are much more socially withdrawn, insecure, and fearful than their peers. Interestingly, they also have low activity level scores. One personality profile was referred to as the Withdrawn High Achiever profile and describes a group of children who are high achieving and have very high achievement motivation. Compared to their peers, they were perceived as more empathetic toward others and as expressing more positive emotionality (happiness, joy). In addition, they were rated as having a low activity level indicating that they are less energetic and physically vigorous than their peers and less likely to engage in gross motor activity such as sports. This low activity score may have another meaning. Because these children are socially anxious, they may be less engaged in a particular social setting because they are trying to understand the social issues in that setting due to their excessive caution. Thus, low activity in social situations may be the result of obsessive rumination about how to act in the situation. Some of these children then may be quite physically active when playing on their own. Children in the Withdrawn High Achiever cluster were perceived by parents and teachers as having a lower tendency than their peers to express irritability or antagonism toward others. They are seldom argumentative. This may result from being particularly apprehensive about the social consequences of these actions. Additional characteristics of the Withdrawn High Achiever group include having excellent control of their attention processes and being more able than their peers to keep their belongings and activities organized. These children are most like Karen, the shy and anxious child I remember from my middle childhood. The second group that was identified from temperament ratings by parents and teachers will be referred to as Withdrawn Low Achievers. The children exhibiting this profile in middle childhood were perceived by parents and teachers to have low levels of academic ability and achievement motivation. In addition to being socially withdrawn, they are much more insecure and fearful than most other children of the

92

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

same age. Like their higher achieving but also withdrawn peers, they are less physically vigorous and energetic than other children their age. They are unlikely to enjoy participation in sports. In addition, these children have a number of behavioral characteristics that make them difficult for parents and teachers to manage. They have a range of problems related to poor self-regulation. For example, they have difficulty focusing their attention and are disorganized. Both of these tendencies serve to inhibit academic performance. They also were perceived to be less capable (or willing) to self-regulate their anger when frustrated, more often engage in rule-breaking behaviors at home and at school, and are more argumentative than their peers. These dysregulated behavioral tendencies make their profiles similar to the poorly self-­ regulated profiles (to be described later), with the exception that the Withdrawn Low Achievers have high levels of social withdrawal and insecurity/fearfulness. The Withdrawn High Achiever behavioral profile is descriptive of a small percentage of children in middle childhood. In our universal sample, 3.9% of the children were in this cluster. The Withdrawn Low Achiever profile also described a small percentage of children (5.1%). The percentages were very similar across the three samples studied, with percentages in both clusters ranging from 3% to 7% in all samples. It is noteworthy that parents and teachers did not isolate a group of children like my acquaintance Craig. Remember that he was socially isolated but was not particularly anxious or socially insecure. He was a brilliant student and simply did not seem to care about interacting very much with other children. This group of children may not have been isolated from the samples we studied because of the way that social withdrawal was measured. Also, this is likely to be a very small percentage of children in middle childhood. Further, it is possible that parents and teachers perceive that children who behave like Craig are socially anxious; thus, these children ended up in the Withdrawn High Achiever cluster. In Western cultures like the United States, being socially withdrawn has a negative connotation and many parents worry about these children perhaps attributing more social discomfort to some of the children than is warranted. Finally, it is possible that the measure of social isolation we used was not well equipped to identify this particular group. Comparisons of mean scores from US and Russian parents and US teachers for the Withdrawn High Achiever profile are presented in Fig. 8.2. The profiles for the US parents and teachers are very similar with teacher giving marginally lower social withdrawal and insecurity ratings. Russian parents had similar average (mean) ratings for these children, but indicated a marginally higher tendency toward irritable and antagonistic behavior, as well as attention problems. The mean scores on the eight cognitive and temperamental characteristics that defined the Withdrawn Low Achiever profile demonstrated more similarity across cultures and parent-teacher comparisons (see Fig. 8.3) than the Withdrawn High Achievers. For both profiles, it is easy to argue both statistically and practically that the behavioral tendencies described in the three samples represented the same patterns of child behavior. There is a large psychological literature on behavioral inhibition, shyness, and social withdrawal among children [3–5]. However, most of this research in done by studying one or a few characteristics in isolation (the so-called variable-centered approach to research). There have been very few studies in the childhood literature

8.2  Parent/Teacher Perception of Temperamental Characteristics

100 90 80 70 60 50 40 30 20 10 0

93

U.S. Parents Russian Parents U.S. Teachers

Fig. 8.2  Comparison of the Withdrawn High Achiever profile obtained from three independent samples (data in percentile form)

100 90 80 70 60 50 40 30 20 10 0

U.S. Parents Russian Parents U.S. Teachers

Fig. 8.3  Comparison of the Withdrawn Low Achiever profile obtained from three samples (data in percentile form)

that have investigated temperamental profiles based on a broad range of characteristics (the person-centered approach). One of the most influential investigations using a person-centered approach was conducted by Thomas and Chess [6], the founders of the modern movement to study temperament in children. They describe three temperament profiles one of which is labeled “Slow-to-Warm-Up.” These profiles were primarily formulated through clinical observation. The “Slow-to-Warm-Up” temperamental constellation of traits is marked by a combination of negative responses of mild intensity to new stimuli and a slow adaptation after repeated contact with new stimuli (new surroundings, new

94

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

toys, or unfamiliar people). These characteristics are clearly similar to those described in our research, although their cluster was not differentiated into subtypes as was the case in our research. This lack of differentiation probably resulted from the age of the children studied by Thomas and Chess who focused on infancy and the early childhood period. The group of children defined by this Slow-To-Warm-Up profile contained 15% of their sample, whereas in our research we found that both clusters of withdrawn children (Withdrawn High Achievers, Withdrawn Low Achievers) when combined accounted for approximately 10% of our universal sample. Many other researchers who have isolated temperament-based profiles of children in the preschool age range have also obtained one or two inhibited groups [7– 10]. Kamphaus and colleagues [11–13] published a series of studies of children in middle children aimed at developing and validating a typology of child behavior problems in children 6–11 years of age. They modeled nine latent profiles based on parental ratings and seven profiles based on teacher ratings. The parental ratings produced two clusters that are related to our two withdrawn clusters. One is labeled “Physical Complaints and Worry.” This group consisted of 8% of the sample and was characterized by high scores on the Anxiety and Somatization (children with many complaints and worries about their physical health). The second cluster labeled “Internalizing Problems” consisted of 9% of the sample. This group had significant elevations on Anxiety, Depression, and Social Withdrawal. They also had difficulties in adapting to new environments. Finally, Scott [14] studied the temperamental profiles of a large sample of twin pairs (mean age of 7). Using 12 different temperament/personality characteristics, they isolated four profiles, one of which was labeled “Dysregulated, Negative Reactive.” This cluster of children was most clearly defined by high levels of fear, shyness, sadness, and inhibitory control (extent to which the child self-regulated negative emotion). This cluster of children also had high scores on the anger scale— suggesting difficulty self-regulating strong anger-related emotions—which was characteristic of the Withdrawn Low Achieving group in our study. In summary, researchers who have studied temperament or behavior problem profiles of children have obtained one or more profiles that were primarily characterized by fearfulness, insecurity, and shyness. The number of children identified who exhibit these characteristics as well as the percentage of the sample exhibiting this pattern has varied due to differences in the traits studied, the age of the children, and the statistical methods used to obtain the latent profiles. Because no prior research team included academic ability and achievement motivation as part of the characteristics they studied, no other group has isolated the two withdrawn profiles that were isolated in our research.

8.3  Demographic Characteristics Analyses of our three samples (US parents, US teachers, Russian parents) separately and in combination indicated that there were no statistically significant differences in the number of boys and girls in middle childhood exhibiting the two

8.4  Academic Ability and Motivation

95

withdrawn temperament profiles. As an indication of socioeconomic status, the education level of parents was available for the US parents and Russian parents. We found no meaningful differences in parental education for either of the withdrawn profiles. Ethnicity/race data were available from US teachers and US parents, and neither group produced statistical or practically meaningful differences. Thus, socioeconomic, ethnicity/race, and child gender played no statistically significant role in the composition of the cluster of children who exhibited the Withdrawn High Achieving and Withdrawn Low Achieving profiles. However, the number of children in each of the withdrawn clusters is small, which limited the ability to find statistically significant effects. It is possible that some differences would emerge if larger samples had been studied.

8.4  Academic Ability and Motivation Parents and teachers provided data relevant to their perception of academic ability and achievement motivation. These ratings contributed to the latent profiles defined as Withdrawn High Achievers and Withdrawn Low Achievers. Children exhibiting the Withdrawn High Achiever profile were perceived by parents and teachers as exhibiting a high level of academic ability as well as very high level of achievement motivation. In contrast, children who exhibited the Withdrawn Low Achiever profile were perceived as exhibiting very low levels of academic ability and achievement motivation. The scores of this latter group were lower than for any other clusters of children we identified. In order to determine how the actual grade point average (GPA) of these two groups of children was associated with these behavioral profiles, available data from the Russian sample was analyzed (for details see Appendix F). Parents reported their child’s grade point average based on the most recent report card. Grades were assigned to Russian students on a 5-point scale with “5” indicating the highest level of achievement. Children who exhibited the Withdrawn High Achievement profile had an average GPA of 4.23 (s.d. = 0.68), which was the second highest average GPA of any profile type (the Well-Adjusted cluster had the highest GPA of 4.42). Girls exhibiting this Withdrawn High Achiever profile had a higher GPA than boys (4.41 vs. 4.00). The GPA of children exhibiting the Withdrawn Low Achievement profile had the lowest GPA of any profile group in the Russian sample (3.39, s.d. = 0.57). Again, girls in the Withdrawn Low Achieving cluster performed at a higher level than boys (3.61 vs. 3.23). Thus, the perceptions of parents of their children’s academic ability and motivation was accurately reflected in their grade point average. With regard to peer perceptions of academic ability and motivation, data from two groups of students in Georgia were analyzed. The two groups were analyzed separately because they had a slightly different age and grade distribution, as well as a different ethnic/race makeup. They also had somewhat different socioeconomic characteristics. In one cohort of students (Cohort A; 451 children), each child nominated from their classmates those that they perceived as academically talented and motivated to achieve. The specific questions were as follows: This person tries hard

96

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

to do good school work (measure of motivation). This person makes good grades, is smart, and usually knows the right answer (measure of ability). The scores given to each child was the number of nominations they received from all participating classmates. Of all temperament profiles, children in the Withdrawn High Achiever cluster had the second highest number of peer nominations related to academic ability and the highest score related to academic motivation. This was very similar to the perceptions of parents and teachers. In contrast, children in the Withdrawn Low Achiever cluster had the lowest number of nominations for academic ability as well as for achievement motivation. This again was in agreement with parental and teacher perceptions. The children in the two withdrawn groups also indicated their own perceptions of their academic ability and motivation. Self-rated data were available for Cohort B of the Georgia student sample (for details see Appendix F). Each child was asked, “If you were to list all of the students in your grade from worst to best in school work, where would you put yourself?” Also children were asked, “How good would you be at learning something new in school?” Responses were “1”  =  one of the worst through “5” = one of the best. The responses to the two scores were summed to obtain the score for each individual.  Children described by the Withdrawn High Achievement profile perceived their academic ability as in the low average range (mean percentile  =  39th) and their academic motivation as being in the high average range (mean percentile = 67th). The self-perception of this group is in marked contrast to that of their parents, teachers, and peers. This is particularly true for perceptions of their academic ability. The insecure and fearful tendencies of these children seems to extend to their impressions of their own academic ability. (More will be said about the low self-perception of ability later in this chapter.) Children who exhibited the Withdrawn Low Achievement profile rated their own academic ability as very close to average (mean percentile  =  45). Self-ratings of academic motivation were marginally above average (mean percentile  =  57th). These scores are also markedly different for the perceptions of parents, teachers, and peers. In this case, the socially withdrawn low achievers have an unrealistically high perception of their abilities and motivation at least in comparison to the perceptions of peers and the adults in their lives. In summary, parents and teachers in the United States and Russia believe that children who exhibit the Withdrawn High Achiever profile have considerable academic talent and have unusually high levels of achievement motivation. Grade point averages of the Russian children support these perceptions. Their peer group also believe that these children are above average in academic talent and have very high levels of achievement motivation. However, the students themselves do not believe their ability is even average, although they perceive that their achievement motivation is modestly above average. Children in the Withdrawn Low Achiever cluster also do not accurately perceive their academic ability and motivation, although their errors are in the opposite direction of their higher achieving and withdrawn peers. That is, they believe they have

8.5  Prosocial Behavior, Compliance, and Likeability

97

average ability and high average achievement motivation, whereas their peers, parents, and teachers perceive their abilities and motivations to be very low. Actual grade point averages of Russian children support the perceptions of peer and adult raters rather than the children’s perceptions of themselves.

8.5  Prosocial Behavior, Compliance, and Likeability Prosocial behaviors, as defined in our research, include positive emotional reactions to people as well as empathetic behavior. It also includes a measure of compliance with adult and peer rules and expectations. Compliant children are socially sensitive and are particularly aware of the expectations of others for what is appropriate behavior. The tendency to engage in all of these prosocial behaviors typically engender positive affective regard (i.e., liking) among one’s peers. As part of the indicators used to empirically develop temperament profiles of children in middle childhood, parents in the United States and Russia as well as a sample of teachers in the United States rated children on their tendency to express positive emotion and their tendency to be empathic (for details see Appendix H). Children in the Withdrawn High Achiever cluster were perceived by parents and teachers as having high levels of happiness and empathy, whereas children in the Withdrawn Low Achiever cluster were perceived to exhibit very low levels of these behaviors. These perceptions became part of the data used to construct temperament profiles. Parent and Teacher Perception of Compliance  All parents and teachers completed a compliance measure that assessed social sensitivity to adult expectation as well as obedience. Compliance ratings were not included in the indicator characteristics that were used to develop temperament profiles for each student. Withdrawn High Achievers were rated by parents and teachers as being unusually sensitive to the demands and expectations of the adults in their environment. They understand the rules and expectations of the adults in their world and are eager to comply. This group has the highest mean rating on compliance of any temperamental profile type (86th percentile). In contrast, Withdrawn Low Achievers had very low compliance scores (13th percentile); they are perceived by their parents as frequently noncompliant with adult expectations and rules set by parents and teachers (for details see Appendix G). Peer Perception of Happiness, Empathy, and Likeability  Do peers of children in these two withdrawn groups perceive large differences between the groups in happiness, empathic, and likability? These data have been graphed in Fig. 8.4. Data for two large cohorts of students are presented separately because these cohorts differ slightly in ethnic and socioeconomic characteristics (for details see Appendix J). Further, the procedures used for collecting peer nominations were somewhat different (see Appendix F).

98

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

These data reveal that Withdrawn High Achiever cluster are viewed by their peers as considerably above average in the expression of positive emotion and in their caring for, and being empathic toward, other children. These ratings from peers, however, are not as high as the scores provided by parents and teachers. Children in the Withdrawn Low Achiever cluster had very low peer nomination scores for expression of positive emotion and empathy. Thus, peers have the same general impression of the differences between these two groups of children as parents and teachers have, but peers’ ratings were less extreme than those of parents. The tendency for children exhibiting the Withdrawn High Achieving profile to be unusually compliant, social sensitive, expressive of positive emotion, and expressive of empathy should lead to the perception by peers that these children are attractive as social partners. To determine the affective regard peers have for these children, students were asked which of their classmates they would like to play with the most and which of their classmates they would least like to play with. A likeability score (sometimes referred to as social preference) was calculated by subtracting the number of “like least” nominations a child obtained from the number of “like most” nominations. The data summarized in Fig. 8.4 indicates that children in the Withdrawn High Achiever cluster were perceived as somewhat above average in likeability. Children in the Withdrawn Low Achiever cluster were perceived as considerably below average in likeability. One outcome in these data regarding children in the Withdrawn High Achiever cluster is important to consider. Although these children were perceived by their peers to express high levels of positive emotion and to be exceptionally empathetic, their likeability score was only marginally above average. Thus, they were perceived to have admirable social characteristics by their peers, but were not as

90 80 70 60 50 40 30 20 10 0

Withdrawn High Achievers Withdrawn Low Achievers

Fig. 8.4  Comparison of prosocial behavioral characteristics of Withdrawn High Achievers and Withdrawn Low Achievers (data in percentile form)

8.6  Social Status

99

attractive as play partners as their prosocial scores might otherwise suggest. This is probably related to the feelings of insecurity and fearfulness that is typically exhibited by this group. Insecurity and fearfulness make this group of children less likely to initiate play opportunities and to be more inhibited around other children. This, then, detracts from their attractiveness as a social partner. The very low rates of prosocial behavior exhibited by the Withdrawn Low Achiever cluster, in addition to the tendency to be irritable, argumentative, and antagonistic likely contributes to peers not liking to play with them as much as with others. As a result of these behavioral tendencies, this group was rated as the least likeable of all seven of the groups defined by their temperament profile. These characteristics, in addition to being less attractive as social partners, are likely to create developmental risks for this group as they continue their social development. Other children simply do not enjoy interacting with them and, thus, avoid close social contact. This avoidance reinforces the tendency to withdraw from social interaction and limits the developmental benefits (e.g., development of social skills) of being well integrated into the peer group. In summary, the two clusters of withdrawn children had profound differences in their prosocial behavior as perceived by parents, teachers, and their peer group. Withdrawn High Achievers are much more empathic, happier, and socially sensitive than their Withdrawn Low Achieving peers. They are also better liked by peers than are their Withdrawn Low Achieving peers. Although Cohort A and Cohort B differed somewhat with regard to ethnic/racial and socioeconomic characteristics, peer ratings of positive emotionality, empathy, and likeability were similar for both groups indicating that the social reactions of others to these two profile types are not affected by the ethnicity/race or social-economic circumstances of their peers.

8.6  Social Status As noted above and also in prior research, shy and withdrawn children experience lower affective regard and are more likely to be rejected by peers than their nonwithdrawn age-mates [15]. In prior research, shy and withdrawn behavior has been examined as an isolated variable rather than as one component of a broader profile; the current research is consistent with past studies, though by showing that children with profiles containing high levels of withdrawal are also disliked. We also sought to examine other aspects of social status, including the reputational aspects of status, of children for whom shyness and social withdrawal are part of a broader profile of temperamental characteristics. Four different aspects of social status in the eyes of peers were measured. These included popularity, leadership, admiration, and social prominence (standing out from the crowd—seen as “cool”). In response to three different questions, students in the third through fifth grades were asked to nominate peers who had the following characteristics (see Appendix F).

100

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

1. Which of your classmates are the most popular at school? Which are the least popular? (The popularity score was calculated by subtracting the most popular score from the least popular score.) 2. This person gets chosen by the others as the leader. Other people like to have this person in charge. (The number of nominations each child obtained was the leadership score.) 3. This person is really cool. Just about everybody in school knows this person. (Total number of nominations was the social prominence score.) 4.  Which person are admired by others. Other people want to be like them. Withdrawn High Achievers had a below average number of nominations on the popularity measure and on the social prominence measure (being well known; see Fig. 8.5). For admiration and leadership, there was an important difference in the number of nominations from students in the two cohorts of students that were studied. In Cohort A, Withdrawn High Achievers received an above average number of nominations as a leader and as someone peers admired, whereas in Cohort B the number of nominations as admired and as a leader were below average. Cohort A contained children of somewhat lower socioeconomic status than Cohort B, and had a large percentage of African American students. It also contained a higher percentage of fifth graders (the oldest age group in both cohorts). It is interesting that in this social environment, Withdrawn High Achievers were admired more and considered to have more leadership potential than was the case for Cohort B. This result is most probably related to the higher percentage of older children in Cohort A for whom achievement may take on more social importance. This result emphasizes the importance of the social context in determining who is perceived as a leader and who is admired; in general, the affective and reputational aspects of social status that a child obtains is not independent of the social context of the peer group. 90 80 70 60 50 40 30 20 10 0

Withdrawn High Achievers Withdrawn Low Achievers

Fig. 8.5  Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on four measures of social status (data in percentile form)

8.7  Influence on Peers

101

Withdrawn Low Achievers had below average scores on all reputational measures of social status. This cluster of children had the lowest number of social status nominations of any behavioral profile. Despite some differences due to characteristics of the peer group, in general, being socially withdrawn is significantly related to the child’s popularity and other measures of social status. Neither withdrawn group was popular nor were they perceived as standing out from the crowd in terms of being well known by a wide range of children.

8.7  Influence on Peers In addition to likeability (affective) and reputational social status, data obtained by Lease and colleagues on the two cohorts of students in Georgia included items assessing the influence each child has on the others in the peer group. Influence was studied in five areas: academics, sports, social trends (e.g., clothing, slang, music), fantasy/make-believe games (e.g., role-playing and computer games), and inappropriate or rule-breaking behaviors (e.g., talking back to teachers). Influence was measured by having same-age peers nominate classmates (Cohort A) or those in their grade (Cohort B) who they thought of as especially influential in each of these areas. Data summarized in Fig. 8.6 indicate that there were important differences in the influence exercised by Withdrawn High Achievers and Withdrawn Low Achievers, but this influence depended on the area of influence and the social context of the school the child attended. In general, both types of withdrawn children had significantly below average influence on their peers in the areas of cultural trends (music, slang, clothing) and in sports. Both groups are socially withdrawn so their below average scores on influence in the area of peer cultural is not surprising. Also, their temperament profiles include a tendency to have low energy levels. This seems consistent with sports not being an area in which they excel and influence others (for details see Appendix I). Children in the Withdrawn Low Achiever cluster had much more influence on other children in the area of instigation of inappropriate behavior in the classroom than did the Withdrawn High Achievers. The Withdrawn Low Achievers have a temperamental profile that includes a tendency to have problems in self-control as well as a tendency toward antagonistic behavior. These characteristics would predict that the Withdrawn Low Achievers would be more engaged in inappropriate behavior in the classroom than would the average child, but it also appears that they influence peers to act similarity. In the area of academics and imaginary games (fantasy role-playing), the influence associated with the withdrawn temperament profiles was moderated by the social context of the peer group. Withdrawn High Achievers were perceived by their peers in Cohort A as having considerable influence on others in the areas of academics and games (selecting the game, getting others to play). However, in Cohort B, they scored below average in all areas of influence. Students in Cohort A were

102

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

100 90 80 70 60 50 40 30 20 10 0

Withdrawn High Achievers Withdrawn Low Achievers

Fig. 8.6  Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on five areas of social influence (data in percentile form)

somewhat older than those in Cohort B, as Cohort A consisted of only fourth and fifth graders while Cohort B consisted of grades 3 through 5. Further, Cohort A had a higher percentage of African American students than Cohort B. What effect these and other differences in the demographic composition of the two student cohorts had on the perceived influence of withdrawn students is unclear. What is clear, however, is that the groups perceived the influence of withdrawn students differently.

8.8  Behavior Problems Data collected by Slobodskaya and colleagues from parents provided a measure of behavior problems using the Strengths and Difficulties Questionnaire (SDQ) [16, 17], a widely used measure to screen for behavioral, social, and emotional problems in children. Four behavior problem scores are obtained from this measure: hyperactivity, emotional problems (primarily anxiety and depression), conduct problems (rule-breaking and antisocial behaviors), and peer problems (alienation from peers). These data are presented in Fig. 8.7. These Russian parents perceived that very few of the children in the Withdrawn High Achiever cluster had problems with hyperactivity or conduct (e.g., rule-­ breaking, antagonistic behavior toward others). Further, this cluster of children had scores for emotional and peer problems that were in the average range. In contrast, children in the Withdrawn Low Achiever cluster, however, had much higher levels of all four behavior problems. This resulted in a total behavior problem score mean in the 91st percentile (for details see Appendix K). These data indicate that the average child in the Withdrawn High Achiever and Withdrawn Low Achiever cluster had very different pattern of behavior problems

8.8  Behavior Problems

103

100 90 80 70 60 50 40 30 20 10 0

Withdrawn High Achievers Withdrawn Low Achievers

Fig. 8.7  Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on four types of behavior problems as rated by Russian Parents

Table 8.1  Percentage of children in the Withdrawn High Achiever and Withdrawn Low Achiever clusters who are perceived to have significant behavior problems Behavior problem Hyperactivity Emotional problems Conduct problems Peer problems Total behavior problem Score

Withdrawn High Achievers 4.8% 31.8 4.8 38.6 9.5

Withdrawn Low Achievers 27.3% 66.2 45.5 66.2 64.5

particularly in the areas of hyperactivity and conduct problems. However, the mean score for either cluster does not indicate how many individuals in the group had significant behavior problems. A significant behavior problem was defined as being at least one standard deviation above the mean of all children studied (i.e., having a score greater than the 84th percentile). This level of problems does not necessarily indicate that they would receive a clinical diagnose, but it is often considered by mental health professionals as being in the borderline region of a clinically significant problem. Table 8.1 summarizes these data. Since 16% of all children in the sample had significant behavior problems based on the way we defined “significant,” it can be seen that the Withdrawn High Achievers had a much lower percentage of children with significant problems in the areas of hyperactivity and conduct than was typical of the Russian children as a whole. However, as expected, they did have a higher percentage of emotional and peer problems based on parental perception than the majority of their peers. The Withdrawn Low Achievers had a high percentage of children with emotional and peer problems and also an above average percentage of children exhibiting uncontrolled activity levels and conduct problems.

104

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

The data regarding behavior problems were obtained from Russian parents, and these same parents completed the behavioral assessments. Therefore, both sets of data were obtained from the same source. Researchers refer to this issue as a lack of independence of the data source. In order to determine if children exhibiting the two withdrawn behavioral profiles had different levels of behavior problems based on independent sources, we analyzed data from children’s peers and the children themselves. Georgia students were asked to nominate children who were particularly socially aggressive (e.g., attempting to exclude others from their group), verbally aggressive (e.g., saying mean things to others), or physically aggressive (e.g., pushes or hits others). Figure  8.8 presents the comparison between Withdrawn High Achievers and Withdrawn Low Achievers regarding how many nominations each group obtained, on average, for these kinds of aggression. Inspection of Fig. 8.8 reveals that the two profiles of withdrawn children differed significantly in their aggression levels based on peer nominations. The Withdrawn High Achievers had a very low mean percentile score on all types of aggression (social, verbal, physical), whereas the Withdrawn Low Achiever cluster was perceived by peers as having above average levels of all types of aggression. These differences between withdrawn clusters were somewhat higher for Cohort A than for Cohort B (for details see Appendix L). One cohort of Georgia students (Cohort A) completed a screening instrument designed to measure self-perceptions of multiple types of behavior problems. Self-­ perceptions of anxiety and depressive symptoms were of particular interest as these relate to the emotional problems reported by Russian parents. We also examined self-reports of interpersonal problems and school-related problems. However, in none of these four areas were their statistically or practically important differences between the Withdrawn High Achievers and the Withdrawn Low Achievers. In all 100 90 80 70 60 50 40 30 20 10 0

Withdrawn High Achievers Withdrawn Low Achievers

Fig. 8.8  Comparison of Withdrawn High Achievers and Withdrawn Low Achievers on measures of social, verbal, and physical aggression as perceived by peers (data in percentile form)

8.8  Behavior Problems

105

areas, mean scores (based on self-reports) for both clusters were very close to the average of the whole sample. Thus, despite parents, teachers, and peers perceiving large differences between these two groups in behavior problem symptoms, the children themselves maintain a self-perception that they exhibit average levels of problematic behavior (for details see Appendix M). For peer and self-perceptions of behavior problems, the data provided described the means of the two withdrawn clusters. Because behavior problems are typically highly skewed (most children have few problems and a few children have many), it is important to determine the number of children in each cluster who exhibit significantly high levels of these problems. As described previously, a significant level of problems was set at a score that was 1 standard deviation above the mean for the sample (greater than the 84th percentile). Table 8.2 presents the percentage of children in both withdrawn groups who had behavior problems scores that were considered significant. With regard to aggressive behavior, none of the children in the Withdrawn High Achiever cluster was viewed by their peers as having significantly high levels of any type of aggressive behavior. A somewhat higher than average (average for the entire sample was 16%) number of children in the Withdrawn Low Achievers category, however, had significant levels of aggressive behavior as viewed by peers. The self-reported levels of anxiety, depression, interpersonal, and school-related problems provide some insight into the nature of the two withdrawn groups. For the Withdrawn High Achiever group, about an average percentage of children (16% is average) reported problems with anxiety and school-related problems in the significant range. A low percentage reported having significant levels of interpersonal problems, but a somewhat higher than typical percentage reported depressive symptoms. In contrast, a smaller percentage of the Withdrawn Low Achiever group reported significant levels of depression, interpersonal, or school-related problems (for details see Appendix M).

Table 8.2  Percentage of children in the Withdrawn High Achiever and Withdrawn Low Achiever clusters who have significant behavior problemsa based on peer and self-report Source Problem type Peer Physical Aggression—Cohort A Verbal—Cohort A Social Aggression—Cohort A Verbal/Physical Aggression— Cohort B Social Aggression—Cohort B Self Anxiety—Cohort A Depression—Cohort A Interpersonal—Cohort A School Related—Cohort A

Withdrawn High Achievers 00.0 00.0 00.0 00.0

Withdrawn Low Achievers 25.8 25.8 9.7 27.6

00.0 14.3 21.4 0.00 14.3

20.7 16.1 9.7 9.7 9.7

A significant problem was defined as a score at or greater than the 84th percentile for the sample of children studied (≥1 s.d.)

a

106

8  Shy and Socially Withdrawn Children: Two Temperament Profiles

References 1. Kagan, J., Reznick, J.  S., & Snidman, N. (1988). Biological bases of childhood shyness. Science, 240, 167–171. 2. Asendorpf, J. (1990). Beyond social withdrawal: Shyness, unsociability and peer avoidance. Human Development, 33, 250–259. 3. Rubin, K.  H., & Coplan, R.  J. (2010). The development of shyness and social withdrawal. New York: Guilford. 4. Rubin, K., Wojslawowicz, J., Burgess, K., Riose-Krasnor, L., & Booth-LaForce, C. L. (2006). The friendships of socially withdrawn and competent young adolescents. Journal of Abnormal Child Psychology, 34, 139–153. 5. Kagan, J. (2012). In M. Zentner & R. L. Shiner (Eds.), The biography of behavioral inhibition (pg. 69–82). New York: Guilford. 6. Thomas, A., & Chess, S. (1977). Temperament and development. New York: Bruner/Mazel. 7. Caspi, A., Moffitt, T. E., Newman, D. L., & Silva, P. A. (1996). Behavioral observations at age 3 years predict adult psychiatric disorders. Archives of General Psychiatry, 53, 1033–1039. 8. Caspi, A., & Silva, P. A. (1995). Temperamental qualities at age 3 predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486–498. 9. Van den Akker, A. L., Dekovic, M., Prinzie, P., & Asscher, J. J. (2010). Toddler’s temperament profiles: Stability and relations to negative and positive parenting. Journal of Abnormal Child Psychology, 38, 485–495. 10. Janson, H., & Mathiesen, K. S. (2008). Temperament profiles from infancy to middle childhood: Development and associations with behavior problems. Developmental Psychology, 44, 1314–1328. 11. DiStefano, C., Kamphaus, R. W., Horne, A. M., & Winsor, A. P. (2005). Behavioral adjustment in the U.S. Elementary school: Cross-validation of a person-oriented typology of risk. Journal of Psychoeducational Assessment, 21, 338–357. 12. Kamphaus, R.  W., Huberty, C.  J., DiStefano, C., & Petoskey, M.  D. (1997). A typology of teacher-rated child behavior for a national U.S. sample. Journal of Abnormal Child Psychology, 25, 453–463. 13. Kamphaus, R.  W., Petosky, M.  D., Cody, A.  H., Rowe, E.  W., Huberty, C.  J., & Reynolds, C. R. (1999). A typology of parent rated child behavior for a national U.S. sample. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 40, 1–10. 14. Scott, B. G., Lemery-Chalfant, K., Clifford, S., Tein, J. Y., Stoll, R., & Goldsmith, H. H. (2016). A twin factor mixture modeling approach to childhood temperament: Differential heritability. Child Development, 87, 1940–1955. 15. Rubin, K. H., Bowker, J., & Gazelle, H. (2010). Social withdrawal in childhood and adolescence: Peer relationships and social competence. In K. H. Rubin & R. J. Copan (Eds.), The development of shyness and social withdrawal (pp. 131–156). New York: Guilford. 16. Goodman, R. (1997). The strengths and difficulties questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581–586. 17. Goodman, A., & Goodman, R. (2009). Strengths and difficulty questionnaire as a dimensional measure of meatal health. Journal of the American Academy of Child and Adolescent Psychiatry, 48, 400–403.

Chapter 9

Fostering the Development of Shy, Withdrawn Children

9.1  Introduction Children who are fearful, insecure, inhibited in unfamiliar situations, and socially withdrawn require different parenting and teaching skills than is the case for children with other temperamental profiles. If these tendencies are not extreme, these behavioral tendencies can have advantages. The also create risks for the child’s development. Before describing some of the issues specifically faced by Withdrawn High Achievers and Withdrawn Low Achievers, let us consider the more general questions regarding assets and liabilities of being an inhibited, somewhat fearful, and socially introverted child. Before beginning this discussion, it is important to know how social withdrawal was measured in our research. Children in middle childhood can exhibit social withdrawal for a number of reasons. This can occur because of (a) fearfulness based on inhibition when placed in a novel situation; (b) a general tendency to be fearful and anxious; (c) experiences with parents, teachers, or peers that are harsh, punitive, or rejection; or (d) a simple lack of a desire to be with other and a preference for solitary activity (low sociability). Unfortunately, our research was unable to differentiate all of these causative factors. The parents and teachers who rated the behavioral tendency of the children we studied simply observed the extent to which the child makes friends easily, has lots of friends, prefers to be alone, and is slow to warm up in new social situations. The measures can best be characterized as a composite of temperamental shyness and low sociability.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_9

107

108

9  Fostering the Development of Shy, Withdrawn Children

9.2  Developmental Assets Being socially withdrawn, in most Western cultures, is often considered a disadvantage. US culture, for example, tends to place high value on verbally assertive and gregarious behavior. These behavioral tendencies in combination with a preference for action versus contemplation, risk-taking versus heed-taking, and certainty versus doubt are characteristic of the extrovert. Many parents and teachers find extraverted, gregarious children to be more socially attractive and as having developmental advantages even into adulthood. This cultural perspective leads many teachers to be particularly concerned if a child is socially withdrawn. However, some of the characteristics associated with being socially withdrawn can have positive effects on the developmental progress of the child. Many writers, scientists, academics, and mental health professionals are more socially withdrawn than their peers. This withdrawal makes possible the long hours of individual work that is required in many walks of life. These are tendencies that foster contemplation and attention to detail [1]. Social inhibition associated with fearfulness (e.g., inhibition to the unfamiliar) can be helpful in the development of self-regulation. One group of researchers lead by Rothbart at the University of Oregon found that fearful infants showed greater empathy, guilt, and shame in childhood [2]. These findings seem to indicate that fearfulness might be related to the development of conscience. Several studies have demonstrated a link between fearfulness and the development of conscience particularly if mothers used gentle socialization techniques [3, 4]. A conscience is an internalized restraint based on social learning. It helps children self-regulate aggressive impulses toward others, and aids in the internalization of familial and school rules. The children exhibiting both withdrawn profiles in our research also had higher than average levels of fearfulness. Fearfulness is linked to enhanced attention to threats. Mild anxiety and fear help the child internalize family and cultural rules because transgressions can lead to parental or teacher punishment, or peer disapproval. Withdrawn children who also have good control of their attention also have particularly strong assets that favor positive development. There has been important work in the temperament literature on the ability to consciously focus attention [5, 6]. This process is sometimes referred to as effortful control. Children with higher levels of effortful control in early childhood have an enhanced ability to wait for appropriate opportunities to act, to resist distraction, to detect errors in their work, to resist temptation, and to attain desired long-range goals. These effects of effortful control are particularly applicable to school performance. This discussion supports the guiding principle of the research reported in this book. The developmental effects of individual differences in behavior are best understood when they are studied in combination; that is, in profile across many characteristics. Social withdrawal when associated with mild fearfulness, good self-­ regulation of attention, and expression of negative emotion can have positive

9.3  Developmental Risks

109

developmental consequences for children. For example, children who exhibit the Withdrawn High Achiever profile were found to have a number of behavioral characteristics that had positive consequences in middle childhood, and are likely to aid their future development. They achieve academically at a high level and have a strong desire to live up to societal demands for performance in school and social interaction. They were very rarely aggressive, antagonistic, or argumentative. All of these characteristics mean that they are socially attractive in many circumstances. These generalizations obtained from our study have been given considerable support in large-scale longitudinal research efforts in which shy children were assessed in early childhood, then again in late childhood, adolescences, and into adulthood. For example, in one of the largest and longest running studies of early childhood temperament, researchers in New Zealand found that adolescents who had been in the Reserved child group (a cluster defined by moderate inhibition-to-­ the-unfamiliar) at age 3 were found to have normal or average behavioral tendencies when assessed in adolescence [7]. Research from the Longitudinal Study of Australian Children [8] found that children who were anxious in childhood but became less anxious and depressed as teenagers were those who had stronger social skills and more positive school experiences [9]. Both studies describe children with similar behavior profiles to the Withdrawn High Achievers.

