The Psychology of Group Perception

Citation preview

THE PSYCHOLOGY OF GROUP PERCEPTION

THE PSYCHOLOGY OF GROUP PERCEPTION Perceived Variability, Entitativity, and Essentialism Edited by

Vincent Yzerbyt Charles M.Judd and Olivier Corneille

Psychology Press New York and Hove

Published in 2004 by Psychology Press 29 West 35th Street New York, NY 10001 www.psypress.com Published in Great Britain by Psychology Press 27 Church Road Hove, East Sussex BN3 2FA www.psypress.co.uk Copyright © 2004 by Taylor & Francis, Inc. Psychology Press is an imprint of the Taylor & Francis Group. This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical or other means, now known or hereafter invented, including photocopying and recording or in any information storage or retrieval system, without permission in writing from the publisher. www.socialpsychologyarena.com Library of Congress Cataloging-in-Publication Data The psychology of group perception: perceived variability, entitativity, and essentialism/ Vincent Yzerbyt, Charles M.Judd & Olivier Corneille, ed[itor]s. p. cm. Includes bibliographical references and index. ISBN 1-84169-061-9 (Print Edition) 1. Social groups. I. Yzerbyt, Vincent. II. Judd, Charles M. III. Corneille, Olivier. HM716.P79 2003 302.3–dc21 2003010696

ISBN 0-203-64497-2 Master e-book ISBN

ISBN 0-203-67690-4 (Adobe eReader Format)

Contents

1

About the Editors

vii

Contributors

viii

Preface

xii

Perceived Variability, Entitativity, and Essentialism: Introduction and Overview Vincent Yzerbyt, Charles M.Judd and Olivier Corneille

1

SECTION 1. CONSTRUCTS AND DEFINITIONS 2

Dynamic Entitativity: Perceiving Groups as Actors Marilynn B.Brewer, Ying-Yi Hong and Qiong Li

19

3

Perceiving the Groupness of Groups: Entitativity, Homogeneity, Essentialism, and Stereotypes David L.Hamilton, Steven J. Sherman and Julie S.Rodgers

30

4

Essentialism and Entitativity: Structures of Beliefs about the Ontology of Social Categories Nick Haslam, Louis Rothschild and Donald Ernst

47

5

The Mental Representation of Social Categories: Category Boundaries, Entitativity, and Stereotype Change Myron Rothbart and Bernadette Park

60

6

Subjective Essentialism in Action: Self-Anchoring and Social Control as Consequences of Fundamental Social Divides Vincent Yzerbyt, Claudia Estrada, Olivier Corneille, Eléonore Seron and Stephanie Demoulin

77

SECTION 2. ANTECEDENTS AND CONSEQUENCES 7

In the Eye of the Beholder: Lay Theories and the Perception of Group Entitativity, Variability, and Essence Jason E.Plaks, Sheri R.Levy, Carol S.Dweck and Steven J.Stroessner

96

8

Components of Homogeneity: A Multiple-Process Model of Social Categorization Karl Christoph Klauer, Katja Ehrenberg and Ingo Wegener

110

9

Forming Stereotypes of Entitative Groups

121

v

Craig McGarty 10

From Basketball to Business: Expertise, Implicit Covariation, and Social Judgment Patricia W.Linville and Gregory W.Fischer

135

11

Gender Outgroup Homogeneity: The Roles of Differential Familiarity, Gender Differences, and Group Size Mark Rubin Miles Hewstone, Richard J.Crisp, Alberto Voci and Zoë Richards

153

12

Group Size, Outcome Dependency, and Power: Effects on Perceived and Objective Group Variability Ana Guinote

166

13

The Acquisition, Transmission, and Discussion of Social Stereotypes: Influences of Communication on Group Perceptions Markus Brauer, Charles M.Judd and Micah S.Thompson

178

14

Culture, Communication, and Entitativity: A Social Psychological Investigation of Social Reality Yoshihisa Kashima

193

15

Group Socialization, Uncertainty Reduction, and the Development of New Members’ Perceptions of Group Variability Carey S.Ryan,Debbie R.Robinson and Leslie R.M.Hausmann

206

16

Entitativity, Group Distinctiveness, and Social Identity: Getting and Using Social 219 Structure Russell Spears, Daan Scheepers, Jolanda Jetten, Bertjan Doosje, Naomi Ellemers and Tom Postmes

17

Social Identity as the Basis of Group Entitativity: Elaborating the Case for the “Science of Social Groups Per Se” Katherine J.Reynolds, Penelope J.Oakes, S.Alexander Haslam, John C.Turner and Michelle K.Ryan

236

18

The Perception of Outgroup Threat: Content and Activation of the Outgroup Schema Tim Wildschut, Chester A.Insko and Brad Pinter

249

19

The Maintenance of Entitativity: A Subjective Group Dynamics Approach Dominic Abrams, José M.Marques, Georgina Randsley de Moura, Paul Hutchison and Nicola J.Bown

268

20

On the Advantages of Reifying the Ingroup Emanuele Castano

283

21

Uncertainty and Extremism: Identification with High Entitativity Groups under Conditions of Uncertainty Michael A.Hogg

298

22

Entitativity and Social Integration: Managing Beliefs about the Reality of Groups Richard L.Moreland and Jamie G.McMinn

311

vi

References

326

Index

371

About the Editors

Vincent Yzerbyt is a professor of psychology at the Catholic University of Louvain at Louvain-la-Neuve, Belgium. Dr. Yzerbyt received his Ph.D. from the University of Louvain in 1990 and has written and edited several books dealing with social cognition as well as journal articles on intergroup relations and stereotyping. He served as Associate Editor of the British Journal of Social Psychology and the Personality and Social Psychology Bulletin. Dr. Yzerbyt is currently the President of the European Association of Experimental Social Psychology. Charles M.Judd is a professor of psychology at the University of Colorado, Boulder. Dr. Judd received his Ph.D. from Columbia University in 1976, and was an assistant professor of psychology at Harvard University until 1981. He is the former editor of the Journal of Experimental Social Psychology, as well as the associate editor of the Journal of Personality and Social Psychology and the Psychological Review. Olivier Corneille graduated from the University of Liège, Belgium in 1992 and obtained his doctoral degree in psychology from the Catholic University of Louvain in 1996. He is currently a Research Associate at the Belgian National Fund for Scientific Research. His research is mainly concerned with the impact of categorization on social perception and behavior.

Contributors

Dominic Abrams University of Kent at Canterbury Nicola J.Bown University of Leeds Markus Brauer CNRS, Université Blaise Pascal Marilynn B.Brewer Ohio State University Emanuele Castano Graduate Faculty, New School University Olivier Corneille Catholic University of Louvain at Louvain-la-Neuve Richard J.Crisp University of Birmingham Stéphanie Demoulin Catholic University of Louvain at Louvain-la-Neuve Bertjan Doosje University of Amsterdam Carol S.Dweck Columbia University Katja Ehrenberg Universität Bonn

ix

Naomi Ellemers Leiden University Donald Ernst Hillsdale College Claudia Estrada Catholic University of Louvain at Louvain-la-Neuve Gregory W.Fischer Duke University Ana Guinote University of Kent David L.Hamilton University of California, Santa Barbara S.Alexander Haslam University of Exeter Nick Haslam University of Melbourne Leslie R.M.Hausmann University of Pittsburgh Miles Hewstone University of Oxford Michael A.Hogg University of Queensland Ying-Yi Hong University of Illinois Paul Hutchison University of Kent at Canterbury Chester A.Insko University of North Carolina, Chapel Hill Jolanda Jetten University of Exeter Charles M.Judd University of Colorado, Boulder Yoshihisa Kashima University of Melbourne Karl Christoph Klauer Universität Bonn Sheri T.Levy State University of New York at Stony Brook Qiong Li University of Maryland Patricia W.Linville Duke University

x

José M.Marques Univerity of Porto Craig McGarty The Australian National University Jamie G.McMinn University of Pittsburgh Richard L.Moreland University of Pittsburgh Penelope J.Oakes The Australian National University Bernadette Park University of Colorado, Boulder Brad Pinter University of Washington Jason E.Plaks University of Washington Tom Postmes University of Exeter Georgina Randsley de Moura University of Kent at Canterbury Katherine J.Reynolds The Australian National University Zoë Richards Cardiff University Debbie R.Robinson University of Pittsburgh Julie S.Rodgers University of California, Santa Barbara Myron Rothbart University of Oregon Louis Rothschild Brown University Mark Rubin University of Newcastle, Australia Carey S.Ryan University of Nebraska at Omaha Michelle K.Ryan The Australian National University Daan Scheepers Leiden University Eléonore Seron Catholic University of Louvain at Louvain-la-Neuve

xi

Steven J.Sherman Indiana University Russell Spears Cardiff University Steve J.Stroessner Barnard College Micah S.Thompson University of Colorado, Boulder John C.Turner The Australian National University Alberto Voci University of Padova, Italy Ingo Wegener Universität Bonn Tim Wildschut University of Southampton Vincent Yzerbyt Catholic University of Louvain at Louvain-la-Neuve