9.3  Developmental Risks Social withdrawal is also associated with developmental risks. First, being temperamentally shy (inhibited in unfamiliar situations) and fearful tends to result in a constriction of one’s social world; high levels of shyness, fearfulness, and insecurity can result in severe social isolation. This set of characteristics minimizes social contact and results in fewer opportunities for children to practice social interactions with a range of people. Thus, withdrawn children typically take longer to learn the complex social skills required to optimally function in the world. The elevated fearfulness and social insecurity that often accompany social withdrawal also limit the expression of talents and skills. To be able to express oneself in a social environment, to perform artistically, to let the world know about your skills as a fisherman, to speak another language, or to care for an aging grandmother requires some degree of asserting oneself. Because of the reticence to be assertive, social inhibition and social withdrawal tends to produce an underestimation of skills and talents from parents, teachers, and peers. One manifestation is that many inhibited children are believed to be less intelligent than their measured intelligence score would indicate [10]. This underestimation is a logical consequence of the child not speaking up and sharing their opinions and insights. In the classroom, inhibited children are often more reluctant than their more outgoing peers to raise their hand in response to a teacher question or in other ways to seek attention. Thus, peers and the teacher have less of an opportunity to see what the child knows. This phenomenon may partially account for the very low estimates

110

9  Fostering the Development of Shy, Withdrawn Children

of academic ability made by parents, teachers, and peers for the Withdrawn Low Achieving cluster of children identified in our study. This can result in an internalization of a lack of confidence in one’s talents and an underestimation of one’s own abilities. This effect was  also observed in our data in that the Withdrawn High Achievers tended to underestimate their intellectual abilities when compared to the actual grades they achieved. Shy, withdrawn children seek safe social circumstances with other people they know well and trust. This kind of selection reduces the risks of being rejected or criticized. Unfortunately, this restriction of social contacts in general, and restriction of the range of individuals with whom one has contact, limits the social impact the child has on others. This, in turn, puts a limit on their social standing among, and influence with, their peers. Shy children are also more likely than their peers to be the target of social exclusion. The withdrawn child initiates fewer conversations with others, and is often content to listen to others talk. This makes them less satisfying social participants, which facilitates social exclusion. When a birthday party or a sleep-over is planned, these children are less likely to be invited. Further, in more extreme cases, withdrawn children are perceived by peers as vulnerable. This perception may be made more likely by their tendency to exhibit less gross motor activity and appear less energetic, an associated characteristic of both the withdrawn clusters. This perceived vulnerability results in an increased likelihood that the child is the object of bullying [11–13]. In response to social exclusion or bullying, withdrawn children may become even more fearful of classmates and withdraw further from peer interaction and school-related activities [14]. We have demonstrated that many withdrawn children have less influence on the behavior of peers than other students in their classroom. Other researchers have observed children in middle childhood in interaction with peers. In these studies, withdrawn children interacting with nonwithdrawn children were observed to have less success when asserting their rights, or in getting others to perform a specific behavior. If this pattern of interaction persists, a lack of success in impacting the behavior of others becomes internalized and fewer attempts are made in the future. This, in turn, leads the child to be more fearful and withdrawn in social situations as they become anxious about their abilities to control the social interactions [15].

9.4  Differences Between Two Types of Withdrawn Children We have demonstrated that there are two quite different behavioral clusters of children who exhibit social withdrawal, social insecurity, and fearfulness. One of the differentiating factors is academic performance and motivation. This differentiation could be related to differences in academic ability (verbal and numeric ability), but we have no direct measure of intelligence so our research cannot directly address this question. However, it is likely that this explanation is far too simplistic. These two groups of socially withdrawn children vary in self-regulation of attention and, of course, the ability to manage distractions and to attend to academic material is

9.4  Differences Between Two Types of Withdrawn Children

111

strongly influenced by this kind of self-regulation. Further, children in the Withdrawn Low Achiever cluster have difficulty controlling expressions of negative emotion and engage in more antagonistic behavior toward peers. These behaviors further exacerbate any tendency toward social isolation; in fact, parental and teacher observations of the social isolation of children may be primarily caused by the darker moods and more aggressive behaviors of the Withdrawn Low Achievers. Thus, differences in academic ability and performance, in self-regulation of attention or irritability/antagonism, or some combination of these factors may be at the root of the differentiation between these two groups of withdrawn children. There is a relatively new body of research that may shed light on the causative factors that differentiate these two groups. For about 40 years, Thomas Boyce, who is Head of the Division of Developmental Medicine at the University of California, San Francisco, has been working on a theory that some children are particularly sensitive to environmental influences while others are much less impacted by positive or negative environmental circumstances. By analogy, the former group is labeled “Orchids.” They have a biologically embedded sensitivity that creates an acute vulnerability to environmental hazards, particularly in the social environment (e.g., poverty, harsh parenting). However, this same sensitivity allows them to be more receptive and to benefit more than less sensitive children when in a nurturing environment (e.g., parents and teachers who offer a rich stimulating environment of experience in a psychological safe space). The contrasting group is labeled “Dandelions.” These children are hardy and show a remarkable ability to thrive in almost any circumstance in which they find themselves [16]. This theory is supported by a common experience. Almost everyone knows of some child who was brought up in poverty or in a chaotic, abusive family, or who experienced poor quality schooling, and yet became a prominent person in their community. It is also a common experience to have a relative or friend who is completely thrown off their developmental path by parental divorce or by insensitive experiences with peers. The most revolutionary aspect of this theory, however, is the understanding that “Orchids” are not simply unusually vulnerable to adversity, but are also unusually sensitive to nurturing environments. Orchid and Dandelion children have been demonstrated to have different health histories. There is a long history in medicine documenting that increased social stress increases the likelihood of respiratory illnesses (cough, colds, sinus infections). However, there is a good deal of variability among children in their disease proneness even among children who experience equal levels of environmental stress. To systematically investigate this issue, Boyce studied how children (3–8  years of age) responded to small laboratory “challenges” such as a drop of lemon juice on the tongue, viewing an emotional video, and being interviewed by an unknown adult. Immediately following the challenge, he and his staff measured heart rate, heart rate variability, blood pressure, and levels of a stress hormone (cortisol) in samples of saliva. They determined that some children were involuntarily and biologically reactive to these laboratory challenges, whereas others were not. He then demonstrated that the highly biologically reactive children (Orchids) were particularly prone to respiratory illnesses, while the less reactive group (Dandelions) were significantly less prone to these common illnesses of childhood [16, 17].

112

9  Fostering the Development of Shy, Withdrawn Children

So how do these findings regarding Orchids and Dandelions relate to socially withdrawn children? Several teams of temperament researchers have intensively studied shy, socially inhibited children and children who exhibit unusual levels of emotional reactivity [18, 19]. Some of these children react to challenging environments poorly, slowly adapting to change, and often withdrawing from novel environments. They also have tendency toward sensory sensitivity. They are more aware of bodily discomfort (e.g., small aches and pains) as well as loud noise or uncomfortable clothing. These children have much in common with the description of the highly sensitive Orchid group. Temperament researchers have also intensively studied negative emotionality. Some children have a strong tendency to expresses negative emotions (crying, screaming) particularly in the face of environmental challenges causing frustration. Negative emotionality occurs most often when the child is prevented from engaging in a desired activity by parental or school rules, or by competition from siblings or peers (e.g., who get more adult attention; who gets to play with a prized toy). Negative emotions can result from a low threshold to the normal hassles of daily life. It can be one indication of a strongly reactive nervous system. Researchers have studied differences in parenting behavior during toddlerhood on children who were assessed for negative emotionality in infancy [20]. The major finding was that parenting was more predictive of behavior problems for children with high levels of negative emotionality than for other children who did not display high levels of negative emotionality in infancy. In other words, the children who exhibited heightened levels of negative emotionality in infancy can be thought of as Orchids, while those who did not can be thought of as Dandelions. Specifically, Orchid infants who were raised by parents who themselves displayed high levels of negative emotionality ended up with high levels of internalizing (e.g., anxiety) and externalizing problems (e.g., conduct problems like antagonistic behaviors toward parents and teachers) at age 3, compared to their Dandelion peers. However, Orchid infants had fewer behavior problems than Dandelions if raised by more positive parents. This kind of research has now been replicated a number of times [21, 22]. Other research groups have looked at specific genes and found that children with a specific variant are more susceptible to environmental influences than those without that variant. For example, the serotonin transporter (5HTT, SLC6A4) is one of the most important regulators of serotonin levels. Serotonin appears to affect and/or regulate a wide range of body functions. It is a neurotransmitter that plays a role in regulating mood. In particular, it has been called the “feel-good” chemical because those who have low levels of serotonin tend to have a low sense of well-being (e.g., have symptoms of depression). Many antidepressants increase levels of serotonin by inhibiting reuptake of the chemical, thus leaving high levels in the nervous system. Serotonin has many other functions including effects on digestion, blood clotting, bone density, and sexual functioning. However, in the present discussion of socially withdrawn children, the effect on fearfulness and irritability are our focus as both are indicators of mood. It turns out that there is a region of the serotonin transporter gene (5HTTLPR) that varies among individuals. A recent review of more than 70 studies has shown that individuals with two short variants are more responsive to environmental influences than are individuals with two long variants in this transporter region [23]. For

9.4  Differences Between Two Types of Withdrawn Children

113

example, children who had been initially raised in an orphanage and then placed with foster parents were assessed for the 5HTTLPR variant as well as for their attachment to their fostering caregiver. Children who had two short variants in the serotonin transporter region had more externalizing behaviors at age 4 if they had an atypical attachment. There was no association between attachment status and externalizing behavior for children carrying at least one copy of the long variant. This demonstrates that children with the two short variants of this gene were particularly sensitive to the attachment relationship with their foster parent [24]. Similar results have been reported for genes regulating the neurotransmitter dopamine [25, 26]. This line of research seems to suggest that children who exhibit both of the withdrawn profiles we have identified are particularly sensitive to their environment. It is possible that the Withdrawn High Achievers have experienced an environment with more opportunities supportive of school achievement, than their Withdrawn Low Achieving peers. In support of this idea, in our data 62.1% of the parents with children in the Withdrawn High Achiever cluster had graduated from college, whereas for the Withdrawn Low Achievers, this percentage was 37.9. In our research, we isolated two clusters of students who differed significantly in achievement and self-regulation of attention and emotion (see Chap. 6). However, parent education levels were very similar. We also identified two clusters of children who were viewed by parents and teachers as being poorly self-regulated (see Chap. 10). They also differed significantly in achievement. Again, education levels of parents were very similar. This pattern of results is strongly supportive of the notion that children who are particularly fearful and shy are more susceptible to environmental conditions than children who are not. In other words, children in the Withdrawn Low Achiever cluster seem to have been particularly hampered in their educational achievement and self-regulatory behaviors by low levels of academic achievement by their parents, whereas children in the Withdrawn High Achiever cluster have had all the advantages surrounding higher socioeconomic status brought about by higher parental educational levels contributing to a more positive schooling experience. One of the most interesting findings from our research was that children exhibiting the Withdrawn High Achiever profile consistently underestimated their academic ability, while children in the Withdrawn Low Achiever cluster overestimated their ability when compared to the assessments made by parents, teachers, and peers. With regard to the Withdrawn High Achievers, it is likely that this group is particularly insecure about their abilities because they live in families of relatively high academic achievement and in families in which such achievement is highly valued. This insecurity may be based on fear of not being able to meet the demands (explicit or implicit) of their family. The cognitive and emotional processes that result in the Withdrawn Low Achievers believing that their cognitive ability  is in the normal range may be explained by a similar comparison process, but in this case parents and family friends may have had less academic success, and thus there is less implicit or explicit pressure of high attainments. An alternative explanation comes from a defensiveness point of view. Children in middle childhood desperately want to fit in, to be one of the group. To think of oneself as having below average academic success is threatening to fitting in. Thus, children in this cluster rate their academic ability as

114

9  Fostering the Development of Shy, Withdrawn Children

being in the average range despite a lower than average grade point average, and a perception by peers, parents, and teachers, that they have low ability. Finally, these two groups of children have very different patterns of problematic behaviors. Problematic behaviors are defined as those that create social difficulties for the child or for others who interact with the child. If these behaviors are severe enough in intensity and stability over time and cause impairments in their ability to function at school, at home, and with peers, they can lead to a diagnosis of psychopathology. In a pattern similar to self-perception of academic ability, self-ratings of internalizing problems indicate that Withdrawn High Achievers have higher levels of depressive symptoms than most of their peers, but Withdrawn Poor Achievers view themselves as having fewer anxieties and depressive tendencies than their peers. Also, Withdrawn High Achievers perceive that they are more maladjusted at school (dislike school and teachers) than the average child their age, while Withdrawn Low Achievers report that they are about average in this respect. If the differential susceptibility hypothesis is correct, then both groups are unusually sensitive to and aware of their family and peer group values and behavior. If the two groups have different levels of expectations from their  family and associate with peers who have different levels of academic achievement and motivation, then it is not surprising that they achieve at different levels and have different ideas about their relative ability. The Withdrawn Low Achievers may attribute their lower achievement in school not to lowered ability but to a lack of motivation to perform at a high level in this environment. Whether children in the Withdrawn Low Achiever cluster have lower academic ability, or simply are conforming to peer and perhaps family expectations regarding school achievement cannot be determined by our research methods. It is perhaps most likely that both of these factors play some role. However, it is also likely that the abilities of this group are underestimated by their peers, parents, and teachers, because their shy, withdrawn behavior produces fewer chances for the world to see what they can do.

9.5  G  eneral Guidelines for Parenting/Teaching Withdrawn Children The following are guidelines for helping withdrawn children take advantages of their behavioral assets and to limit the risks associated with the two profiles we have identified. These guidelines apply most directly to fear-based social withdrawal (inhibition to the unfamiliar; fear of peer evaluation). However, many of the guidelines also apply to the few children in this cluster who simply have little interest in the social world. 1. The first goal for the parent and the teacher is to understand the characteristics of socially inhibited and withdrawn children, and the biologically influenced nature of the behavior. It is particularly important to understand that fear-based responses to new activities or new social contexts are not willful. They are not under conscious control.

9.5  General Guidelines for Parenting/Teaching Withdrawn Children

115

2. Through understanding, parent and teachers can learn to truly value the behavioral style the child exhibits. All behavior profiles have positive attributes as well as risks. Valuing the assets of cautious, sensitive, and shy children is aided by the understanding that these behavioral tendencies work well in many situations. If parents do not learn to value the behavioral tendencies of the shy child, there will be a tendency to push for more gregarious behavior. In the United States, parents often push their shy children to be more assertive, to initiate more conversations, and to have more social contacts. Continually pushing these goals simply does not work. Further, it is likely to create resentment toward parents and teachers, and to increase social withdrawal. Research from the Australian Temperament Project has clearly documented that shy children identified in early childhood had reduced levels of shyness in early adolescence if their parents were warm and nurturing, and did not push too soon for more gregarious behavior. If parents were less child focused, if they made child management procedures that made the child feel guilty or anxious, there was a tendency later in development toward less independence and more social withdrawal [9]. This outcome is consistent with the differential susceptibility hypothesis outlined above, which indicates that these children are particularly sensitive to environmental pressures. 3. The goal of parenting or teaching socially inhibited children is not to change these general characteristics. The goal is to help the child cope with the adverse aspects of being socially inhibited. This process often involves learning to cope in specific situations through repeated graduated exposure in a supportive environment. For example, teachers want to help children learn to stand up in front of their classmates and present an idea or a project. This can be an overwhelming experience for shy, withdrawn children in middle childhood. But these skills can be learned. The authors have known world-class researchers who are temperamentally shy and socially inhibited. They are very sensitive about the attitudes of their peers regarding their work. These tendencies work against being a good public speaker, but public speaking is a necessary skill for these professionals. Many are teachers; in addition, they must be able to present their ideas at professional conferences. Some become outstanding public speakers, although many never develop high-level skills. For those that do, the development of the skills results from lots of practice in safe and nurturing environments. Interestingly, for some these high-level performances seem to come at some psychological and physiological cost. For example, one outstanding speaker known to the first author spends a long time in preparation for public presentations and then after the speech finds he is physically exhausted by the experience. This seems to indicate that the skills have been learned, but the performance goes against his natural tendencies. In developing the social skills of inhibited children, patience and unusual sensitivity on the part of the parent/teacher is required. Repeated practice in a safe, relatively stress-free environment is optimal. This progressive process may simply begin with observing others in a particular situation. The normal coping style of shy children is to spend time observing others when placed in an unfamiliar

116

9  Fostering the Development of Shy, Withdrawn Children

situation. This is a safe and anxiety-free way to begin to feel comfortable in the social group. After a while, the child may become more confident about joining in. But the child has to be given time to simply be an observer. To continue the example of public performances, exposing the child to the performances of other children (e.g., music, theater) as an observer can be particularly helpful in overcoming performance anxieties. Similarly, learning to join with others in a social situation can be facilitated if children are able to observe their parents or other children engaging in socially skilled behavior. Pressing them to engage before they are read can backfire, causing setbacks. Many shy children will find social interaction with one or two other friends much easier than interaction in larger groups. Having a friend with whom one feels safe can play an important role in helping the child learn social skills. Parents and teachers can be helpful by creating situations that facilitate these kinds of social interaction. 4. Parents and teachers should be aware that how they interact with children who are particularly sensitive to their social environment may have long-lasting and powerful effects on their future development. Many socially inhibited and withdrawn children are unusually sensitive to the emotions of their parents, teachers, and peers. Anger and other forms of disapproval expressed toward some children will soon be forgotten. Anger expressed toward socially sensitive children will have a longer half-life. The emotions that anger engenders in the child will continue for a longer time, as will the circumstances surrounding the angry outburst by care-takers. This means that milder forms of parental punishment (e.g., scolding) will have the same impact for these children as stronger punishments will have for the more resilient child.

References 1. Cain, S. (2013). Quiet: The power of introverts in a world that can’t stop talking. New York: Crown. 2. Rothbart, M. K., Ahadi, S. A., & Hershey, K. L. (1994). Temperament and social behavior in childhood. Merrill-Palmer Quarterly, 40, 21–39. 3. Kochanska, G. (1995). Children’s temperament, mothers’ discipline, and security of attachment: Multiple pathways to emerging internalization. Child Development, 66, 597–615. 4. Kochanska, G. (1997). Multiple pathways to conscience for children with different temperaments: From toddlerhood to age five. Developmental Psychology, 33, 228–240. 5. Posner, M.  I., & Rothbart, M.  K. (2007). Educating the human brain. Washington, DC: American Psychological Association. 6. Rothbart, M. K., Posner, M. I., Rueda, M. R., Sheese, B. E., & Tang, Y. Y. (2009). Enhancing self-regulation in school and clinic. In D. Cicchetti & M. R. Gunnar (Eds.), Minnesota symposium on child psychology: Meeting the challenge of translational research in child psychology (Vol. 35, pp. 115–158). New York: Wiley. 7. Caspi, A., & Silva, P.  A. (1995). Temperamental qualities at age three predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486–498.

References

117

8. Vassalio, S., & Sanson, A. (2013). The Australian temperament project: The first 30 years. Australian Institute of Family Studies: Australian Government. 9. Letcher, P., & Olsson, C. (2013). Early adolescent outcomes from childhood. https://aifs.gov.au/publications/australian-­t emperament-­p roject-­f irst-­3 0-­y ears/ Australian-­temperament-­project#r130. 10. Martin, R.  P., & Holbrook, J. (1985). Relationship of temperament characteristics to the academic achievement of first grade children. Journal of Psychoeducational Assessment, 3, 131–140. 11. Erath, S., Flanagan, K., & Bierman, K. (2007). Social anxiety and peer relations in early adolescence: Behavioral and cognitive factors. Journal of Abnormal Child Psychology, 35, 405–416. 12. Gazelle, H., & Ladd, G. (2003). Anxious solitude and peer exclusion: A diathesis-stress model of internalizing trajectories in childhood. Child Development, 74, 257–278. 13. Rubin, K., Wojslawowicz, J., Burgess, K., Rose-Krasnor, I., & Booth-LaForce, C. L. (2006). The friendships of socially withdrawn and competent young adolescents. Journal of Abnormal Child Psychology, 34, 134–153. 14. Rubin, K., & Coplan, R.  J. (2010). The development of shyness and social withdrawal. New York: Guilford. 15. Nelsen, L. J., Rubin, K., & Fox, N. (2005). Social and nonsocial behaviors and peer acceptance: A longitudinal model of the development of self-perception in children ages 4 to 7 years. Early Education and Development, 20, 185–200. 16. Boyce, W. T. (2019). The orchid and the dandelion. New York: Knopf. 17. Boyce, W. T., et al. (1995). Psychobiologic reactivity to stress and childhood respiratory illnesses: Results of two prospective studies. Psychosomatic Medicine, 57, 411–422. 18. Kagan, J., Reznick, J.  S., & Snidman, N. (1988). Biological bases of childhood shyness. Science, 240, 167–171. 19. Chess, S., & Thomas, A. (1986). Temperament in clinical practice. New York: Guilford. 20. Belsky, J., Hsieh, K., & Crnic, K. (1998). Mothering, fathering, and infant negativity as antecedents of boys’ externalizing problems and inhibition at age 3: Differential susceptibility to rearing influence. Development and Psychopathology, 10, 301–319. 21. Belsky, J., Friedman, S. L., & Hsieh, K. H. (2001). Testing a core emotion-regulation prediction: Does early attentional persistence moderate the effect of infant negative emotionality on later development. Child Development, 72, 123–133. 22. Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influences. Psychological Bulletin, 125, 885–908. 23. van Ijzendoorn, M. H., Belsky, J., & Bakerman-Kranenburg, M. J. (2012). Serotonin transporter genotype 5HTTLPR as a marker of differential susceptibility? A meta-analysis of child and adolescent gene-by-environment studies. Translational Psychiatry, 2, e147. 24. Humphreys, K.  L., Zeenah, C.  H., & Drury, S.  S. (2015). Serotonin transporter genotype (5HTTLPR) moderates the longitudinal impact of atypical attachment on externalizing behavior. Journal of Developmental and Behavioral Pediatrics, 36, 409–416. 25. Bakermans-Kranenburg, M.  J., & van Ijzendoorn, M.  H. (2011). Differential susceptibility to rearing environment depending on dopamine-related genes: New evidence and a meta-­ analysis. Development and Psychopathology, 23, 39–52. 26. Marsman, R. L., Oldehinkel, A. J., Ormel, J., & Buitelaar, J. K. (2013). The dopamine receptor D4 gene and familial loading interact with perceived parenting in predicting externalizing behavior problems in early adolescence: The tracking Adolescents’ individual lives survey (TRAILS). Psychiatry Research, 209, 66–73.

Chapter 10

Poorly Self-Regulated Children: Two Temperament Profiles

10.1  Introduction My memory (RM) of experiences in middle childhood is filled with examples of irritable, argumentative, and aggressive children. I was particularly vigilant about determining who was irritable and had the potential to be verbally or physically aggressive. I had a neighbor we will call Frank. He was sometimes pleasant to be around, but on occasion he could be aggressive. He seemed angry much of the time and hung around with what I would have called at the time “tough kids.” At one point in my adolescence, he called my name at a party, I turned around and he punched me in the face. He would have been very difficult in a fight and I was so taken aback that I just walked away. I was embarrassed. He had no particular argument with me. This seemed to be instrumental aggression; that is, his motivation seemed to be to show off for a particularly popular girl. Another neighbor, whom I will call Lew, also had an angry streak, particularly if things did not go well in athletic activities. He was a fine varsity-level basketball player in high school, but he and I played a lot of basketball during our elementary school years. He was better than I was, but when I had a lucky day and pulled ahead in a game, he often got angry, started to foul me, or said I was cheating in some way. He was generally moody and very competitive. I knew his father put a lot of pressure on him to be a good athlete and, even at the time, I felt he was reacting to this pressure. His motivation for playing the game was to win no matter what; on the other hand, my motivation was to play competently, but winning and dominating the opponent were not primary motives. Lew and I had a clear personality difference. Children who are irritable and aggressive toward peers are also often antagonistic and push back against the rules set down by parents and teachers. For both of these neighborhood acquaintances I have described, they had a tendency to get into trouble with the law. My father was a policeman at the time, and I learned that Frank had some minor difficulties (fighting, public intoxication) that I became aware of

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_10

119

120

10  Poorly Self-Regulated Children: Two Temperament Profiles

when we both were in high school. Both Frank and Lew were average students. Reports from friends in the community indicated that Frank became a productive and respected engineer. I did not hear much about Lew after we left high school. My experiences in middle childhood with aggressive acquaintances and with those who tended to break the rules set down by parents and teachers indicate that children of this age are keenly aware of these behavioral traits in others. However, it was not clear to me at the time if parents and teachers were aware of these tendencies. Many of these aggressive interactions took place outside of the vicinity of adults. The research reported here was designed to answer several questions with regard to children who have difficulty controlling negative emotion and are aggressive. First, based on the perceptions of parents and teachers, is there one or more behavioral profiles that identify children with these tendencies? Second, what characteristics in addition to negative emotionality (the temperamental tendency toward anger and general irritability) are included in the profile? If one or more profiles are isolated by our statistical modeling methods, how do peers of children who exhibit the profile(s) respond to these children?

10.2  Different Profile Types Our research on large samples of US parents and teachers, as well as Russian parents, found that a scale designed to measure negative emotionality (i.e., tendency to express anger and related emotions and behaviors) was substantially related to the tendency to be antagonistic to adult rules and to be argumentative. (For details of these measurement issues, see Appendix B.) Therefore, an aggregate, higher-order indicator was constructed that was labeled the Irritability/Antagonism scale. Three profiles were isolated through our statistical modeling procedures (latent profile analysis) in which there were substantial elevations on the Irritability/ Antagonism scale, as well as high levels of motor activity and poor regulation of attention. However, one of these profiles included very few children and did not occur in all three samples. Therefore, only two profiles with a higher percentage of children from the universal sample (when all three samples were combined) will be described in detail. A brief discussion of the third profile will be presented at the end of the chapter. The two profiles that were identified in all of our samples were differentiated primarily by adult perception of the academic ability of the children. The cluster of children exhibiting one of these profiles was labeled “poorly self-­regulated higher ability” and the cluster of children exhibiting the other was labeled “poorly self-regulated lower ability.” Figure 10.1 presents the two profiles in percentile form. The data presented is based on the combined sample of ratings from US and Russian parents as well as US Teachers (2359 children). Children in both profiles exhibit high levels of gross motor activity, have poor control of their attention (are highly distractible and disorganized), and exhibit high levels of irritability and antagonism. We label both

10.2  Different Profile Types

100 90 80 70 60 50 40 30 20 10 0

121

Higher Ability Lower Ability

Fig. 10.1  Behavior profiles of two types of poorly self-regulated children: higher academic ability and lower academic ability (data in percentile form)

profiles as fitting into the broad category of “poorly self-regulated children.” Self-­ regulation has been shown to be related to several functional systems of the brain, some of which are part of the system related to executive function, including working memory [1]. However, many brain regions and neurochemical processes have been implicated in the complex set of behaviors related to self-regulation [2, 3]. Lower levels of self-regulation of motor activity, attention, and emotion affect academic achievement, motivation, and general educational progress [4–6]. In both profiles, academic achievement motivation is perceived by parents and teachers as being low. On the positive side, both groups are socially engaged, with the poorly self-regulated higher academic ability group having particularly low levels of social withdrawal. Each of the poorly self-regulated profiles is descriptive of a small percentage of children in middle childhood in the general population. In our universal sample, 6.2% of the children were in the high-average academic ability cluster and 6.3% were in the lower academic ability cluster. Both profiles have many characteristics in common with children who are diagnosed with attention deficit hyperactivity disorder (ADHD). Children with ADHD have problems with academic motivation and are characterized by having high levels of poorly controlled gross motor activity and attention problems. They also tend to be disorganized. Further, they are more prone to irritability and antagonistic behavior than are their peers [7]. However, while these profiles have similarities to children with ADHD, both profiles represent a normal variation among children in middle childhood. The profiles represented in Fig. 10.1 are of the average child in the cluster. Thus, some exhibit less extreme examples of the profiles and some more extreme. A few of those who exhibit the more extreme forms may be given a diagnosis of ADHD based on a thorough clinical examination, but such a determination is based on the extent to which the behavior pattern exhibited results in significant dysfunction at home and school. A competent diagnosis is not simply based on a set

122

10  Poorly Self-Regulated Children: Two Temperament Profiles

of behaviors or a trait-based profile. (For a more detailed discussion of the implications of our research for clinical diagnosis, see Chap. 13.) Figures 10.2 and 10.3 provide comparisons of profiles for each of the samples studied. It can be seen that there was considerable agreement across these samples in the configuration of the two profiles. Other researchers have found personality profiles that have similar characteristics to the ones obtained in our research. One of the most influential of these investigations was conducted by Thomas and Chess [8]. They describe three temperament profiles one of which is labeled “the difficult child.” The group is biologically irregular, has a negative withdrawal response to new stimuli, is slow to adapt to change, is emotionally intense, and is often in a negative mood. This pattern of behavior was measured in early childhood. The constellation of behaviors clearly shares the emotional intensity and negative mood aspects of the profiles found in our study in middle childhood. About 10% of their sample displayed this pattern of behavior which is close to the 12% found when both clusters of children exhibiting the poor self-regulated profiles we obtained when the samples are combined. Asendorpf and van Aken obtained mother and teacher ratings of personality characteristics using a method of sorting descriptive adjectives [9]. Measurement took place at ages 4 and 10. Three groups of children were identified using this method. One was labeled “under-controlled children” and it represented 19% of 10-year-old sample. This group was energetic and lively, and frequently expressed negative feelings directly. They were also described as stubborn. One of the most influential studies of behavioral profiles in children is the Dunedin Multidisciplinary Health and Development Study [10]. Three-year-old children participated in a 90-minute testing session involving assessment of cognitive and motor ability. Following the testing session, the examiner rated the child’s behavior on 22 behavioral characteristics. Five homogeneous and mutually 100 90 80 70 60 50 40 30 20 10 0

U.S. Parents U.S. Teachers Russian Parents

Fig. 10.2  Comparison of the poorly self-regulated higher ability profiles obtained from US parents and teachers and Russian parents (data in percentile form)

10.2  Different Profile Types

100 90 80 70 60 50 40 30 20 10 0

123

U.S. Parents U.S. Teacher

Fig. 10.3  Comparison of the poor self-regulated lower ability profile obtained from US parents and teachers (data in percentile form)

exclusive clusters were identified. One group was labeled “under-controlled” (10.4% of the sample; 62% male) and these children were described as irritable, impulsive, and not persistent in solving problems. They had difficulty sitting still and had large, rapid changes in mood. At age 18, adolescents who were in the under-controlled group at age 3 described themselves as danger-seeking, impulsive, prone to respond with strong negative emotions to everyday events, and were enmeshed in adversarial relationships. By age 21, they were more likely to have a psychiatric disorder than were persons who had been placed in the other groups at age 3. They were also more likely to have been diagnosed with multiple disorders. The under-controlled group was 2.9 times as likely to be diagnosed with antisocial personality disorder, 2.2 times as likely to be recidivistic offenders, and 4.5 times as likely to be convicted for a violent offenses compared with children in this large sample [11]. Kamphaus and colleagues [12–14] published a series of studies aimed at developing and validating a typology of child behavior using measurement tools designed primarily to assess behavior problems. The measurement instrument used was the Behavior Assessment System for Children (BASC), which is now the most widely used measure of behavior problems in the public schools of the United States. This measure also includes scales that measure more positive attributes such as leadership and social skills. One study [14] was based on the US normative sample for the Parent Rating Scales of the BASC. The sample included 2029 children of ages 6–11. The research team isolated nine clusters of children including one group with similarity to those identified in our universal sample as poorly self-regulated. This cluster of children was referred to as the Disruptive Behavior Problem Group (9% of the sample; 74% male). This cluster of children had very high scores on the aggression, hyperactivity, and conduct problems scales and also had important elevations on the attention

124

10  Poorly Self-Regulated Children: Two Temperament Profiles

problems and depression scales. With the exception of the depression scale, a construct we did not use to identify profiles, this cluster had much in common with the two groups we isolated in our model using temperament-based dimensions as opposed to behavior problem indicators. In summary, several other research groups who have utilized temperament, personality, and/or behavior problem indicators of behavioral tendencies have identified 9% to 19% of their samples as having a behavior profile similar to the pattern observed in our samples. In our study, two groups were isolated, whereas in other studies only one profile had similar characteristics. This difference is primarily attributable to our use of parent and teacher perceptions of academic ability (e.g., intelligence) and achievement motivation as indicator variables in our analyses, which were useful in differentiating between the two types of poorly self-regulated children.

10.3  Demographic Characteristics Analyses of our three samples, in combination, indicated that there were significantly more boys than girls exhibiting the poorly self-regulated higher ability profile (61.9% males). This finding is consistent with most research on attention, hyperactivity, and conduct problems in children [12, 13]. Interestingly, there were no gender differences for the poorly self-regulated lower ability profile. This may indicate that the two poorly self-regulated profiles have a different etiology. In the case of the lower ability profile, attentional problems and other self-control problems may be primarily the result of problems associated with academic ability. For these children, most traditional schooling may not fit their cognitive needs and abilities. Many of these children may fall into the category referred to as “slow learners.” These children do not have diagnostic intellectual disabilities, but simply are much more challenged by the academic demands in the traditional classroom. With regard to ethnic differences, children in the US parent and teacher samples were divided into three groups based on parental report: European American, African American, and Other. (The number of minority children other than African Americans was too small in these data sets to provide meaningful analyses, so they all were placed in the “other” category.) There was no indication of an overrepresentation of any race/ethnic group in the two poorly self-regulated clusters.

10.4  Academic Ability and Motivation Parents and teachers provided data relevant to the perception of academic ability and achievement motivation. These ratings contributed to the latent personality profiles defined as poorly self-regulated higher ability and poorly self-regulated lower ability. Children exhibiting the higher ability profile were perceived by parents and teachers as exhibiting an average to above-average academic ability, while children

10.4  Academic Ability and Motivation

125

in the lower ability cluster were perceived as having significantly below-average academic ability. Both groups were perceived as having low academic motivation. How accurate were the parental and teacher estimates of academic ability and motivation? Our research did not have access to individually administered intelligence test data. However, in the Russian sample, the grade point average (GPA) of each student was available, and school grades are, of course, a reflection of academic ability and motivation. Parents reported the grade point average based on the most recent report card. Grades were assigned to Russian students on a 5-point scale with “5” indicating the highest level of achievement. Children who exhibited the poorly self-regulated higher ability profile had an average GPA of 3.89 (s.d. = 0.54). This score was marginally below the average GPA of the Russian sample. Based on all information obtained from Russian parents including their ratings of academic motivation, it seems likely that the parents believe that their children in this cluster are under-achieving, as they were perceived to be somewhat above average in ability. (No data were available for the lower ability profile as this profile did not occur in the Russian sample.) The next question that we addressed was, “Did the student peers of children in the two poorly self-regulated groups perceive these children in the same way that parents and teachers perceived them?” With regard to peer perceptions of academic ability and motivation, data from two cohorts of students in the US teacher sample were analyzed. The two cohorts were analyzed separately because the cohorts had somewhat different socioeconomic and ethnic characteristics. Each child nominated classmates that they perceived as academically talented and motivated to achieve. The children in both cohorts were asked to choose from the following options: ‘This person makes good grades, is smart, and usually knows the right answer.’ ‘This person tries hard to do good school work.’ Scores were given to each child based on the number of nominations they received from all of their classmates. Self-reports of their own perceptions of their academic ability and motivation were available for cohort B of the Georgia student sample (446 children). Each child was asked: ‘If you were to list all of the students in your grade from worst to best in school work, where would you put yourself?’ ‘How good would you be at learning something new in school?’ Responses were as follows: “1” = one of the worst through “5” = one of the best. The responses to the two scores were summed to obtain the self-assessment of academic ability. Academic motivation was assessed with one question. “For me being good at my schoolwork is … (1 = not important; 5 = very important).” Figure 10.4 presents the combined parent/teacher assessment of academic ability and motivation based on parent and teacher perceptions (the universal model) for the poorly self-regulated higher ability cluster of children. These assessments are compared to the peer assessments of ability and motivation as well as self-ratings. This figure reveals that the combined parent/teacher assessment of academic ability was somewhat above average while the assessment of academic motivation was very low. Peers did not make a meaningful distinction between ability and motivation, and viewed this cluster of children as being below average on both ability and motivation. When these same children rated themselves, a similar pattern emerged. These children perceived that their ability and motivation were below average.