Preface

This book emerged out of the enthusiasm that followed a meeting held in Louvain-la-Neuve, Belguim in July 1999. The meeting, devoted to the topic of group perception and cognition, was organized under the auspices of the European Association of Experimental Social Psychology. By all indications the meeting was a huge success. We were fortunate to be able to bring together 25 leading researchers from all over the world. All 25 presented their latest work on the important issues of group perception, and these presentations led us to focus on the particular issues surrounding the central themes of this book: Entitativity, Perceived Variability, and Essentialism of social groups. We realized during the course of the meeting that these central themes were closely intertwined, that there were very different schools of thought about these constructs, and that the interchange that occurred at the meeting and beyond might fruitfully advance thinking about these issues as central to our understanding of group perception. This is not to say that the discussions at the meeting, or subsequently, resolved all disagreements or led to some overarching agreement about the relations of these three constructs, their antecedents, and their consequences. Indeed, as the pages in this book indicate, there remain significant and theoretically important divergences in the points of view espoused on these issues by the leading researchers in the area. Yet what was clear is that the discussion at the meeting and beyond led to major clarifications in points of view and a recognition of the theoretical issues along which viewpoints diverged. Not all was divergence of opinion, however. For instance, there was no doubt that all at the meeting agreed about the excellence of Trappist breweries and of Belgian cuisine. Additionally, and importantly for the present volume, we all agreed that continuing dialogue on these important issues, without or with the assistance of Belgian ales and cuisine, would be beneficial. After returning to our respective institutions and laboratories, a large number of participants contacted us as the meeting’s organizers and suggested that further dialogue might be useful in terms of a collection of papers that we might publish jointly. This would not only facilitate further communication among the people at the meeting, but it might also expose a

xiii

broader audience to the issues that we had found so compelling. In the months that followed, we asked participants at the meeting to contribute to the present edited volume, and we also approached a number of people whose work was clearly relevant but who, for various reasons, were unable to attend the meeting. The end product is the present edited volume which brings together 22 chapters authored by no fewer than 62 researchers coming from the leading research centers in North America, Europe, and Australia. The book would not have been possible if it were not for the help and support of numerous people. We would like to take this opportunity to thank the graduate students and staff members who contributed to making the small group meeting a memorable event both scientifically and socially. In particular, we would like to thank the undergraduate and graduate students of the social psychology division of the Université Catholique de Louvain at Louvain-la-Neuve, David Bourguignon, Michel Desert, Stephanie Demoulin, Muriel Dumont, Claudia Estrada, Steve Rocher, as well as our division secretary, Anne-Françoise Cabiaux. We also want to acknowledge the generous sponsorship of the European Association of Experimental Social Psychology, the Belgian Fonds National de la Recherche Scientifique, and the Department of Psychology of the Catholic University of Louvain at Louvain-la-Neuve. We are also indebted to Allison Mudditt, then at Psychology Press and Stacy Mayill for their enthusiastic response to the project. As publisher, Psychology Press offered its full backing at all stages of the process. And, as editors, we also want to thank all of the authors of the chapters in this volume for having the wisdon to give us their distinguished contributions and the patience to bear with us as this project took unreasonable amounts of time. Last but not least, we would like to thank our families and, especially, our partners in life, Isabelle, Liz, and Florence for their constant support.

1 Perceived Variability, Entitativity, and Essentialism Introduction and Overview VINCENT YZERBYT Catholic University of Louvain at Louvain-la-Neuve CHARLES M.JUDD University of Colorado OLIVIER CORNEILLE* Catholic University of Louvain at Louvain-la-Neuve *Also at the Belgian National Fund for Scientific Research

This his volume represents an attempt to provide an integration of three different lines of research in social psychology. These three are reflected in our choice of title: “Perceived Variability, Entitativity, and Essentialism.” Each of these three terms emerged from distinct lines of work, much of it done by contributors to this volume. As we will see, these three terms, when applied to the perception of social groups, refer to related concepts. Our goal in launching this volume was to clarify their definitions, their relations, their causes, and their consequences. One line of work, the Entitativity part of our title, derives from a seminal theoretical article that Donald T.Campbell published in 1958 on the perceptual reality of social groups. He suggested that groups may vary in the extent to which they are perceived as being a “real thing” or an entity, and he coined the rather unfortunate term entitativity to refer to this characteristic. He went on to speculate about cues that perceivers might use to infer the entitativity of groups. Such cues were thought to include the extent to which group members are similar to each other, whether group members function in a coordinated manner in pursuit of shared goals, whether group membership is stable and group boundaries relatively fixed, and whether group members are physically near each other. Although Campbell offered no data in support of his ideas about these perceptual cues, a number of subsequent researchers have carried his work on group entitativity forward (e.g., Abelson, Dasgupta, Park, & Banaji, 1998; Brewer & Harasty, 1996; Dasgupta, Banaji, & Abelson, 1999; Hamilton, Sherman, & Lickel, 1998; Lickel, Hamilton, Wieczorkowska, Lewis, Sherman, & Uhles, 2000; S.J.Sherman, Hamilton, & Lewis, 1999). This work has clearly demonstrated that indeed groups do differ in the extent to which they are perceived to be entities and that many of the perceptual cues that Campbell outlined are in fact attended to in forming such perceptions. Independent of this work on groups as entities, stereotyping researchers have been interested in the strength of group stereotypes, focusing in particular on perceptions of group variability, the second term in our title.

2

THE PSYCHOLOGY OF GROUP PERCEPTION

Presumably, if perceivers have particularly strong stereotypes about given social categories, then they should see relatively little diversity among the members of such categories. The attributes associated with the category as a whole should also be associated with nearly every category member, allowing strong inferences to be made from the group to individual members and vice versa. Because of the fact that stereotypes have traditionally been associated with significant social categories (e.g., gender or ethnicity), nearly all of this work has focused on the perceived variability of such categories rather than other sorts of social groups (e.g., task groups, intimacy groups). Additionally, most of the work on perceived variability has focused on one particular determinant of perceived group variability, namely whether the target group is an ingroup or an outgroup. A robust, although not universal finding is the outgroup homogeneity effect, whereby perceivers judge ingroups to be more variable than outgroups (e.g., Judd & Park, 1988; Linville, Fischer, & Salovey, 1989; Park & Rothbart, 1982; Park & Judd, 1990; Quattrone & Jones, 1980). More recently, research has focused on the ways in which power, status, and category size (minority versus majority) may moderate these ingroup/outgroup differences in perceived variability (e.g., Simon & Brown, 1988; Lorenzi-Cioldi, 1988; Guinote, Judd, & Brauer, 2002). The third line of work that we seek to integrate derives from work in cognitive and developmental psychology on categorization processes. This work has differentiated between two different kinds of categories, typically labeled natural kinds and human artifacts. The former are categories where there is an essence, often biological in nature, that defines and dictates category membership. The latter consists of categories that are constructed by perceivers in the absence of any inherent or biological basis for categorization. The prototypic natural kind category is a species (e.g., leopards, elephants). Human artifact categories are exemplified by object categories, such as chairs or tables. Cognitive and developmental psychologists have argued that perceivers routinely dis tinguish between these two different sorts of categories and reason about them in rather different ways (e.g., Gelman, 1988; Hirschfeld, 1996; Keil, 1989). Based on this work, social psychologists have recently become interested in the extent to which social groups and categories are “essentialized” (the third term in our title) or seen as natural kinds. For instance, Rothbart and Taylor (1992) and Yzerbyt and colleagues (Yzerbyt, Corneille, & Estrada, 2001; Yzerbyt, Rocher, & Shadron, 1997; Yzerbyt & Rogier, 2001) have argued that perceivers often treat social categories as natural kinds, assuming there exists some underlying essence that unites category members. Haslam and colleagues (N.Haslam, 2002; N.Haslam, Rothschild, & Ernst, 2000) have explored the components and consequences of essentialized views of groups. And Levy, Plaks, and colleagues have examined individual differences in the tendency to see social categories as natural kinds (Levy, Plaks, Hong, Chiu, & Dweck, 2002; Levy, Stroessner, & Dweck, 1998; Plaks, Stroessner, Dweck, & Sherman, 2001). A few years ago, when we decided to initiate this volume, it seemed evident to us that the central constructs from these three research traditions were related. Groups that were perceived as entities probably also were groups where the perception of variability was low and were also probably more likely to be essentialized. Yet the literature to integrate theoretically these three was remarkably absent. Our hope is that this volume now fills this void. The chapters in this volume are grouped into two sets. Those in the first we consider to be more definitional and theoretical in nature. Here authors who have been active contributors to one or more of the above traditions have attempted to theoretically delineate these constructs and their interrelations. The chapters in the second set are a bit more empirically oriented, and they tend to explore the causes and consequences of entitativity, perceived variability, and essentialism in the perception of groups. In the sections that follow in this chapter, we briefly provide a road map for chapters in both sections. In so doing,

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

3

we also hope to provide our own perspective on the theoretical integration of these three terms and their research traditions. SECTION ONE: DEFINITIONAL CONSIDERATIONS We see the chapters in this section as struggling with the issues of (a) what exactly do we mean by the terms perceived variability, entitativity, and essentialism, and (b) how are these related to each other, both theoretically and empirically. As will be clear on a close reading of these chapters, our authors have rather different perspectives on these two issues and argue for theoretical integration in rather different ways. Although this diversity ultimately threatens our goal of providing a quick and easy integration of these traditions and literatures, it does mean that the reader must attempt his or her own unique theoretical understanding, and we see this as a very positive outcome indeed. The initial chapter in this section, by Marilynn Brewer, Ying-yi Hong, and Qiong Li, focuses in particular on one of the central attributes that D.T.Campbell (1958) recognized as a cue in the perception of entitativity: whether or not a group has a set of shared goals and purposes and acts in a coordinated manner toward those. Brewer et al. argue that the perception of similarity or homogeneity of group members is not a sufficient condition for the perception of group entitativity. Rather, entitativity is inferred either when the group is a seen as a natural kind (essentialized) or when it is seen as acting in a coordinated manner on common goals and purposes. And these two bases for the perception of entitativity have somewhat different consequences, with the perception of essence influencing more judgments about individual group members and the group prototype, and the perception of agency or common goals influencing emotional reactions to the group (and its purposes) as a whole. Beyond these theoretical considerations, Brewer and colleagues report some data from cross-national samples in which perceptions of both a group essence and group agency are shown to predict judgments of entitativity. Although they report some cultural differences, the bottom line is that both are influential in both cultures. Like Brewer et al., David Hamilton, Steven J.Sherman, and Julie S.Rodgers primarily focus on the perception of group entitativity and how it relates to the other constructs of essentialism and perceived homogeneity. They argue that these three concepts, while they are certainly positively correlated, are not redundant. Further they argue that the exact relations among them depend on the type and function of the group involved. Specifically, their empirical work has differentiated among social categories, task groups, intimacy groups, and loose associations (e.g., people waiting for a bus together), and the relative degree of entitativity, homogeneity, and essentiality for all four types of groups. They suggest that intimacy groups (e.g., families) are the highest on entitativity, but typically do not have strong stereotypes associated with them. They suggest that social categories may be essentialized, but they are actually fairly low on perceived variability compared to the other types of groups. For task groups, they argue that perceptions of entitativity are relatively high, but again essentialism and perceived homogeneity may be relatively lower than for other types of groups. The bottom line in the Hamilton et al. chapter is that while these three constructs are related, they are clearly discriminable, with different causes and consequences. By focusing on the four different types of groups that their research has uncovered, they characterize the differences between entitativity, perceived homogeneity, and essentialism primarily by discussing how they vary (and covary) between types of groups, rather than how they vary (and covary) across groups within a given type. This is a somewhat different approach than that followed by the authors of the other chapters in this section.

4

THE PSYCHOLOGY OF GROUP PERCEPTION

The Brewer et al. and Hamilton et al. chapters seem most focused on entitativity and its relations to other constructs. Nick Haslam, Louis Rothschild, and Donald Ernst, on the other hand, clearly come out of the research tradition on essentialism and the properties that differentiate essentialized groups from groups that are less likely to be seen as “natural kinds.” They start their chapter with a fascinating tour of the history of thinking about essentialism, tracing the origin of the concept in the linguistic and philosophical traditions. They then report on a series of empirical studies that have explored the characteristics of essentialized groups. From this work, they argue that there are two components to essentialism, with groups varying more or less independently on them both. Characteristics included in the first component have to do with traditional notions of what constitutes a “natural kind”: they are groups that are immutable, natural, historically invariant, have sharp boundaries, and necessary features. The second component, they argue, is made up of characteristics traditionally associated with entitativity. Characteristics such as homogeneity, inductive potential (i.e., inferring what someone is like from group membership and vice versa), and membership based on shared inherent properties are associated with the second component. Starting from a theoretical interest in defining the characteristics of essentialized groups, Haslam, Rothschild, and Ernst thus come to the conclusion that there are two components: one having to do with groups being seen as “natural kinds” and the other having to do with the perceived entitativity of groups. Thus, they see entitativity as one component of essentialism, and, within entitativity, perceived homogeneity is one factor that leads to the perception of entitativity. This is clearly, then, a sort of nested structure, with perceived homogeneity as a condition for entititativity and entitativity as a condition for essentialism. While Brewer et al. and Hamilton et al. start from the point of view of explaining entitativity, and Haslam et al. start from the point of view of explaining essentialism, Mick Rothbart and Bernadette Park emerge much more squarely out of the stereotyping literature. Accordingly, their initial question is focused on the process by which stereotypes may be changed, and in this regard they necessarily focus on factors that affect whether individual group members who disconfirm a group stereotype will be functionally integrated into the group and thereby lead to change. They suggest that whether a disconfirming group members is functionally included in or excluded from the group depends on both the perceived variability of the group and the degree to which the group is seen as an entity. They further argue that entitativity is not simply the same thing as perceived variability. Group entitativity is also affected by the extent to which a group is seen as acting on common goals or its degree of agency (consistent with Brewer et al.). Finally, they suggest that essentialized groups are a subset of groups that are seen as entitative, having both high perceived similarity and a sense of agency. All of this leads them to suggest that disconfirming group members will be functionally excluded, and stereotype change inhibited, when group boundaries are strong. And this is most likely to be the case in entitative groups that possess shared attributes and common goals, and are essentialized. The final chapter in this section, by Vincent Yzerbyt, Claudia Estrada, Olivier Corneille, Eléonore Seron, and Stéphanie Demoulin, also starts by asserting that these three constructs are not one and the same thing. But they argue in particular for a different relationship between entitativity and essentialism than that articulated by others. They suggest that entitativity is the observable or phenotypic attribute of “groupness,” whereas group essence is the underlying or genotypic attribute that results in the phenotypic expression. They then show, in a series of reported studies, that when essentialism is manipulated, entitativity and perceived similarity are consequences. Thus, perhaps like Rothbart and Park, they suggest that essentialism is the most fundamental of the three concepts, that essentialism requires perceived similarity and entitativity, and that perceived similarity and entitativity are broader constructs and, as such, can be found with nonessentialized groups.