126

10  Poorly Self-Regulated Children: Two Temperament Profiles 70 60 50 40 Academic Ability

30

Academic Motivation

20 10 0 Parents & Peers-CohortPeers-Cohort Self-Cohort Teachers A B B

Fig. 10.4  Comparison of parental (US and Russian), teacher, peer, and self-ratings of academic ability and motivation for children exhibiting the poorly self-regulated higher ability profile (data in percentile form)

45 40 35 30 25 Academic Ability

20

Academic Motivation

15 10 5 0 Parents & Peers-CohortPeers-Cohort Self-Cohort Teachers A B B

Fig. 10.5  Comparison of parent/teacher, peer, and self-perceptions of academic ability and motivation for the poorly self-regulated lower ability cluster (data in percentile form)

Figure 10.5 presents that data for children in the poorly self-regulated lower ability cluster. Adult assessment indicated that this cluster had very low academic ability and motivation. However, the peer group perceived these children as having about as much ability and motivation as the poorly self-regulated higher ability group. Further, self- assessments were consistent with the peer group assessment. Thus, the students did not differentiate between the two profiles in terms of academic ability the way that parents and teacher did. In summary, based on grade point average, peer assessment, and self-report from students themselves, both the poorly self-regulated groups were performing at a

10.6  Prosocial Behaviors and Likeability

127

below-average level (around the 30th percentile on average) and their academic abilities and motivations were viewed as being in the same range. Parents and teachers viewed the abilities and motivations of these two groups to be quite different and made a much greater differentiation between ability and motivation than did peers and the children themselves. However, all raters perceive these students to be performing academically at below-average levels.

10.5  Compliance In all families and in educational institutions, there are explicit rules and norms for appropriate social behavior of children. These rules and norms may differ by culture and socioeconomic status, and they are influenced by the personalities of the individual adults who interact with the children. However, there are clear individual differences in the extent to which children meet these expectations. One way to measure these individual differences is to ask the adults in the situation how compliant their child (or student) is with behavioral rules and norms. All parents and teachers completed a multi-item compliance measure that assessed social sensitivity to adult expectation as well as obedience. Since these two groups of children were defined as poorly self-regulated in terms of emotional expression, activity level, and ability to focus attention, we expected that they would be perceived as being below average in general compliance with adult rules. Further, we expected that the children in the poorly self-regulated higher ability cluster would have somewhat higher compliance scores than those in the poorly self-­ regulated lower ability cluster. When the US and Russian parent and US teacher samples were analyzed separately, all samples rated the compliance of the poorly self-regulated higher ability cluster of children at about the 30th percentile and the poorly self-regulated lower ability cluster at the 9th percentile. There was also considerable uniformity in these ratings across the three samples when analyzed separately. Thus, children in the poorly self-regulated higher ability cluster, while somewhat below average in compliance, were viewed as substantially more compliant than their peers in the poorly self-regulated lower ability cluster.

10.6  Prosocial Behaviors and Likeability Prosocial behaviors, as defined in our research, include positive emotional reactions to people as well as empathetic, caring responses to others. As part of the indicators used to empirically develop personality profiles of children in middle childhood, parents in the US and Russian samples, as well as a sample of US teachers, rated children on their tendency to express positive emotions (happiness, joy) and their tendency to be empathic (to express understanding and helpfulness toward others who were sad or having other difficulties). Children in the poorly self-regulated

128

10  Poorly Self-Regulated Children: Two Temperament Profiles

higher ability cluster were perceived by parents and teachers as having below-­ average levels of expressions of happiness and empathy for others, and children in the poorly self-regulated lower ability cluster were perceived by adults as exhibiting very low levels of these behaviors. Measures of these prosocial tendencies were part of the higher-order indicator variables used to identify the temperament profile that defined each cluster. The question we addressed in our research was, “Do peers of children perceive the prosocial tendencies of these children the same way parents and teachers rated them?”. These data have been presented in Fig. 10.6. Data for two large cohorts of students in the US teachers’ sample are presented separately, because these cohorts differed slightly in age distribution and ethnic and socioeconomic characteristics (for details regarding the measure of these tendencies, see Appendix H). In addition to measures of prosocial tendencies, we also obtained peer-report measures of how much they like and dislike to play with their same-age classmates, and these reports were used to create a score for each child (see Appendix H). This measure was considered as a concrete indication of the prosocial tendencies of the child. Figure 10.6 indicates that there were important cohort differences in peer perceptions of prosocial behavior and likeability. For children in cohort A, peers perceived the poorly self-regulated higher ability cluster as marginally below average in exhibiting positive emotion, but in the average range for expressions of empathy. This cluster was also perceived as average in likeability. In cohort B, both the empathy and likeability ratings were significantly below average. For the poorly self-­ regulated lower ability cluster, these children were significantly below average in both cohorts in all aspects of prosocial behavior as well as likeability (around the 30th percentile). Cohort A contains somewhat older students (a higher percentage in the fifth grade) than cohort B, and a higher percentage of the children are African 60 50 40 30 20 10

Higher Ability

0

Lower Ability

Fig. 10.6  Peer perceptions of prosocial behavior and likeability for the poorly self-regulated children in two student cohorts

10.7  Social Status

129

American children. How each of these factors affected these differences in peer perceptions of prosocial behavior remains unclear. There is one important caveat to the description just given. There was a large gender difference in expressions of prosocial behavior of boys and girls. In every analysis, girls in the higher ability cluster exhibited more prosocial behaviors than boys, while in the lower ability cluster, boys exhibited more prosocial behaviors and were more likeable than girls. Because the number of children in these clusters is small, division by gender produces even smaller groups of children. Thus, these data are only suggestive, but this kind of interaction of gender by ability level for poorly self-regulated children deserves further study.

10.7  Social Status In addition to likeability, a dimension of social status indicating peers’ affective regard for the child, we investigated four different aspects of reputational aspects of social status, using data from the US teacher sample obtained by Lease and colleagues: popularity, leadership, admiration, and social prominence (standing out from the crowd). Students in the fourth and fifth grades (cohort A) nominated classmates for the following four items, whereas students in the third through fifth grades were asked to nominate those from their grade level (see Appendix J). 1. Which of your classmates are the most popular at school? Which are the least popular? (a popularity score is created by subtracting the score for least popular from the score for most popular). 2. This person gets chosen by the others as the leader. Other people like to have this person in charge. 3. This is a person who others in class admire. Other children want to be like this person and to be around him/her. 4. This person is really cool. Just about everybody in school knows this person. The data from these two sizeable cohorts of Georgia students presents a relatively consistent and revealing picture of the reputational status of children who exhibit both profiles of poorly self-regulated behavior (see Fig. 10.7). (For details regarding the measure of these tendencies, see Appendix J.) The most interesting ­finding is that children in both clusters were near average in their social status across all of these assessed aspects of reputational status. While the children who exhibited the poorly self-regulated lower ability profile had slightly lower levels of social status in all four areas than the children who exhibited the poorly self-­ regulated higher ability profile, their social status was not as low as their very low levels of prosocial behavior and likeability as described above. Further, the children in the poorly self-regulated higher ability cluster were perceived by their peers as having above-average leadership qualities and as more cool and more well known than many of their peers (between the 60th and 70th percentile on average). There were no important differences between cohorts of children in the perception of these children.

130

10  Poorly Self-Regulated Children: Two Temperament Profiles 80 70 60 50 40 30 20 10 0

Higher Ability Lower Ability

Fig. 10.7  Comparison of social status scores of two clusters of poorly self-regulated children in two student cohorts

How do poorly self-regulated higher ability children obtain levels of social status within their peer group that seems at odds with their academic achievement, their prosocial behaviors, and their likeability? The answer seems to lie in their social integration and their assertiveness. These children are not shy. Further, they are perceived as not often feeling insecure or fearful. Their teachers and parents rate them as less insecure as a group than all other clusters we isolated with the exception of the well-adjusted high achieving cluster. This group is not afraid to push the boundaries of adult expectation. Finally, they are probably viewed by peers as slightly dangerous. It is clear that they are perceived as emotionally labile (irritable) and more frequently expressive of negative emotion than are their peers. So their mood and behavior are probably carefully monitored by more fearful children. Thus, other children know them and talk with others about them. Children in this group are not necessarily the child that others feel comfortable with as a play companion, but they are perceived as dominant and socially prominent.

10.8  Influence on Peers Reputational status in the peer group is likely to be related to the amount of social influence children have over others. Thus, because children who exhibit the poorly self-regulated higher ability profile are perceived as leaders and are well known to other children, they seem likely to exercise influence on the behavior of others in their peer group. However, their influence may vary across different types of social interactions or activities. Data obtained by Lease and colleagues on the two cohorts of students looked at the influence each child has in five areas  as perceived by peers: academics, sports, social trends (e.g., clothing, slang, music), make-believe games (e.g., role playing and computer games), and inappropriate or rule-breaking behaviors (e.g., talking back to teachers). (see Appendix I). The results are presented in Fig. 10.8.

10.9  Behavior Problems

100 90 80 70 60 50 40 30 20 10 0

131

Higher Ability Lower Ability

Fig. 10.8  Comparison of social influence of two clusters of poorly self-regulated children in two student cohorts (data in percentile form)

This graph makes it clear that the influence that children have differs by  the nature of their peer group. Children in the poorly self-regulated clusters had more influence in all areas in cohort A than in cohort B. This is particularly true of influencing others in in the areas of academics, cultural trends (e.g., clothing, music, slang), and make believe or computer games. The largest difference between the cohorts was in the area of academic influence. In cohort B which appeared to be a somewhat more academically oriented set of schools, the academic influence of poorly self-regulated children was much lower than in cohort A. The students in cohort A were drawn from schools that served somewhat lower socioeconomic families than those in Cohort B. Thus, in this context, higher levels of expression of negative emotion, being antagonistic toward authority as well as being inattentive, and disorganized and poor control of motor activity were positively related to the influence the child had on other children. In both cohorts, both types of poorly self-regulated children were perceived by their peers as having the most influence on others with regard to behavior considered inappropriate by the teacher. These children led others in “fooling around” when the teacher was not looking or when he/she left the room. Also, these children seem to engage in antagonistic behaviors more frequently than their peer such as “talking back to teachers.”

10.9  Behavior Problems Russian parents responded to a measure of behavior problems using the Strengths and Difficulties Questionnaire (SDQ), a widely used measure to screen for behavioral, social, and emotional problems in children [15, 16]. Four behavior problem

132

10  Poorly Self-Regulated Children: Two Temperament Profiles

scores are obtained from this measure: hyperactivity, emotional problems (primarily anxiety and depression), conduct problems (rule breaking and antisocial behaviors), and peer problems (alienation from peers). When the Russian sample was analyzed, the only poorly self-regulated cluster that was isolated was the average ability group. Thus, no data are available for the poorly self-regulated low ability cluster. For the poorly self-regulated higher ability cluster, a high percentage of these children were perceived by parents to have problems with hyperactivity (50%) and conduct (38.9%). Both of these percentages were much higher than the rate for all Russian children studied (16%). On the other hand, a lower percentage of this cluster of students obtained significant scores for parent-rated emotional problems (5.6%) and peer problems (11.1%). (Significant scores were those above the one standard deviation above the mean – above the 84th percentile.) Parental measurements of behavior problems are likely to be affected by their perceptions of the temperamental characteristics of their children. Since the poorly selfregulated children were rated by parents as highly active, irritable and antagonistic, and have problems with attention, it is not surprising that they also perceived a high percentage of these children as having significant levels of hyperactivity and conduct problems (antagonistic responses to authority as well as rule breaking and aggression). Thus, we needed to know the extent to which independent sources (student peers and the children themselves) thought that they had significant behavior problems. In the US teacher data, students were asked to nominate children in their classroom who were particularly socially aggressive (e.g., attempting to exclude others from their group), verbally aggressive (e.g., saying unkind things to others), or physically aggressive (e.g., pushes or hits others). Figure 10.9 shows the comparison between poorly self-regulated average ability students and poorly self-regulated lower ability students on these measures of aggression. As somewhat different measurements of aggression were obtained from two separate cohorts of students (i.e., questions were asked in a somewhat different manner), these data are separated by cohort. Analysis of Fig.  10.9 reveals that children who exhibit the two temperament profiles described as poorly self-regulated were perceived by their peers as more physically, verbally, and socially aggressive than most of their peers. Further, the poorly self-regulated lower ability cluster had higher scores on all three types of aggression than the poorly self-regulated higher ability group. The lower ability group had aggression scores in the 70th to the 90th percentile on average. This pattern of results was consistent across the two student cohorts that differed slightly in age, social class, and racial/ethnic make-up. Cohort A also completed a screening instrument designed to measure self-­ perceptions of multiple types of behavior problems. We focused on self-perceptions of anxiety and depressive symptoms as these relate to the emotional problems reported by Russian parents. Since the two poorly self-regulated groups had high scores on aggression, particularly the children with lower academic ability, we also examined self-perceptions interpersonal problems. Poorly self-regulated children (on average) did not self-report unusual levels of anxiety, depression, or interpersonal problems. Specifically, their self-assessment scores were very close to the average of the scores of all children (the 50th percentile).

133

10.9  Behavior Problems 100 90 80 70 60 50 40 30 20 10 0

Higher Ability Lower Ability

Fig. 10.9  Comparison of two clusters of poorly self-regulated children in two student cohorts on peer perceptions of three types of aggression (data in percentile form)

Table 10.1  Percentage of children in the two poorly self-regulated clusters who have significant behavior problemsa based on peer and self-report Source Peer

Self

Problem type Physical aggression – Ab Verbal – A Social aggression – A Verbal/physical aggression – Bb Social aggression – B Anxiety – A Depression – A Interpersonal – A School related – A

Poorly self-regulated Higher academic ability 25.0 20.0 30.0 15.6 28.1 05.3 05.3 00.0 21.1

Lower academic ability 56.4 51.3 28.2 32.5 42.5 21.6 21.6 11.8 23.5

Significant problems were defined as scores at or greater than the 84th percentile for the sample of children studied (> = 1 sd) b “A” signifies Georgia students in cohort A; “B” signifies Georgia students in cohort B a

Further, the mean scores of the average ability and lower ability clusters were very similar. Thus, despite parents, teachers, and peers perceiving these two clusters to have large differences in behavior problem symptoms, the children themselves maintain a self-perception that they exhibit average levels of problematic behavior. The data just provided describe the mean scores of the two profiles of poorly selfregulated children. Because behavior problems are typically highly skewed (most children have few problems and a few children have many), it is important to determine the percentage of children within each profile who have significantly high levels of these problems. “Significant” was defined as a score that was one standard deviation above the mean for the sample (greater than the 84th percentile). Table 10.1 presents the percentage of children in both clusters who had significant behavior problems scores.

134

10  Poorly Self-Regulated Children: Two Temperament Profiles

With regard to aggressive behavior, a somewhat higher-than-average (greater than 16%) percentage of children exhibiting the poorly self-regulated higher ability profile were reported by peers to be aggressive in all its forms (social, verbal, physical). However, a much higher percentage of children exhibiting the poorly self-­ regulated lower ability obtained peer nomination scores that were in the significant range. Self-assessments of anxiety, depression, and interpersonal problems  indicated that a very low percentage of children in the poorly self-regulated higher ability cluster had significantly high scores.

10.10  The Highly Achieving, Highly Emotional Profile The eight-profile model of ratings by US parents and teachers revealed a small group of academically talented children who exhibited high levels of negative emotionality. This group included 4.6% of the parent-assessed sample and 3.3% of the teacher-assessed sample. The profile is graphed in Fig. 10.10. This profile describes a group of children who are bright and highly motivated to achieve in school. This cluster of children is further distinguished by high ratings on the irritability/antagonistic scale as well as on the insecure/fearful scale. Thus, this group displays both aspects of the temperamental construct of negative emotionality. They have difficulty controlling their responses to frustration which appears to manifest as anger and being argumentative, and they also have difficulty controlling their negative emotionality in response to situations that illicit fear or social insecurity. This profile is theoretically interesting as well as of clinical interest. However, because this profile did not occur in the Russian parent-rated sample when analyzed 100 90 80 70 60 50 40 30 20 10 0

U.S. Parents U.S. Teachers

Fig. 10.10  Comparison of US parent and US teacher profiles for the poorly regulated highly emotional cluster (data in percentile form)

References

135

separately, and because in the aggregate sample this profile included a very small number of children (1.9%), we did not analyze it further. More research with larger and more culturally diverse samples is required to determine the meaningfulness and generalizability of this profile.

References 1. Hofmann, W., Friese, M., Schmeichel, B.  J., & Baddeley, A.  D. (2011). Working memory and self-regulation. In K. D. Vohs & R. F. Baumeister (Eds.), Handbook of self-regulation: Research, theory, and applications (2nd ed., pp. 204–225). New York: Guilford. 2. White, L. K., Lamm, C., Helfinstein, S. M., & Fox, N. A. (2012). Neurobiology and neurochemistry of temperament in children. In M.  Zentner & R.  L. Shiner (Eds.), Handbook of temperament (pp. 347–367). New York: Guilford. 3. Bell, M. A., & Deater-Deckard, K. (2007). Biological systems and the development of self-­ regulation: Integrating behavior, genetics, and psychophysiology. Journal of Development and Behavioral Pediatrics, 28, 409–420. 4. Martin, R. P. (1989). Activity level, distractibility, and persistence: Critical characteristics in early schooling. In G. A. Kohnstamm, J. E. Bates, & M. K. Rothbart (Eds.), Temperament in childhood (pp. 451–462). New York: Wiley. 5. McClelland, M.  M., & Cameron, C.  E. (2011). Self-regulation and academic achievement in elementary school children. In R. M. Lerner, J. V. Lerner, E. P. Bowers, S. Lewin-Bizan, S. Gestsdottir, & J. B. Urban (Eds.), Thriving in childhood and adolescence: The role of self-­ regulation processes. New Directions for Child and Adolescent Development, 133, 29–44. 6. Nelson, B., Martin, R. P., Hodge, S., Havill, V., & Kamphaus, R. (1999). Modeling the prediction of elementary school adjustment from preschool temperament. Personality and Individual Differences, 26, 687–700. 7. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, D.C.: American Psychiatric Publishing. 8. Thomas, A., & Chess, S. (1977). Temperament and development. New York. 9. Asendorpf, J. B., & van Aken, M. A. G. (1999). Resilient, over controlled, and under-­controlled personality prototypes in childhood: Replicability, predictive power, and the trait-type issues. Journal of Personality and Social Psychology, 77, 815–832. 10. Caspi, A., & Silva, P. A. (1995). Temperamental qualities at age 3 predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486–498. 11. Caspi, A., Moffitt, T. E., Newman, D. L., & Silva, P. A. (1996). Behavioral observations at age 3 years predict adult psychiatric disorders. Archives of General Psychiatry, 53, 1033–1039. 12. DiStefano, C., Kamphaus, R. W., Horne, A. M., & Winsor, A. P. (2005). Behavioral adjustment in the U.S. Elementary school: Cross-validation of a person-oriented typology of risk. Journal of Psychoeducational Assessment, 21, 338–357. 13. Kamphaus, R.  W., Huberty, C.  J., DiStefano, C., & Petoskey, M.  D. (1997). A typology of teacher-rated child behavior for a national U.S. sample. Journal of Abnormal Child Psychology, 25, 453–463. 14. Kamphaus, R.  W., Petosky, M.  D., Cody, A.  H., Rowe, E.  W., Huberty, C.  J., & Reynolds, C. R. (1999). A typology of parent rated child behavior for a national U.S. sample. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 40, 1–10. 15. Goodman, R. (1997). The strengths and difficulties questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581–586. 16. Goodman, A., & Goodman, R. (2009). Strengths and difficulty questionnaire as a dimensional measure of meatal health. Journal of the American Academy of Child and Adolescent Psychiatry, 48, 400–403.

Chapter 11

Fostering the Development of Poorly Self-­Regulated Children

11.1  Summary of Temperamental Characteristics We have identified two temperament profiles in middle childhood that are characterized by (1) high levels of (often undirected) energetic motor behavior, (2) frequent and intense expressions of negative emotion (e.g., screaming, crying), (3) high levels of socially antagonistic behavior (e.g., disobedience, argumentative behavior), (4) poor ability to attend (e.g., difficulty concentrating on learning tasks), and (5) problems with disorganization (e.g., loses belonging). Most children in middle childhood occasionally lose their belongings, argue with the parents, become intensely emotionally upset, and run around the house. However, the children exhibiting the two profiles that we identified are viewed by parent, teachers, and peers as exhibiting these behaviors more often than most other 8- to 12-year-old children. A few generalizations can be made about the two groups of children we have isolated. First, they have more academic ability in the eyes of parents and in their own estimation than is reflected in their achievement. However, their academic motivation is low, as perceived by all reports (parents, teachers, peers). They also exhibit less empathy for other children’s misfortunes than it is typical for their age group. This is particularly true of boys and for children in the lower academic achievement profile. They perceive themselves to be more aggressive than other children and this perception is shared by their peer group. The aggression that is exhibited is primarily verbal or involves social exclusion (higher for girls), and they also engage in more pushing and other physically aggressive acts (higher for boys) than do other peers. One of the more interesting outcomes of our research was to show that both clusters of poorly self-regulated children had more influence on other children than the average child. However, the influence is primarily negative. This influence spanned several areas from academic achievement to influencing other children to behave badly in the classroom.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_11

137

138

11  Fostering the Development of Poorly Self-Regulated Children

The poorly self-regulated children we identified were classified into two groups: those with average to above-average academic ability and achievement and those who had lower academic ability. The personality characteristics of these two groups had different effects on their social environment. The more academically able children exhibited fewer behavior problems.

11.2  The Meaning of Poor Self-Regulation Currently, in American child psychology, children who exhibit this pattern of behaviors are thought of as having a poor ability to self-regulate their behavior. This view has been heavily influenced by the theoretical view of Rothbart and colleagues [1, 2] who define temperament as constitutionally based individual differences in reactivity and self-regulation. The reactivity component of this theory makes the assumption that individual differences in energy level, attention, and expression of emotion are linked to the physiological excitability of the neurological system. Self-­ regulation refers to those psychological processes that enable the individual to modulate their automatic, involuntary reactivity. Derryberry and Tucker [3] describe four motivational brain systems that respond to environmental stimulation to increase excitability; individuals are assumed to have large individual differences in the sensitivity of these systems. They include (a) an appetitive system responsible for sensitivity to rewards (e.g., food), (b) the defensive or fearful motivational system that is responsible for sensitivity to signals of harm, (c) the frustration/aggression system that is responsible for cueing defensive aggression, and (e) the affiliative system that is responsible for nurturing and regulating social interaction. The self-­ regulatory components are assumed to be controlled by three attentional systems. The most important of these in the Rothbart and Derryberry model is the anterior attentional system that is assumed to regulate effortful control which involves conscious efforts to regulate the sensitivity of the motivational systems. If we consider the appetitive system, some individuals are particularly sensitive to rewards associated with food. They may not attend to cues that they have eaten enough to satisfy their biological needs, so they continue to eat because it brings pleasure. Not surprisingly, persons who have this constitutional predilection tend to be overweight. (There are many reasons to be overweight and this is only one factor.) However, self-regulation through effortful control would involve being aware of these tendencies and controlling consumption of food through calorie counting, limiting the amount of snack foods in the home, or any number of other behaviors. The poorly self-regulated profiles we have developed from parent and teacher descriptions of child behavior are likely to be reflective of children who are particularly sensitive to reward. This sensitivity may explain problems with sustained attention and their relatively poor academic achievement. Children in both poorly self-regulated profiles are socially engaged; social interaction is a source of reward for these children. In support of this idea, we found that children exhibiting these profiles were socially prominent; that is, they were considered to be “cool” and were

11.2  The Meaning of Poor Self-Regulation

139

well known to others in at school. Being sensitive to cues of reward and finding social interactions rewarding means that it is likely that these children find social interactions on their phone or paying attention to others in classrooms much more rewarding than trying to read and learn difficult material. Further, getting the attention of others is rewarding, so when the teacher leaves the room, it is likely to be much more fun for these children to “fool around,” to be silly, or to engage in other inappropriate behavior. We found that children in the two poorly self-regulated profiles were strongly motivated to engage in inappropriate behavior  (by their own admission) when the teacher leaves the classroom and, in turn, were viewed by their peers as highly influential in leading others to engage in such inappropriate classroom behavior. Being particularly sensitive to reward also helps explain why children exhibiting these two profiles are easily frustrated. Children who exhibit high levels of irritability/antagonism, activity level, and attention/organization problems have difficulty complying with family and school behavioral rules and norms. School and parental rules often frustrate the desire for some form of immediate gratification. For example, a child who wants to spend the night with a friend might not be granted parental permission due to family plans. This situation is likely to result in arguments, anger, or vigorous expressions of frustration (slamming doors) because it frustrates access to social rewards. Peers may also frustrate the path to a reward and be the focus of anger and aggression. The frustration/aggression system is another system of Rothbart’s motivational components that seems to be implicated in the behavior profiles of children who are poorly self-regulated. Negative emotionality is at the heart of the tendency to be irritable and to engage in aggressive behaviors. The negative emotional component might be expressed in the form of anger because an approach to some potential reward has been blocked. Researchers have found that the tendency to experience anger has not only emotional components that contribute but also several cognitive contributors. These include a tendency to attribute hostile intentions to the behavior of others in ambiguous situations (e.g., my parents are unfair and my teacher doesn’t like me). There is also a tendency to ruminate about the perceived provocations of other people. These cognitive and perceptual biases can lead to hostile and argumentative behavior and tend to increase the probability of a child to respond with anger and aggression [4]. Children with these tendencies are often seen as having a difficult temperament and as undercontrolled [5, 6]. But as the Derryberry and Rothbart theoretical model points out, the behavior that is expressed in any environmental circumstance (e.g., at home or at school) is not only a function of the sensitivities of different motivational systems but also the ability of the child to self-regulate these tendencies. Self-regulation, according to this model, has automatic components such as the orienting response (e.g., looking toward an unexpected noise) and fear (e.g., immediate response to seeing a snake). But most attention in this model is placed on effortful control, the voluntary control of behavior and emotion. Effortful control is defined as the ability to inhibit a dominant response in order to perform a subdominant response. The notion is that a child who has a tendency to get out of their seat in the classroom to go talk to another

140

11  Fostering the Development of Poorly Self-Regulated Children

child would be exercising effortful control if they controlled this impulse by saying to themselves, “I can talk to her after class” and then refocusing on their reading material. Effortful control as a form of self-regulation is most clearly evident when the child controls a dominant response when there is no adult monitoring of the behavior. It is also likely that self-regulation is a function of the maturity and functional level of several central nervous system processes often described as executive functioning mechanisms. These involve attention focusing, working memory, and cognitive inhibition functions. These brain functions begin their long maturation process in the second year of life, but this maturation is not complete until later adolescence or early adulthood. Thus, some problems with self-regulation in children may be due to physiological issues related to brain maturation. Support for this idea comes from observations of aggression in children at different ages. Low to moderate aggression is relatively common early in the preschool years, which then gradually declines from the fourth year of life onward [7]. This follows the time of brain development in areas related to executive functioning. Further, several researchers have found that irritability/antagonism is associated with poor central nervous system functioning that supports effortful control [8, 9]. We have labeled the clusters of children who exhibit high levels of motor behavior, distractibility, and irritability/antagonism as being poorly self-regulated. However, there is no way to know from the current data whether these children have particularly high sensitivity to reward, have unusually strong affiliative motivations, or whether they have low levels of self-regulation. All the explanations are possible. In the psychological literature, there is a tendency to place the emphasis on effortful control and other self-regulatory mechanisms because these skills, in part, seem to be the result of experience and can be taught. Expressions of anger, behavioral antagonism, and excess gross motor activity typically decrease as development progresses from early childhood through late adulthood. Such changes are often attributed in part to changes in self-regulation through socialization [10]. However, there is evidence from the aggression literature that high rates of aggression during the preschool years are predictive of increasingly frequent and intense aggressive behavior later in childhood, adolescence, and early adulthood [11, 12]. Is this persistent form of aggression due to very strong biologically based tendencies toward aggression, does this result from problems in brain maturation in areas associated with executive functioning and self-regulation, or is it the result of poor parental socialization practices (to mention only some of the possibilities). The jury is still out.

11.3  Impulsive Extraverts Extraversion is a broad personality trait described as a tendency to be sociable, expressive, high-spirited, lively, socially potent, and physically active. Extraverts also are characterized by an eagerness to approach rewarding situations [13]. A

11.4  Differences Between Profiles

141

brief look at the two profiles we identified with poor self-regulation indicates that these children exhibit many behavioral tendencies characteristic of extraversion. That is, children described in both profiles were rated as highly active and social (low to average levels of social withdrawal). We have argued that they are likely to be sensitive to cues of reward. They also are reported by peers to be among the most socially prominent in their peer group. One way to think of the two clusters of children who are poorly self-regulated is to think of them as impulsive extraverts. By impulsive, we mean the tendency to act on immediate urges before considering the possible negative consequences. For an impulsive child, long-term goals are not a priority or they consider the goals less important than the reward associated with the impulse. Many parents have had the experience of asking a child after they have engaged in a forbidden or dangerous activity, “What were you thinking?” The answer almost always is that the child was not thinking; they were simply responding. “The money was left on the counter and I just took it.” “Richard called me a name so I hit him.” “I told the teacher that the assignment was stupid because it was.” What is missing most often with children in these situations is the ability to slow down the dominant response tendency long enough to consider the consequences. Most children have limited access to the things they desire, including preferred foods (e.g., candy, soft drinks), preferred play opportunities, attention of peers and adults, etc. Limited access might be due to lack of opportunity, but it also might be due to parental or teacher restriction or lack of attention by parents and teachers to the child’s needs. Under these circumstances, acting in such a way that maximizes the current opportunity to obtain these desired circumstances is a logical and adaptive thing to do. Often this will look like an impulsive act. Thus, behavior that is often referred to as a lack of self-control of motor behavior, attention, and emotional expression may in fact be an adaptive response to an environment that is unresponsive to the child’s needs. Several studies have shown that children will adjust their delay of gratification in rational ways depending on the probability of receiving a reward later and the magnitude of the reward [14, 15]. This discussion of the impulsivity, its causes and relations to current thinking about extraversion and self-regulation, indicates that there are multiple perspectives on the psychological basis for the behavioral profiles we have labeled “poorly self-­ regulated.” Currently, the thinking of the temperament research community, particularly influenced by the work of Rothbart and colleagues, has the most well-developed conception of the psychological processes involved.

11.4  Differences Between Profiles The discussion to this point has addressed the general behavioral tendencies of children who have poor self-regulation. However, our research found that parent and teacher ratings of temperament identified two different profile types. Children exhibiting either profile are characterized by high levels of motor activity, high

142

11  Fostering the Development of Poorly Self-Regulated Children

levels of negative emotionality, and poor control of attention. However, the two profiles describe children who are different in significant ways. First, parents and teachers attribute higher academic ability to one cluster of children than the other. Thus, the profiles are labeled as ‘higher achieving poorly self-regulated’ and ‘poorly achieving poorly self-regulated’. Further, children in the higher achiving profile exhibit average levels of prosocial behavior (empathy for others and expression of happiness and joy), whereas children in the lower achieving profile have significantly lower levels of prosocial emotional expression. Also, the level of irritability and antagonism expressed by the higher achieving group, while above average, is still lower than the high level observed in the lower achieving group. Finally, the higher achieving group has a very low score on social inhibition and withdrawal. They are highly social, while the lower achieving group is perceived to be average in social engagement. In support of the idea that poorer academic ability and/or achievement is central to the differences between these two profile types, the data from the Russian sample on academic achievement is illuminating. The poorly self-regulated higher achievers had a grade point average of 3.89 on a five-point scale which is near to the average for all of the Russian samples (girls somewhat higher than average and boys somewhat lower), while the average grade point of the poorly self-regulated lower achievers was 3.24. This is the lowest average of any of the seven profiles. These behavioral profile differences mean that socializing children exhibiting the higher achieving profile will be easier than is the case for the lower functioning cluster. Whether the behavioral differences between these clusters of children are driven by learning ability (academic ability, social intelligence) or whether the poorly achieving group is alienated from the school situation by their irritability and antagonism is unclear.

11.5  Parental Influence on Self-Regulation The seven-profile model of child behavior indicates that parents and teachers in the USA and Russia are aware of individual differences in self-regulation of attention, motor activity, and expressions of negative emotionality. Can different parenting practices help these children develop better self-regulation of these kinds of behaviors? Many experts agree that parenting plays a role in the development of self-­ control [16, 17] and there is a body of research supporting this hypothesis [18, 19]. Most research on parenting and self-regulation has been done with children in early childhood. This raises questions about effects during middle childhood and adolescence. The general consensus is that parenting within middle childhood and adolescence that includes monitoring of youth’s behavior, use of consistent discipline, parental warmth and support, and a close bond between parent and child is related to better self-control. Inconsistent discipline, harsh parenting, and physical punishment, in conjunction with a poorer relationship between the parent and child, are

11.5  Parental Influence on Self-Regulation

143

associated with lower levels of self-control in early and middle childhood. As children get older, the effects of parenting have been shown to be less clear cut. These less clear results have been interpreted as being consistent with the changing influence of parents and peers as development proceeds from early childhood through adolescence. In early childhood, children are much more dependent on parents for help with emotional regulation, goal setting, scheduling of activities, and general behavioral guidance. During late middle childhood and adolescence, youth often feel a need to pull away from the controls of parents and look to their peer group for norms regarding goals, decision-making, and norms for self-regulation. Thus, parental control over self-regulation is reduced. An alternative explanation is that during early and middle childhood, the child has fewer opportunities to choose environments that are consistent with their genetically influenced temperamental tendencies than is the case in adolescence. With increased independence, new opportunities for choosing activities and peer partners lead to more choices that are congruent with genetically driven temperamental tendencies. Jian-Bin Li at the Education University of Hong Kong in conjunction with other colleagues has done a review of 191 articles in which the relationship between parenting and self-control among children in late middle childhood through late adolescence (ages 10–22) was studied [20]. Self-regulation during this period is particularly important; youth in this developmental period show increased prevalence of risk-taking behavior that can have long-term consequences including risky sexual behavior, drug experimentation, risky use of motor vehicles, and many others. The researchers categorized parenting practices and parent-child relationships into three categories: positive parenting, negative parenting, and quality of parent-­ child relationships. The results demonstrated that parenting is associated with adolescent self-control when both are measured at the same age. The correlation, however, was low (r = 0.20). Further, in studies in which parenting practices were measured first at a different age than the measurement of self-control, parents still predicted adolescent self-control but at a lower level (r = 0.16). Interestingly, adolescent self-control also significantly predicted subsequent parenting practices (r = 0.16) and the effect was equal to the prediction of self-control from parenting. Finally, all three of these effects were very similar across cultures, ethnicities, the age of the adolescent, and the gender of the children studied. The take-home message from this rigorous review of many similar studies is that parents have an effect on the self-control exercised by children in late middle childhood through adolescence, but the effect is small. The finding that self-control also affected (to the same degree) parenting practices is an alternative interpretation. The effect is driven by the child’s behavior. Self-controlled children receive more positive parenting because they tend to lead parents to engage in positive parents and closer parent-child relationships. Thus, the data indicate that the effect can operate in both directions. This broad kind of study which looks at all kinds of children leaves many questions unanswered. One of the most important from our perspective is, “Do some children respond much more favorable to positive parenting than others?” As we have seen in discussions in previous chapters of the differential susceptibility

144

11  Fostering the Development of Poorly Self-Regulated Children

hypothesis [21], it is very likely that some do. The use of a temperament profile model like the one proposed in this book could go a long way toward determining if different clusters of children respond to parenting in different ways. This would be particularly important for determining if the two poorly self-regulated clusters have a different response to positive parenting and how much difference particularly positive parenting makes for both of these clusters.

11.6  General Guidelines for Parenting and Teaching The following are guidelines for helping poorly self-regulated children take advantages of their behavioral assets and to limit the risks. (For much more comprehensive discussions of parenting of children with characteristics similar to the two poorly self-regulated profiles, see McMahon & Forehand [22] and Forehand & Long [23]). 1. The first goal for the parent and teacher is to understand the characteristics of poorly self-regulated children. It is particularly important to understand that reward sensitivity, irritability/antagonism, and poor attention control are often not under the child’s conscious control. However, skills can be learned that moderate the negative aspects of these tendencies. 2. Through understanding, parents and teachers can learn to truly value many aspects of the behavioral style the child exhibits. All behavior profiles have positive attributes as well as risks. Valuing the assets of assertive, active, risk-taking children is aided by the understanding that these behavioral tendencies work well in many situations. The pattern of intense and frequent expressions of negative emotion and gross motor activity, as well as disorganization (poor ability to plan; impulsivity) and inattention, often creates difficulties for the adults and peers. For example, in classroom with many children under the supervision of one (sometimes two) adults, all of these behaviors are disruptive and make it difficult for the teacher to manage the attention of other students. These disruptions often cause the adults in schools to view children with these profiles in ways that hamper appreciation for their positive characteristics. Children who are active, intense, and social have many behavioral assets. First and foremost, peers attend to children who exhibit these tendencies. They are socially prominent and often admired, although sometimes they are feared due to their social or verbal aggressive tendencies. However, these behaviors mean that children who have these temperamental characteristics have leadership potential. Observations of the behavioral styles of many successful entrepreneurs such as trial lawyers, military leaders, politicians, and football coaches reveal that they have behavioral characteristics related to our temperament profiles that contribute to their social prominence. They succeed, in part, because they are assertive, active, intense, and social.

11.6  General Guidelines for Parenting and Teaching

145

These characteristics also decrease their vulnerability to victimization and aggression from others. I remember a mother of a sometimes-aggressive, always assertive, and competitive 8-year-old, who told me how proud she was of his behavioral style. She had grown up in a lower middle-class area of a major city and was sometimes bullied by others. She told me that she wanted her boy to be able to stand up for himself. Impulsive responses under some circumstances can also be adaptive. I knew of an academic who was visiting a developing country that was in the midst of armed attacks by several revolutionary groups. While boarding a small commuter plane, someone rolled a hand grenade down the aisle and ran. This professor picked up the hand grenade and threw it out of the plane door before anyone else moved. Most people were so stunned by the weapon rolling down the aisle that they froze or recoiled. When interviewed, the professor said it was an instinctive response done without much thought. In this situation, there was no time to evaluate options, he just reacted. In many dangerous situations, similar impulsive responses may be highly adaptive. Of course, being assertive, sometime aggressive, social, active, and emotionally intense are not enough to guarantee positive adaptation. Being a quick learner, have socially attractive skills (e.g., athletic, musical), and being able to modulate one’s initial response pattern through the exercise of self-control will allow these characteristics to lead to more positive adaptations. Nevertheless, children in middle childhood who exhibit the two temperamental profiles we have identified should be viewed as having real behavioral assets. 3. The goal of parenting or teaching children is not to change these general characteristics. A child who is highly active, emotional intense, has oppositional tendencies, and has difficulty regulating attention in middle childhood will spend a life-time learning to modulate these tendencies so that they are expressed in the most adaptive way. The goal of parental and educational training, then, is to help the child begin to learn these skills. This process often involves parents and teachers helping the child learn to select environments that foster appropriate skills. This can be done in two ways. First, parents and teachers can help select environments that allow for highly active, emotionally intense expression in a highly social context. Participation in sports is an example of such a context for middle school children with these characteristics. Drama and music experiences may also provide a good fit for their behavioral tendencies. However, all children in this age group have to learn to cope with environments that are not a good fit for their temperamental structure. Schooling that involves being in a classroom with many other children is a particularly bad fit. This environment has many social temptations that are positive reinforcing, is filled with distracting stimulation, and has an increased probability of social conflict. In such schooling environments, small modifications can be made that can be helpful. Working on projects in groups where the child can have their social needs met and may be able to positively show leadership can be helpful to some of these children. Being given time

146

11  Fostering the Development of Poorly Self-Regulated Children

during the day for movement can also be particularly helpful. Unfortunately, in most schooling for children of this age in the USA, such opportunities (e.g., recess periods) are being systematically reduced or eliminated currently. Having a multi-media curriculum presentation with films, response activities, and small group discussion has many advantages for those children with poor self-regulation. The worst fit for middle childhood children who exhibit poorly self-regulated behavior profiles is one in which they are asked to quietly sit and direct their attention toward reading material. For many, this material has very little stimulus value. It is a low intensity reward even if they have good reading skills. For those with poor reading skills, it is a punitive environmental context. In short order, attempts at social interactions begin (whispering, passing notes) and negative interactions with teachers often follow. In the home, similar situations occur on long car rides or during formal occasions like sitting through religious ceremonies. Adaptive parents learn to limit the length of such occasions, to provide frequent breaks (e.g., in car rides), to provide high-­ intensity distractions (e.g., video games), etc. 4. Avoid the cycle of escalating coercion The primary method used in homes and schools to attempt to control poorly self-­ regulated children is simply punishment. It is used in the hope that the child will learn to anticipate the punishment and think before they act and refrain from the punished behavior on the next occasion. Interestingly, there is a good deal of research showing that children with the tendencies described by our two poorly self-­ regulated profiles are not very responsive to punishment, particularly milder forms of punishment used by most parents and teachers. Punishing children with poor impulse control doesn’t often work well. In part, they simply respond impulsively without thinking about the consequences of their behavior. Further, the consequences are not as impactful and the immediate reward (e.g., laughter of peers in the classroom) of the inappropriate behavior. There is a natural tendency for an inappropriate behavior to lead to a warning. This is then followed by more inappropriate behavior followed by harsher verbal threats and eventually by an expression of anger by the parent or teacher. Often with enough anger, the child stops the inappropriate behavior which is reinforcing to the parent. This in turn leads to an increased tendency toward anger and yelling on the part of the adult in the future. In the process, the adult has modeled exactly the behavior that they want least from the child. They have shown the child that anger and aggression work and they now have created a child who quickly becomes angry and hostile when asked to comply even in response to the simplest request. If the adult in this situation is not self-regulated, the anger and hostility of both parties continue to escalate. Poorly self-regulated children have a natural tendency to respond to frustration with heightened negative emotionality. They are prone to anger. It is easy to enter into a cycle of escalating anger with children who exhibit the poorly self-regulated profiles. This cycle of anger and hostility between parent and child damages the relationship and eliminates one of the most important tools in helping the child learn self-control.

References

147

5. Avoid giving in to the child’s expressions of anger Children who express high-intensity negative emotion often, particularly anger, create a punitive environment for parents and teachers. Their behavior can be disruptive to the family and school life. Thus, there is a tendency to give in to demands in order to reduce the angry behavior. Unfortunately, this reinforces that behavior, making it more likely to occur again and again. Having reasonable rules that are consistently enforced with reasonable consequences without anger on the part of the parent or teacher is the key. This can be very difficult. It often involves removing the child from the situation and putting them in a place where they can calm down. 6. Keeping a caring, positive relationship with poorly self-regulated children Perhaps the strongest leverage a parent or teacher has with poorly self-regulated children is a positive, caring relationship. Some of their natural tendencies make this difficult. However, many parents and teachers enter into a style of persistent behavioral correction with hundreds of repetitions over time. This does nothing but erodes the relationship between the parent (teacher) and the child. Finding behaviors that can be praised, demonstrating how to deescalate conflict through modeling appropriate behavior, and truly valuing the behavioral assets presented by the child can help build and maintain a positive relationship.

References 1. Rothbart, M. K. (2011). Becoming who we are: Temperament and personality in development. New York: Guilford. 2. Rothbart, M. K., Ahadi, S. A., Hershey, K. L., & Fisher, P. (2001). Investigation of temperament at three to seven years: The children’s behavior questionnaire. Child Development, 72, 1394–1408. 3. Derryberry, D., & Tucker, D. M. (2006). Motivation, self-regulation, and self-organization. In D. Chicchetti & D. J. Cohen (Eds.), Developmental psychopathology: Vol. 2. Developmental neuroscience (2nd ed., pp. 502–532). New York: Wiley. 4. Dodge, K., Bates, J., & Pettit, G. S. (1990). Mechanisms in the cycle of violence. Science, 250, 1678–1683. 5. Chess, S., & Thomas, A. (1987). Origins and evolution of behavior disorders. Cambridge, MA: Harvard University Press. 6. Caspi, A., & Silva, P. A. (1995). Temperamental qualities at age 3 predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486–498. 7. Alink, L.  R., Mesman, J., van Zeill, J., Stolk, M.  N., Juffer, F., Koot, H.  M., Bakermans-­ Kranenburg, M. J., & van Ljzendoorn, M. H. (2006). The early childhood aggression curve: Development of physical aggression in 10- to 50-month-old children. Child Development, 77, 954–966. 8. Bell, M. A., & Deater-Deckard, K. (2007). Biological systems and the development of self-­ regulation: Integration behavior, genetics, and psychophysiology. Journal of Developmental & Behavioral Pediatrics, 28, 409–420. 9. Posner, M. I., & Rothbart, M. K. (2007). Research on attention networks as a model for the integration of psychological science. Annual Review of Psychology, 58, 1–23. 10. Halverson, C. F., Jr., Kohnstamm, G. A., & Martin, R. P. (Eds.). (1994). The developing structure of temperament and personality from infancy to adulthood. New York: Erlbaum.

148

11  Fostering the Development of Poorly Self-Regulated Children

11. Aguilar, B., Sroufe, L.  A., Egeland, B., & Carlson, E. (2000). Distinguishing the early-­ onset/persistent and adolescence-onset antisocial behavior types: From birth to 16 years. Development and Psychopathology, 12, 109–132. 12. Moffitt, T.  E., Caspi, A., Harrington, H., & Milne, B.  J. (2002). (2002). Males on the life-­ course-­persistent and adolescence-limited antisocial pathways: Follow-up at age 26 years. Development and Psychopathology, 14, 179–207. 13. Shiner, R., & Caspi, A. (2012). Temperament and the development of personality traits, adaptions, and narratives. In M.  Zentner & R.  L. Shiner (Eds.), Handbook of temperament (pp. 497–518). New York: Guilford. 14. Kidd, C., Palmeri, H., & Aslin, R. N. (2013). Rational snacking: Young children’s decision-­ making on the marshmallow task is moderated by beliefs about environmental reliability. Cognition, 126, 109–114. 15. Lee, W. S. C., & Carlson, S. M. (2015). Knowing when to be ‘rational’: Flexible economic decision making and executive function in preschool children. Child Development, 86, 1434–1448. 16. Bridgett, D. J., Burt, N. M., Edwards, E. S., & Deater-Deckard, K. (2015). Intergenerational transmission of self-regulation: A multidisciplinary review and integrative conceptual framework. Psychological Bulletin, 141, 602–654. 17. Sameroff, A. (2010). A unified theory of development: A dialectic integration of nature and nurture. Child Development, 81, 6–22. 18. Davis, M., Bilms, J., & Suveg, C. (2017). In sync and in control: A meta-analysis of parent-­ child positive behavioral synchrony and youth self-regulation. Family Process, 56, 962–980. 19. Pallini, S., Chirumbolo, A., Morelli, M., Bajocco, R., Laghi, F., & Eisenberg, N. (2018). The relation of attachment security status to effortful self-regulation: A meta-analysis. Psychological Bulletin, 144, 501–531. 20. Li, J.  B., Willems, Y.  E., Stak, F.  M., Dekovic, M., Bartels, M., & Finkenauer, C. (2019). Parenting and self-control across early to late adolescence: A three-level meta-analysis. Perspectives on Psychological Science, 14, 967–1005. 21. Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influences. Psychological Bulletin, 135, 885–908. 22. McMahon, R.  J., & Forehand, R.  L. (2005). Helping the noncompliant child (2nd ed.). New York: Guilford Press. 23. Forehand, R., & Long, N. (2010). Parenting the strong-willed child. New York: McGraw Hill.

Part III

Stability, Causes, and Implications for Diagnosis

Part III is designed to answer some of the most often asked questions about behavioral tendencies observed in childhood. First, Chap. 12 addresses the question of stability of behavioral tendencies. Every parent wants to know, for example, if their irritable and antagonistic 8-year-old is likely to be irritable and antagonistic at age 12, 16, and 30. This chapter briefly reviews the extensive data on the stability of individual temperamental characteristics in addition to summarizing the results of research on stability of temperament-based profiles. Chapter 13 tackles the complex factors that contribute to individual differences in cognitive, social and emotional behavior in children. Emphasis is given to individual differences in temperamental characteristics that play a role in the lives of typical children in middle childhood. Genetic and environmental factors, and the interaction of these factors, are discussed. Chapter 14 addresses the importance of considering broad temperamental profiles when making diagnostic decisions about the presence of psychiatric conditions. This chapter also offers an alternative to the medical model of disease for many behavioral problems of children. Finally, Chapter 15 presents a summary of the most important findings of our research as well as some of the limitations of this research program.

Chapter 12

Stability of Temperament Traits and Profiles

12.1  Introduction Our study of the temperament-based profiles of middle childhood does not provide any information on the stability of the profiles across different ages. All the data analyzed were cross-sectional in nature; that is, information on the behavioral tendencies of a given child was obtained at one point in time. Such data cannot address the important question of the stability of cluster assignments over time. If a child is perceived to exhibit a given profile at age 6, for example, we need to know the likelihood that that child will exhibit the same profile at 8, 10, 12, or 18. This issue is complicated by the probability that the profile structure is likely to get somewhat more differentiated as the child matures. Specifically, in adolescence, we would expect some of the seven profiles identified in middle childhood to differentiate. Such an event is anticipated based on the results of preliminary analyses of adolescent data by our research team. Further, in a preliminary analysis (data not presented here), a five-profile model seemed most appropriate in early childhood. Thus, profile differentiation seems likely with advancing age. Unfortunately, there is no longitudinal research currently available which used the same temperament-based behavioral traits that we have studied. Thus, the stability of child behavior as defined by the seven-cluster model cannot be directly addressed. However, there is a large literature relating early assessed temperament to later outcomes that implies some form of continuity, as well as many studies of the stability of individual temperamental traits and related personality traits across different age groups. Also, there are a few studies of the stability of temperamental profiles. All of this research leads to the conclusion that temperament-related behavioral tendencies have a substantial stability. It is instructive to look at a few specific findings.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_12

151

152

12  Stability of Temperament Traits and Profiles

12.2  Long-Term Prediction There is a growing body of research that demonstrates that temperamental traits assessed in early childhood are predictive of important outcomes that occur later in life. Studies have shown that temperament assessed during the preschool years, for example, is predictive of (a) the need for psychological and psychiatric care later in life [1–5], (b) moral development [6], (c) school adjustment [7, 8], (d) violence in late adolescence [9], (e) personality in adults [10–12], (f) economic and political discontent in adulthood [13], and (f) political orientation (progressive/conservative) in young adults [14, 15]. Such findings make a case for the importance of early appearing behavioral tendencies and imply some form of behavioral stability. However, when behavior observed in childhood is found to predict behavior in adolescence and adulthood, there is an implication that either (a) the behavior pattern observed early in life is relatively stable and contributes to events at each age that leads to the outcome or (b) the early behavior sets in motion a chain of events that creates a cascade of behaviors (e.g., a feedback loop between the child and the environment) resulting in the later outcome. In the latter case, a child who exhibits unusually high levels of negative emotionality as preschooler might trigger harsh parenting which, in turn, fosters poor mental health outcomes (including negative affect) in adolescence. In this interpretation, a strict definition of behavioral stability would not be indicated. However, if this child continued to exhibit high levels of negative emotionality throughout development, then the outcome would be indicative of stability. As we will see, there is strong evidence that temperament traits are at least moderately stable from the preschool years to adulthood. Thus, the probability that a set of behavioral tendencies in early childhood has little stability–yet plays a role in schooling or mental health outcomes later in life–is unlikely.

12.3  Mean Age Differences in Temperament-Related Behaviors When researchers investigate the stability of a behavioral trait, one approach involves investigating the average level (for a population of children) of the trait as children increase in age. For example, there is an increase in expression of negative emotionality around age 3, followed by relatively constant levels in middle childhood, which then increases again in early adolescence. Then, anger and irritability gradually decrease throughout adulthood [16]. Activity level, on the other hand, is at its highest level in early childhood and decreases almost linearly with age after that point [17]. The ability to self-regulate one’s attention and emotion increases rather dramatically during the preschool period, which then continues to increase more gradually through middle childhood and adolescence. As development continues into adulthood, measures of brain activity indicate that the brain circuitry underlying self-regulation becomes more focal and refined which contributes to efficiency

12.4  Stability of Individual Differences in Temperamental Traits

153

[18]. Empathy and prosocial behavior tend to increase from the preschool period to adolescence, at which time there seems to be a small decrease [19]. This brief discussion indicates that behavioral tendencies in children fluctuate as a function of age. These data, however, do not address individual differences in relation to the mean levels of most children. If Johnny has increased his expressions of empathy and consideration of others from age 7 to age 12,  this does not tell us if he was among the lowest in his age group at age 7 and still is low at age 12 compared to his peers or is above average at age 12. It is this kind of individual stability – stability in level compared to the level of same-age children – that is of primary interest in the current discussion. In other words, developmentally normative change is expected. The question is whether the child’s relative level of a trait, compared to others of the same age, is stable.

12.4  S  tability of Individual Differences in Temperamental Traits In order to determine if individual differences are stable, the most common procedure is to measure a particular behavioral tendency in a group of children at one time and then measure them again months or years later. Simply put, a correlation is calculated between the two sets of scores. These correlations represent the extent to which the individuals’ rank (relative position in the group of children studied) at a given time is similar to their rank at a subsequent time. The correlation is 1.00 if the rank order of all individual remains the same from time 1 to time 2. If the correlation is in the 0.60–1.00 range, it is generally interpreted to indicate strong stability. Correlations in the range of 0.30–0.59 would generally be considered moderate, and lower correlations (0.29–0.00) would indicate lower stability. It is worthwhile to note that these correlations are not affected by mean level changes for the entire group of children at different ages. Thus, the correlation for activity level between ages 3 and 10 could be high (e.g., 0.70), even though the average level of activity at ages 3 and 10 is quite different at these ages. In general, these stability correlations are higher when the interval between assessments is short (a few weeks to a few months between assessments). For example, short-term stability within the infancy/toddler or preschool period is typically in the moderate range (0.30–0.60), while short-term stability within middle childhood or adolescence is higher. For example, the one-month stability of irritability/ anger in the toddler and preschool period is typically in the 0.30–0.60 range [20], while the stability correlations in middle childhood and early adolescence are typically in the 0.50–0.70 range [21]. Stability in adulthood is at about the same level as in adolescence. Of most interest, however, is longer term stability (1 year or longer). One of the most illuminating studies of stability of parent-rated temperament was conducted by Diana Guerin and colleagues [22] on approximately 100 children who were

154

12  Stability of Temperament Traits and Profiles

assessed first at age 2 and then periodically until age 12. Three-year stabilities of activity level, approach (behavioral inhibition scored in the opposite direction), task persistence, distractibility, and negative mood from age 2 to 5 ranged from 0.36 to 0.48. Three-year stabilities from age 5 to age 8 ranged from 0.37 to 0.55. Thus, three-year stabilities in the age range of 2–8 were quite similar. However, four-year stabilities from age 8 to age 12 were significantly higher ranging from 0.57 to 0.71. These data and the data from short-term stability studies illustrate that temperamental characteristics had moderate stability in the toddler and preschool years across a three-year period but had higher stability in middle childhood. Similar findings have recently been reported by other researchers [23]. The results of many studies of longitudinal stability might be criticized because of the modest sample size or a failure to study child or parental factors that might play a role in stability ratings. However, Bronstein and colleagues [24] studied temperamental stability from ages 3 through 6 in a very large sample of British families (approximately 10,000). They measured four temperament traits: activity level, negative emotionality (tendency to be fussy, cry), shyness (inhibition with unfamiliar people), and sociability (tendency to prefer the company of others), using maternal reports. Stability was assessed from ages 3–5, 5–6, and 3–6 years. All stabilities ranged from 0.46 to 0.73. Longer term stability (ages 3–6) was lower across all traits than short-term stability, but the decline was modest. The stability observed for boys and girls was very similar, as was stability of firstborns and later born children. Changes in maternal anxiety and depression had very minor effects on child ratings of temperament. Older mothers and more educated mothers produced somewhat higher stabilities; however, the effects of these maternal characteristics were also modest. The strength of this study was its very large sample size and the investigation of effects of a variety of child and maternal characteristics that have sometimes been hypothesized to be associated with ratings of child behavior. However, as noted, child and maternal characteristics had very little effect on the stability of the ratings of temperamental characteristics. One aspect not explored by these research efforts is the relative stability of children at different levels of temperamental dimensions. That is, temperamental characteristics are normally distributed (on a “bell curve”). The majority of children have mid-level scores while small numbers of children have extreme scores (high or low). There is good theoretical support for the idea that children at the extreme ends of the continuum exhibit more stability than do those in the middle. The argument goes like this. Children who exhibit very high levels of negative emotionality or are highly prosocial, for example, tend to create their own social environments. Their tendencies have an outsize impact on the social responses of others. It is difficult for parents and teachers not to find the child who frequently exhibits high levels of anger to be difficult to manage. Children who exhibit high levels of prosocial behaviors are seen as more likeable by peers. In the case of high levels of irritability/ antagonism, parental reactions might be to react with intense emotion and harsh punishment or to avoid some types of interactions with the child altogether. Thus, this child creates an aversive social environment for himself/herself, which reinforces the initial tendency. This creates an environment that leads to more

12.5  Stability of Profiles

155

behavioral stability. Similar processes happen for highly prosocial children. Based on this idea, children at the high ends of these tendencies should be more stable than those in the middle range. This outcome has been observed by some researchers. Prior and colleagues [25], reporting on the results of a nationwide study of Australian children from infancy to adolescence, found that there was more stability of traits for children at the extreme ends of the temperament distribution than in the middle ranges.

12.5  Stability of Profiles There are very few studies that have looked at profile stability of children as they develop. At least two types of stability are important to consider. First, for example, is whether the profile structure is the same for 5-year-olds and for 12-year-olds. By structure we simply mean that the same number of profiles is found to describe the ratings of parents (or teachers) of children at different ages and these profiles have similar characteristics (similar mean levels on each of the behavioral traits studied, similar variation around the mean, etc.). A second kind of stability concerns the extent to which an individual is defined by a given profile at different ages. This kind of stability, sometimes referred to as classification stability, could be assessed by determining the percentage of children who maintained the same profile at different ages. Because this kind of research is new to the study of behavioral tendencies in children, and because such research is very expensive, there are only a few studies that report on both types of stability. One recent investigation of both types of stability during the infancy and toddler periods studied 561 children who were adopted in early infancy from four regions in the USA [26]. Adoptive parents (ratings of mothers and fathers were averaged) rated temperamental characteristics of their children at 9, 18, and 27 months of age. Characteristics assessed in toddlerhood were activity level, anger, fear, interest, and pleasure, with similar characteristics in infancy. Four profiles were found at both 18 and 27 months, while five were isolated in infancy. We will consider here only the four profiles obtained both at 18 and 27  months. They were labeled (1) positive reactive (slightly below average levels of activity, anger and fear, with slightly above average levels of interest and pleasure); (2) negative reactive (slightly above average levels of activity, anger, and fearfulness, with below-average interest and pleasure); (3) fearful (low activity level, low anger, high average fear, low interest and pleasure), and (4) active reactive (very high levels of activity and above average levels of anger and pleasure, and low average fear and interest). Results indicated that 78% of the positive reactive infants retained that classification at 27 months. The negatively reactive cluster also showed significant stability, with 51% retaining their classification, as did infants in the fearful profile (55% retained the same profile in toddlerhood). However, most children in the active reactive profile at age 18 months transitioned into two other profiles by 27 months: 34% transitioned into the positive reactive profile, 37% changed into the negative

156

12  Stability of Temperament Traits and Profiles

reactive profile, and only 29% remained in the active reactive profile. Of note, not one child who was classified as active reactive at age 18 months had transitioned in to the fearful profile at age 27 months. This study shows that the profile models identified for these traits varied from only modestly stable to very stable. Some instability is to be expected, given that the changes in measurements as well as the major changes in development that take place across this age range. Despite this, some profiles isolated in infancy and toddlerhood showed strong stability from infancy to 27  months. This kind of result raises many research questions, among them being, “Are some temperament-based profiles more stable than others?” Another study, conducted in Norway, is one of the strongest studies of this kind. The researchers studied 921 children at 18  months, 30  months, 4–5  years, and 8–9 years of age [27]. Measures were obtained from maternal reports of activity level, sociability, emotionality, and shyness. A five-profile classification was determined to fit all age levels adequately. (It must be noted that the statistical procedures used in this project were designed to produce the same structure at all ages.) The five profiles were named as follows: (1) confident (high activity level, high sociability, low emotionality, low shyness), (2) unremarkable (slightly below-average scores on all dimensions), (3) uneasy (above-average shyness, emotionality, and activity level, average sociability), (4) inhibited (high shyness levels, low sociability, and activity level), and (5) undercontrolled (high levels of emotionality, sociability, and activity). Significant individual stability was found throughout development. The stability of children in the undercontrolled, confident, unremarkable, and inhibited profiles was statistically significant (i.e., the percentage of children who retained their classification was greater than would be expected by chance). Uneasy children at younger ages tended to end up in the inhibited classification at older ages. The take-home message from this study is that parents of children in the toddler and preschool years can identify individual differences that result in profiles that are relatively stable into middle childhood. This is particularly evident for the confident, unremarkable (average), inhibited, and undercontrolled profiles. A recent study illuminates the stability of self-reported behavioral profiles in adults. A group of French researchers [28] studied adults (sample size was more than 500; average age approximately 35 years) living in Quebec. They were administered a questionnaire designed to measure six emotional-related behavioral tendencies [29]: (a) seeking (exploring, striving for solutions to problems, positively anticipating new experiences); (b) caring (nurturance, feeling soft-hearted toward animals and people in need, feeling empathy); (c) playfulness (joy, having fun, humor, laughter); (d) fear (anxiety, feeling tense, worrying, struggling with decisions, ruminating); (e) anger (being hot-headed, easily irritated and frustrated, expressing anger verbally or physically); and (f) sadness (experiencing separation distress, feeling lonely, thinking about loved ones and past relationships, and feeling distress). These characteristic behavioral tendencies have a striking similarity to characteristics that we studied in middle childhood, although they were given somewhat different names. Using latent profile analysis procedures that were similar to the procedures used in our study, they found three broad profiles: (1) low negative

References

157

emotions (24% and 27% women and men, respectively), (2) balanced emotions (65% and 64%), and (3) highly emotional (12% and 9%). With regard to stability over the four-year period between assessments, these researchers calculated the probability of individuals staying in their classification from time 1 to time 2. The stability of profile 1 (low negative emotions) was 0.98 for women and 1.00 for men. The stability was somewhat lower for profile 2 (balanced emotions): 0.85 for women and 1.00 for men. The stability for profile 3 (highly emotional) was again very high (0.96 for women and 0.95 for men). These data indicate that almost all these adults remained in their classification across 4  years of assessment based on these self-ratings. It is important to note that the methods used to measure the temperaments of children and adults are quite different. For children, they are rated by others (parents, teachers, peers). For adults, most research projects like those cited above are based on self-report. These two methods may result in estimates of different stabilities. In particular, the estimates of self-ratings may be more stable than those provided by observers. However, there is a consensus among psychologists that adult temperament/personality is more stable than childhood temperament.

References 1. Chess, S., & Thomas, A. (1987). Origins and evolution of behavior disorders. Cambridge, MA: Harvard University Press. 2. Caspi, A., Henry, B., McGee, R. O., Moffitt, T. E., & Silva, P. A. (1995). Developmental origins of child and adolescent behavior problems: From age three to age fifteen. Child Development, 66, 55–68. 3. Caspi, A., Moffitt, T. E., Newman, D. L., & Silva, P. A. (1996). Behavioral observations at age 3 years predict adult psychiatric disorders. Archives of General Psychiatry, 53, 1033–1039. 4. Al Clark, L. (2005). Temperament as a unifying basis for personality and psychopathology. Journal of Abnormal Psychology, 114, 505–521. 5. Dougherty, L.  R., Klein, D.  S., Drubin, C.  E., Hayden, E.  P., & Olino, T.  M. (2010). Temperamental positive and negative emotionality and children’s depressive symptoms: A longitudinal prospective study from age three to age ten. Journal of Social and Clinical Psychology, 29, 464–490. 6. Kochanska, G., Murray, K., Jacques, T., Koenig, A. L., & Vandegeest, K. (1996). Inhibitory control in young children and its role in emerging internalization. Child Development, 67, 490–507. 7. Blair, C. (2012). School readiness: Integrating cognition and emotion in a neurobiological conceptualization of children’s functioning at school entry. American Psychologist, 57, 111–127. 8. Nelson, B., Martin, R. P., Hodge, S., Havill, V., & Kamphaus, R. (1999). Modeling the prediction of elementary school adjustment from preschool temperament. Personality and Individual Differences, 26, 687–700. 9. Henry, B., Caspi, A., Moffitt, T. E., & Silva, P. A. (1996). Temperamental and familial predictors of violent and nonviolent criminal convictions: Age 3 to age 18. Developmental Psychology, 32, 614–623. 10. Caspi, A., & Silva, P. A. (1995). Temperamental qualities at age 3 predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486–498.

158

12  Stability of Temperament Traits and Profiles

11. Eisenberg, N., Guthrie, I.  K., Murphy, B.  C., Shepard, S.  A., Cumberland, A., & Carlo, G. (1999). Consistency and development of prosocial dispositions: A longitudinal study. Child Development, 70, 1360–1374. 12. Pulkkinen, L. (2017). Human development from middle childhood to middle adulthood. New York: Routledge. 13. Lewis, G.  J. (2018). Early-childhood conduct problems predict economic and political discontent in adulthood: Evidence from two large longitudinal United Kingdom cohorts. Psychological Science, 29, 711–722. 14. Block, J., & Block, J. (2006). Nursery school personality and political orientation two decades later. Journal of Research in Personality, 40, 734–749. 15. Fraley, R. C., Griffin, B. N., Belsky, J., & Roisman, G. I. (2012). Developmental antecedents of political ideology: A longitudinal investigation from birth to age 18 years. Psychological Science, 23, 1425–1431. 16. Deater-Deckard, K., & Wang, Z. (2012). Anger and irritability. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 124–144). New York: Guilford. 17. Strelau, J., & Zawadzki, B. (1997). Activity as a temperament trait. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 83–105). New York: Guilford. 18. Rueda, M. R. (2012). Effortful control. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 145–167). New York: Guilford. 19. Knafo, A., & Israel, S. (2012). Empathy, prosocial behavior, and other aspects of kindness. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 168–179). New York: Guilford. 20. Putnam, S. P., & Rothbart, M. K. (2006). Development of short and very short forms of the children’s behavior questionnaire. Journal of Personality Assessment, 87, 102–112. 21. Kim, J., Mullineaux, P. Y., Allen, B., & Deater-Deckard, K. (2010). Longitudinal studies of stability in attention span and anger: Context and informant effects. Journal of Personality, 78, 419–440. 22. Guerin, D. W., Gottfried, A. W., Oliver, P. H., & Thomas, C. W. (2003). Temperament: Infancy through adolescence, the Fullerton longitudinal study. New York: Springer. 23. Neppi, T.  K., Donnellan, M.  B., Scaramella, L.  V., Widaman, K.  F., Spilman, S.  K., Ontai, L. L., & Conger, R. D. (2010). Differential stability of temperament and personality from toddlerhood to middle childhood. Journal of Research in Personality, 44, 386–396. 24. Bornstein, M. H., Hahn, C. S., Putnick, D. L., & Pearson, R. (2018). Stability of child temperament: Multiple moderation by child and mother characteristics. British Journal of Developmental Psychology, 37, 51–67. https://doi.org/10.1111/bjdp.12253. 25. Prior, M., Sanson, A., Smart, D., & Oberklaid, F. (2000). Pathways from infancy to adolescence: The Australian temperament project 1983–2000. Melbourne, Australia: Australian Institute of Family Studies. 26. Beekman, C., Neiderhiser, J. M., Buss, K. A., Loken, E., Moore, G. A., Leve, L. D., Ganiban, J. M., Shaw, D. S., & Reiss, D. (2015). The development of early profiles of temperament: Characterization, continuity, and etiology. Child Development, 86, 1794–1811. 27. Janson, H., & Mathiesen, K. S. (2008). Temperament profiles from infancy to middle childhood: Development and associations with behavior problems. Developmental Psychology, 44, 1314–1328. 28. Orri, M., Pingault, J. B., Rouguette, A., Lalanne, C., Falissard, B., Herba, C., Côté, S. M., & Berthoz, S. (2017). Identifying affective personality profiles: A latent profile analysis of the affective neuroscience personality scales. Scientific Reports, 7, 4548. https://doi.org/10.1038/ s41598-­017-­0478. 29. Davis, K. L., & Pankseep, J. (2011). The brain’s emotional foundations of human personality and the affective neuroscience personality scales. Neuroscience Biobehavior Review, 35, 1946–1958.

Chapter 13

Why Children Exhibit Different Behavioral Patterns

13.1  Nature and Nurture: A Little History The history of psychology has witnessed wide pendulum swings in explanations of individual differences in behavior. There have been periods in which genetics and other physiological phenomena (nature) were dominant, and there have been periods in which environmental explanations (nurture) dominated this discourse. From the beginnings of psychology as a discipline, through the 1930s, many psychologists leaned in the direction of nature. In its most simplistic form, criminologists looked to differences in physique or structure of the skull (phrenology), for example, to explain criminal behavior. In its most terrifying form, this line of thinking was viewed by some in political power as supporting social class and ethnic/racial prejudice. A grossly simplified and misunderstood genetic rationale was the bases for the eugenics movement which resulted in the extermination of entire populations of human beings in the 1940s. Of course, most social scientists as well as most thoughtful human beings even during the darkest period of the 1930s took a much more nuanced view of genetic differences in human beings. In the 1940s through 1970s, under the influence of behaviorism, the pendulum swung to environmental explanations. Many social scientists had endorsed and researched environmental explanations for child behavior prior to that time. However, under the influence of demonstrations of operant learning in animals and related work with children, parents were advised to think of their child as a blank slate, a completely malleable template on which they could create the kind of behavior they wanted their child to exhibit. The behavior would be molded through the judicious application of behavioral learning principles. More recently under the influence of breakthroughs in behavioral psychology and molecular genetics, a more nuanced understanding of the influence of nature (genetics) and nurture (environmental influences) has progressively come to light. In broad terms, the issue is currently framed in terms of the relative influence of

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_13

159

160

13  Why Children Exhibit Different Behavioral Patterns

behavioral predispositions (nature) and environmental factors on behavior in specific circumstances (nurture), although the distinction between nature and nurture has begun to fade as we will see. Because of the proliferation of research in evolutionary theory, genetics, and child behavior, only a cursory summary of this work is possible. The following is an attempt to synthesize some of the contemporary thinking about influences on common individual differences in the behavior of children.

13.2  Estimates of Genetic Effects: Behavior Genetics Currently there are two different approaches to the study of genetic effects on behavior. One method is referred to as behavioral genetics or quantitative genetics. Through the study of people who have different relationships with one another (cousins, siblings, monozygotic twins – identical, and dizygotic twins– fraternal), comparisons are made of the similarities in behavior across varying degrees of biological relatedness. In this research, no DNA is collected. A typical study will find that identical twins (they share 100% of their DNA) are more similar in their behavior than are fraternal twins (they share 50% of their DNA on average), and both of these relationships show more similarity in behavior than do unrelated individuals. If the differences between identical and fraternal twins are substantial and both are substantially different from unrelated individuals, the heritability of the behavior is high. Heritability is calculated on a 0.00–1.00 continuum. A behavior determined to be highly heritable, like academic intelligence, has a heritability score in the 0.60–0.80 range. If people who vary in their relationships to one another do not show differences in behavior, then heritability will be low (close to 0). Heritability is interpreted as a proportion; it is the proportion of the variability observed in a sample of children that can be explained by differences in genetic similarity. Heritability is an estimate of the genotype of the individual. The measured behavioral pattern (anxiety, leadership, or irritability) is referred to as the phenotype and is composed of genetically related factors as well as environmental factors. Using these techniques, the heritability of most temperamental and later developing personality characteristics is moderate (0.30–0.60). This indicates that from one-third to more than one-half of the variation in broad behavioral traits observed in a large group of people is due to genetic factors. Two types of environmental effects are also estimated using behavioral genetics techniques: shared and non-shared environment. Shared effects occur when siblings or twins are raised in the same family. It includes a lot of shared manifestations of social class, cultural patterns, educational resources in the home, and marital stability of the parents, for example. The non-shared environment is associated with differences in parental treatment of children in the family, sibling interactions, effects of accident or illness on individual children, and influences outside the home such as peers. Thus, a behavioral trait might have a heritability of 0.50, a shared environment effect of 0.20, and a non-shared environmental influence of 0.30. That is, 50%

13.3  Estimates of Genetic Effects: Molecular Genetic

161

of variation between individuals in the behavior is due to genetic factors, 20% due to shared environmental factors, and 30% due to non-shared genetic factors. In general, behavioral genetics research on broad temperament and personality traits indicates that a higher percentage of individual differences in behavior are accounted for by genetic differences than by shared and non-shared environmental factors. Shared effects on behavior tend to decrease with age as children move away from shared family influences [1, 2]. Thus, in adolescence and adulthood, the predominant environmental influence is non-shared, randomly occurring events. Current estimates indicate that from 30% to 50% of the variation in behavior in the general population is due to chance. Chance plays a major role in the trajectory of a life. Such environmental events include major illness, accidents, or war. Some of these events occur during fetal development and have a fundamental effect on cognitive ability and temperamental tendencies. These include exposure to bacterial pathogens, toxins, or protein calorie deprivation during fetal development or during the first few years of life [3]. Other effects include seemingly small social rejections and limitations placed on social and learning participation that accumulate over time into a cascade of detrimental events. Such events can be caused by the luck of being born into a lower social class or into a racial/ethnic group that is socially marginalized. In summary, hundreds of behavior genetic studies in child behavior have shown that 30–60% of the variation observed in cognitive and temperamental/personality characteristics is associated with genetic factors. As determined by this method, a sizeable proportion of temperament-related behavior is due to innate predispositions. Robert Plomin, a pioneer in behavior genetics, summarizes this research by saying “DNA isn’t all the matter, but it matters more than everything else put together” (Plomin, 2020, p. 10 [4]). Not all researchers would agree with this statement, but there is a consensus among all researchers that individual differences in behavior among children are substantially affected by genetic factors.