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

5

This brief summary of these chapters (admittedly not the summary that these authors themselves might have provided) clearly points to diversity in both the author’s starting points and the conclusions they eventually reach. Yet we believe that there are threads of consensus that are apparent across the chapters, and hopefully they show the way to a more integrative account of the theoretical relationships among entitativity, perceived variability, and essentialism. First, no one argues that these are one and the same construct. Groups vary in the extent to which they are perceived to be real real or meaningful, the extent to which their members are seen to be diverse, and the extent to which they are seen to have an essence that defines them as a “natural kind.” Although these perceptions certainly covary, they are not one and the same thing. A necessary corollary of this is that these three characteristics of groups have at least somewhat different causes and consequences. The factors that cause a group to be seen as entitative, that lead to perceptions of homogeneity, and that result in an essentialized view of group membership are not one and the same. And yet we would suggest that all three of these perceptions have the potential for influencing each other. Many of the studies that are summarized in these chapters suggest that each of these perceptions has potential implications for the others. Manipulations of essence lead to higher perceived similarity and higher entitativity. Groups that come to be seen as less variable are more likely to be seen as entitative, and so forth. Though not identical constructs, and though potentially having different causes and consequences, they nevertheless are causally related to each other. And the causal relations are certainly reciprocal and nonrecursive. If we grant this, then the next obvious question concerns whether these constructs share a nested or hierarchical structure of relations or whether they are simply overlapping sets, capable of influencing each other in a probabilistic manner. Here, we are taken by many of the ideas contained in Figure 1 of Rothbart and Park. Consistent with Brewer et al., it seems that entitative groups derive their “realness” either from the perceptions of similarity of members or from a sense of agency and purpose. And the conjunction of the two is particu larly likely to lead to group entitativity. But essentialism seems to us to be more than just another partially overlapping set of groups. Groups that are essentialized seem to us to necessarily have a realness to them that cannot be denied, and that realness inevitably leads to strong and impermeable group boundaries with high perceived genotypic similarity. In this sense, essentialism seems to have a rather different conceptual basis from the other two constructs, in that it implies attributional consequences that the other two do not. Perceived homogeneity focuses on the extent to which group members share attributes. Entitativity focuses on the extent to which a group is seen as real or coherent. Essentialism derives from both of these but it seems to us to imply more. It entails not just a description of a group and what makes it a group, but also a theory about why group membership matters and what it implies. If a group is essentialized, then we are permitted to make inferences about why this group exists (why it is an entity) and why group members are similar to each other. We are not saying simply that the group is an entity and that its members are similar, but we are implying an attributional process for the origins of entitativiy and similarity: this is a group and its members are similar to each other because they all share an essence that is invariant and unmutable. Essentiality, in other words is not just the conjunction of similarity and entitativity. It really implies a theory about the origins of similarity and entitativity. It implies that the perceiver has a “natural kind” explanation for the existence of the group and for why its members are similar to each other. It suggests that the perceiver attributes groupness and similarity to internal, dispositional, and immutable characteristics of group members. Turning to the distinction made by Hamilton et al. among different types of groups, it seems to us particularly interesting that it is often social categories (rather than task groups or intimacy groups) that are imbued with essentialistic qualities. And then the interesting question for us concerns not the differences

6

THE PSYCHOLOGY OF GROUP PERCEPTION

among these types of groups (categories versus intimacy groups versus task groups) but the conditions under which social categories are essentialized. This focus on social categories, and the question of when they are essentialized, enables us to effectively bridge the three diverse traditions that we identified at the start of this preface and that gave rise to our notions of entitativity, perceived variability, and essentialism. Strong stereotypes, prejudice, discrimination, and hostile intergroup relations have been issues at the core of social psychology since its founding. And from our point of view, notions of entitativity, perceived variability, and essentialism take on particular interest to the extent that they help us understand these social problems. Strong stereotypes occur when the members of social categories are viewed as very similar to one another and when the boundaries that differentiate those categories are seen as sharp and fixed. And then when perceivers go the additional step of essentializing those categories, then the definition of a social stereotype takes on a very different aspect: category membership becomes inevitable and immutable. Category members resemble each other because of their unchangeable and true essence. And the boundaries that divide groups and that lead to social conflict are seen as fixed forever as a natural part of the world, just like the lines that separate different species. Such essentialized social categories, it seems to us, are what leads to the most severe forms of intergroup conflict and hostility, and ethnic cleansing, or ethnic genocide, when it has occurred historically, has certainly been accompanied by an essentialistic rhetoric. From this point of view, what becomes important is the evolution of essentialistic thinking about social categories: the factors that lead to more essentialistic views of social categories, given perceived entitativity and similarity. And this leads us to want to understand the dynamic relations among these constructs and the factors that lead to changes in all three, within the same social categories, over time and across perceivers. The chapters in the first section of the book, discussed above, really focus more on defining the three constructs and to a lesser extent on their dynamic interrelations. The chapters in the second set, which we now briefly summarize, are concerned more with the various factors that lead to the perception of group entitativity, group homogeneity, and essentialism. They also stress some of the consequences of entitativity and essentialism. And from our point of view, when talking about social categories, these are the important questions to be addressed in the attempt to understand why it is that groups and people conflict and how such conflicts might be ameliorated. SECTION TWO: VARIATIONS ON THE THEME As we have argued above, the perspective prevailing in some of the definitional chapters is that the perception of homogeneity, entitativity, and essentiality depends largely on the particular group that is being appraised (Hamilton et al., this volume; Haslam, this volume; but see Yzerbyt et al., this volume). This target-based approach has proven fruitful in stressing some major differences among groups. At a very general level, two images of social groups seem to emerge. Whereas some groups elicit the belief that their members are closely connected, well organized, and working together to reach some joint outcome, others are associated with the idea that similarity of group members is the key feature, some deep underlying marker making it difficult if not impossible to change or deny group membership. From this point of view, a perceiver’s impression of a group depends on the particular characteristic of that group. With a few exceptions, the chapters in this second section tend to stress a rather different approach, one that we see as most promising. Rather than assuming that groups have defining characteristics that affect how they are perceived, most authors of the chapters in this second section concentrate on possible structural and social factors that may shape a perceiver’s propensity to see any given group as a coherent entity or as a looser set of people. A neat illustration of this perspective is the chapter by Jason E.Plaks, Sheri R.Levy, Carol S.Dweck, and Steve J.Stroessner. These authors build a strong case for an individual