13.3  Estimates of Genetic Effects: Molecular Genetic There are two approaches to molecular genetic studies of the DNA of individuals as it relates to behavioral traits. One approach is referred to as the candidate-gene method in which specific genes are investigated based on some theory of a possible connection to a specific behavior. For many reasons, this technique has produced only a few meaningful associations between individual genes and behavior. The most important reason for the many failures of this type of research is that most complex behaviors like intelligence, activity level, irritability, and the other behaviors of interest in child development are likely to be caused by a large number of genes, each of which has a very small effect [1, 2, 5]. The second and more productive approach to date is referred to as the genome-­ wide association method. A large number of genes are tested and those that are associated with the trait of interest are aggregated to create a heritability index. In a

162

13  Why Children Exhibit Different Behavioral Patterns

recent study of loneliness in adults, for example, the heritability calculated using this method ranged from 0.14 to 0.27 [2]. This is a lower estimate than those obtained from studies of loneliness using behavior genetics approach which was 0.37–0.55. This difference is due to the fact that the heritability estimates from the two approaches are based on different statistical procedures; thus, heritability in the two approaches is defined differently. At the current time most research using the genome-wide association method has resulted in heritability estimates for intelligence, social, and emotional variables in the range of 0.10–0.20 [1]. But as this technique has become more refined, the estimates have increased. Many researchers believe that this method will continue to find increasing levels of genetic effects as the number of genes studied increase and the samples of persons studied increase in size. The bottom line from behavior and molecular genetic studies is that human behavioral tendencies like temperamental traits have a clear genetic component. Unshared environmental effects tend to be randomly occurring events in the lives of children that typically contribute from 30% to 50% to individual differences in behavioral traits [1, 5].

13.4  Epigenesis Genetic effects described to this point may leave the reader with an oversimplified idea of how genes affect behavior. Several additional considerations are crucial in understanding genetic effects. First, the genome does not directly affect behavioral traits. Genetic information (DNA) is transcribed into RNA in each cell which in turn produces proteins. These proteins form enzymes and neurotransmitters, for example, that contribute to individual differences in behavior as they affect a wide range of physiological processes including processes governing brain activity. However, research on how specific RNA products contribute to specific tissue functions and specific behavioral traits is in its infancy. Second, genes do not set a pattern of behavioral tendencies that are unchangeable or unresponsive to environmental effects. Gene activity is altered throughout the development of the individual’s life through the interaction of genetic processes with environmental influences. Epigenetics refers to mechanisms that affect the activity of DNA but do not change DNA itself [6]. Some epigenetic effects are related to the activity of the so-called “timing genes.” For example, the physiological processes that contribute to puberty are genetic and are “turned on” by the action of one set of timing genes on other genes. This example is illustrative because the timing of puberty has been shown to be related to environmental stress. For example, highly trained female athletes have a later onset of puberty than typical females, and in some cases training regimes have to be reduced or eliminated for puberty to occur. This effect has been most often observed in elite female gymnasts. This example shows that sometimes epigenetic effects are reversible; that is, as the physiological stress is reduced, the effects of the genetic process returns to normal levels.

13.5  The Differential Susceptibility Hypothesis

163

13.5  The Differential Susceptibility Hypothesis As if the processes described to this point are not complicated enough, it has become clear in the past two decades that some individuals are more responsive to environmental influences than others. This research is based on a theoretical foundation referred to as the differential susceptibility hypothesis [7]. While this hypothesis has been described previously in this volume, it deserves perhaps further elaboration. To illustrate this hypothesis and related research findings, some children are referred to as “dandelions” and some as “orchids”; that is, some children are relatively unaffected by their environment (dandelions) and some are very sensitive, for better or worse, to environmental influences (orchids) [8]. The most interesting aspect of this hypothesis is that the individuals described as orchids will develop positive outcomes more readily than their peers in supportive environments and they also develop more negative outcomes in stressful or non-supportive environments. Dandelion children, however, are less affected by either positive or negative environment events. To be clear, in highly stressful and chaotic environments such as those experienced in war, all children are likely to experience adverse behavioral outcomes. However, the effects are likely to be far less for some children than they are for others. Those most affected will be those who have a genetic propensity to be particularly sensitive to environmental effects. Imagine the following example. A researcher is studying groups of children in which half are selected because they are dandelions (less sensitive) and half are orchids (highly sensitive). She then divides each group into those who have a high predisposition to express negative emotionality (e.g., anger) and those who do not. She then measures the school environments for each of the four groups of children and divides these environments into highly supportive, average, and non-supportive. The researcher then measures the aggression that all the children are exhibiting. This hypothetical example is illustrated in Fig. 13.1. The graph indicates that for highly sensitive children (orchids), the effects of the environment have a major effect on their level of aggression at both levels of genetic predisposition toward anger. For children who are less sensitive (dandelions), there is a clear genetic effect depending on the level of the predisposition toward anger, but there is less of an effect of the environment. Note that the level of aggression for the highly sensitive children under conditions of highly supportive environment was even lower than that of children with low environmental sensitivity. This hypothetical example is meant to illustrate the basic principles of the differential susceptibility hypothesis. Research of this kind was conducted by Obradovic and colleagues [9]. They studied differential reactivity tendencies at the physiological level by measuring a child’s cortisol response (a stress-related hormone) during a visit to the research laboratory. They also measured family income with stress being thought to increase with lowered family income. The outcome measures of interest were skills in self-­ regulation of attention, behavior, and emotions. The most important finding was that self-regulation skills were positively related to family income (higher the  family

164

13  Why Children Exhibit Different Behavioral Patterns 3.5 3 2.5 2

Unsupportive

1.5

Average Highly Supportive)

1 0.5 0 Orchids Orchids Dandelions Dandelions (High Anger) (Low Anger) (High Anger) (Low Anger)

Fig. 13.1  Hypothetical outcome of aggression scores for highly sensitive (orchids) and less sensitive children (dandelions) and for children who have high and low genetic predispositions toward anger and are experiencing three levels of environmental support

income, the higher the levels of self-regulation), but only for child who had a strong cortisol response (highly reactive orchids). For the dandelion-like children there was no relationship between the socioeconomic status of their family and self-regulation. In another study, a Norwegian and American team of researchers studied peer problems in preschool aged children (ages 4–6) [10]. Parents provided assessments of the hyperactive and impulsive tendencies of their children. Teachers assessed the level of peer problems in a preschool setting. Finally, a specific gene region (5-HTTLPR) that is linked to serotonin transcription and transmission was studied, which is known to regulate both mood and cognition. Two variations of the gene are known, the long and the short variants. Children with the short variant (orchids) have been shown in other research to be more affected by stress than children with the long variant (dandelions). In two studies, these researchers showed that for children with the short variant (orchids), peer problems at age 6 were predicted by parental ratings of hyperactivity and impulsivity obtained 2 years prior. However, for children with the long variant (dandelions), there was no relationship between peer problems and hyperactivity/impulsivity. This study again underscored the importance of genetic effects on individual differences in susceptibility to environmental influences as they alter the effects of the environment on behavior.

13.6  Genetic and Environmental Effects Are Correlated Let’s consider one more important complication in the understanding of the interplay of nature and nurture. There are a variety of ways through which genetically driven behavioral tendencies actually alter the environment that the child experiences. First, children seek environments that are supportive of their genetic

13.6  Genetic and Environmental Effects Are Correlated

165

tendencies. Consider the child who has a tendency to learn abstract material more quickly than his peers (is academically intelligent). If given appropriate opportunities to read, for example, this child will gravitate toward reading material and will find it fun. The child asks to be read to and asks parents for books. Thus, the genetic tendency fosters an environment that is rich in words and reading material, enhancing reading skills and the use of words. In this way, the child has created an environment for himself that fosters the tendencies set up by his genetic makeup. This is a much different environment than the one created for a child who does not find reading fun, but who instead asks parents for sporting equipment due to a genetic tendency toward high levels of physical activity. Second, the social environment responds to individual differences in a manner that often reinforces them. The child who exhibits high levels of irritable/antagonistic is often fussy as an infant and has more intense and prolonged temper tantrums as a preschool child. These behaviors tend to illicit less warmth from parents and teachers, and even harsh, rejecting verbal responses indicating disapproval. The child, then, has created a caretaking environment that tends to increase his irritability because the adults in his world find his behavior objectionable. This might engender comments like, “Why do you yell at me so much.” “You hate me.” The peer group also tends to engage in more rejection of irritable children because they are often not fun to be around. This increases the frustration and irritability of the child with these tendencies. The irritable child, then, has created for himself an environment that is much different than the one experienced by the emotionally positive child who has a sunny disposition. Others enjoy being around the sunny and positive child and tend to respond to them with warmth and social praise. Third, parents and children are genetically related. Thus, parents create environments for themselves that tend to foster the same genetic tendencies in their children. Parents who are socially withdrawn probably have fewer social gatherings at their home than parents who are socially outgoing. This means that the children in a family prone to social withdrawal have fewer possibilities to interact with other people, particularly adults who are not family members. Because they have less experience with strangers, they do not have the opportunity to develop some social skills and do not feel as comfortable with strangers as does the child who has outgoing, socially active parents. If the parent and the child have the same genetically influenced propensities to be shy (which is likely), these propensities will be enhanced simply because they share an environment with their parents. All three of these effects (correlations between genes and the environment) complicate the distinction between genetic tendencies and environmental effects on behavior. They have confused much of the interpretation of psychological research. As Robert Plomin has pointed out, “Most measures of the environment used in psychology show substantial genetic influence. What looks like environmental effects in correlations between ‘environmental’ measures and psychological traits are actually genetic effects” (Plomin, pp. 185–186 [1]). However, it is also true that when environmental variables are not measured carefully in studies, or not all relevant environmental variables are included, genetic effects can appear stronger than they are.

166

13  Why Children Exhibit Different Behavioral Patterns

13.7  Other Considerations The Age Effect  One other finding from genetic research is worth mentioning. The heritability of cognitive and temperament-related behavioral tendencies increases throughout the life span. Thus, the older one gets, the more likely it is that an individual will resemble their parents or their brothers and sisters in temperamental tendencies. This increasing similarity is less likely for other types of individual differences such as attitudes, beliefs, or consumer preferences. So, if you are selecting a mate, it is probably wise to look at the temperamental tendencies of the parents of a potential mate. This may give a better indication of the behaviors the person is likely to exhibit over the long-haul. The Random Nature of Environmental Events  As we have described above, many environmental events occur on a more or less random basis. The death of a child’s parent, the family moving from one location to another, or serious illness are examples of randomly occurring events. Many random events affect the behaviors exhibited by children as they develop. However, a large number of these effects may be temporary. They may change the individual’s behavior pattern from their natural (genetically influenced) tendencies, but after a period of perturbation, the behaviors may return to the level prior to the disruption. This pattern has been seen in research on dieting. Body weight can be changed by many types of dieting regimes, but often the individual gains the weight back after a period of time, returning to a weight more like their original weight (referred to as a return to baseline). However, there is good research showing that some people successfully keep the weight off by changing their behavior in a number of ways. (Body weight is influenced by many factors; we are using a simplified example.) The initial choice of going on a diet may have been influenced by idiosyncratic events like a particularly embarrassing event that stimulated concern for one’s weight. It could have been brought about by a doctor’s pointed comment about the association of body weight with heart disease. All individuals, including children, are strongly affected by random events, and these tend to account for about as much variation in individual differences as known genetic effects.

13.8  Evolution and Individual Differences in Behavior Evolutionary theory indicates that behaviors that enhance the probability of survival to reproductive age and that enhance the probability of reproduction would tend to be maintained in the genome. Behaviors that do not enhance reproductive success are eliminated. We have shown in the research reported in this book that irritability/ antagonism and social withdrawal tend to reduce the social attractiveness of children in middle childhood. Many other researchers have also documented this effect. We have reviewed studies indicating that these effects are relatively stable; thus, the

13.8  Evolution and Individual Differences in Behavior

167

child with these tendencies often grows up to be an adult with these tendencies. Evolutionary theory might be thought to indicate that irritability/antagonism and social withdrawal would reduce one’s chances of surviving into adulthood and/or to reduce the chances of successfully finding a partner with whom to produce offspring. Why then do we see in the human population such large individual differences in the behaviors we have measured? To overly simplify a complex set of findings in evolutionary psychology, in the distant past, individuals at all levels of the eight broad behavioral tendencies we have studied have existed in an environment in which those tendencies have been adaptive. Consider fearfulness. Human beings are relatively weak and slow creatures compared to other animals. We cannot outrun a lion or physically overcome a hippopotamus in a one-on-one confrontation. To be able to survive on the savannas of Africa, in places where there were many strong and fast predators, high level of vigilance driven by fear was adaptive. In situations in which the environment is changing (e.g., dwindling food supplies), or where there is a probability that other human beings might take your food supplies by force, fearfulness is less adaptive. In these circumstances, more adventuresome and aggressive behavior would make reproductive success more likely. The first author recently observed a nature film that illustrated how a mixture of temperamental styles among members of a group enhances the survival of the group. Currently, in remote regions of France, wolves have migrated in from Italy. (All wolves had been hunted to extension in France until recently.) The influx of wolves has caused considerable damage to flocks of sheep that graze freely in mountainous regions of France. Some breeds of dogs have been used to protect the sheep; they do not herd the sheep they just travel with them. Recently, researchers, using cameras attached to the dogs and other remote sensing technology, have discovered that the dogs in this situation have a social hierarchy in which the most aggressive, dominant dogs travel in the center of the sheep herd, and the more timid, fearful dogs travel in the periphery. This social organization works well because the fearful dogs are constantly on the alert for danger and sense the approach of wolves first. When they bark and growl in response to approaching wolves, the more dominant and aggressive dogs come out from the flock to help chase away the wolves. The differentiation in temperament among the protective dogs has come to be valued by the shepherds as it has led to more success in controlling wolf predation on the sheep. This example shows how in this social group, individuals with different temperamental tendencies make unique contributions to the health and safety of the group. A similar circumstance could be imagined for early human beings on the African savanna. Thus, optimal adaptation to an environment would have been enhanced by having individuals in the social group with different temperamental characteristics. It is the goodness-of-fit of the behavioral tendencies of the individual and the specific aspects of the environmental niche in which they find themselves that determine survival and reproductive success. Because human beings have lived in many different environmental circumstances for the past 50,000  years, these specific adaptations have been encoded in our DNA and have resulted in individual

168

13  Why Children Exhibit Different Behavioral Patterns

differences in genetically driven behaviors. DNA is a code of effective environmental adaptation. Problems in child behavior, then, is a partial reflection of a behavior pattern that in the child’s deep genetic history worked well, but now may not be a good fit in their contemporary environmental niche.

13.9  Evolution and the Seven-Profile Model To help understand a temperament-based model of individual differences in single characteristics and in profiles of these characteristics, it may be helpful to think of individual differences as talents. Some children can learn to play the trumpet, can learn to solve differential equation, or can learn to appreciate poetry more quickly than others. This is the common use of the term “talent.” Talented people simply learn the new task easily and quickly, and often become better at using these skills in the future. Individual differences in traits related to emotion, attention, and motor behavior can also be thought of as talents. For example, some people have a natural tendency to care for others; they more easily understand the point of view of others and can sense the sadness or anger that others feel. This is a particularly valuable talent in the social world. Some children have the ability to focus their attention on learning material and not impulsively move from one stimulating event to another in quick succession. Attentive children have an important cognitive talent that is likely to have long-term implications for their academic and work success. All children have fears and the family and/or the culture attempts to build in some adaptive fears by teaching children about things that are dangerous (e.g., look both ways before crossing the street because cars are dangerous). Some children learn to be careful more quickly than do others and to regulate their fears so that they are useful (not getting hit by a car) but not overwhelming (avoiding roads and cars). Other children become overwhelmed by their fears and by all those things in the environment that their culture has taught them to fear. The children who can moderate their fears have a talent for self-regulation that helps them when they have to get an injection from their doctor, get their first haircut, or have to deal with a bully in middle school. On the other hand, the child who has very little fear may have a talent for adventuresome activities and for reducing the likelihood of being bullied. For each of these “talents,” there is a huge range of individual differences. A few children are very talented in one or more areas, and a few children have a great deal of difficulty learning the necessary skills in one or more areas. Most children are in the middle range. Given the same environmental condition, each group will respond differently due to differences in their talent structure. We posit that there are a set of behavioral individual differences that can be considered talents. This set includes the talent for (a) academic learning, (b) developing empathy and expressing positive emotion in the presence of others, (c) vigorous motor activity, (d) self-regulation of negative emotion and antagonistic behavior, (e) self-regulation of attention, (f) affiliation with other human beings (sociability), and (g) self-regulation of fear and feelings of insecurity. We settled on this list of

13.9  Evolution and the Seven-Profile Model

169

individual differences because, in free descriptions of parents and teachers, these characteristics are most often mentioned. Further, this list of characteristics, in one form or another (at one level of abstraction or another), has been considered by temperament researchers who have looked at early-appearing individual differences in childhood. Finally, some researchers have developed lists of individual differences in adults that bear a striking resemblance to the characteristics we chose to study in childhood. For example, coming from a different theoretical position, Panksepp has described six basic motivations corresponding to regions of the brain that affect human motivation. Based principally on research using deep brain stimulation, the six basic motivations are labeled seeking, caring, play, fear, rage/anger, and sadness/panic. The similarity of these basic motivations to most lists of temperament characteristics is substantial and obvious [11, 12]. Many researchers find that there are robust genetic links between traits [1, 13]. Researchers who study behavior in children find that there are substantial correlations between some sets of behavioral tendencies. Researchers who study psychopathology have for decades described tendencies toward comorbidity; that is, children who have one form of behavior problem (on type of diagnosis) have a tendency to have other forms of behavior problems. Each of these often-replicated findings supports the meaningfulness of describing child behavior as a profile instead of describing a child in terms of a list of discrete and independent traits. These profiles are descriptive of groups of children (a cluster) who exhibit similar levels of a number of behavioral tendencies. Many children who have an above-average tendency toward distractibility and disorganization also have an above-average tendency toward irritability and social antagonism. Self-regulation of attention and irritability/antagonism seem, on the surface, to be very different behaviors. But because they are correlated, relatively large groups of children can be statistically identified that are homogeneous with regard to these traits. For example, in our research, two groups were found who had low levels of self-regulation, and both groups had elevations on tendencies toward poor self-regulation of attention as well as increased levels of irritability/antagonism. The two groups were different with regard to the average level of academic ability and motivation. Interestingly, academic ability and motivation had a lower correlation with control of attention and irritability than attention control and irritability had with each other. In summary, many environmental factors affect the behavior of children. Social class, minority status, culture, as well as parental and teacher behavior are some examples. However, how children respond to these circumstances is moderated by genetically driven individual difference in temperamental characteristics. Some children, even under severely stressful environments, flourish, are confident, and are psychologically healthy. Some children, despite the most benign and privileged circumstances, do not flourish, lack confidence, and have significant psychological difficulty. Inattention, irritability, antagonism, unhappiness, and poor scholastic achievement occur in all racial groups, all social classes, and all cultures. These are behavioral responses linked to genetically driven temperament characteristics. The range of talent in any racial/ethnic group, in any social class, and in any culture is

170

13  Why Children Exhibit Different Behavioral Patterns

enormous. These talents not only affect how quickly a child may learn to do algebra or to play the piano but also affect how quickly they will learn self-regulation of emotion, attention, and motor activity, how readily they will express empathy for others, how often they express joy, and how many friends they have. These talents, particularly when viewed as a profile of related talents, have not been well studied in psychology and the other social sciences. But new research techniques are increasing our understanding of these differences, which will lead to better, more individualized parenting and teaching.

References 1. Plomin, R. (2018). Blueprint. Cambridge, MA: MIT Press. 2. Spithoven, A. W. M., Cacioppo, S., Goossens, L., & Cacioppo, J. T. (2019). Genetic contributions to loneliness and their relevance to the evolutionary theory of loneliness. Perspectives on Psychological Science, 14, 376–396. 3. Martin, R. P., & Dombrowski, S. C. (2008). Prenatal exposures: Psychological and educational consequences for children. New York: Springer. 4. Association for Psychological Science. (2020). Robert Plomin receives Grawemeyer award for behavioral genetics research. Observer, 33, 10. 5. Saudino, K. J., & Wang, M. (2012). Quantitative and molecular genetic studies of temperament. In M. Zentner & R. L. Shiner (Eds.), Handbook of temperament (pp. 315–346). 6. Lester, B. M., Conradt, E., & Marsit, C. (2016). Introduction to the special section on epigenetics. Child Development, 87, 29–37. 7. Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influences. Psychological Bulletin, 135, 885–908. 8. Boyce, W. T. (2019). The orchid and the dandelion: Why some children struggle and how all can thrive. New York: Knopf. 9. Obradovic, J., Portilla, X. A., & Ballard, P. J. (2016). Biological sensitivity to family income: Differential effects on early executive functioning. Child Development, 87, 374–384. 10. Stenseng, F., Li, Z., Belsky, J., Hygen, B.  W., Skalicka, V., Guzey, I.  C., & Wichstrom, L. (2018). Peer problems and hyperactivity-impulsivity among Norwegian and American children: The role of 5-HTTLPR. Child Development, 89, 509–524. 11. Davis, K. L., & Montag, C. (2019). Selected principles of Pankseppian affective neuroscience. Frontiers in Neuroscience, 12, 1025. 12. Panksepp, J., & Biven, L. (2012). The Archeology of mind. New York, NY: Norton. 13. Mukherjee, S. (2016). The gene: An intimate history. New York: Scribner.

Chapter 14

Diagnostic Implications of Temperament-­ Based Profiles

14.1  The Narrowing Conception of Normal The primary rationale for developing a model of common temperament-based behavioral profiles is to provide a framework for understanding individual differences in children. We have claimed that research on individual temperamental and personality characteristics has resulted in hundreds of behavioral traits. As a consequence, parents, teachers, and many researchers have a difficulty forming an integrated and comprehensive understanding of the behavioral propensities of children. One of the consequences of this lack of understanding is that parents, teachers, and professional diagnosticians tend to view individual differences in children through the lens of the medical model of pathology, which is often inappropriately applied to normally occurring individual differences. This chapter examines these issues. In the past few decades, parental conceptions of what is normal have been especially influenced by descriptions of diagnostic categories of psychiatric disorders described in newspapers, magazines, computer-based information sources (e.g., blogs), and television talk shows. This is, in part, the consequence of real advances in professional understanding of the mental health problems of children. Thus, many parents have heard something about disorders commonly known as autism, attention deficit hyperactivity disorder, learning disability, or depression. Many parents will also have looked up symptomatic descriptions of these disorders on the internet. This understanding, no matter how incomplete, adds to the anxiety many parents and teachers have about the individual differences they observe in the behavior of their children. This pathological emphasis often leads to a focus on behavior viewed as problems and as potential signs of pathology. Growing up in the 1950s and 1960s, my parents and their middle-class friends had a much simpler set of concepts to refer to when considering child psychopathology. Primarily, they worried about indications of “mental retardation,” now more appropriately referred to as intellectual disability. They observed the same

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_14

171

172

14  Diagnostic Implications of Temperament-Based Profiles

individual differences in children that parents have always observed, but their concept of what was normal was broader than that of many contemporary parents. They had biases of course about what behaviors were optimal or most desirable (“She is a delightful child, so nice.” “He is a stinker, a real problem for his parents.”), but they seldom talked about their children or the children of others from the perspective of psychiatric abnormality. In addition to increased exposure to clinical diagnostic categories, parental perceptions, and those of educators, are also affected by cultural pressure for optimal behavior and high levels of achievement. My (RM) parents wanted their children to be happy, obedient, and “nice” and to do well in school. Getting into highly selective K-12 schools or prestigious colleges, for example, was not part of their concern; nice if it occurred, but not mandatory. After the Second World War, the economy of the USA was expanding, jobs were plentiful, and the middle class had increasing material comforts. Being nice, “getting along,” and doing reasonably well in school were not a bad strategy. But the world my parents experienced is a different world than the global competitive world of the twenty-first century. Parents of all socioeconomic circumstances are now more concerned about academic preparation and about building a resume of achievements for their child, beginning in early childhood. Specialized instruction from an early age in languages, mathematics, science, and even the arts and sports is progressively more common. In order to profit from the advanced and specialized instruction that many parents consider to be important for their children’s future, there is increasing pressure for their child to exhibit the motivation, social skills, and self-regulation that optimizes learning. Parental concerns about such behavioral tendencies are not unfounded. As we have seen throughout this volume, there is a relationship between temperament-­ based behavioral characteristics and success in school. Three consequences of increased awareness of psychological pathologies as conceptualized in a medical model, as well as increased societal pressures for achievement in childhood, include (a) increased rates of diagnoses of mental disorders in children, (b) increased utilization of special education services based on these diagnoses, and (c) increased prescription rates for psycho-active medication for children. Let’s look at the data in each of these areas.

14.2  I ncreasing Rates of Diagnoses of Mental Disorders and Special Education Placement There has been a dramatic increase in the number of children receiving diagnoses and special education services in several categories of mental disorders during the past four decades. For example, the rates of diagnosis of autism spectrum disorder (ASD) have rapidly increased. In an analysis of responses to the National Survey of Children’s Health in the USA [1], an independent nationally representative

14.2  Increasing Rates of Diagnoses of Mental Disorders and Special Education…

173

telephone survey of households with children, parents in 2007 reported that 1.16% (1 in 86 children) of their children received a diagnosis of ASD. The same survey conducted in 2012 found that the prevalence was 2.00% (1  in 50 children). This represents an increase in 5  years of 72%. Data from the 2014 National Health Interview Survey, another well-established and authoritative survey of American households, produced an even higher estimate of the prevalence of ASD at 2.24% [2]. Increased rates of diagnoses of autism spectrum disorder have had a dramatic effect on rates of enrollment in special education classes for ASD. In the 1990–1991 schoolyear, the enrollment in ASD classes was zero as children with these characteristics, particularly those with severe symptoms, were often placed in special classes with children who had intellectual disabilities. By the 2000–2001 schoolyear 93,000 children in the USA were enrolled in special classroom or programs for ASD. By the year 2010–2011, enrollment was approximately 417,000. Just considering the increase from 2000 to 2010, this was an increase of 348% [3]. Data from the 2017 to 2018 schoolyear indicated that 670,000 children (10% of all 6.7 million children in the USA in special education) are diagnosed and receiving special education services for ASD [4]. Similarly, rapid increases in the diagnosis of attention-deficit/hyperactivity disorder (ADHD) have also occurred during the past several decades. The percentage of children diagnosed with ADHD increased steadily from 1997 to 2006 [5] and increased by 42% from 2003–2004 to 2011–2012 [6]. Children with ADHD are included in the diagnostic label “Other Health Impaired” for educational purposes in the US. During the 1980s and most of the 1990s, this category included only children with chronic diseases that had a significant detrimental effect on school participation and learning such as neurological disorders (e.g., seizure disorders), cancers, digestive disorders, and health conditions not included in other special education categories. In the late 1990s, ADHD was added to the Other Health Impaired category. As a result, enrollment increased from 59,000 in the 1990–1991 school year to 303,000 by 2000–2001, an increase of 413%. Enrollment continued to increase through the 2010–2011 school year to 716,000 children, an increase of 136% in the 10  years from 2000 to 2010 [3] (National Center for Educational Statistics, 2013). The latest statistics indicate that 19% (1,273,000) of all 6.7 million children in special education are included in the Other Health Impaired category, the vast majority of whom have a diagnosis of ADHD [4]. Most of the increase in enrollment in special education programs occurring during the past four decades has been for the following four diagnostic categories: learning disabilities, autistic spectrum disorder, attention-deficit/hyperactivity disorder, and developmental delay. There has been little change in the rates of special education enrollment in disabilities related to sensory difficulties (e.g., vision, hearing), orthopedic difficulties, or neurological difficulties that do not involve mental disabilities. Driven primarily by these four categories of disability, the overall enrollment (in all 13 categories of special education services) has risen from 9.95% in 1980 to 12.89% in 2010 of all children enrolled in public education [4]. Since 2010, the increases have continued to a rate of 14% in 2018.

174

14  Diagnostic Implications of Temperament-Based Profiles

14.3  Prescriptions for Psychoactive Medication The shrinking conception of what is normal, and consequent increases in medical and psycho-educational diagnoses, has resulted in an increasingly large number of children receiving psychoactive medication in the past three decades. One robust study [7] provided estimates in 2014 of the prevalence of use of psychoactive drugs in children and adolescents. Using a database of individual prescriptions from approximately 33,000 retailers, estimates of prescriptions rates for the three most commonly used groups of psychotropic medication (simulants, antidepressants, and antipsychotics) in the USA were made for four age groups: ages 3–5 (0.8%), ages 6–12 (5.4%), 13–18 (7.7%), and 19–24 (6.0%). If this sample (n = of 6,351,000) were projected to  current US census population  levels,  (roughly 330,000,000), 209,000 children age 3–5, 3.4 million ages 6–12, 3.4 million adolescents 13–18, and 3.0 million young adults ages 19–24 would be estimated to receive one or more of the three types of medications studied. These data indicate that roughly 19.9% of all children of ages 0–18 are receiving some form of psychoactive medication. A study of prevalence of psychotropic drugs in children and adolescents in 23 countries [8] (mostly in Europe) was 15.3% for ADHD medications, 6.4% for antidepressants, and 5.5% for antipsychotics. It is difficult to know if this is a higher percentage than the 19.9% for the USA as many children receive both an ADHD medication and an antidepressant for example. Further, criteria for diagnosis of some psychiatric conditions are different in Europe than in the USA. However, a good estimate is that about 20% of all children in the USA and Europe receive some type of psychotropic medication. Unfortunately, these studies did not indicate changes in rates of prescriptions over time. One study [9] indicated a moderate increase in the use of ADHD medications (based on parent report) from 1999 to 2014, but no increase for antidepressants or antipsychotics. However, there has been a remarkable increase in the use of psychotropic drugs in children from the 1970s to the current time.

14.4  A Problem or Progress? In many respects, the increasing numbers of children being diagnosed and receiving specialized treatment for behavioral and emotional problems are signs of progress. No one who is knowledgeable about the suffering of children who have significant mental disorders wants to return to the ways these children were treated in the 1950s and before. Many children with ADHD, for example, were thought of merely as having a discipline problem and families were blamed for not providing appropriate parenting. It is now widely known that these children have problems that are grounded in their genetics and neurobiology present from the time of birth. Advances in research on child psychopathology by the psychological and psychiatric communities during the past seventy years have been substantial. Much more is known about the genetics, physiology, and social consequences of a wide range of mental disorders. There has also been progress in the treatment for mental

14.4  A Problem or Progress?

175

disorders in the areas of both pharmacology and behavioral interventions. This is a humanitarian success story of significant proportions. Unfortunately, this success has come with a number of associated costs. One cost is that treatment is linked to a diagnosis. In other words, in order for families to receive help for a behavior or mental health problem, at least in the USA, the child must first be defined as having a medically defined abnormality. Receiving a diagnosis is often the first step toward getting treatment, and it also has the potential for substantial negative consequences. Primary among these is the damage to the self-­ image of the child and damage to the perceptions held by parents, teachers, and the peers of the child. Most mental disorder diagnoses for children are made in the age range 7–13. This is a critical time for self-understanding and for finding one’s place in the social world as we have demonstrated in the research we have reported in this volume. Carrying a diagnosis of a mental disorder often comes at that time of life when fitting in and being like ones’ peers is most important. Feeling different may have many adverse effects ranging from increasing social isolation to lowered expectations for the future. In addition to problems of self-perception, it may also create inappropriately negative expectations of future life goals on the part of the child’s parents and educators. Diagnoses are also linked to treatment by funding. Diagnosis is the key step in obtaining insurance reimbursement for medications. It is also linked to the manner in which physicians are reimbursed for their services by insurance companies. Similarly, in the educational arena, in order for children to receive special education services, they must be given a psycho-educational diagnosis. A diagnostic classification allows the local school district to receive the additional funds that are necessary for the support of the special education programs. Supplying medications for mental disorders is a very lucrative business for pharmaceutical companies. It is enhanced by the practice of health insurance reimbursements being higher and easier to obtain for drug treatment than for behavioral therapy. The enormous money involved (over $40  billion dollars in 2012  in the USA) has created its own problems. The pharmaceutical industry has found that sales are strongly enhanced by marketing to the general public. From 1996 to 2005, the drug industry tripled its spending on marketing, including a fivefold increase in direct-to-consumer advertising. The effect is that patients request advertised drugs when visiting the physician. Further, when patients request advertised medications, they are much more likely (some reports say 17 times as likely) to receive one or more new prescriptions than are patients who did not request any medications [10]. Aggressive marketing has fueled the off-label prescriptions of antipsychotic drugs for a growing list of mental health disorders, particularly in the case of children. Off-label uses are those uses that have not been adequately researched to insure effectiveness and safety. Therefore, off-label uses are not sanctioned by the Federal Drug Administration. Individual physicians may prescribe a drug for an off-­ label use under specific circumstances, but pharmaceutical companies are prohibited by the FDA from promoting drugs for off-label uses. During the past 10 years, a number of successful law suits have been filed against pharmaceutical companies for this practice.

176

14  Diagnostic Implications of Temperament-Based Profiles

Another problem involved in diagnosis and treatment of children is that in the USA, less than half of the prescriptions for psychotropic drugs for children are prescribed by child psychiatrists. This is primarily due to the shortage of child psychiatrists in the USA. Most prescriptions for psychotropic medications are written by pediatricians who have very little training in developmental psychology and child psychiatry.

14.5  Have We Lost the Concept of “Normal”? In addition to problems related to socially stigmatizing children and monetary incentives for diagnosis by physicians and drug companies, the emphasis on “abnormality” has other less obvious effects. In the realm of child behavioral research, funding from large-scale research projects is often linked to pathology. It is easier to obtain research funds after a school shooting, for example, because the public and politicians want to find early signs of mental illness that is often considered by the public and the news media as being fundamental to the cause of this form of violence. It is much more difficult to obtain funds to study normal individual differences in children. Further, the focus on diagnoses and psychopathology can inappropriately shift parental, educational, and research decisions from the remaining 80–90% of the child population. This leads to the expectation that if children do not have a mental disorder, they are normal and do not need special societal attention. In education, for example, this fosters the idea that typical children don’t need an individualized plan of interventions to effectively foster their behavioral or emotional development. The emphasis on pathology also fosters the notion that all children in the “normal” range are similar. This reinforces the simplifying idea that one parenting style, one method of discipline, or one education program will optimize developmental outcomes. This notion affects children in numerous ways. For example, curriculum changes in education most often test the effect of the “new program” on all children. It is rare that behavioral individual differences in children are considered as a variable that might moderate the effects of the “new program.” If the new curriculum has a positive effect on one-third of the children, but no effect on the remaining 67%, then the program is likely to be evaluated as a failure because it produced no significant effect for the entire school population. In fact, it may have been very successful for some children with certain characteristics while being ineffective for children with other characteristics. Children have varying talents for learning, for controlling their aggressive tendencies, managing their fears, and regulating their attention. Some children who are very talented in this regard exhibit these skills early in development. Other children are less talented and must have many more repetitions of experiences in order to learn these skills. Many parents and teachers have no difficulty understanding individual differences in cognitive, athletic or artistic talent, yet children also exhibit a similar range of talent with regard to their social and emotional lives and these talents are less well understood.

14.6  The Medical Model and Common Problems of Children: A Mismatch

177

14.6  T  he Medical Model and Common Problems of Children: A Mismatch The medical model is based on pathologies that predominantly have a single cause. Injury (e.g., broken arm due to a fall) and infection (e.g., influenza) are classic examples. Some pathology is also caused by genetic abnormalities in one or a few genes (e.g., Huntington’s disease). There are a large number of such disorders but they all are very rare. In all these cases, you either have the pathology or you don’t. Diagnosis, then, is made on a qualitative, binary basis. However, there is no single gene for ADHD, for anxiety, for depression, for learning disabilities, or for most forms of lower academic performance. Years of research that aimed at discovering the single gene abnormality causing these common conditions has resulted in failure [11]. Instead, it is now clear that all these common problems are influenced by hundreds if not thousands of genes, each of which makes only a very small contribution to the likelihood that an individual will experience any of these common problems. The genes that are associated with these problems turn out to be associated with normal variation in individual differences in the human population. Some of us inherit only a few of the thousands of genes associated with ADHD and have a low probability of ever manifesting the symptoms regardless of the environments that we experience. Some of us inherit more of these genes and if we were raised in families characterized by chaos and poor self-regulation training, then we might manifest many of the symptoms of ADHD. For some children who had the unfortunate accident of inheriting many of the genes related to ADHD, they are likely to manifest these problematic behaviors in a wide range of environmental circumstances. Only in environments designed to be particularly therapeutic for the development of self-regulation of emotion, attention, and motor-activity will these children be able to manage these propensities. In addition to genetic issues, additional considerations support the idea that the tendency to experience psychopathology is on a continuum; that is, there is a generalized tendency toward problematic behavior. Two broad types of behavior problems have been observed in children for the past 70  years. These are labeled “internalizing” problems (e.g., phobias, anxiety, and depression) and “externalizing” problems (e.g., conduct disorders, ADHD, and opposition defiant disorder). Children within each of these problem groups are likely to exhibit symptoms of the other diagnoses in the category. If they do exhibit more than one such disorder they are then said to be “comorbid.” Further, a common finding is that children who exhibit externalizing disorders often have clinically significant levels of internalizing problems. Thus, comorbidity occurs across these two categories of problems as well as within. Statistically, the correlation between scores on a measure of externalizing and internalizing problems is often found to be close to 0.50. This indicates that there is a tendency for children with any type of psychopathology to have an increased probability of exhibiting other problematic behaviors. This understanding has fostered the idea that there is a general tendency (labeled “p”) to exhibit

178

14  Diagnostic Implications of Temperament-Based Profiles

problematic behavior that is much like the general intelligence dimension (“g”) assessed by intelligence tests. In other words, there are many forms of intelligence, but these forms are related and can be understood as a general propensity to learn quickly. Research indicating that there is a general tendency to exhibit behavior problems leads to questions about how individual differences in temperament and personality might relate to this tendency. Jennifer Tackett and colleagues have investigated the associations between personality variables and a general pathological tendency in a number of studies [12–14]. One such investigation [15] of 695 children of ages 7–13 years demonstrated that the general personality factor of neuroticism (made up of items related to fear as well as irritability) had a correlation of 0.76 with internalizing problems (e.g., anxiety and depression) and 0.49 with externalizing problems (e.g., conduct problems, aggression, ADHD). Further, a general pathology factor (P) was developed encompassing both internalizing and externalizing problems and was found to correlate 0.81 with the neuroticism factor. These are very high correlations and they indicate that the personality dimension of neuroticism (principally made up of temperamental variables of fear and irritability) is measuring much the same thing as the propensity to manifest childhood behavior problems. This result strongly supports the idea that common childhood psychopathology is on a continuum, and in fact, is on the same continuum as normal individual differences in temperament/personality. What we consider abnormal are behaviors that are extreme on the normal distributions of individual differences in temperament and personality. Data from molecular genetics, analysis of comorbidity in childhood behavior problems, and relationship of behavior problems to individual differences in temperament/personality, all indicate that the medical model in which diagnosis is based on a qualitative difference between health and illness makes no sense for these common developmental problems of childhood. Genetic propensities toward problems related to obesity, reading problems, ADHD, anxiety, drug and alcohol abuse, conduct problems, and even schizophrenia and perhaps some symptoms related to autism are normally distributed. Thus, they exist on a continuum. We all have some probability of manifesting these problems from time to time, but there are large individual differences in these probabilities. Thus, drawing a line in the normal distribution that indicates that those below the line are “normal” and those above the line are “abnormal” is not meaningful. Yet, this is what school psychologists, clinical psychologists, pediatricians, and child psychiatrists have to do every day in order to obtain treatment for children in distress. The best of these diagnosticians makes an assessment of the extent to which the child’s current behavior and emotions are adversely affecting their life. That is, the decision about whether to treat someone is made on the goodness-of-fit between the child’s behavior and their ability to cope with the social and educational demands of school, and the level of distress the child is experiencing in this environment and in the home. This forces the diagnostician to consider a continuum of distress felt by the child, as well as the extent to which the behavior exhibited is distressing and disruptive to teachers and peers at school. But how much distress, how much family

14.7  The Seven-Profile Model and Problematic Behavior

179

conflict, or how much school failure is required before a diagnosis is made and a treatment instituted. In the end, this depends on family and school expectations as well as cultural norms for appropriate behavior. It is these societal norms to a greater extent than medical considerations that often determine if a child will be given a diagnosis. If parents indicate they are highly distressed by the child’s irritability, aggression, and disobedience, and the school also has concerns about these behaviors, then it is a problem because these parties find the behavior to be a problem. This sense of distress gets communicated to psychological and medical authorities, and there is a high probability that some diagnosis (e.g., conduct disorder) will be made. This argument, in no way, indicates that families, teachers, and children in distress do not need professional help. The problem is simply that we have tied providing help to children to a medical diagnosis. This implies that a child who is diseased, and is abnormal. This is an unfortunate and misleading conception of the mechanisms involved in behavior problems. It often contributes to roadblocks to getting treatment for the child.

14.7  The Seven-Profile Model and Problematic Behavior The authors of this book believe that a temperament-based model of individual differences in children like the one we have described could have a positive effect on how childhood psychopathology is understood and treated. Four potentially salutary effects are described. 1. Parents, teachers, and mental health professionals need help in understanding the range of normal individual differences in behavior. Due to advances in understanding psychopathology and the subsequent increases in awareness by parents and school officials of these pathologies, many parents have lost sight of the extent of normal variation in the cognitive processes (attention, talents for learning different aspects of academic material), emotional reactivity (including both positive and negative emotions), and social behaviors among children. When a child is not meeting parental or school expectations, this misunderstanding of normal variation often results in conceptualizing the child as having a pathology. The range of individual differences within the normal range is much larger than it is generally believed. These individual differences are normally distributed, with a few children being at the extremes of the distribution (having a very low tendency to exhibit the behavior or a very high tendency), but most children are somewhere in a middle range. However, the actual behavior exhibited by children in any given environment depends on the fit between the child’s general behavioral tendencies and the demands placed on the child by the environment. If the fit is poor (a highly active child is not given the opportunity to engage in strong motor movement; an irritable child is often spoken to by parents or teachers in an aggressive,

180

14  Diagnostic Implications of Temperament-Based Profiles

argumentative manner), problematic coping strategies on the part of the child are likely to occur. When these poor fit occasions occur, the child’s natural temperament-driven tendencies will be exaggerated. The irritable poorly self-regulated child will become more irritable and the shy child will become more reclusive. In such circumstances, they will need help from the adults around them and sometimes from mental health professionals in order to optimize their adjustments to these challenging environments. Increased understanding of the range of normal individual differences might also have the salutary effect of reducing the number of referrals for psychological and psychiatric evaluations. There is a shortage in almost all cultures of mental health professionals. For example, in the public schools of the USA, psychologists are constantly trying to keep up with a tidal wave of referrals from teachers and parents for psycho-educational assessment. Child psychiatrists are in extremely short supply and wait times to obtain an appointment can be months. These professionals need to have the time to focus their energies on the most critical cases, particularly emergencies that require immediate attention. Helping referral sources have a better understanding of “normal” could have a positive effect toward more appropriate professional mental health activities. 2. Knowledge of normal individual differences helps parents, teachers, mental health professionals, and others understand the probabilities of particular behavior difficulties that a given child is likely to manifest. In a recent report, a group of researchers reviewed the literature on the prevalence of mental health problems over the life span [16]. The researchers focused on studies of children who had been assessed for academic ability (intelligence) and behavior on multiple occasions over a number of years. In some cases, the research included children who had been assessed at age 3, for example, and were repeatedly assessed through middle age. These researchers found, as many others had found in recent years, that from 50% to 90% (percentages vary due to the nature of the research methodology) of all children will have some diagnosable mental health problem during their life span. This figure is at odds with the understanding of the general public (i.e., most parents and teachers). Most people feel that mental health problems are rare and the majority of people are free of mental health problems throughout their lives. However, this body of research indicates that our general cultural understanding is in error. One of the implications of this kind of finding is that mental health and behavior problems should be thought of more like the common cold than like a permanent abnormality (i.e., schizophrenia or autism). That is, we are accustomed to understanding that our somatic health fluctuates over time. We are all likely to get a cold, perhaps even have pneumonia. But this distressing condition is then treated and we recover. In many cases, we recover with no professional medical intervention. Our research shows (see Chap. 4) that some children who exhibit the  well-­ adjusted behavior profile, and who generally have low probabilities of experiencing behavior and emotional problems, report problems with symptoms of depressions in middle childhood. Our research also demonstrated that some children in all of the

14.7  The Seven-Profile Model and Problematic Behavior

181

seven profiles were having significant behavioral or emotional difficulties. Understanding how children who exhibit commonly occurring behavior profiles occasionally have significant problems may help parents and teachers better understand the nature of mental health problems without relying on the medically oriented nomenclatures of pathology. The profiles we have identified are relatively common in the general populations of children. In this sense, each profile identifies a normal group of children. Many of these children, at one time or another will adapt poorly to some set of environmental circumstances. For most of these children, the poor adaptation will be temporary. Unfortunately, they often cannot receive professional help unless they are given a diagnosis indicating that they are abnormal (i.e., they exhibit a pathology). For a number of reasons already pointed out above, this is an unfortunate and inappropriate process. Needing help for what is often a temporary behavior problem should not require this kind of label. We have reported elsewhere in this volume that temperamental characteristics assessed in early childhood have repeatedly been found to predict mental health problems throughout the life span. A recent study of the relationship of temperament to mental health takes a revealing look at the other side of the equation. These researchers attempted to understanding that rare group of persons who never experienced a diagnosable mental health problem [16]. In the Dunedin study of children in New Zealand, 17% of the entire sample of 968 persons who had been followed from age 3 through age 30 had never met criteria for a diagnosable mental health problem. What were the characteristics of these persons who had enduring mental health? To the surprise of the researchers, they were not born into well-to-do families, and they did not have markedly better physical health nor did they have higher intelligence than the group of persons who did experience mental health problems. The characteristics that most clearly differentiated this group were their advantageous temperamental style and a family history indicating that few relatives had mental disorders. Their temperamental style was characterized by fewer emotional problems (e.g., fears, irritability), low levels of social withdrawal, and age-­ appropriate levels of self-control. These characteristics were measured at age 3. As adults they report superior occupational attainment and greater file satisfaction, and their relationships with others were of higher quality. This study points out that some lucky individuals (a rare few), who had a predisposition toward self-regulation of emotion and attention, enjoyed being with others, were not easily emotionally upset (even when assessed during their preschool years), and had a favorable history of mental health through age 30. The profiles we have identified in our seven-cluster model provide a framework for understanding the differential probabilities for life outcomes. For example, some portion of the well-adjusted and high-average self-regulators are likely to have enduring positive mental health. Each of the other groups have increased probabilities of having behavioral and mental health problems throughout life, and the types of problems that are most likely vary from profile to profile. However, it is also likely that children with each of these profiles in middle childhood will have normal or typical lives, in which long periods of relative happiness will be punctuated with occasional life difficulties.

182

14  Diagnostic Implications of Temperament-Based Profiles

3. Understanding normal behavioral tendencies can greatly enhance the meaningfulness of a mental health diagnosis. When a behavior problem occurs and a referral is made for an evaluation by a mental health professional, one step that could broaden the kinds of information obtained in a typical psychological/psychiatric assessment of a child is to include information on a full range of normal temperament-related behavioral tendencies. This is seldom done in current practice. The vast majority of assessment is directed at determining if some threshold of “abnormality” has been reached so that the given diagnostic label can be applied. This is a short-sighted strategy that limits information about factors that will affect the prognosis of the maladaptive behavior and the appropriateness and effectiveness of any given treatment. For example, academically talented children with significant ADHD symptoms have a far different developmental trajectory than children who are below average in academically related cognitive abilities who also have these symptoms. Our research indicates that bright children with ADHD-like behaviors are far more socially integrated and are often considered to have above-average leadership capabilities by their peers. Less academically talented children, who exhibit poor self-­ regulation of attention and more frequent expression of irritability, are much more often socially isolated and given less status by their peer group. Likewise, children who are socially inhibited and fearful but who are academically talented have a much different social and schooling experience in middle childhood than children exhibiting the same level of social inhibition and withdrawal, but who are below average in academic abilities. Second, understanding normal individual differences helps determine the type of environment that is likely to be most therapeutic for a particular child. Many children who have tendencies toward social withdrawal, regardless of diagnostic classification (learning disability in reading; some children with autism), will profit most from a therapeutic process that involves working with teachers and groups of other children that are warm and accepting. Children who have reading problems, and also have normal tendencies toward poor self-regulation of attention and expression of negative emotion, need an environment in which distractions are kept to a minimum. A social environment involving working with other children, in most cases, doesn’t work well with poorly regulated children because the behavior of others is distracting. Third, even if a child is determined to meet diagnostic criteria for a particularly mental or behavior disorder, communicating with the child, parent, and teacher about the positive characteristics of the child can make the adjustment to an abnormal condition much more positive. Parents of children with cerebral palsy know that their child has a disabling condition. However, to be made aware that their child’s sunny, positive temperamental disposition is a tremendous advantage in facilitating treatment and optimal development, can be a useful and ameliorative piece of information. This joyful disposition will facilitate the interaction with occupational and physical therapists, for example, because working with the child is fun and rewarding. It is a much more difficult task for these professionals to work with a child who is irritable. An irritable child with cerebral palsy has two

14.7  The Seven-Profile Model and Problematic Behavior

183

disabilities, one which is obvious (cerebral palsy) and the other is their strong tendency toward irritability whose effects require professional and skilled therapists with particular skills in working with children who are easily frustrated and often find therapy emotionally difficult. 4. Researchers, particularly those who evaluate parenting or educational interventions, should find that the outcomes are better understood if a profile model of child behavior is used as a moderating variable in their research. Many programs designed to evaluate parenting styles or educational intervention programs find relatively weak effects. However, part of the difficulties in showing success of such programs is that the intervention may be been quite successful for some children, of no benefit for others, and may have even had a negative effect on others. In this circumstance, the overall effect (when all children are considered) is negligible. If children are measured on a wide range of cognitive and temperamental characteristics (as we did in the current seven-profile model) and assigned to clusters based on profile similarity, then the potential differential effects based on child characteristics can be found; that is, the effects of any intervention that is designed to change the behavior of children is likely to be different for children with different behavioral tendencies. This approach to research on child behavior is consistent with the current models of medical intervention (sometimes referred to as personalized medicine) in which different medications are administered depending on the genetic characteristics of the patient. Some patients do not metabolize a given medication well. It simply has no effect because the active ingredients are not assimilated physiologically in a way that is useful to the patient. Other patients cannot tolerate a medication due to adverse side effects. It is no different with behavior change programs for children; that is, due to fundamental differences in the child’s behavior, some groups of children will be unable to respond positively to almost any therapeutic intervention. Others may respond particularly well. Only when researchers look for these moderating effects based on behavioral predispositions are the effects of intervention programs likely to be meaningfully understood. In summary, the profile model that we have described is viewed as one step toward helping parents, teachers, and mental health professionals understand individual differences in children. A better understanding of the ways normal children differ in their cognitive abilities, their emotion expressions, and their behavior propensities should help caretakers better cope with the behavior problems that arise and limit the tendency to view normal behaviors of children as being symptoms of a disease process. If the child, parents, and school personnel are feeling stressed by the child’s behavior, and this behavior is been relatively consistent over time, then the help of mental health professionals may be sought. These professionals will engage in a diagnostic process that would be significantly enhanced by a formal and broad-spectrum analysis of the normal behavioral tendencies of the child. This kind of information is currently rarely obtained in a formal, scientifically justified way. Such information would aid most diagnostic and treatment decisions made by mental health professionals.

184

14  Diagnostic Implications of Temperament-Based Profiles

References 1. Blumberg, S. J., Bramlett, M. D., Kogan, M. D., Schieve, L. A., Jones, J. R., & Lu, M. C. (March 20, 2013). Changes in prevalence of parent-reported autism spectrum disorder in school-aged U.S. children: 2007 to 2011–2012. National Health Statistics Reports, 65. (U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics). 2. Zablotsky, B., Black, L.  I., Maenner, M.  J., Schieve, L.  A., & Blumberg, S.  J. (2015). Estimated prevalence of autism and other developmental disabilities following questionnaire changes in the 2014 National Health Interview Survey. National Health Statistics Reports, 87. (U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics). 3. National Center for Educational Statistics. (2013). The condition of education, 2013. U.S. Department of Education. 4. Office of Special Education Programs. (2018). The condition of education: A letter from the Commissioner. U.S. Department of Education, Individuals with Disabilities Education Act 5. Pastor, P. N., & Reuben, C. A. (2008). Diagnosed attention-deficit/hyperactivity disorder and learning disability: United States, 2004–2006. National Center for Health Statistics. Vital Health Statistics, 10(237). http://w2ww.cdc.gov/nchs/data/series/sr_10/SR10-­237.pdf. 6. Visser, S.  N., Zablotsky, B., Holbrook, J.  R., Danieloson, M.  L., & Bitsko, R.  H. (2015). Diagnostic experiences of children with Attention-Deficit/Hyperactivity disorder. National Health Statistics Reports, 81, 1–7. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics. 7. Sultan, R. S., Correll, C. U., Schoenbaum, M., King, M., Walkup, J. T., & Olfon, M. (2018). National patterns of commonly prescribed psychotropic medications in young people. Journal of Child and Adolescent Psychopharmacology, 28, 158–165. 8. Piovani, D., Clavenna, A., & Bonati, M. (2019). Prescription prevalence of psychotropic drugs in children and adolescents: An analysis of international data. European Journal of Clinical Pharmacology, 75, 1333–1436. 9. Hales, C., Kit, B. K., Gu, Q., & Ogden, C. I. (2018). Trends in prescription medication use among children and adolescents—United States, 1999–2014. Journal of the American Medical Association, 319, 2009–2020. 10. Smith, B. L. (2012). Inappropriate prescribing. Monitor, 43, 36. 11. Plomin, R. (2018). Blueprint: How DNA makes us who we are. Cambridge, MA: MIT Press. 12. Soto, C. J., & Tackett, J. L. (2015). Personality traits in childhood and adolescence: Structure, development, and outcomes. Current Directions in Psychological Science, 24, 358–362. 13. Tackett, J. L. (2006). Evaluating models of the personality-psychopathology relationship in children and adolescents. Clinical Psychology Review, 26, 584–599. 14. Tackett, J. L., Lahey, B. B., van Hulle, C., Waldman, I., Krueger, R. F., & Rathouz, P. J. (2013). Common genetic influences on negative emotionality and a general psychopathology factor in childhood and adolescence. Journal of Abnormal Psychology, 122, 1142–1153. 15. Brandes, C. M., Herzhoff, K., Smack, A. J., & Tackett, J. L. (2019). The p factor and the n factor: Associations between the general factors of psychopathology and neuroticism in children. Clinical Psychological Science, 7, 1266–1284. 16. Schaefer, J. D., Caspi, A., Belsky, D. W., Harrington, H., Houts, R., Horwood, L. J., Hussong, A., Ramrakha, S., Poulton, R., & Moffitt, T. E. (2017). Enduring mental health: Prevalence and prediction. Journal of Abnormal Psychology, 126, 212–224.

Chapter 15

Limitations, Major Findings, and Implications

15.1  Limitations In these days when it is particularly difficult to determine what is “true” and what is not, it is particularly important to indicate the limitations of the research on which much of the information presented in the volume is based. All research has limitations and this is particularly true of the social sciences due to the complexity of the questions under investigation.

15.1.1  All Models Are Wrong to Some Extent The goal of our research is to present a model of temperament-based individual differences in children at one age level (ages 8 through 12). More specifically, the goal was to develop a taxonomy of the most commonly occurring temperamental profiles. The word “taxonomy” in this context is meant to indicate that the profiles developed through our research procedures would capture the full range of commonly occurring non-pathological behavioral tendencies. This was an ambitious goal, as there have been only a few other studies on which to base our procedures and interpretations. One important consideration in understanding the approach taken in this volume comes from the idea of a model. In the case of our study, we attempted to capture the most essential aspects of parent and teacher assessments of the behavior of more than 2000 children. Children were assessed on eight broad characteristics. This means we had more than 16,000 pieces of data at our disposal (eight characteristics per child). Through statistical methods we determined that seven behavioral profiles could be extracted from these descriptions that replicated across the three samples studied (US parents, US teachers, and Russian parents).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4_15

185

186

15  Limitations, Major Findings, and Implications

Our research illustrates a truism about models: They are simplifications. No model can explain all the data in the data set. For example, we described each temperament profile using average scores for each of the eight characteristics, calculated from the scores of all children who were placed in that cluster. However, any given individual in the cluster will deviate to some degree from that average score. Thus, there is variation in characteristics even within a profile. The children described by the profile have similar scores on the characteristics. Their profile scores are closer to those in that cluster of children than they are to children exhibiting any other profile, but they are not all the same. The profile then is a simplification of the scores of children defined by that profile. Thus, all models are wrong to some extent. What is important is how useful the model is in helping human being understand complex phenomena. Consider, for example, the model of the atom that most of us were taught in high school. It involved electrons circling around a central nucleus (proton and neutron). This model was helpful because it had similarities to the planetary orbits around the sun and so was easily understood. Further, it helped explain important aspects of chemical reactions. However, we now know that the structure of the atom is much more complex than this simple model would suggest. The model was wrong, but it was useful. There are many other examples of models that affect our lives on a daily basis. Weather prediction is perhaps the one most often encountered. Through consideration of jet stream patterns, atmospheric temperature in various regions of the world, ocean temperatures, etc., predictions are made about the chances of rain in a given area in a specific time period. This model makes predictions in terms of the probability that rain will occur. This means the prediction is known to contain error. It is an oversimplification of all the events that affect weather, but it is a useful model. This discussion of models leads directly to a practical consideration. If a parent rated their 9-year-old son on the temperament measure that we used, their scores on all eight indicator variables (cognitive and temperamental characteristics) could be used to determine which one of the seven profiles their scores would be most similar to. However, it is likely that they would not be perfect fit. For example, while having many characteristics of well-adjusted high achievers, they might have some characteristics of withdrawn high achievers. Understanding the developmental trajectory of this child might require understanding the assets and risks of both clusters of children.

15.1.2  More or Less than Seven Profiles May Prove Useful The number of profiles obtained through a statistical process, no matter how sophisticated the procedures used, will depend on the specific variables that are included in the model. We chose eight broad temperamental characteristics that a good deal of previous research indicated were important in lives of children in early and middle childhood. However, a different set of characteristics is likely to result in a different number of profiles. One researcher might include physical attractiveness as judged by peers as an indicator, whereas another researcher might include

15.1 Limitations

187

individual differences in socio-cultural variables such as parental income or ethnic/ racial characteristics. If these characteristics were included in the current model, it is likely that a different number of profiles would be isolated. Profile structure will also be affected by the sample of children studied. The attempt in the current study was to obtain observations of parents that were representative of the population of parents in the USA who had children in middle childhood. Given the size and representativeness of the sample studied, this was a major strength of our project. In addition, a sizeable sample of teacher observations of individual differences was studied, although this sample was not representative of the US population; all the children were from one region in the state of Georgia. Finally, we obtained a sample of parental observations of children from Russia. This was done in an attempt to understand how generalizable the structure was across cultural differences. This effort was a first attempt at studying cultural differences in profile structure using the indicator variables we studied. While we demonstrated that there was substantial similarity in profile structure, this comparison was limited by the modest size of the Russian sample (about 550 children that we studied). For studies of latent profile structures, this is not a large sample. Also, this sample came from one geographical area in country. We believe that this study has provided a model of individual differences in middle childhood that has real potential to be useful. But we make no claim that the seven profiles defined by the eight characteristics we studied will prove to be the most useful model. Most models of complex phenomena, like child behavior, become more precise over time as better measurements are obtained and more characteristics of interest are included.

15.1.3  L  ongitudinal Changes in Profile Stability Were Not Studied We have documented that other researchers have found that temperament-related behaviors studied individually have moderate stability. For example, children who exhibit more social withdrawal than their peers as 3-year-olds tend to be more socially withdrawn than their peers at age 5, 10, 15, and even as adults. A few studies have shown that children who exhibit a given temperamental profile in early and middle childhood tend to exhibit related behaviors as adults. Thus, children who exhibit a profile that is primarily defined by poor self-regulation of attention, negative emotions, and antagonistic behaviors have been found to exhibit more problems related to alcohol abuse and gambling as adults. These important studies have begun a process of documenting behavioral continuity over-time understood as profiles of temperament-related tendencies, but our understanding is still very limited. Does the nature of profile structure (number of profiles, descriptive characteristics) change as children develop? What little evidence that now exists indicates that this is likely. Most probably the number of profiles will increase as behavioral tendencies become more complex with increasing age and developing capabilities of children.

188

15  Limitations, Major Findings, and Implications

One important question regarding stability involves the extent to which a child maintains his/her behavior profile across developmental stages. What little data is available indicates that early in life there is considerable change, but with increasing maturity, there is increasing probability that a person will exhibit a similar profile. None of these issues were investigated in our project. The hope is that the project reported here provides useful information on which to build future research efforts. Longitudinal research of this type is extremely difficult and expensive; studies like ours can provide information about which characteristics are most important to include in longitudinal work. An appropriate study of individual and structure stability of profiles in children will require 2000–5000 children at minimum. Undertaking studies of this size requires a substantial commitment of funds (millions of dollars). Such allocations have been made in recent studies in Australia and in Finland. Such research holds great promise for understanding the issues of profile stability.

15.1.4  Latent Profile Analysis Currently Has Limitations Latent profile analysis (LPA) was used as the primary tool for determining the nature and number of profiles that best describe individual differences of children in middle childhood. This is a useful and highly touted tool, but the method is not highly developed; it is still in its infancy. In large data sets using continuous variables (in our case, temperamental characteristics that are normally distributed across a wide range), most of the typically used criteria to determine the number of profiles in a given data set provide no clear indication of the exact number that is optimal. Further, the criteria in some cases provide contradictory indications. (For details and a number of examples, see Appendix E.) The further development of LPA and related tools takes time and considerable software development to get to the stage that it can be broadly utilized by general researchers. Currently, due to the limitations of these tools, the researcher has to make a number of decisions in the process of data analysis based on theory. All statistical outcomes must be evaluated based on theory, but with further evolution of LPA, more definitive statistical guidelines will no doubt evolve.

15.1.5  Models Are Only as Good as the Measurements Made There are many factors that contribute to measurement quality, and the measurement tools used in the current study have been studied in a number of contexts and refined through a number of iterations. However, measurement of the perceptions of any parent or teacher is subject to a variety of idiosyncratic biases (mental health status of the observer, experience with children, the understanding of what is typical behavior for a child of a certain age). Some of these biasing effects can be reduced by combining the assessments of parents and teachers. In the current project, measurements were obtained from US parents, US teachers, and Russian parents. However, US parents and teachers did not observe the same children. No data were

15.2  Summary of Findings

189

available from Russian teachers on the children who were assessed by Russian parents. Thus, it was not possible, given our available data, to combine assessments of parents and teachers for the same child. There is a long history of research on parent-teacher agreement and important differences in perception are often obtained. However, recent research by Major and colleagues [1] has shown that many of these differences are related to the level of the profile rather than the contour of the profile. In our research, primary consideration was given to the topographic contours of the profile rather than the level of the profile. Specifically, mean level differences in profiles of parents and teachers in the USA were eliminated by calculating scores on each characteristic in relation to the mean of each sample separately (scores were centered at the sample level). The same procedure was used when comparing profiles from Russian parents to those obtained from US parents and teachers. The focus of the research was on individual differences, not on average differences between parent and teacher ratings. However, if the same children had been assessed (rated) by parents and teachers in the current study, issues of profile level and contour could have investigated.

15.2  Summary of Findings We believe that despite the limitations just mentioned, the research we have reported makes a number of meaningful contributions to the current understanding of individual differences in children. The most important of these contributions are listed below. 1. Parents and teachers have a similar understanding of individual differences of children in middle childhood. Seven profiles of cognitive and temperament-­ related characteristics of children in middle childhood were extracted from assessments of US parents, Russian parents, and US (Georgia) teachers. Each of these profiles, extracted from the three sets of assessments, were shown to be similar in topographic contour (the means of each characteristic for a given profile were similar for the groups of raters). This is the first study to show this level of inter-rater agreement on temperament profile structure in middle childhood where both parents and teachers were used. It is also the first study to demonstrate profile equivalence for parents from different cultural contexts. 2. Temperament-based profiles were demonstrated to relate to how compliant the child is to adult expectations at home and school. Parents and teachers rated the level of compliance each child demonstrated regarding rules and normative expectations at home and at school. The measure of compliance was very strongly related to profile type. 3. Temperament-based profiles were found to significantly relate to ratings of behavioral problems exhibited by children. Parents and teachers also rated several types of behavior problems exhibited by children. Children in different clusters (defined by profiles) were found to have significantly different rates of aggression, hyperactivity, emotional problems, conduct problems, and problems with peers. Specifically, measures of these types of problems were significantly related to profile type, although not all profiles were different from one another on any given problem.

190

15  Limitations, Major Findings, and Implications

4. Temperament-based profiles were found to be significantly related to achievement as assessed by grade point average. For the Russian sample, profile type was significantly related to grade point average. This result indicates that the perceptions of these parents regarding the temperament-related characteristics of their children were related to their school achievement in meaningful and practically significant ways. 5. Temperament-based profiles based on parent and teacher perceptions were significantly related to the perceptions of academic ability and achievement motivation as rated by peers in school. The perceptions of same-age peers in school regarding academic ability and achievement motivation were obtained from peers who were unaware of the ratings of teachers. Yet the perceptions of peers at school were closely related to the perceptions of teachers in the areas of academic ability and achievement motivation. This indicates that by middle childhood, there is a “community” consensus about the academic talents and motivations exhibited by children. 6. Temperament-based profiles were significantly related to the perceptions of the social status, likeability, social influence, and social prominence of their peers. Current findings demonstrate that temperament-based profiles are related in important ways to the social environmental experiences of children. In middle childhood, the attitudes and behaviors of peers toward one another play an important role in the socialization and development of self-perceptions of children, and temperamental characteristics play a role. 7. Temperament-based profiles were significantly related to the perceptions of peers regarding amount and type of aggression exhibited by children. Children were asked to nominate other children who were verbally, socially (excluding others from the group), or physically aggressive. Profile types (based on parent and teacher perceptions) significantly predicted peer measures of aggression. 8. Temperament-based profiles were significantly related to self-perceptions of motivations. Students rated their own academic motivation and motivations to engage in inappropriate behavior in school. Profile type was significantly related to these self-perceptions in meaningful ways.

15.3  Implications 15.3.1  S  imilarity in Profile Structure Across Raters and Cultural Environments Reinforces the Ideas that Temperament-Related Individual Differences Are a Natural Part of the Human Condition We have shown striking similarity in the profile structure of the eight temperament-­ related behavior characteristics that we studied. This result supports the idea that individual differences in these behaviors are readily observed in the home and in the school setting. Further, they are observed in different cultural settings. Of course,

15.3 Implications

191

we have studied only two cultural groups. But our results (and the results of other studies) lead to the notion that these profile structures are a natural part of the human condition; that is, these measures tap individual differences that are commonly observed regardless of the sex of the child, their socioeconomic circumstances, or their minority/majority status. It is important in this context to understand what we have not demonstrated. We have shown that the contours of the seven profiles we found are very similar. These contours are defined by differences from the mean level of each of the eight behavioral characteristics studied. However, we have not shown that other characteristics of the profiles are similar across rater type or cultural setting. In fact, our data indicates that there are sizeable differences in the variability around these means for parents and teachers, and for parents in different cultural settings. Further, the percentage of the samples exhibiting each profile type in different samples is not equivalent. For example, the teacher profile for the well-adjusted profile contained about 50% more children than the US parent profile. These differences could result from differences in sample characteristics or other technical issues. However, it is likely that even if these profiles have a high generalizability across setting (home and school) and culture, that setting and culture will play a significant role in how a given profile of behavior is evaluated. For example, it is known that culture plays a role in the value placed on social inhibition. Eastern cultures find inhibition and cautiousness in children a more attractive trait than do those in Western cultures. Social sensitivity, one aspect of social inhibition, has been found to be related to various forms of social maladjustment in Canadian children of ages 4–8 years but is related to better adjustment in children of similar ages in Shanghai [2]. Thus, values placed on specific levels of behavioral characteristics are likely to vary by cultural settings and to have different effects on adjustment based on these values. These values, in turn, may play a role in the initial assessment of the behavioral characteristic and on the social implications of that characteristic for future behavior.

15.3.2  The Social Environment Stabilizes Behaviors The results of this project provide data that helps explain why early appearing individual differences in children become progressively more stable as the child matures. One such mechanism is that the social world of the child (parents, teachers, peers) in middle childhood (ages 8–12) is coming to a consensus about his/her behavioral characteristics. This is a period in development in which the child (a) increasingly comes into contact with children outside the home, (b) has the cognitive capabilities to begin to understand the social world, and (c) must perform difficult cognitive and self-regulatory tasks in an increasingly demanding environment. Children in the age range of 3–7 years have begun to experience contact with others outside the home, are asked to perform cognitive tasks, and have to exercise some level of self-regulations. However, in middle childhood, the range of peers the child encounters and the social demands greatly increase. Further, their developmental capacities to begin to make sense of this complex and demanding world are increasing.

192

15  Limitations, Major Findings, and Implications

As children in middle childhood cope with all of these increasing demands, those around them are forming impressions of, expectations for, and predictions about how they are likely to behave. Our research has made the assumption that these impressions are based primarily on characteristics related to cognitive ability, achievement motivations, physical vigor, irritability/antagonism, self-regulation of attention, social inhibition, social withdrawal, and the tendency to feel insecure and fearful. We based these assumptions on extensive research showing that parents, teachers, and others view these characteristics as being important in interactions within the social world. Further, individual differences in most of these temperamental characteristics can be observed very early in development—some in infancy, others in toddlerhood and the preschool period. Temperament research indicates that differences in genetic makeup, and in early environmental experiences (including the inter-uterine environment), foster behavioral individual differences. But these differences by middle childhood begin to be encoded by the social environment and a general consensus begins to form. Your mother may believe you are the brightest 3-year-old that was ever born, but by age 8, based on various types of feedback from others in the community (including teachers), a more realistic impression often emerges. Peers observe how well a child performs in school and begins to form an impression of a peer’s academic capability. Through interactions with others, the child’s self-image is formed. All of these processes begin to converge and stabilize through social consensus. It must be emphasized that not all teachers or peers agree on how a given child is likely to behave. What we have shown in this research is that the average of the perceptions of one’s peers regarding each of the eight characteristics studied agrees with what the average parent and teacher would perceive. Some children, perhaps best friends, would have quite a different view. Still, through constant talking among peers, a general consensus is likely to form. As the social community begins to predict how you are likely to behavior, this reinforces the original behavioral tendencies of the child. Thus, in addition to genetic and congenital influences that foster individual differences in behavior, the child has created his own social environment through his behavioral tendencies. The way and extent to which those in the environment, in turn, react to those behaviors tend to further stabilize those early individual differences.