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

7

difference approach that complements the view according to which entitativity or perceived variability primarily derives from the intrinsic features of the target groups. To be sure, most if not all social psychologist would agree that the stimulus is not the sole factor in the perception equation and that the perceiver plays a most critical role. Still, according to Plaks and colleagues, most contemporary efforts may have somewhat underestimated the extent to which group perception also resides in the eye of the beholder. People’s a priori beliefs about human nature, whether chronic or more transitory, should likely orient their perception of entitativity and homogeneity. In line with this idea, Plaks and colleagues identify two contrasting lay theories that underlie quite divergent perspectives for understanding individuals as well as groups. Whereas “entiteists” view personal characteristics as fixed entities that are unlikely to change over time and they expect a high degree of consistency in people’s behaviors, “incrementalists” prefer to see personal characteristics as dynamic and open to modification, and they situational forces that impinge on people’s actions. Plaks and colleagues provide an impressive body of evidence to suggest that these two views translate into distinct takes on group perception. Studies indeed reveal that, compared to “incrementalists,” “entiteists” more readily apply psychological traits to groups, perceive less within-group variability and greater between-groups differences, form and endorse stronger stereotypes, and process information about the group members in such a way that less attention is given to the counterstereotypic evidence. Entity theorists not only use and preserve their stereotypic beliefs more so than “incrementalists,” they also are more likely to believe that innate factors rather than shared environmental experience are accountable for shaping group members’ characteristics (see also C.L.Martin & Parker, 1995; Yzerbyt & Rogier, 2001). In contrast, Plaks and colleagues argue, “incrementalists” are more sensitive to the presence of shared goals as a cue to group coherence. There are a number of links that can be drawn between the contribution of Plaks and colleagues and other chapters in this volume. In particular, we would point to the overlap between their analysis and Brewer’s distinction between groups organized around some essence and those organized around some goals. Also, we see interesting parallels between Plaks et al.’s suggestions and those made by Social Identityy Theorists (SIT) about stability and permeability, entity theorists living in a world where groups are clearly segregated and the social structure is seldom challenged and “incrementalists” believing the social landscape is largely open to revision. In addition to perceiver differences, it is clear that a number of other factors affect the perception of group homogeneity, entitativity, and, possibly, essence. Most of the remaining chapters help us identify these factors. To be sure, these factors can be grouped in various ways. One way that we find rather convenient in organizing the various chapters builds on the extent to which the perceiver plays an active role in the construction of entitativity. At one end of the spectrum, one can identify a set of factors to which the perceiver simply responds. He or she plays no real active role or does not engage in what could be seen as truly motivated or strategic group perception and cognition. Familiarity with a group is a good illustration of this kind of factor: Differences in group familiarity may affect the perception of group entitativity and homogeneity, but these differences are not actively sought out or strategically motivated on the part of the perceiver; rather he or she just happens to have more or less knowledge about the group, and this affects perceptions. Going to the opposite end of the continuum, the perception of entitativity might be seen to result from a deliberate line of actions and cognitive strategies. For instance, if the perceiver is motivated to think of his or her own group as needing to realize some personally important goal, to reach a specific end, then actions may be undertaken to organize the group toward that purpose, and more entitative group impressions will

8

THE PSYCHOLOGY OF GROUP PERCEPTION

ensue. At this end of the continuum, then, the perceiver actively and strategically affects the group itself in ways that lead to changes in how entitative it is seen to be. Halfway along the continuum, one might find that processes purposively engaged in by the perceiver might end up in promoting the perceived entitativity of the groups even though this was not the perceiver’s primary goal. For instance, sometimes people talk with each other about their own groups and other groups. The goal of communication is not to strategically affect the perceived entitativity of those groups. But it may still have these unanticipated outcomes. In the remainder of this introductory chapter, we would like to walk you through this continuum from one pole to the other, discussing each of the chapters in turn. Needless to say, this tour only provides a highly subjective and necessarily partial account of the content of the various chapters. Having reviewed all the chapters, we will then come back to a number of issues that allow us to compare and to contrast more specific contributions. A first factor that may affect the extent to which perceivers see groups as being coherent wholes is the very act of categorization. By slicing the social environment into different zones of coherence, perceivers may well end up overestimating the cohesiveness of the groups they have categorized. In short, the mere act of categorization is likely to promote similarity among the members of same category and, as a consequence, entitativity of the category. This observation is of course far from new, and the single most famous empirical demonstration of the phenomenon dates back to the study conducted by Tajfel and Wilkes (1963). The original paradigm has come in many disguises. The most popular version is the so-called “Who said What?” paradigm in which participants observe a discussion between a limited number of members comprising two distinct social categories, black versus whites, males versus females, overestimators versus underestimators, etc. At the end of the discussion comes a surprise memory test in which observers are presented with individual statements and asked to assign them to the correct speaker. Categorization is assumed to have affected the perceptions of homogeneity if errors of recognition more frequently involve within-category confusions than between-category confusions. In their chapter, Karl Christoph Klauer, Katja Ehrenberg, and Ingo Wegener take issue with this working assumption and provide a thorough examination of the various components that may be responsible for memory errors in experimental settings that rely on the “Who said What?” paradigm. Instead of comparing the number of between-categories and within-category confusions, the authors suggest that the paradigm be altered by adding a set of distracters to the list of statements comprising the memory test. Doing this would allow researchers to disentangle the impact of a number of independent cognitive processes, some directly relevant to the issue of homogeneity and some not, and to examine the impact of various contextual factors thought to affect the emergence of homogeneity. The relevant cognitive processes are person memory (Is there a memory of the person who made the statement?), category memory (Is there a memory of the category of the person who made the statement?), and category guessing (Are there cues that can be used to infer the category membership of the person making the statement even though category membership cannot be recalled?). Whereas accurate person memory within a category is indicative of heterogeneous processing of the discussants, both category memory and category guessing suggest rather different processes indicative of homogeneity in perception. Klauer and colleagues present a series of studies aimed at validating the relations between the process indicators and the underlying mental operations. For instance, a manipulation of the salience of the categories only affected the category memory parameter and not the other indices. Beyond establishing the convergent and discriminant validity of the various process measures, the authors examine the impact of a series of important contextual factors on group perception such as relative group size and crossed categorization. Their findings lead them to argue for the dissociation between various facets of category-

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

9

based processing, possibly allowing a reconciliation of contradictory findings with regard to the impact of cognitive load in social perception. Keeping with the discussion of some fundamental processes involved in the categorization processes, Craig McGarty considers how prior expectations about the existence of group differences interacts with stimulus information one receives about groups to affect the extent to which those groups are seen as entitative. He argues that the mere existence of prior beliefs that says that groups differ from each other is sufficient to trigger the formation of group stereotypes and the perception of groups as entities, assuming that in fact there is sufficient information received about those groups to enable effective categorization. McGarty uses these general ideas to offer an insightful explanation for illusory correlation results, suggesting some nice extensions to Fiedlers (1991) and E.R.Smiths (1991) account based on information loss and unreliability of judgment. McGarty’s explanation suggests that the typical illusory correlation results are due to the expectation that the groups must differ in some way and then by the availability of sufficient information about the two groups to permit that expectation to result in rather different stereotypic beliefs. In a similar vein, McGarty suggests that perceivers may hold prior expectations about the degree to which social categories are essentialized. And, consistent again with his theoretical position about the constraints that exist between expectations and the available stimulus information, he suggests that these expectations may permit essentialized beliefs for some social categories more than for others. Continuing with the idea that the perceived entitativity of groups may emerge simply from extraneous factors that involve little strategic or motivated reasoning on the part of the perceiver, Patricia W.Linville and Gregory W. Fischer extensively discuss the role of differential familiarity with groups as a factor that affects their perception. Interestingly, rather than examining a variety of prototypic attributes thought to characterize groups in general, these authors concentrate on a somewhat neglected facet of the mental representation of groups, namely the perception of covariation among the features describing category members (but see Judd & Lusk, 1984; Park, Ryan, & Judd, 1992). This is an important consideration because, as Linville and Fischer argue, perceived covariation is likely to be related to the degree to which people make distinctions among group members. Moreover, there is no reason to expect any strong link between the perceived covariation among a set of attributes and the perceived variance along each individual attribute. As much as we have learned from closer scrutiny of the perception of single feature variability, it may be fruitful to examine the perception of multiattribute covariability. Linville and Fischer’s contribution revolves around two main ideas. First, perceptions of covariation generally overestimate the actual level of covariation, and this tendency is stronger at lower levels of familiarity. Supportive evidence for this conclusion comes from several studies using various broad social categories such as age, gender, occupation, or even race groups. When people have extensive contact with members of the target group, when they are members of the group, or when they are otherwise knowledgeable about the group, their judgments entail lower perceived covariation among the various attributes. Additional work on highly entitative groups such as basketball teams confirms this pattern and additionally indicates that familiarity and expertise allow people to make less biased estimates of objective feature covariation. Interestingly, Linville and Fischer’s strategy for collecting information about perceived covariation is highly implicit in that participants are not directly asked to make covariation judgments. Rather, perceived covariation is computed from judgments that are given along trait dimensions. Second, Linville and Fisher suggest that implicit perceptions of covariation play a role in social judgment inasmuch as a higher degree of covariation should be associated with more extremity in judgments. Again, data from high and low entitative groups seem to be consistent with this conjecture. Linville and Fischer conclude by saying that perceived covariation need not be related in any simple way to group entitativity.

10

THE PSYCHOLOGY OF GROUP PERCEPTION

Depending on the basis for entitativity, either similarity or interdependence, one might expect more or less perceived covariation. Unlike Linville and Fischer, who focus almost exclusively on familiarity, Mark Rubin, Miles Hewstone, Richard J.Crisp, Alberto Voci, and Zoë Richards additionally tackle the factors of group size and group status and examine how all three factors play a role in the emergence of the outgroup homogeneity effect, i.e., the tendency to perceive the outgroup as being more homogeneous than the ingroup. On the basis of the evidence obtained in minimal group studies and with gender groups, Rubin and colleagues question the status of differential familiarity as a necessary cause for the outgroup homogeneity effect to emerge. In fact, and contrary to both the differential familiarity hypothesis and Lorenzi-Cioldi’s position on gender effects, gender outgroup homogeneity (OH) can be seen as a modest yet reliable phenomenon (Park & Judd, 1990). The authors then engage in a critical analysis of Lorenzi-Cioldi’s interpretation of the larger OH effect among men than among women. On the basis of the accumulated evidence, they conclude that members of high status groups, and thus men, perceive greater outgroup homogeneity effects. In contrast, members of low status groups, and women in particular, show no such tendency. Rubin and colleagues also discuss the so-called minority group homogeneity effect as well as the interactive effect of group status and group size. In the second part of their chapter, Rubin and colleagues present a research program aimed at disentangling the impact of these three factors on the perceived dispersion of members of gender groups. Their studies reveal that group size is indeed more strongly associated than status or gender with a tendency to perceive outgroup homogeneity, familiarity playing no causal role in the process. Interestingly, however, measures that make reference to individual group members (dispersion rather than global similarity) and targets that stress interpersonal comparisons (individual targets rather than abstract categories) suggest that men are more prone than women to enhance their personal identity, a tendency that facilitates the emergence of outgroup homogeneity among men. Ana Guinote’s chapter presents yet another perspective on the role of group size and group status on the perceived homogeneity of groups, adding the factor of power. The empirical work that is being reviewed rests on two main hypotheses. For one thing, Guinote argues that the position of individuals in the social structure directly affects their perception of control. This perception of control, in turn, influences the extent to which perceivers will attend to the variability of the groups out there. Being able to count on numbers, status, and power thus leads people to see other groups in a more homogeneous manner. For another, Guinote holds that perceived control also has some bearing on the actual spontaneity in behaviors, that is, people’s objective heterogeneity. Members of large, high-status, or otherwise powerful groups are likely to act in more deliberate ways and thus be inherently more variable than members of small, low-status, or powerless groups. These two phenomena converge so that the members of subordinate groups come across as less variable than the members of dominant groups. A series of studies are presented that lend support to these two hypotheses. With the two following chapters, we continue the journey along our continuum. After concentrating on homogeneity and entitativity as the accidental result of the peculiarities of people’s perceptual and cognitive processes or their mere position in the social structure, we now turn to work that assigns a more active role to social perceivers. Indeed, a number of contributions see homogeneity, entitativity, and possibly essence, as emergent products of social interaction. As a case in point, Markus Brauer, Charles M.Judd and Micha S.Thompson tackle the important question of whether communication about a group may indeed affect the perceived entitativity of a target group. Building upon Park and Hastie’s (1987) distinction between instance-based versus abstraction-based stereotype acquisition, the first study uses a rumor transmission paradigm to show that people end up holding more extreme and less variable views about a target group when they learn about the group only

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

11

from other people than when they form their impression solely on the basis of first-hand behavioral observations. Brauer and colleagues further show that this communication effect is not restricted to recipients of communication but that people asked to tell others about their impressions of a group similarly form more stereotypic views than people simply asked to study the materials. Relying on actual group communication settings, two additional studies confirmed the fact that these two effects combine to polarize stereotypes about a target group but only when all group members possess an equal (small) share of counterstereotypic information. Clearly, this chapter provides convincing evidence regarding the role of communication in the formation and preservation of stereotypes and, as such, stresses the social nature of stereotypic beliefs (for a related set of issues, see McGarty, Yzerbyt & Spears, 2002). In the following chapter, Yoshihisa Kashima further stresses the idea that beliefs about groups are social constructions. Indeed, this contribution focuses on the psychological processes involved in the production and reproduction of social reality. A central tenet of Kashima’s analysis is that communication contributes to providing a sense of realness to whatever is the topic of the exchange. Importantly, communication takes place between people in the context of a larger cultural background. In line with the theme of the present book, this phenomenon applies to social categories as much as to anything else. That is, within a given cultural context, communication makes social categories become entitative and real. As a consequence, Kashima argues, people tend to endow social categories with some essence. A noteworthy feature of Kashima’s argument is the distinction between the notion of entitativity as proposed by D.T. Campbell, which rests on a series of “measurable” perceptual clues and refers to the ontology of groups, and the concept of essentialism, which more directly relates to the psychological sense of realness and is highly sensitive to the occurrence of communication (see also Yzerbyt, Corneille, & Estrada, 2001). Several pieces of empirical evidence allow Kashima to argue that communication indeed not only produces more polarized and homogeneous impressions of groups but also leads perceivers to see the characteristics of the target group to be more unalterable. Additional cross-cultural work also shows the nonnegligible role of culture in the attribution of essence to social entities. Kashima’s chapter is particularly important in that it simultaneously stresses the active part people play in constructing social reality and the critical role of communication in providing the psychological sense of that reality. The remaining chapters in the book carry us to the more strategic end of the continuum. For many of the authors of this third category of chapters, entitativity comes across as a feature that is being monitored, relied upon, used, and even sought out and constructed. To be sure, we note a progression in the extent to which the authors subscribe to such a functional perspective in their treatment of group entitativity. Starting with the idea of a genuine interest in the perception of regularities among the group members, one ends up with some indication that people are indeed doing their best to construct the perception that a group is a coherent social entity. Carey S.Ryan provides a good demonstration of people’s special attention to aspects of group homogeneity and entitativity. Her chapter focuses on the process by which individuals become socialized into groups, specifically concentrating on the ways in which socialization affects the perception of group variability, particularly of the ingroup. According to Ryan et al., new group members initially prefer to pay attention to the similarities of individuals within the ingroup at the expense of their idiosyncrasies. Why is this? According to Ryan, discovering what the common themes are among ingroup members allows new members to deal with their uncertainty and lack of knowledge about the group and maximizes their chances of acceptance. At the intergroup level, such a group-level approach helps them to map the social landscape and to clarify the distinctive features of their ingroup. At the intragroup level, the construction of a schematic understanding of the group provides a means to reassure the other members that one is a good recruit. Ryan’s model is thus essentially predicting a stronger need for a homogeneous view of the ingroup