15.3.3  R  esearch on Individual Differences in Temperament-­Related Behavioral Tendencies Strongly Implies that Relying on One Parenting or Educational Approach Toward Behavior Management Will Be Ineffective if Not Harmful Fearful, insecure children, who are shy (inhibited in new environments), for example, tend to be sensitive to expectations of adults. They are particularly sensitive to punishment, even mild punishments. Some portion of shy, fearful children are

15.3 Implications

193

particularly prone to environmental influences in general, including positive as well as negative conditions (they are susceptible). Many are highly compliant to adults and tend to look to adults for approval. Thus, they thrive under conditions of warmth and support, and when parents and teachers understanding of their behavioral tendencies. If these conditions are not met, they tend to have a poor development outcome. Poorly self-regulated children, on the other hand, are particularly sensitive to cues of reward and often become angry when the path to the reward is blocked. They are relatively more rewarded by the social world of their peers, because it often leads to social rewards. This group is relatively unaffected by the minor punishments that are used by most parents and teachers. Thus, they are perceived by adults in their world as non-compliant. Some portion is simply less responsive to environmental influence in general, whether this feedback is positive or negative (they are less susceptible). These two examples indicate that optimal educational experiences and behavioral management of children with these two patterns of behavior are drastically different. Most socially sensitive parents and teachers are aware of these differences and to a greater or lesser degree attempt to treat these two types of children differently. However, there is a natural tendency to treat all of one’s own children the same. This is particularly true in schools where issues of fairness and justice create a press toward uniformity of treatment. But the world is moving toward individualized consuming (creating more and more options), individualized advertising, and individualized medicine. The latter example is of interest because individualized medicine is based on the idea that not all people process (metabolize) the medication in the same way or have the same tendencies toward particular pathologies. Therefore, a given medication does not have the same effect on all people. Further, some groups because of their genetic propensities have a greater likelihood of contracting specific diseases. However, for many reasons, we tend to treat all our children in schools in the same way, unless they have a medical or psycho-educational diagnosis. Any individualization is left to the natural instincts of the teacher. Some do a wonderful job, many do not. Most teachers receive excellent instruction on how to create curricular materials but very little instruction on child development, in general, and on individual differences, in particular. Many parents have a very poor understanding of normal individual differences. A first step toward individualization would be for researchers and evaluators of parent-based and school-based interventions to test the differential effects of each behavioral or academic intervention on children who have different cognitive abilities and temperamental profiles. A profile model like the one we have proposed could be helpful in this regard. In this way, the effects of the natural predispositions of the children would serve as a moderator variable, allowing for the discovery of exactly which groups of children profit from the intervention and which do not. A similar approach could be taken with parent intervention programs.

194

15  Limitations, Major Findings, and Implications

15.3.4  A  ppropriate Clinical Diagnosis of Behavior Problems Requires an Understanding of the Normal (Non-­pathological) Behavioral Tendencies of the Child Good clinicians have always known that making decisions about whether a child meets some set of criteria for a particular mental or behavioral disorder does not in itself allow for appropriate treatment. Consider the example of a diagnosis of a viral infection. Once a test determines that the individual has the corona virus, for example, the next question is: What is the person’s general health status? Do they have chronic lung problems? Do they have allergies to specific medications? Are they cognitively and behavioral capable of following an appropriate therapeutic regime? In a specific example of a behavioral problem that has reached the point of being detrimental to the child’s developmental progress, a child of age 9 might be given a diagnosis of ADHD.  How parents and teachers go about devising an appropriate program for this child should be made based on some of the following considerations: Does this child have the cognitive ability to cope with the schooling demands they currently are facing? How happy is the child on most days? Is the child considerate and empathetic toward others? Is the child shy when first meeting new people? Does the child become frustrated easily and express negative emotion in these circumstances? The answer to these questions and several others will relate to how much they enjoy schooling, their friendship patterns, and their social status; how well they are liked by teachers; and what other kinds of sub-clinical behavioral problems the child might have. Bright children who enjoy learning in spite of attention problems and have a high activity level have a much different prognosis than children who struggle at school. The optimal intervention for academically talented children is likely to be very different than the intervention for children who learn at a slower rate than their peers.

15.3.5  A  llowing Children to Be Who They Are Is Likely to Have the Most Positive Effect on Their Developmental Path The LGBTQ movement has opened the eyes of many to the harm caused by forcing children and adults into overly restrictive behavioral roles. The women’s movement, in all its forms, has liberated women to develop skills and play any of the roles in the society that are available to men. Similar movements toward racial and religious equality have fought for equal opportunity and tolerance for differences. The profiles we have modeled regarding the behavior of children have a similar meaning. Valuing these patterns of individual differences is particularly important as they interact with differences in cultural, religion, or sexual orientation. For example, an African American adolescent in the USA who has a tendency toward poorer

References

195

self-­regulation of emotion may be at much higher risk in confrontations with authority figures (e.g., have discipline problems in school and have harmful interactions with legal authorities like the police) than his more self-regulated peer of the same cultural background. Children in all cultures, in families with all religious or political orientations, and with parents who are of the same sex or different, exhibit a wide range of learning ability, achievement orientation, motor vigor, irritability, shyness, insecurity, fearfulness, and desire to be with others. In our research, we found very little effect of socioeconomic status, or ethnicity/race on the behavioral tendencies we measured. Thus, all human beings exhibit these social and emotional differences. Yet, some kinds of behavioral predispositions are typically undervalued. This results in attempts by well-meaning adults to push for changes in children toward cultural norms that often fail, and in fact, exacerbate behavioral problems. Valuing individual differences in social and emotional predispositions does not mean that parents and teachers do not have to enforce rules that keep the child from harming others or themselves. In fact, parents and teachers have the obligation to help the child learn the skills to exercise self-control of behavioral tendencies that do not work well in some environments and to help the child find those environments where their natural tendencies do work well. All the behavioral profiles we have outlined work well in some environments, given some specific circumstances, but they all do not work well in others. Helping children find their environmental niche and helping them cope more effectively with the difficult task of controlling one’s behavior in situations where your natural predispositions do not fit well are the challenging roles of the parent and teacher.

References 1. Major, S.  O., Seabra-Santos, M.  J., & Martin, R.  P. (2015). Are we talking about the same child? Parent-teacher ratings of preschoolers’ social-emotional behaviors. Psychology in the Schools, 52, 789–799. 2. Chin, X., Liu, J., Ellis, W., & Zarbatany, L. (2016). Social sensitivity and adjustment in Chinese and Canadian children. Child Development, 87, 1115–1129.

Appendix A: Study Participants

The research reported in this volume is based primarily on the analysis of data from three samples. Samples are described in terms of who provided the data and the country where the children lived. Two separate groups of children from the United States were studied. One group was a sample of children described by their parents, and a second group was a sample of children described by their teachers. The group of students whose behavioral tendencies were described by their teachers also provided descriptions of some of their own behaviors. Further, each child in the teacher sample responded to measures describing the behaviors of their peers at school. In order to help determine if the results of our research on US children could be generalized to children in other countries, comparisons were made to a third sample of children from Russia whose behavior was assessed using parent reports.

US Parents Data from US parents were collected using internet brokers that specialize in bringing together persons who are willing to do computer-based tasks at home with persons who have tasks (in our case questionnaire-based measures of child behavior) that can be presented via a computer. Data on child temperament and personality characteristics in the US parent sample were collected by the primary author (RM). Data were obtained using two different survey companies (Mturk & Survey Monkey) who obtain data from panels of people who are approximately representative of the national population. These companies develop panels of persons who respond to surveys for a small fee. This kind of data collection procedure has been used widely in the social sciences (Table A.1). Data were collected from all states in the United States, including Alaska and Hawaii. The original sample contained 3500 children from age 3 through 18. In this report, we will describe the analyses of parental descriptions of 912 children in the

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4

197

Appendix A: Study Participants

198 Table A.1  The number of children studied by sample

Source of the data Basic samples  US parents  Russian parents  US teachers  Student peers

Original samplea

Analyzed sample

1150 538 913 473/460b

912 538 913 473/460

The original sample indicates the data collected that meet criteria for inclusion. The analyzed sample of data provided by US parents was reduced by selecting a subsample that most closely approximated the characteristics of the US census in 2010 and the demographic characteristics of the teacher sample b There were two student samples; due to child absences at the time of assessment, many analyses included approximately 450 students a

Table A.2  Comparison of characteristics of US parent-rated sample to the US population Characteristic levels Gender  Male  Female Ethnicity of child  African American  Asian American  European American  Latin/Hispanic American  Other Education level of respondent  High school graduate or less  Some college or technical school  Graduated from college degree  Professional or advanced degree  Missing data

US populationa

US parent sample

51.1% 48.9

51.0 49.0

12.6% 4.8 63.7 16.3 2.6

9.9 5.0 54.1 8.2 22.8

44.7% 27.9 18.0 9.2 0.0

16.8 36.1 40.2 6.9 0.2

Based on 2010 Census data

a

age range of 8 through 12 years. This subset of the data in the 8–12 age range from parent was selected to match the sample size and demographic characteristics of the teacher sample (described below). The sampling procedure was designed to be representative of parents with children in the United States. This goal was approximated although the sample had a somewhat higher education level than the population of the country based on 2010 census data (see Table A.2 for details). The middle childhood period (ages 8–12) of development was given priority for several reasons. First, during this period of time, the child’s behavioral tendencies (temperamental and personality characteristics) are stabilized more fully than at earlier ages. For example, if behavior is measured twice 1 year apart, the results will

Appendix A: Study Participants

199

Table A.3 Geographic distribution of US parent-­ rated samples: percentage of respondents by region Regions New Englanda Middle Atlantic South Central Deep South Upper Mid-West Northern Plains Southern Plains Mountain Southwest Pacific Coast

US census 4.7% 15.6 9.7 13.2 15.0 3.8 14.6 1.9 5.3 16.2

8–12 3.5% 14.6 10.8 13.9 18.1 3.6 13.5 1.2 5.8 15.0

New England  =  Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, Vermont Middle Atlantic = Delaware, Maryland, New Jersey, New York, Pennsylvania, Washington, DC South Central  =  Kentucky, North Carolina, Tennessee, Virginia, West Virginia, Deep South  =  Alabama, Florida, Georgia, Mississippi, South Carolina, Upper Mid-West  =  Illinois, Indiana, Michigan, Ohio, Wisconsin North Plains = Iowa, Minnesota, Nebraska, North Dakota, South Dakota, South Plains  =  Arkansas, Kansas, Louisiana, Missouri, Oklahoma, Texas, Mountain = Idaho, Montana, Nevada, Wyoming, Southwest = Arizona, Colorado, New Mexico, Utah, Pacific Coast = California, Oregon, Washington

a

be more similar for children measured at 8 and again at age 9 than if the same measurements were obtained at age 3 and 4. Second, middle childhood is a critical period in the social development of the child. During these years, individual differences begin to interact with the demands of the world outside the home, and the child now has the cognitive capacity to begin to understand and to label aspects of this social world. Further, this is the period in which the effects of individual differences on schooling behaviors, peer interactions, and achievement are beginning to have a significant effect (Table A.3).

Teacher Sample Data were collected by Lease and colleagues from 52 teachers in ten schools in the north central part of the State of Georgia. Each teacher responded to the measures that were designed to measure temperamental and personality characteristics of children in middle childhood. Data were available on 913 children who were in grades 3 through 5, and the analyzed data was limited to children between the ages of 8 and 12 years of age. Because data were collected in two different waves from schools who served students with somewhat different demographic characteristics, the two groups of students were labeled Cohort A and Cohort B. Cohort A students were from families with moderately lower socioeconomic circumstances, were somewhat older (4th and 5th grade students), and served a higher proportion of African American students than Cohort B (contained 3rd through 5th grade

200

Appendix A: Study Participants

students). Both samples included roughly 450 children. There was minor variation from one measurement to another due to child absences at the time of assessment (see Table A.4).

Russian Parent Sample Finally, we wanted to determine the comparability of the profiles obtained from US parents with those from parents in another cultural setting. Data from a large Russian sample that had been collected by Helena Slobodskaya and colleagues at the Siberian Branch of the Russian Academy of Medical Sciences were made available for analysis. From this sample, behavioral measurements from 538 children in the age range of 8 through 12 were analyzed. These children resided in the center of the Russian Federation in or near the city of Novosibirsk, the third largest city in the country. The parents who described the behavior of their children were from a broad range of socioeconomic backgrounds. However, on average the parents were somewhat more highly educated than the general Russian population. The vast majority had a college degree or some amount of postgraduate training in academic institutions or technical institutes (see Table A.5). All three samples analyzed in this set of studies were sufficiently large to meet the purposes and statistical requirement of the data analytic procedures used. The samples include representatives of the groups of individuals that are most influential in the lives of children in middle childhood. That is, information was obtained from parents and teachers regarding the behavioral tendencies of children, and the children themselves provided a wide range of information regarding how they perceived their own behavior. Finally, peer perceptions of behavior, social status, and other variables were analyzed. This allowed for the evaluation of the extent to which behavioral profiles that were isolated affected the social lives of children in middle childhood. Table A.4 Demographic characteristic of the two cohorts of Georgia students

Characteristic Child sex  Boys  Girls Grade level  Third  Fourth  Fifth Ethnicity  African American  European American  Other

Cohort A

Cohort B

47.4% 52.6

47.9% 52.1

00.0% 37.3 62.7

22.4% 34.1 43.5

41.9% 55.0 3.1

13.0% 77.3 9.7

Appendix A: Study Participants Table A.5  Characteristics of the Russian parentrated sample

201 Characteristic Gender  Male  Female Respondents  Mothers  Fathers  Other relatives Education of mothers  14 years or less  More than 14 years  Missing data Mean age  Mother  Father

Russian sample 52.4 47.4 82.3% 8.6 9.1 57.5% 37.3 5.2 34.3 36.9

Appendix B: Measurement Methods

The Survey of Individual Differences of Children and Adolescents (SIDCA [1]) was the measurement tool used initially to develop behavioral trait scores for the US parent-rated sample. It was designed to measure 15 characteristic behavior patterns of children. This measure was derived from the 144-item version of the ICID [2]. The ICID was derived from an extensive study of descriptions of parents in six countries. The SIDCA is a second-generation instrument, which was designed to be shorter than the ICID and to provide a nationally representative normative sample of US parental ratings of children and adolescents. The specific items initially selected were the six items of each scale of the ICID that had the highest correlation with the total score of the scale. This was determined by averaging the item-total correlations from several published and unpublished analyses. The instrument used to measure the 15 behavioral traits in the teacher sample was a short form of the ICID. This measure was very similar in content to the published short form of the ICID by Deal and colleagues (ICID-S [3]). Since items present on the experimental form used in the teacher sample were present in the published Deal short form and were also present in the long form of the ICID used in the Russian sample and the SIDCA used in the US parent sample, the Deal short form items were extracted from all measures and were used as the final measures analyzed in all samples. This procedure was designed to create a comparable set of items across all samples. Items are responded to on a 7-point scale based on relative frequency of occurrence. Ratings are made by comparing the child’s behavior to that of an average child of similar age (1 = much less than average; 4 = average; 7 = much more than average). Although the same items were used in the calculation of scale scores in all samples, several differences in the procedures followed in these data collections are noteworthy. First, data were collected in two different languages, as the instrument responded to by Russian parents was translated into Russian from English. Second, for the Russian sample data were collected in a traditional paper-and-pencil format as was the data collection for US teachers. For the US parent sample, data were

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4

203

204

Appendix B: Measurement Methods

collected via a computer-based process. Third, the order of items on questionnaires was constant for the Russian and US teacher-rated samples, while for the US parentrated sample the items were randomly ordered for each participant. These measurement differences, as well as cultural and context (classroom versus home) differences, make the comparisons between the latent profile models obtained from these samples a conservative test. If a similar structure is obtained from all three samples, we can be assured that the results are robust across these different measurement techniques. Many research efforts in the social sciences involve multiple measurements of the constructs of interest. In the current example, 14 of the 15 constructs measured were analyzed. (The Compliance scale was eliminated from the latent profile analysis because of high correlations with several other scales.) Many trait measures were moderately correlated, and a few were highly correlated. All 14 indicator scores could be used in the latent profile analysis (LPA), or some form of aggregate of these scales (e.g., principal component or some other factor analytic method) could be used. Many researchers utilize factor score as indicators in LPA, but some specialists in latent profile analysis argue that factor scores “soak up the variance” in component measures and create a generalized measure that limits the number of profiles that are found in the analysis. Consider the general personality factor labeled “Neuroticism”. Aspects of this factor appear in nearly all personality measures for adults and are related to the irritability/antagonism dimension studied by temperament researchers in children. This factor in adults has been found to be related to how stress prone an individual is, how happy they tend to be, and how satisfied they are with their lives. However, the factor is made up of a number of facets including anxiety, angry hostility, depression, self-consciousness, impulsivity, and vulnerability. These facets correlate with one another in the range of .31 to .64 (median r = .47) [4]. Weiss and Deary [4] have shown that for individuals who respond with a high score on most items of the measure tend to have poorer mental and physical health. However, individuals who have high scores on elements of the scale related to anxiety and vulnerability tend to have better health outcomes. Similarly, a recent paper by Fournier et al. [5] has shown that in two samples, a sample of depressed adults and a community sample, there were clear advantages in separating components of the neuroticism factor in understanding the relationship of the factor and its components to depression and anxiety. Specifically, some of the specific components accounted for additional variance in interpersonal function over and above general negative affectivity, acute depressive, and anxiety symptoms. These studies indicate that specific aspects of broad factor scores can have very different psychological meaning despite being correlated. These considerations led us to investigate the utility of factor scores versus scale scores in the latent profile analysis of our data. We conducted analyses of the US data set and found that the use of factor scores reduced the specificity, theoretical meaningfulness, and number of profiles that were isolated in the analysis. However, some of the 14 indicator measures were so highly correlated that they approached the reliability of the constituent measures. Table B.1 presents the correlations between scales making up the aggregate scales for all three samples. It can be seen that most scales included in the aggregates were correlated above .60. The primary exception was the lower correlations among the scales measuring negative emotionality, antagonism, and strong-willed behavior in the Russian parent sample.

Appendix B: Measurement Methods

205

Table B.1  Correlations of scales chosen to create aggregate measures Scales Intelligence and Openness to Experience Positive Emotionality and Considerate of Others Negative Emotionality and Antagonism Negative Emotionality and Strong-Willed Antagonism and Strong-Willed Distractibility and Disorganized Inhibition with Social Withdrawal

US parent .68 .73 .60 .62 .58 .63 .59

US teachers .76 .79 .82 .69 .66 .70 .76

Russian parents .69 .72 .41 .53 .54 .64 .55

Table B.2 Reliabilitya,b of aggregate and single scales used as indicator variables for US parents, Russian parents, and US teachers Indicator Academic Ability Achievement Motivation Prosocial Emotion and Behavior Activity Level Irritability/Antagonism Poor Attention Regulation Inhibition/Social Withdrawal Insecurity/Fearfulness

US parents .88 .69 .85 .84 .85 .84 .85 .77

Russian parents .80 .75 .84 .83 .79 .77 .73 .71

US teachers .93 .91 .91 .84 .95 .89 .90 .84

Internal consistency reliability of items making up each indicator, calculated as alpha coefficients Items were based on the Deal short form for all three samples

a

b

As a result of these analyses, we constructed five aggregate measures that were composites of scores on substantially correlated scales. New aggregates were constructed from the items from the constituent scales. The composites are described in Chap. 3. The five aggregated measures and three original scales that were not substantially correlated with any other scale were used as the eight indicator measures in our analysis. We also did analyses in which all 14 of the original scale scores were used as indicator variables. The specificity and theoretical meaningfulness of the resulting models when compared to the eight indicator models were very similar. There are advantages of combining scales that are substantially correlated in addition to theoretical meaningfulness. The aggregated measure is a more reliable measure than measurement made of the individual scales. It is likely to be more stable across different situations in which the child finds himself/herself and across developmental periods. Table B.2 presents the reliability coefficients (alpha) for the eight scales, including the composite scales used as the primary measures in the analysis. The derivation of profile models from these eight temperament-related variables is based on the notion that there are moderate correlations among behavioral constructs of children. Therefore, it is useful to know for each of the major samples studied the nature of the correlations between the behavioral variables once the composite scales were created. Tables B.3, B.4, and B.5 present the correlations among the eight indicator variables for US parents, Russian parents, and US teachers.

Appendix B: Measurement Methods

206

Most correlations were quite similar across samples. However, inspection of the correlations among measures obtained from teachers indicates that they had a much higher correlation between the Academic Ability and Achievement Motivation scales than those obtained from parents (both US and Russian). They also had higher correlations between the academic ability and motivations scales and the scale measuring poor self-regulation of attention. Most other correlations were quite similar to those obtained from US and Russian parents. Table B.3  Correlations among temperament-related indicators: US parentsa Indicator AchAa AchM ProS ActL IrrA PoorAR InhW InsF

AchA 1.00

AchM .52 1.00

ProS .58 .42 1.00

ActL .28 .19 .36 1.00

IrrA −.23 −.22 −.50 −.07 1.00

PoorAR −.31 −.54 −.31 −.03 .51 1.00

InhW −.39 −.20 −.50 −.52 .26 .13 1.00

Insf −.36 −.18 −.34 −.30 .47 .40 .47 1.00

n = 913 a AchA Academic Ability—Intelligence, Openness to Experience; AchM Achievement Motivation; ProS Prosocial Emotion and Behavior—Positive Emotionality, Empathy; ActL Activity Level; IrrA Irritability/Antagonism—Negative Emotionality, Antagonism, Strong-Willed; PoorAR Poor Attention Regulation—Distractibility, Disorganization; InhW Inhibition/Social Withdrawal— Inhibition to the Unfamiliar, Sociability reversed scored; InsF Insecure/Fearful

Table B.4  Correlations among temperament-related indicators: US teachers Indicator AchAa AchM ProS ActL IrrA PoorAR InhW InsF

AchA 1.00

AchM .82 1.00

ProS .51 .67 1.00

ActL .28 .14 .14 1.00

IrrA −.36 −.55 −.75 .14 1.00

PoorAR −.72 −.82 −.56 −.08 .49 1.00

InhW −.46 −.42 −.51 −.58 .23 .41 1.00

Insf −.52 −.47 −.31 −.37 .21 .50 .58 1.00

n = 913 a AchA Academic Ability—Intelligence, Openness to Experience; AchM Achievement Motivation; ProS Prosocial Emotion and Behavior—Positive Emotionality, Empathy; ActL Activity Level; IrrA Irritability/Antagonism—Negative Emotionality, Antagonism, Strong-Willed; PoorAR Poor Attention Regulation—Distractibility, Disorganization; InhW Inhibition/Social Withdrawal— Inhibition to the Unfamiliar, Sociability reversed scored; InsF Insecure/Fearful

Appendix B: Measurement Methods

207

Table B.5  Correlations among temperament-related indicators: Russian parents Indicator AchAa AchM ProS ActL IrrA PoorAR InhW InsF

AchA 1.00

AchM .57 1.00

ProS .51 .62 1.00

ActL .44 .25 .36 1.00

IrrA −.14 −.44 −.40 .03 1.00

PoorAR −.38 −.64 −.35 −.11 .50 1.00

InhW −.46 −.21 −.38 −.57 .14 .23 1.00

Insf −.29 −.17 −.21 −.31 .28 .36 .47 1.00

n = 538 a AchA Academic Ability—Intelligence, Openness to Experience; AchM Achievement Motivation; ProS Prosocial Emotion and Behavior—Positive Emotionality, Empathy; ActL Activity Level; IrrA Irritability/Antagonism—Negative Emotionality, Antagonism, Strong-Willed; PoorAR Poor Attention Regulation—Distractibility, Disorganization; InhW Inhibition/Social Withdrawal— Inhibition to the Unfamiliar, Sociability reversed scored; InsF Insecure/Fearful

Appendix C: Statistical Issues

Introduction Based on parent and teacher responses, children ages 8 through 12 were given a score on each of 15 behavioral traits measured. There was wide variation among children with each sample studied. All behavior traits or tendencies measured were normally distributed. While 15 traits were measured, only 14 were used to form the eight indicator variables in our analyses. The Compliance scale correlated highly with a number of other scales, and so it made little unique contribution to profile differentiation and was eliminated from the latent profile analysis. At first glance, it seems possible that in a large sample (hundreds of children) every combination of scores across the eight behavioral characteristics would occur. Further, you might think that every profile (patterns across all eight measures) would be equally likely. In fact, research over the past 30 years has shown that some score profiles occur frequently, some are very rare, and some do not occur at all. This happens because behavioral traits are not all independent of one another; they are often moderately correlated as we have seen in Tables B.3, B.4, and B.5. The result is that some score profiles occur much more frequently than others. The goal of the analysis was to determine the number of profiles that best fit this complex set of behavioral tendencies. Stated another way, the goal of the research was to determine the number of profiles that were implicit in the behavioral ratings of parents and teachers. Another goal was to determine if the profiles obtained from parents and teachers in the United States were similar, and the extent to which these profiles were similar to those obtained from Russian parents. Within the realm of statistical methods for determining individuals with similar profiles, two general approaches can be used: traditional clustering methods or latent profile analytic methods. Traditional clustering procedures rely on distance measurements between individuals (on multiple measures) as the primary method of classification (i.e., those who are the shortest distance from one another are

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4

209

210

Appendix C: Statistical Issues

placed in a class where distance is measured by the difference in scores). Latent profile analysis relies on the probability that an individual is placed in a specific class based on estimated parameters (e.g., group means) as the primary method of classification. More precisely, cases in the same latent group are similar to each other because their responses are generated by the same probability distribution [6]. Our research has relied on latent profile analysis (LPA) procedures for estimating models of clusters of individuals with similar scores on multiple characteristics. (The procedure is referred to as latent profile analysis when continuous measurements are used as was the case in the current research or as latent class analysis when categorical data are used). Latent clusters are defined as unobservable subgroups of individuals that share a pattern of behavior (a behavioral profile). The goal of latent profile analysis is to determine the smallest number of latent clusters that is sufficient to account for the associations observed among the measured variables. Each cluster contains individuals that are relatively homogeneous with regard to their scores on the eight behavioral characteristics measured.

Specific Statistical Techniques Latent profile analysis was implemented using the mixed models subroutine of Mplus 7.3 [7]. Following established procedures [8], we estimated models typically with three through nine clusters and chose the best-fitting model based on multiple criteria. Primary consideration was given to replication of the best-fitting model. Two different types of replication were considered. The first type was replication across random starts; then we considered replication across different types of samples. With regard to replication across random starts, LPA is done by beginning with one randomly selected individual then this case is compared to others in the data set. The process of probability estimation can be influenced by this initial randomly selected individual and a small group of children with similar profiles. This means that the analysis may not produce the same results when repeated even with the same data set and the same model specification. This problem is referred to as the local maxima problem. Thus, the analytic method must control for this issue. The software program that was chosen (MPlus) allows for multiple random starts to the analysis and bases the final categorization on best fit (lowest log-linear value) from all these random starts. Further, the number of replications of this best fit across all random starts is printed. We typically ran a given model using 400 random starts and 200 optimizations. If the best-fitting model replicated at least 20 times, we considered the model fit to be acceptable. After this criterion and several others described below were assessed and considered, we then ran multiple models (from 3 to 9 clusters, typically) on other samples. If similar results occurred, we then considered the model to replicate. (See further explanation below.) For models that replicated across random starts, we then

Appendix C: Statistical Issues

211

considered three criteria to help decide which model provided the most complete specification of the latent structure of the data sets we were studying. These were the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) and the sample size-adjusted Bayesian Information Criterion (BICADJ). These indices are based on the log-linear value, but these are adjusted for sample size and the number of parameters that are estimated. The BICADJ is a more conservative index than the AIC and the BIC. Generally, as models with 3 through 9 clusters were fit to our data, all or some of these criteria would get smaller. Under ideal circumstances, all criteria would get progressively smaller then begin to increase. If this situation occurs, the model with the lowest values would be selected, all other considerations being equal. If the information criteria continue to decrease across all models, as was the case for most of our analyses, these information criterion indexes were plotted across models with increasing numbers of clusters. Typically, what occurs in these plots is that the information criteria decrease sharply as the number of clusters in the model increase up to a point at which time the decrease becomes more gradual. The point at which this change in rate of decrease occurs is referred to as the “elbow” of the plot. This deflection point has been used by some researchers to determine the optimal model [9, 10]. The point after which the slope flattens indicates the optimal model. Several other characteristics of the model fit were considered. The quality of latent class membership classification was examined. Each individual has a probability of being assigned to each cluster and often there is some probability of being assigned to more than one cluster. For example, in a four-cluster model, the individual might have an 80% probability of assignment to cluster 3, a 15% probability of assignment to cluster 1, a 5% probability of assignment to cluster 2, and a 0% probability of placement in cluster 4. The individual is assigned to the cluster with the highest probability. These probabilities are then averaged across all individuals and the mean probability for the cluster is determined. Nagin (reported in Muthen and Muthen [7]) indicates that when the probability of the assignment to the highest probability cluster is greater than .70, the classification is acceptable and we utilized this criterion. In the vast majority of our best-fitting models, classification probabilities were greater than .80. We also considered the index of entropy. This is an overall index of the utility of classification based on classification uncertainty. Only models that had an entropy index greater than .75 were considered in our analyses; this has been described as a high level by other researchers (Clark, reported in Muthen and Muthen [7]). All the models selected as the best fit in our analyses had indices of entropy greater than .80. Finally, we used two ratio tests that are often used in latent class and latent profile analyses to determine if a model had a significantly better fit with the data than the model with one less cluster. These are the Lo, Mendell, and Rubin likelihood ratio and the bootstrap likelihood ratio test. Several researchers have not found these tests to be useful in that the former is often too conservative and the latter too liberal [11]. However, the outcomes of these tests were studied for each model tested.

212

Appendix C: Statistical Issues

Our research was aimed at isolating the most common behavioral trait profiles identified by parents and teachers. Thus, we were not focused on the very low incidence handicapping conditions (e.g., mental retardation) that occur with low frequency in the general population. We, therefore, set a size limit on the smallest cluster of children that we would consider of 2% of the sample. This limit also helped with the replication and meaningfulness of the cluster models because small clusters may occur due to randomly occurring measurement issues or due to local maxima as discussed above. This reduces the probability that they will replicate. In order to find the best model for a set of data, several simplifying assumptions are often used. One is to restrain the model in such a way that the variability of scores around the mean (i.e., the standard deviation) is the same for all clusters in the model. The second simplifying assumption is that the correlation among the variables within each cluster is zero. In fact, both of these assumptions are only approximately true for any given set of data. The researcher has the choice of modeling these parameters (cluster variance, scale correlations within clusters) in addition to the more typical parameters that are modeled (e.g., cluster means). However, as the number of parameters estimated increases, the information criteria increase, which may reduce the number of clusters that can be identified by the model. We examined our data with regard to the standard deviations of indicator variables for each cluster in our best-fitting models and determined that they were similar enough to allow for this assumption. Further, we examined the correlations among the indicator variables for our best-fitting models and found that the correlations were seldom above .25. Therefore, we allowed the simplifying assumption of no correlations among indicators. Before submitting data to LPA, raw scale scores for each sample were centered at the mean of the sample. Thus, the average child rating on a given variable in Russia was given the same score as the average child rating on the same variable in the United States. These centered scores were then transformed to a T-score (M = 50; SD = 10) or a z-score (M = 00.0; SD = 1). The rationale for centering these scores was that this analysis was focused on individual differences and not on culture-­ specific or source-specific (parents vs. teachers) effects on mean levels. Thus, the focus was on variation from the culture-specific mean. Data provided by teachers were somewhat more complicated than that provided by parents because teachers provided ratings of more than one child. This raised the possibility that teachers could have different average scores for all the children they rated; that is, some teachers might rate all their students as more intelligent (on average) than other teachers. We made the simplifying assumptions that students were assigned at random to classes (at least the homeroom classes where the assessments were made) so data were centered at the mean score for each teacher. Again, our emphasis was on individual differences in children, not on differences in teacher biases or perceptions of children. Missing data were handled in two ways. In cases where at least half of the items of the scale were present, missing item scores was estimated from the mean of the remaining items. In cases where more than half of the items were missing from the scale score, the scale score was defined as missing. In model estimation, missing scale scores were estimated by the full information maximum likelihood approach.

Appendix C: Statistical Issues

213

 etermining Similarity of Latent Profile Structures D Across Groups Until recently, there was no well-developed and documented procedure that allowed for the comparison of profile structures of known groups. One of the primary goals of our research was to determine if a similar number of profiles of indicator characteristics could be found in ratings by parents in two cultural settings, and between parents and teachers. A procedure has been published by Morin and colleagues [12] that makes this possible. The processes we used to determine the similarity in LPA structure between US and Russian parents are described. The same procedure was used in comparing ratings by parents and teachers in the US sample. The first step in determining the similarity of latent profile structures for US parents and Russian parent was to create models from 3 through 9 clusters for the two data sets separately. The analysis was done using one simplifying assumption that the correlation of indicators within clusters was zero. Inspection of all model fit criteria and the other criteria established for this study indicated that an eight-cluster model seemed to fit both sets of data best. Morin et al. refer to the finding of the same number of latent clusters across known groups as confirming the configural similarity of the models. Thus, we considered there to be support for the configural similarity of the eight-cluster model. The next step in determining the similarity of the models was to determine the relative fit of the model under the assumption that the means of all indicators for each cluster were different (they were not constrained to be equal) for the US and Russian samples. In this analysis, the samples are combined, and known groups, namely, the two parent groups are specified. Then progressively the equality of means for each cluster was tested to see if it enhanced the fit. This was done for all eight profiles. The model in which all eight profiles were constrained to be equal had the best-fit indices (BIC, BICAdj). Thus, structural (to use Morin et al. terminology) similarity was confirmed. Once structural similarity has been supported, then the model can be tested for similarity of variance, and similarity of size of clusters (probabilities). In both cases, cluster similarity was not achieved. The results of these analyses are illustrated for the parent-teacher and US-Russian parent comparisons in Appendix E. Example of Mplus Input Code for the Eight-Cluster Solution The following is an example of the code used to do multiple group analyses in MPlus of LPA models. In this example, comparisons are made between the Russian and US parent samples for the eight-cluster model. For other explanations of this procedure see Supplements for Similarity in Latent Profile Solutions by Morin et al. [12]. Title: This is an example of latent profile analysis with continuous indicators ages 8–12, US and Russian parents; Data: File is “c:\Users\rpmartin\Desktop\Sidca Raw Data Final2\USRuss m ­ erge\ usruss.dat”;

214

Appendix C: Statistical Issues

Format is f12.0 7f8.0 16f8.2; Variable: Names are id g school grade sex age source ethnic intopenr achr posconr actr negantsr disdisr inhsocwr insr intopent acht poscont actt negantt disdist inhsocwt inst; Note: The data set contained scores on the eight indicators in raw and T-score form. The variable g is the grouping variable identifying both samples. In addition, data are available for several identifying and demographic variables. Missing = all (99); Note: Statement indicates that missing data was coded as 99. Usevariables intopent acht poscont actt negantst disdist inhsocwt inst; Note: In this analysis, the T-score variables were used as indicator. knownclass = cg (g = 1 g = 2); Classes = cg(2) c(8); Note: In the above two line, g is the number of known classes, with the US parent group set to 1 and the Russian set to 2. The next line indicates that 16 profile clusters will be modeled, with 8 clusters for each known class. Analysis: processor = 3; Note: This statement speeds up the analysis time by using three processors simultaneously. Type = Mixture; Starts = 800 400; stiterations = 400; Note: In this example, 800 random startswere specified as were 400 optimizations and 400 iterations Model: %overall% Note: This overall statement indicates that the class sizes are freely estimated in the two samples. Only seven statements are given below in this eight-cluster model since only k-1 statements are required. c#1 on cg#1; c#2 on cg#1; c#3 on cg#1; c#4 on cg#1; c#5 on cg#1; c#6 on cg#1; c#7 on cg#1; Note: Class-specific statements below are defined followed by parameter (means, variances) estimates of the indicator variables. Means are indicated by brackets ([

Appendix C: Statistical Issues

215

]), with no symbol in front of variances. Values in parentheses indicate parameters that are estimated to be equal (m1-m8) when compared across groups. The analysisillustrated here is for the means to be freely estimated but the variances are constrained to be equal across profiles and countries. %cg#1.c#1% [intopent acht poscont actt negantst disdist inhsocwt inst] (m1-m8); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#1.c#2% [intopent acht poscont actt negantst disdist inhsocwt inst] (m9-m16); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#1.c#3% [intopent acht poscont actt negantst disdist inhsocwt inst] (m17-m24); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#1.c#4% [intopent acht poscont actt negantst disdist inhsocwt inst] (m25-m32); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#1.c#5% [intopent acht poscont actt negantst disdist inhsocwt inst] (m33-m40); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#1.c#6% [intopent acht poscont actt negantst disdist inhsocwt inst] (m41-m48); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#1.c#7% [intopent acht poscont actt negantst disdist inhsocwt inst] (m49-m56); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#1.c#8% [intopent acht poscont actt negantst disdist inhsocwt inst] (m57-m64); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#2.c#1% [intopent acht poscont actt negantst disdist inhsocwt inst] (m25-m32); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); Note: This statement indicates that means for clusters 25 through 32 of the US sample (cg#1.c#4) are constrained to the equal the Russian sample for (cg#2.c#1).