12

THE PSYCHOLOGY OF GROUP PERCEPTION

at early stages of group membership. As time passes, a greater increase in perceived dispersion is found for ingroups than for outgroups, meaning that outgroup homogeneity should emerge over time. Data from a longitudinal study with new members of sororities lend strong support to these conjectures. Ryan et al. also presents experimental evidence confirming that people who join new groups and are uncertain about their entry in the group primarily search for information about the group rather than about the individual members of the group. As a whole, the data presented by Ryan et al. are unique in examining dynamic impressions of groups over time. They show that forming an individualized view of group members may not be the most functional way to gain access to and become a good member of a group. Instead, constructing some global understanding of one’s group by spotting the similarities is an efficient strategy, at least in the initial phases of membership. Only with the passage of time does the focus on the formation of an adequate representation of the group as a whole give way to some concern for individual differences within the group. Russell Spears and colleagues also focus on homogeneity as a valued characteristic of the group. Adopting a social identity and self-categorization framework, these authors focus on the concept of distinctiveness rather than entitativity. According to them, distinctiveness may well entail a stronger emphasis on the motivational role of the social self and indeed seems to be more directly tied to the consideration of an intergroup context, two features which also have a number of consequences at the measurement level. Having said this, Spears and colleagues set out to examine the different functions of group distinctiveness. In a nutshell, distinctiveness is thought to provide group members with a sense of identity (“who are we?”) as well as with a way of dealing with unfavorable social status (“what can we do?”). As far as the identity function is concerned, Spears and colleagues make a number of suggestions that are strongly reminiscent of other views presented in this volume. Specifically, they note that the perception of the distinctiveness of a group can be as much a reflection on the nature of things as a reaction to the lack of structure and clarity in the social environment. Indeed, empirical evidence is presented showing that group members may sometimes be motivated to differentiate their group from other groups. As to factors that promote such a response, Spears and colleagues suggest that the type of groups and the visibility of other group members play a key role here. Turning to the instrumental function, Spears and colleagues propose that distinctiveness is likely to be useful in the pursuit of group goals that require coordinated group action. Not surprisingly, the presence of a group project should be critical in promoting the willingness of group members to emphasize the coherence and cohesiveness of the group. Interestingly, the authors note that these two functions of distinctiveness are complementary: Distinctiveness/entitativity as ends is a prerequisite for distinctiveness/entitativity as means. Spears and colleagues then go on to present an impressive series of studies that focus on the way people’s level of identification with the group moderates the search for distinctiveness, be it for identity or instrumental reasons. The idea that group entitativity may serve a group in promoting collective action is also very much present in Katherine J.Reynolds, Penelope J.Oakes, Alex Haslam, John C.Turner, and Michelle K.Ryan’s contribution. Still, these authors propose quite a different take on a number of themes that have been tackled in previous chapters. First, like Plaks and colleagues and a few other chapters in this collection, they emphasize that the emergence of entitativity is both a function of characteristics of the target groups and of the perceivers. Concerning the observed groups, the notion of comparative fit is used here to refer to the regularities observed in the social environment. Turning to the perceivers, Reynolds and colleagues stress the role of normative fit and perceiver readiness in shaping people’s tendency to appraise the social situation in terms of intergroup rather than interpersonal relations. Clearly, the notion of normative fit shares a number of features with that of naive theories. As for perceiver readiness or accessibility, one can

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

13

distinguish more chronic preoccupations linked to social identification and more transient concerns deriving from the context. Reynolds and colleagues then go on to propose that structural factors such as the level of power that people enjoy and the nature of the relationship between the groups, i.e., the security of the groups status and the permeability of the group boundaries, are likely to play a major role in the emergence of entitativity. Taking issue with the findings showing that low status groups should be the ones seeing high status groups as heterogeneous, Reynolds and colleagues provide intriguing evidence indicating that low status groups confronted with impermeable and secure relations more readily engage in stereotyping as they derogate the outgroup and see it as a coherent whole to a greater extent than the members of the high status group do. As in the chapters of Guinote and Rubin et al., the work by Reynolds and colleagues makes clear that the perceived homogeneity of a group is highly dependent on the larger social context and that structural variables play a key role in triggering dynamic perceptions of entitativity. In their own way, these last two chapters stress the importance of entitativity in intergroup relations. When observers perceive a group as a coherent whole, they are likely to engage in different behaviors than when they perceive the group as a much less consistent entity. What happens when perceivers draw the conclusion that they are facing a real group rather than a collection of individuals is precisely the question addressed by Wildschut, Insko, and Pinter. Summarizing an impressive program of research, these authors build a case for the idea that groups come across as more threatening than individuals. The repeated observation of the so-called discontinuity effect in mixed-motive situations, i.e., groups are more competitive than individuals, led these authors to investigate in more detail both the content of the outgroup schema as well as the factors that contribute to its activation. As for content, Wildschut and colleagues provide convincing evidence that there is a negative schema of the outgroup consisting of beliefs or expectations that intergroup interactions are competitive, unfriendly, deceitful, and aggressive. So, for instance, communication about future moves in a mixed-motive game is less credible and persuasive when it emanates from groups rather than individuals. People also seem to believe, both at an explicit and at an implicit level, that intergroup interactions are less agreeable and more abrasive than interindividual interactions. They expect interactions with groups to be more competitive than with individuals, and, indeed, discussions involving groups include more statements about distrust than discussions between individuals. Finally, people not only experience interactions with groups as being more competitive but memory comes into play and worsens the picture even more. Turning to the activation issue, Tim Wildschut and colleagues argue that the outgroup schema will be activated whenever an aggregate of individuals is seen as an entity. A key factor in triggering the image of an entitative group is that people’s behaviors and outcomes are seen to be interrelated in producing the groups response, which the authors call procedural interdependence. Such procedural interdependence is of course reminiscent of D.T.Campbell’s (1958) notion of common fate and, Wildschut and colleagues argue, may be fostered by a variety of factors. For instance, when people are confronted with a series of individuals known to rely on consensus in order to reach a group decision, these individuals are likely appraised as a social entity. As a result, fear increases and competition steps up. All in all, the message emerging from this line of work indicates that outgroup fear may be very difficult to reduce. The fact that Wildschut and colleagues paint such a somber picture of entitative outgroups should not be interpreted as a sign that entitativity is a negative feature altogether, one that groups should see as undesirable and try to avoid. As a matter of fact, intuition and empirical evidence alike confirm that entitative outgroups are frequently seen in a negative light (Abelson, Dasgupta, Park, & Banaji, 1998), but this is not the whole story. Often, seeing a group as one coherent block is exactly what perceivers would prefer if collective action is to be initiated or if discrimination and conflict is to be rationalized. It may thus be desirable and

14

THE PSYCHOLOGY OF GROUP PERCEPTION

indeed most important that a group be perceived as an entity. In fact, for quite a few contributions in the book, the take-home message goes exactly in this direction. Interestingly, the distinctive feature is that these chapters all turn their attention to the degree of perceived homogeneity of the ingroup as opposed to the perception of the outgroup or of both groups. A good illustration of this perspective can be found in the chapter by Dominic Abrams, Jose M.Marques, Georgina Randsley de Moura, Paul Hutchinson, and Nicola J.Bown. These authors provide a thorough review of their recent empirical efforts on their subjective group dynamics approach. This line of research, which is an extension of Marques, Yzerbyt, and Leyens’s (1988) earlier work on the black sheep effect, is mainly concerned with people’s reactions to ingroup deviance. The subjective group dynamics model holds that people will simultaneously strive to maximize intergroup distinctiveness and to validate ingroup norms. Whereas the former goal is associated with a high level of category processing, the latter goal implies instead a fair degree of intragroup differentiation. Because people value membership in groups that are entitative, they should feel threatened in the presence of deviant group members who imperil the coherence of the group. Abrams and colleagues present the results of several studies showing that counternormative ingroup deviants are indeed more harshly evaluated than equally counternormative outgroup members. Moreover, they provide evidence that the phenomenon is amplified in group threatening situations. Clearly, a lack of entitativity encourages group members, especially the highly identified ones, to derogate the antinorm deviant, possibly in an attempt to try and restore the perception of group coherence and a sense of subjective reality of the group and the ingroup norms. The idea that people may want to see the ingroup as entitative is taken a step further in Emanuele Castano’s chapter. A first central idea of the chapter is both simple and far-reaching in its implications. If people like to see their ingroup as being rather entitative, they may react more positively to an ingroup that displays coherence as opposed to an ingroup for which there is a lower level of groupness. That is, entitativity occupies the role of an independent variable. Along with a number of colleagues, Castano conducted a series of studies supporting the idea that group members like it when their group is entitative. Specifically, people were found to identify more with groups that were more entitative. A noteworthy feature of this program of research is the use of a wide variety of group features used to trigger entitative perception of the group in the first place. Be it through boundedness, common fate, similarity, or salience, all the features identified by Campbell succeeded in fostering an entitative representation of the ingroup. The second part of Castano’s chapter examines two possible reasons for people’s positive reaction toward an entitative ingroup. The first reason refers to what Castano calls the terrestrial value of entitativity: Groups that are more coherent are endowed with higher intentionality. In other words, there is a sense of purposiveness that entitative groups have that make them threatening if they are outgroups and attractive if they are ingroups. The second reason is the so-called celestial value. Building upon some intriguing empirical research in which terror management theory plays a central role, Castano suggests that membership in a close-knit social entity allows individuals to extend themselves through space and time and releases them from a sense of finitude about their own lives. The chapter concludes with a discussion of the relevance of this work for a series of hot issues in political psychology and international relations. In the next contribution, Michael A.Hogg addresses a number of themes that clearly mirror the issues dwelt upon by Abrams and colleagues and by Reynolds and colleagues. They also extend some of the issues addressed by Castano. The chapter examines the reasons that may lead people to join highly entitative groups, which he calls “totalist” groups. One possible way to account for the seductive power of more entitative as opposed to less entitative groups, Hogg argues, is that such groups are particularly functional in addressing people’s needs to resolve uncertainty. The point of departure of Hogg’s line of reasoning is that,