216

Appendix C: Statistical Issues

%cg#2.c#2% [intopent acht poscont actt negantst disdist inhsocwt inst] (m33-m40); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#2.c#3% [intopent acht poscont actt negantst disdist inhsocwt inst] (m17-m24); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#2.c#4% [intopent acht poscont actt negantst disdist inhsocwt inst] (m9-m16); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#2.c#5% [intopent acht poscont actt negantst disdist inhsocwt inst] (m1-m8); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#2.c#6% [intopent acht poscont actt negantst disdist inhsocwt inst] (m41-m48); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#2.c#7% [intopent acht poscont actt negantst disdist inhsocwt inst] (m49-m56); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); %cg#2.c#8% [intopent acht poscont actt negantst disdist inhsocwt inst] (m57-m64); intopent acht poscont actt negantst disdist inhsocwt inst (v1-v8); Output: sampstat cinterval svalues tech1 tech7;

 ppendix D: Profiles of Child Behavior: A Brief A Review

Categorization Based on Theoretical or Clinical Judgement Physicians, psychologists, and other clinicians have often created categories of child behavior. For most of the history of child psychology, these behavior categories have been based primarily on clinical judgment and have focused on categories of psychopathology. The diagnostic systems documented in the International Classification of Disease maintained by the World Health Organization and the Diagnostic and Statistical Manual of Mental Disorders by the American Psychiatric Association are systematizations of these processes. In recent decades, these clinician-­based judgments have been augmented by a variety of empirical research efforts, but the foundation of these categories is based on clinical judgment. The focus of the current research effort is on documenting categories of behavior that lie primarily in the normal range of behavior; that is, those individual differences that can be observed in nonclinical samples. One of the most influential studies of this type was the New  York Longitudinal Study [13, 14]. The researchers measured nine temperamental characteristics of 141 infants and preschoolers then followed these children into early adulthood. The behavioral characteristics measured were as follows: Activity Level, Rhythmicity (regularity of bodily functions such as sleep, feeding, and elimination schedules); Approach/Withdrawal (the nature of the initial response to new stimuli), Adaptability (speed and ease of adapting to change); Threshold of Responsiveness (intensity level of stimulation that is necessary to evoke a response); Intensity of Reaction (the energy level of a response), Mood (amount of pleasant, joyful behavior as contrasted with unpleasant, crying, or unfriendly behavior); Distractibility; and Attention Span/Persistence (length of time a particular activity is pursued and the continuation of that activity in the face of obstacles).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4

217

218

Appendix D: Profiles of Child Behavior: A Brief Review

They then created three behavioral profiles from these nine temperamental characteristics based primarily on clinical judgment: The Easy Child (40% of the sample), The Difficult Child (10%), and The Slow-To-Warm-Up Child (15%). The Easy Child group was biologically regular, approached rather than avoided new stimuli and people, adapted easily to change, often exhibited a positive mood and expressed emotions with moderate intensity. The Slow-To-Warm-Up Group was distinguished by a combination of negative responses of mild intensity to new stimuli and slow adaptation to new stimuli and people. The Difficult Group was biologically irregular, had a negative withdrawal response to new stimuli, was very slow to adapt to change, was emotionally intense, and was often in a negative mood. While these conceptual groupings were mutually exclusive, the three groups did not encompass all the children in their sample. Approximately 35% of their sample did not meet criteria for inclusion in one of these three groups. The profiles were  created when the children were 3  years of age. Follow-up analyses found that children in the Difficult and Slow-to-Warm-Up groups required more support from family, school personnel, and mental health workers in order to develop appropriately than did children in the Easy group. Further, as development proceeded, children in the two higher risk groups had higher rates of mental illness.

Categorization Based on the Q-Sort Method Block and Block [15, 16] developed a measurement method in which parents, teachers, or clinicians would sort cards that were descriptive of a specific behavior into categories based on how characteristic the behavior was of the child. This measurement method has been used by a number of other researchers [17–19] . The study by Asendorpf and van Aken will be summarized to illustrate this method. The researchers utilized a modified version of this procedure adapted to German on a sample of children initially assessed at ages 4 through 6. Specifically, at the end of each school year, the child’s main preschool teacher sorted cards with 54 adjectival descriptors into 9 categories (1  =  extremely uncharacteristic, 9 = extremely characteristic). Teachers were instructed to sort exactly 6 items into each of the nine categories. At age 10, mothers provided data using the same procedure. Sorts at ages 4 through 6 were averaged, then factor analyzed. A factor solution produced a resilience factor, an over-controlled factor, and an under-controlled factor. Children were assigned to a profile based on their factor scores and discriminant analysis. Characteristics of the three groups were as follows: (1) Resilient children (49% of the preschool sample, 52% of the 10-year-old sample) were described as attentive, competent, skillful, curious, exploring and planful. These children also had low scores on cries easily, immature, and anxiety-related descriptors. Girls were somewhat over-represented in this group. (2) Over-controlled children (21% of the preschool sample, 28% of the 10-year-old sample) were described as considerate of others, helpful and cooperative, obedient, neat, and orderly. They were sought out by others, had low scores on aggression, were confident, and liked to compete. Boys

Appendix D: Profiles of Child Behavior: A Brief Review

219

and girls were equally represented in this group. (3) Under-controlled children (31% of preschool sample, 19% of 10 year-old sample) were described as energetic, lively, and they expressed negative feelings directly. They were also described as stubborn. Boys were over represented in this group.

Categorization Based on Traditional Cluster Analysis Although there are a variety of techniques for creating clusters of children based on cluster analysis, the techniques referred to here as traditional cluster analysis utilize a distance metric to create relatively homogeneous group of children based on a number characteristics. Scale scores were created using a variety of methods. One clustering method uses Euclidian distance (the square root of the sum of the squared difference between indicator variables). One widely used approach using this method is Ward’s agglomerative procedure, which begins by isolating the two children with the most similar scores (smallest Euclidian distance) on all behavioral characteristics, then finds the next two children with the most similar scores, and continues this process until a small designated number of clusters has been obtained. A number of researchers have used these techniques [20–23]. One of the most influential studies of this type is the Dunedin Multidisciplinary Health and Development Study. A complete birth cohort born during 1 year in Dunedin, New Zealand, was intensively studied. This cohort was reassessed at ages 3, 5, 7, 9, 13, 15, 18, and 21 years, and at this writing continues to be studied. At age 3, 1023 children participated in a 90-minute testing session involving cognitive and motor tasks. Following the testing, the examiner rated the child’s behavior on 22 behavioral characteristics. When these 22 characteristics were submitted to a traditional distance-based cluster analysis, five clusters of 3-year-old children were identified. The characteristics of these groups were as follows: (1) Confident children (27.5% of the sample; 52% male) were zealous, eager to explore the assessment materials, and adjusted quickly to the assessment situation. (2) Well-adjusted children (39.6% of the sample; 48% male) had scores on all behavior scales that were within normal limits. (3) Reserved children (14.8% of the sample; 48% male) were timid and somewhat uncomfortable in the testing session, although this response was not extreme and did not interfere with the assessment. (4) Inhibited children (7.8% of the sample; 40% male) exhibited extreme social reticence and fearfulness, and had difficulty concentrating on tasks. (5) Under-controlled children (10.4% of the sample; 62% male) were described as irritable, impulsive, and were not persistent in solving problems. They had difficulty sitting still and had large changes in mood. At age 18, adolescents who were in the Under-Controlled group at age 3 described themselves as danger-seeking, impulsive, prone to respond with strong negative emotions to everyday events, and enmeshed in adversarial relationships. Adolescents who were in the Inhibited group described themselves as over-controlled, harm-­ avoidant, and nonassertive. Adolescents who had been in the Well-Adjusted,

220

Appendix D: Profiles of Child Behavior: A Brief Review

Confident, and Reserved child groups at age 3 were all found to have normal or average behavioral tendencies. By age 21, young adults who had been placed in the Inhibited and Under-­ Controlled groups at age 3 were more likely to have a psychiatric disorder than persons who had been placed in the other three groups. Further, the psychiatric disorders resulted in more impairment. They were also more likely to have been diagnosed as having multiple disorders. The Under-Controlled group was 2.9 times as likely to be diagnosed with antisocial personality disorder, 2.2 times as likely to be recidivistic offenders, and 4.5 times as likely to be convicted for a violent offense. Also, males who had been in the Under-Controlled group at age 3 were more likely to have been convicted of violent offenses [22].

Categorization Based on Behavior Problems A sizeable body of research has focus on clustering children based on parent or teacher responses to measures of problematic behavior. The measurements used in these studies lie somewhere between normal or typical child behavior (e.g., temperamental or personality variation in community samples) and psychopathology (e.g., measures of known clinical syndromes). While often conceptualized as measures of symptoms of psychopathology, behavior problem measures also share items with temperamental measure. Kamphaus and colleagues [24–26] published a series of studies aimed at developing a typology of child behavior using measurement tools designed primarily to assess behavior problems. The measurement instrument used was the Behavior Assessment System for Children (BASC), which is now the most widely used measure of behavior problems in the public schools of the United States. This measure also includes scales that measure more positive attributes such as leadership and social skills. One study was based on the US normative sample for the Parent Rating Scales of the BASC [26]. The sample included 2029 children ages 6–11 rated by parents residing in 116 sites representing different regions of the United States. The scales used as indicator characteristics were as follows: Aggression (verbal or physical), Conduct Problems (antisocial, rule-breaking behavior), Hyperactivity, Anxiety, Depression, Somatization (tendency to be overly sensitive to minor physical problems), Attention Problems, Atypicality (behaviors that are immature, odd, or associated with psychosis), and Withdrawal (tendency to avoid social contact). These scales were designated as clinical scales. Three other scales were used that were designated as Positive Adaptive Skills. These included Adaptability (ability to adapt readily to changes in the environment), Leadership (ability to work with others to accomplish goals), and Social Skills (skills necessary for interacting with peers and adults). This research team used a two-step distance-based clustering procedure: Ward’s clustering followed by the K-means procedure. Based on theoretical and clinical

Appendix D: Profiles of Child Behavior: A Brief Review

221

issues, a nine-cluster model was selected. (1) The Well-Adapted Group (15% of the sample; 57% female) was characterized by high scores on all adaptive scales (Leadership, Social Skills, and Adaptability) and low scores on all clinical scales. (2) The Adapted Group (11% of the sample; 53% male) had high scores on the adaptive scales (particularly Leadership and Social Skills), average scores on most clinical scales, and a low score on Attention Problems. (3) The Minimal Problems Group (17% of the sample; 57% female) had low scores on all clinical measures and were average on adaptive scales. (4) The Average Group (25% of the sample; 62% male)—had near average scores on all characteristics, with a low score on Withdrawal, indicating they were socially integrated. (5) The Physical Complaints/ Worry Group (3% of the sample; 56% female) had elevations on Anxiety and Somatization, but in all other respects these children were average. (6) The Internalizing Problems Group (9% of the sample; 63% female) had average scores on most scales but significant elevations on Anxiety, Depression, and Social Withdrawal. (7) The Attention Problem Group (8% of the sample; 58% male) had average scores on all clinical scales with two exceptions: they had an elevated score on Attention Problems and were characterized by low scores on all three of the adaptive scales. (8) The Disruptive Behavior Problem Group (9% of the sample; 74% male) had very high scores on the Aggression, Hyperactivity, and Conduct Problems scales, and also had important elevations on the Attention Problems and Depression scales. (9) The General Psychopathology-Severe Group (3% of the sample; 59% male) had very high scores on Aggression, Hyperactivity, Conduct Problems, Depression, Attention Problems, and Atypicality and very low scores on the adaptive scales. Kamphaus et al. had access to data on the number of children in each cluster who had previously been diagnosed with behavioral, emotional, or academic problems. For most of the clusters, 5% or less had a diagnosis. Four clusters had higher levels of prior diagnoses: The Disruptive Behavior Problem Group (9%), The Physical Complaints and Worries Group (10%), the Attention Problem Group (13%), and the General Psychopathology-Severe Group (27%). A similar analysis strategy based on measures of preschool readiness was used by Konold and Pianta [27] and Sabol and Pianta [28]. The same procedures were used to cluster the normative teacher sample for the BASC (Kamphaus et al. 1997). The sample consisted of 1227 children ages 6–11. The instrument was the teacher form (TRS-C) of the BASC, which had the same scales as the parent form with two exceptions: on the teacher form, there was a scale for study skills (organizational abilities and study habits) and another scale, which measured learning problems (difficulties understanding and completing school work). Seven clusters were isolated: (1) Well Adapted (34% of the sample, 61% female); (2) Average (19% of the sample, 52% female); (3) Mildly Disruptive (12% of the sample, 70% male); (4) Disruptive Behavior Disorder (8% of the sample, 78% male); (5) Learning Problems (12% of the sample, 60% male); (6) Physical Complaints/Worries (11% of the sample, 60% female); and (7) Severe Psychopathology (4% of the sample, 67% male). Similar results were obtained from

222

Appendix D: Profiles of Child Behavior: A Brief Review

a recent investigation of teacher-based measurements using the BASC by Orpinas et al. [29]. Even though fewer clusters could be reliably isolated from teacher-rated data than from parental data, there was striking similarity in the patterns of behaviors across raters. The biggest differences were that no internalizing problem group could be isolated from teacher data as well as no minimal problem group and no adaptive group was isolated. On the other hand, teacher ratings produced a learning problems group that was not found in parent ratings. The results of the research by the Kamphaus group are similar to those of Curry and Thompson [30], Beg, Casey, and Saunders [31], and Bulotsky-Shearer, Fantuzzo, and McDermot [32], although the latter two groups studied more restricted samples (e.g., preschoolers and low-income children, respectively).

Categorization Based on Latent Profile Analysis (LPA) A recent paper by Scott et al. [33] identified profiles of a large sample of twin pairs (787; 7- and 8-year-olds) using mother- and father-reported temperament. Using latent profile analysis procedures, they found four profile types: (1) regulated, typical reactive (33.9% of the sample), (2) well-regulated, positive reactive (30.7%), (3) regulated, surgent (19.0%), and (4) dysregulated, negative reactive (16.4%). All profiles were found to be heritable. This twin sample consisted of a mixture of monozygotic and dizygotic twins, allowing for estimates of heritability. Heritability ranged from .28 for the “regulated, typical reactive” cluster to .91 for the “well-­ regulated, positive reactive” cluster. Beekman [34] developed a profile using LPA procedures from temperamental characteristics of 561 children who were adopted in early infancy from four regions in the United States. Adoptive parents (ratings of mothers and fathers were averaged) rated temperamental characteristics at 9, 18, and 27 months. Characteristics assessed in toddlerhood were Activity level, Anger, Fear, Interest, and Pleasure with related characteristics assessed in infancy. Four profiles were found at both 18 and 27 months, while five were isolated in infancy. We will consider here only the four profiles obtained both at 18 and 27 months. They were labeled (1) Positive Reactive (slightly below average levels of activity, anger and fear, with slightly above average levels of interest and pleasure); (2) Negative Reactive (slightly above average levels of activity, anger, and fearfulness, with below average interest and pleasure); (3) Fearful (low activity level, low anger, high average fear, low interest and pleasure); and (4) Active Reactive (very high levels of activity and above average levels of anger and pleasure; low average fear and interest). One of the strongest aspects of this study was the investigation of the stability of the profiles obtained when the children were in infancy and later in toddlerhood. Results indicated that for Positive Reactive infants 78% retained that classification at 27 months. The Negatively Reactive cluster also showed significant stability with 51% retaining their classification with a similar pattern occurring for the infants in

Appendix D: Profiles of Child Behavior: A Brief Review

223

the Fearful Profile (55% retained the same profile in toddlerhood). However, most children in the Active Reactive Profile at age 18 months transitioned to three other profiles at 27  months (Positive Reactive—34%, Negative Reactive—37%, and Active Reactive—29%. Of note, not one child who was classified as Active Reactive at age 18 months transitioned in to the Fearful profile at age 27 months.

Our Expectations Based on This Review Regardless of the method used to isolate homogenous clusters and the age of the children participating in the research, there are many striking similarities in the outcomes of these research efforts. From all analyses, a well-adjusted, over-­ controlled, and under-controlled group was observed, although in many studies these broad categories were subdivided into more precisely defined groups. Further, in most studies a group defined by average scores was isolated, which seem to be a useful differentiation from children with exceptional adaptation and those with more adjustment problems. Based on these results, using an item pool that focuses on temperamental and personality characteristics in children ages 8–12, we expect to find that the model that fit our data best  would be somewhere between 5 and 9 clusters of children. Conceptually, these profiles are expected to isolate four broad groups: well adapted and socially/emotionally skilled children; children who are near average scores on all measures; inhibited/fearful children with average scores on most other measures; and poorly self-regulated children who have difficulty controlling attention or controlling their expression of negative emotions and antagonistic behaviors. Since more than four clusters are expected to be isolated, some of these groups will be thought of as subtypes of the four major types.

 ppendix E: Results of the Latent Profile A Analyses

US Parent-Rated Sample One of the goals of the research was to determine if the profiles produced by the latent profile analysis (LPA) of ratings of children in middle childhood by parents in the United States, parents in Russia, and teachers in the United States were comparable. Before comparing samples, the best-fitting model was determined for each sample separately. The first step in this procedure was to determine the best-fitting model for the US parent sample considered separately. Models were tested for three through nine clusters to determine the model that most parsimoniously fit the parental ratings of child behavioral traits. Table E.1 presents the results of these analyses. Focusing on the three information criteria studied (Akaike Information Criterion, Bayesian Information Criterion—BIC, Bayesian Information Criterion adjusted for sample size, BICADJ), all three decreased with each subsequent model indicating that each model had progressively better fit with the data. We also inspected the entropy index. This is an index of the quality of the cluster separation and the eight-cluster model had the highest index value. All values above .80 are considered good. In addition, we set two other criteria. The first was that we would eliminate from consideration those models in which the smallest cluster included less than 2.0% of the sample. Again, the eight-cluster model met this criterion. We then considered the cluster with the lowest classification probability in each model. We set a lower limit of .70 for this value, and all models met this criterion. Two ratio tests are often used in latent class and latent profile analyses to determine if a model had a significantly better fit with the data than the model with one less cluster. The Lo, Mendell, and Rubin likelihood ratio test indicated that the model with three clusters was significantly better than the model with two clusters. However, all other models made a nonsignificant improvement according to this test. In addition, we calculated the bootstrap likelihood ratio test. This test indicated

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4

225

226

Appendix E: Results of the Latent Profile Analyses

that each successive model was a significant improvement over the previous one. These results indicate that neither index was helpful in deciding on the best model. This result has been found by other researchers [35] (e.g., Orpinas, Raczynski, Bandalos, Peters, & Colman, 2014). The final method used to determine the best fit was to plot the information criteria for all models and look for an elbow in the graph. While the Akaike criterion seemed to level out around the eight-cluster model, the BIC and BICADJ were less conclusive (see Fig. E.1). Based on all these considerations as well as the fit of the model to theoretically meaningful categories of child behavior, the eight-cluster model was chosen. Table E.2 presents the mean for each of the eight indicator variables for the eight-cluster model. Clusters are given a label of a number followed by a letter. The number corresponds to the broad category, and the letter indicates a profile type within that category. It should be noted that standard deviations around the means were assumed to be the same for each profile. These data are summarized in Table E.9. Table E.2 presents the indicator means for each cluster of the eight-cluster model. Children exhibiting profile 1a were labeled Well-Adjusted High Achievers. Two clusters of children with scores predominantly in the average range were isolated. Those in cluster 2a are referred to as High Average Self-Regulators while those in cluster 2b as Low Average Self-Regulators. Two profiles emerged describing children who were significantly more socially withdrawn than their peers. Children in cluster 3a are referred to as Withdrawn High Achievers and cluster 3b as Withdrawn Lower Achiever. Finally, three clusters of children were identified who exhibited high levels of distractibility, high activity levels, and/or had a tendency to exhibited negative emotionality, antagonism, and argumentative behavior. Children in cluster 4a are referred to as Poorly Regulated Higher Ability Under-Achievers, and those in cluster 4b as Poorly Self-Regulated Lower Achievers. Children in cluster 4c are referred to as Poorly Self-Regulated High Achievers.

Russian Parent Sample After determine the best-fitting model for the US parent sample, we applied the same procedures and the same statistical criteria to the Russian parent sample. The sample analyzed consisted of 538 children from central Russia in the age range 8–12. Table E.3 presents the results of these analyses. Focusing on the three information criteria studied (AIC, BIC, BICADJ), all three decreased with each subsequent model indicating that each model had progressively better fit with the data. We also inspected the entropy index. This is an index of the quality of the cluster separation, and the eight- and nine-cluster models had the highest index value. In addition, we set two other criteria. The first was that we would eliminate from consideration those models in which the smallest cluster included less than 2.0% of the sample. In this analysis, this would eliminate all models with more than six clusters. We then

Appendix E: Results of the Latent Profile Analyses

227

considered the cluster with the lowest classification probability in each model. We set a lower limit of .70 for this value and all models met this criterion. The Lo, Mendell, and Rubin likelihood ratio test indicated that the model with four clusters was marginally better than the model with three clusters although it not significantly better (p  2.0 is good i LMR = Lo, Mendell, and Rubin likelihood ratio test j BLRT = bootstrap likelihood ratio test a

Appendix E: Results of the Latent Profile Analyses

231

Table E.2  Indicator mean scores for the eight-cluster modela: US parent sample Indicator Intelligence/Openness Achievement Orientation Prosocial Emotion Activity Level Irritability/Antagonism Distractibility/Disorganization Inhibition/Social Withdrawal Insecure/Fearful  Cluster Sized

Clusters 1ab 61.4c 61.5 63.5 60.5 37.9 37.4 37.1 36.7 10.0%

2a 52.7 51.9 54.0 50.8 45.5 46.0 47.0 45.6 32.6

2b 44.1 46.7 43.5 45.0 53.0 51.9 56.2 55.0 37.1

3a 61.3 63.8 55.5 40.6 43.4 41.1 64.1 54.9 2.9

3b 34.5 35.7 33.0 39.1 66.2 63.9 68.9 65.1 2.7

4a 54.1 40.4 56.2 58.1 51.4 62.5 38.3 43.6 4.2

4b 44.3 38.1 41.7 56.6 61.1 66.7 47.4 54.0 6.0

4c 59.3 62.3 56.5 58.4 62.7 54.5 43.2 58.1 4.6

n = 912 Indicators were 8 scale scores based on the Deal short form scoring. Means calculated from data (not estimates) b Profile labels: 1a—Well-Adjusted High Achievers; 2a—High Average Self-Regulators; 2b—Low Average Self-Regulators; 3a—Withdrawn High Achievers; 3b—Withdrawn Lower Achiever; 4a— Poorly Regulated Higher Ability Under-Achievers; 4b—Poorly Self-Regulated Lower Achievers; 4c—Poorly Self-Regulated High Achievers c Underlined means are more than .5 standard deviations below the mean (< 30 percentile); bold means are more .5 standard deviations above the mean (> 70 percentile). These are presented simply to aid in visualizing cluster differences d The percentage of children in the sample who exhibited the profile a

Table E.3  Change in latent profile indices from three- to nine-cluster models: Russian parent-­ rated samplea Model 3 clusters 4 clusters 5 clusters 6 clusters 7 clusters 8 clusters 9 clusters

Nparb 34 43 52 61 70 79 88

AICc 31096 30942 30851 30782 30724 30681 30638

BICd 31241 31126 31074 31044 31024 31019 31015

BICSAe 31133 30990 30909 30850 30802 30769 30736

Entf LCPg .75 .87 .77 .84 .81 .80 .82 .81 .83 .82 .83 .78 .84 .78

SmallCSh 18.2% 11.9% 3.7% 3.5% 1.5% 1.5% 1.3%

LMRi 0.09 0.06 0.20 0.56 0.28 0.43 0.46

BLRTj .0001 .0001 .0001 .0001 .0001 .0001 .0001

N = 538 Indicators were 8 scale scores based on the Deal short form scoring b Npar = number of free parameters that are estimated in the model c AIC = Akaike Information Criterion d BIC = Bayesian Information Criterion e BICSA = Bayesian Information Criterion adjusted for sample size f Ent = Entropy: An index of cluster separation; ≥.80 is good g LCP = Of all the clusters in the model, the lowest classification probability; ≥.70 is good h SmallCS = Percentage of the sample in the smallest cluster; > 2.0 is good i LMR = Lo, Mendell, and Rubin likelihood ratio test j BLRT = bootstrap likelihood ratio test a

Appendix E: Results of the Latent Profile Analyses

232

Table E.4  Indicator mean scores for the eight-cluster model: Russian parent samplea Indicator Intelligence/Openness Achievement Orientation Prosocial Emotion Activity Level Irritability/Antagonism Distractibility/Disorganization Inhibition/Social Withdrawal Insecure/Fearful  Cluster Sizec

Clusters 1ad 62.1b 60.7 62.5 60.8 43.1 40.5 35.5 38.1 11.3

1b 57.8 68.5 63.1 50.9 37.7 33.1 52.3 46.8 5.8

2a 51.1 52.8 50.0 48.2 44.6 45.1 48.7 46.9 22.3

2b 44.0 43.9 44.4 46.0 53.7 54.1 54.2 53.0 40.1

3a 55.7 58.0 56.9 44.1 49.1 54.6 63.5 61.5 5.0

3b 32.2 33.5 34.7 36.2 59.1 64.7 62.5 60.4 3.7

4a 52.6 32.9 42.8 66.9 69.3 66.8 38.4 39.3 1.5

4c 56.5 50.1 53.7 62.7 56.0 54.7 41.9 50.9 10.2

n = 538  Indicators were eight scales in T-score form; means calculated from data (not estimates) b Underlined means are more than .5 standard deviations below the mean ( 70 percentile). These are presented simply to aid in visualizing cluster differences c The percentage of children in the sample who exhibited the profile d Profile labels: 1a—Well-Adjusted High Achievers; 2a—High Average Self-Regulators; 2b—Low Average Self-Regulators; 3a—Withdrawn High Achievers; 3b—Withdrawn Lower Achiever; 4a— Poorly Regulated Higher Ability Under-Achievers; 4b—Poorly Self-Regulated Lower Achievers; 4c—Poorly Self-Regulated High Achievers a

Table E.5  Change in latent profile indices from three- to nine-cluster models: US teacher samplea Model 3 clusters 4 clusters 5 clusters 6 clusters 7 clusters 8 clusters 9 clusters

LLb −25515 −25372 −25266 −25184 −25100 −25016 −24963

Nparc 34 43 52 61 70 79 88

Rep.d 400 280 66 97 87 146 66

AICe 51099 50830 50636 50489 50340 50189 50102

BICf 51263 51037 50886 50783 50677 50569 50526

BICAg 51155 50900 50721 50589 50455 50318 50246

Ent.h .88 .85 .85 .84 .85 .86 .85

LCPi .92 .87 .79 .76 .78 .79 .78

Smallj 24.1 17.3 8.1 6.2 5.1 3.1 3.3

n = 913 a Data were centered for each classroom; were in T-score form calculated from Deal et al. items b LL = Log likelihood c Npar = number of free parameters that are estimated in the model d Replications across 400 random starts e AIC = Akaike Information Criterion f BIC = Bayesian Information Criterion g BICA = Bayesian Information Criterion (adjusted for sample size) h Ent. = Entropy, an index of cluster separation; ≥.80 is good i Of all the clusters in the model, the one with lowest classification probability; ≥.70 is good j Cluster with the smallest size; >2.0 is good

Appendix E: Results of the Latent Profile Analyses

233

Table E.6  Indicator mean scores for the eight-cluster model: US teacher samplea Indicator Intelligence/Openness Achievement Orientation Prosocial Emotion Activity Level Irritability/Antagonism Distractibility/Disorganization Inhibition/Social Withdrawal Insecure/Fearful  Cluster Sizec

Clusters 2a 1ad 63.2b 53.2 63.8 54.2 61.7 54.8 55.7 52.2 41.1 46.2 35.8 46.7 37.7 46.0 38.2 46.7 15.7 20.6

2b 45.6 46.3 47.8 45.5 50.6 53.5 55.3 55.1 32.2

3a 59.0 62.6 56.4 41.4 41.6 38.3 56.1 52.3 5.8

3b 35.1 36.4 37.1 37.2 56.9 60.8 64.5 61.9 6.7

4a 49.4 42.8 50.2 65.3 54.9 60.8 40.3 48.4 4.8

4b 37.7b 35.3 39.3 54.3 61.6 63.3 51.3 52.4 8.9

4c 60.1 62.5 56.1 59.4 61.3 54.5 45.2 60.8 3.3

n = 913 Indicators were eight scales in T-score form; means calculated from data (not estimates) b Underlined means are more than .5 standard deviations below the mean (< 30 percentile); bold means are more .5 standard deviations above the mean (> 70 percentile). These are presented simply to aid in visualizing cluster differences c The percentage of children in the sample who exhibited the profile d Profile labels: 1a—Well-Adjusted High Achievers; 2a—High Average Self-Regulators; 2b—Low Average Self-Regulators; 3a—Withdrawn High Achievers; 3b—Withdrawn Lower Achiever; 4a— Poorly Regulated Higher Ability Under-Achievers; 4b—Poorly Self-Regulated Lower Achievers; 4c—Poorly Self-Regulated High Achievers a

Table E.7  Fit results for the eight-cluster model: US parents, teachers, and Russian parent combined (universal) sample Model Three groups means freeg US parent & teacher means equalg US teacher & Russian parent means equalg US & Russian parent means equalg All three groups equal meang All three groups equal means  Variance & proportions

LLa −68806 −69017 −68929 −68883 −69101

nparb 239 175 175 175 111

−69293 81

AICc 138091 138385 138209 138117 138423

BICd 139469 139394 139218 139126 139063

BICAe 138710 138838 138662 138570 138711

138750 139217 138960 .885

A combined sample of US parent, teacher, and Russian parent ratings; n = 2359 b LL = model log likelihood c npar = number of free parameters estimated d AIC = Akaike Information Criterion e BIC = Bayesian Information Criterion f BICA = Bayesian Information Criterion adjusted for sample size g Profile variances and probabilities (proportions) are free across groups a

Entropyf .895 .889 .896 .895 .890

234

Appendix E: Results of the Latent Profile Analyses

Table E.8  Indicator mean scores for the eight-cluster model: universal samplea Indicator Intelligence/Opennessc Achievement Orientation Prosocial Emotion Activity Level Irritability/Antagonism Distractibility/Disorganized Inhibition/Social Withdrawn Insecure/Fearful  Cluster Sized

Clusters 1ab 63.0c 63.5 62.9 58.2 40.2 36.5 37.2 37.8 12.1

2a 54.1 54.0 54.4 51.7 45.4 45.1 46.3 45.4 23.4

2b 45.4 46.6 45.8 45.8 51.8 52.5 54.8 53.8 41.0

3a 57.7 64.2 57.1 41.1 42.2 39.7 60.8 55.4 3.9

3b 35.2c 36.2 35.3 37.3 59.8 61.5 65.3 62.2 5.1

4a 53.6 44.5 51.6 62.9 55.3 58.9 39.8 46.6 6.2

4b 40.4 35.4 40.4 56.1 61.8 64.5 49.4 52.2 6.3

4c 60.7 63.4 55.2 58.8 62.5 55.1 45.9 60.7 1.9

A combined sample of US parent, teacher, and Russian parent ratings; n = 2359 Profile labels: 1a—Well-Adjusted High Achievers; 2a—High Average Self-Regulators; 2b—Low Average Self-Regulators; 3a—Withdrawn High Achievers; 3b—Withdrawn Lower Achiever; 4a— Poorly Regulated Higher Ability Under-Achievers; 4b—Poorly Self-Regulated Lower Achievers; 4c—Poorly Self-Regulated High Achievers c Scores are expressed in T-score form (mean = 50; standard deviation = 10). Underlined means are more than .5 standard deviations below the mean (70 percentile). These are presented simply to aid in visualizing cluster differences d The percentage of children in the sample who exhibited the profile a

b

Table E.9  Standard deviations of profile means for the eight-cluster model: all samples Indicator Intelligence/Openness Achievement Orientation Prosocial Emotion Activity Level Irritability/Antagonism Distractibility/Disorganized Inhibition/Social Withdrawn Insecure/Fearful

US parents 0.74 0.73 0.67 0.81 0.74 0.70 0.67 0.75

Teachers 0.45 0.38 0.51 0.65 0.54 0.41 0.53 0.60

Russian parents 0.42 0.35 0.51 0.66 0.55 0.38 0.47 0.51

Universal sample 0.67 0.55 0.70 0.75 0.74 0.59 0.63 0.72

Note: These are estimated standard deviations (in z-score form) under the assumption that all eight profiles in each sample had the same variance

 ppendix F: Academic Ability and Achievement A Motivation

Academic Grade Point Average The question addressed in the following analyses was: Is there a meaningful relationship between profiles generated from parents and teachers, and measures of academic ability as well as achievement motivation? If parent- and teacher-based assessments of behavioral characteristics play a meaningful role in the academic life of the child, then there should be different outcomes for children depending on their behavioral profile. The question was tested first for academic grade point average for data collected by Slobodskaya and colleagues. Parents reported the grade point average for academic subjects as reported on the last report from the school (Table F.1; Fig. F.1). An analysis of variance with behavioral profile (generated from the universal sample) and child gender serving as the independent variables and grade point average serving as the dependent variable resulted in significant effects for both profiles and gender. Children in the Well-Adjusted, High Average Self-Regulator, and Withdrawn High Achievers had higher grade point averages than children in the Low Average Self-Regulator, Withdrawn Lower Achiever, Poorly Self-Regulated Lower Achiever clusters (see Table F.2). Girls consistently (across profile type) performed academically at a higher level than boys.

Peer Perception Peer perception data from students in US schools (Georgia) regarding academic ability and motivation were collected by Lease and colleagues. Since there with two cohorts of students providing these data who had somewhat different demographic characteristics (see Appendix A), indicators of academic ability and achievement motivation were first entered into a univariate linear models analysis of variance in © The Author(s), under exclusive license to Springer Nature Switzerland AG 2020 R. P. Martin et al., Temperament and Children, https://doi.org/10.1007/978-3-030-62208-4

235

236

Appendix F: Academic Ability and Achievement Motivation

which profile assignment and cohort were treated as independent variables in order to determine if there was a significant cohort effect or a cohort by profile effect. No cohort effect was found, so the data were combined and the analysis of variance treated profile type and child sex as independent variables. In order to control for increase probabilities of finding significant outcomes due to chance (because a large number of independent variables were analyzed), the alpha level was set at .01. The Gabriel post hoc procedure was used to determine separation between profile types summarized in this report by reporting homogeneous subsets of profiles for each outcome. One item made up this measure academic ability and a separate item was the measure of achievement motivation. The item was responded to by each member of a classroom in the form of nominations of children who fit the description. As many children can be nominated as the student feels meets the description set by the item. Scores were standardized for each classroom studied. Thus, the mean score for each classroom was set at 00.0, with a standard deviation of 1.00. 1 . This person makes good grades, is smart, and usually knows the right answer. 2. This is a person who tries hard to do good school work. Children exhibiting the Well-Adjusted profile (cluster 1a) as well as the Withdrawn High Achiever profile (cluster 3a) were perceived by their peers as having the highest academic ability (see Table F.3) and achievement motivation (see Table F.4). Children in the Withdrawn Lower Achiever and Poorly Self-Regulated Lower Achiever clusters (3b and 4b) were perceived by their peers as having the lowest academic ability and motivation. These analyses indicated that the perceptions of parents and teachers as reflected in the profile types, and the perception of student peers reflected in peer nominations shared common perceptions of the abilities and motivation of students in each profile type. In other words, all relevant social groups had a common understanding of academic ability and motivation.

Self-Perception Students in the 3rd through the 5th grade in the State of Georgia (Cohort B) were asked to rate their own academic ability and motivation. The specific questions were as follows: Academic Ability Measure 1. If you were to list all of the students in your grade from worst to best in school work, where would you put yourself. (Rated on a five-point scale from “one of the worst” to “one of the best”). Achievement Motivation Measure 2. For me, being good at my schoolwork is … (“not at all important” to “very important” on a five-point scale).

Appendix F: Academic Ability and Achievement Motivation

237

Self-perceptions of academic ability and motivation was only modestly related to profile type and self-perceptions (see Tables F.5 and F.6). Thus, children at this age seem less aware of their own academic abilities and motivations than their parents, teachers, and peers. They seem not associate their school performance (the probable basis for parent, teacher, and peer perceptions) with their academic ability and motivation. Grade Point Average Sample: Russian parents reporting of school assigned grades (n = 537) Statistical test: General linear models analysis of variance (SPSS) Effects: Profiles: F = 13.87; p