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

15

contrary to what social identity theory would want us to believe, self-enhancement is likely not the sole reason underlying group identification. As a vast literature indicates, people are also concerned with acquiring a clear sense of who they are, how they are supposed to behave, and what they should expect from others. Social identification with a group that sends out a clear message regarding these different issues offers an ideal means of reducing subjective uncertainty. In other words, the process of depersonalization with respect to group prototypes that are concentrated, focused, simple, and unambiguous constitutes a most effective way to reduce uncertainty. Such group prototypes are typically associated with groups that are highly distinctive and strongly hierarchical, that is, totalist groups. In support of this conjecture, Hogg summarizes a series of experimental and correlational studies suggesting that an increase in or high level of uncertainty leads people to join groups, and in particular to join groups that are highly entitative, and to identify more strongly with those groups. Such findings, Hogg argues, suggest that the documented success of some totalist groups may well derive from people’s attempts to deal with the increased levels of uncertainty they face in the social environment. Given the impact of a group’s entitativity on perceivers, it comes as no surprise that people may want to keep an eye on the kind of image their own group sends out to others. This is the basic question that Richard L.Moreland and Jamie G.McMinn address in their chapter. The authors start by noting that the current meaning of the word entitativity refers mainly to the perception of a group as being real or not, despite the fact that D.T.Campbell coined the word to refer to the actual rather than perceived reality of groups. Given the contemporary emphasis on perception, the authors propose the concept of “social integration” to refer to the objective reality of a group. Building upon the fact that the levels of entitativity and social integration of groups may sometimes correspond but also diverge, Moreland and McMinn take us on a fascinating analysis of the ways by which people may manage how real their groups is seen to be. A number of different situations may lead people to present their group as more entitative and real than it is. Presumably, declining levels of integration, such as recurrent conflicts within a group, be it a large social entity such as a company or a political party or a smaller group such as a couple, may lead some of its members to exaggerate the strength of the group. As the chapters by Castano and Hogg would suggest, an image of a healthy group may also help to gain new members. Moreland and McMinn review a series of cases in which people may instead want to make the group seem less real that it actually is. Indeed, bot dyadic associations, such as extramarital affairs, office romances, or homosexual relationships, and groups involving larger numbers of people, such as subversive groups, provide evidence aplenty that people are not always keen to let others notice the existence of links among them. Interestingly, although efforts may be made on the part of group members to manage the impression others may have regarding their group, Moreland and McMinn also consider the possibility that group members are sometimes less concerned with the perception of the groupness than with the actual level of social integration of their group. Ironically, a prime strategy here may well be to rely on entitativity. The fact that human beings are social creatures by nature, the authors note, should indeed make it more easy to persuade observers about the presence than the absence of a group. This final chapter confirms, if this was still needed, that entitativity is a most powerful determinant of people’s reactions to social targets. FINAL THOUGHTS Over the last decade, scholars of intergroup relations have made impressive progress in their understanding of the antecedents and consequences of perceived group variability. More recently, this interest generalized to the study of perceived group entitativity and essentialism. Empirical and theoretical advances on the latter two concepts have significantly improved our understanding of important social issues such as the

16

THE PSYCHOLOGY OF GROUP PERCEPTION

development and maintenance of prejudice and discrimination toward groups or the nature of social behavior directed toward (or emerging from) ingroup members and newcomers. As research developed at an ever increasing rate on these notions, it became clear that efforts toward conceptual integration should be made. As we already noted, current evidence suggests that perceived variability, entitativity, and essentialism tend to overlap with each other, both at the conceptual and empirical levels, yet they are also distinct constructs in many ways. Importantly but problematically, scholars in the field have both defined and operationalized these constructs in a variety of different ways. Inevitably, this divergence has meant that research conducted by one laboratory on the relations among these three constructs might be less than useful to other laboratories, where different definitions and operationalizations were used. This distressing state of affairs meant that laboratories were speaking past each other rather than to each other, with the consequence that little conceptual progress was being made through abundant data were accumulating. It was this sad state of affairs that prompted us to edit this volume, with the goal of providing a forum where the diverse points of view could speak to each other. It seemed to us that only through such an exchange could we arrive at some common definitions and perspectives, and that these would lead hopefully to some substantial scientific progress on these concepts. In addition to addressing these theoretical and methodological concerns, we also wanted to document the impressive amount of work that has already been conducted on the role of perceived variability, entitativity, and essentialism in social perception and intergroup relations. We chose to organize these research contributions by focusing on the extent to which the social perceiver is seen to be relatively passive or active in the social construction of group perception. At the one end, perceptions of group variability and entitativity are seen to derive from structural conditions that affect perception and to which the perceiver simply responds. At the other end, perceived variability and entitativity are affected by motivated and even strategic processes that the perceiver actively engages in. Certainly these chapters could have been organized in other ways, along other dimensions. One advantage of the present structure, however, is that it makes clear that the perception of group variability, entitativity, and essentialism can have both negative and positive social consequences. For instance, enhanced perceived entitativity and essentialism facilitate people’s social identification to, or socialization within, groups. However, the search for social identification and integration may occasionally lead people to join totalist groups, or it can also more generally contribute to the development not only of ingroup favoritism but also of outgroup derogation. In this sense, we need to distinguish between the processes involved in perceiving variability, entitativity, and essentialism and their consequences. The processes per se really do not have evaluative implications. That is, perceptions of group variability, entitativity, and essentialism in and of themselves, and the factors that influence these, are neither desirable nor undesirable. But it is the content and interpretation that are put upon these perceptions by the social world that have evaluative implications. Seeing a group as less variable and more entitative is a good thing if one is a member of the group and if one is able to take collective action in pursuit of desirable ingroup goals. On the other hand, outgroups that are less variable, more entitative, and more essentialized are more likely to be seen as effective in acting on their goals and thus, to the extent that there exists conflicts between groups, more likely to be targets of discrimination and intergroup hostility. The various contributions to this book represent a fascinating theoretical and empirical journey into the word of intergroup relations. Our hope is that this volume will stimulate further theoretical and empirical work within social psychology on the fundamentals of group perception. The issues that are raised in this volume extend beyond social psychology, however, and therefore we hope that neighboring disciplines may also find the issues addressed to be provocative. Indeed, one of the obvious lessons of the advances made in the study of group perception is its potential relevance both in terms of concepts and methods for such fields

PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM

17

as political science, sociology, and anthropology. Finally, we also believe that the chapters assembled here will prove useful for social practitioners and decision makers who are generally concerned about social beliefs and interactions. The goal of improving intergroup relations is one that requires efforts from both academics and community-based organizations. We ultimately hope that the present contribution will foster exchanges between these social groups as well.

Section 1 Constructs and Definitions

2 Dynamic Entitativity Perceiving Groups as Actors MARILYNN B.BREWER Ohio State University YING-YI HONG University of Illinois QIONG LI University of Maryland

When D.T.Campbell (1958) introduced the term entitativity to the social science literature, he was interested in the ontological status of social groups as units of perception. The criteria he suggested— similarity, proximity, common fate, (correlation), prägnanz (closure)—were derived from principles of Gestalt psychology and can be understood in two different ways. On the one hand, these criteria can be viewed as properties (actual or perceived) of groups themselves—properties that underlie or legitimate the perception of an aggregate as an entitative social unit. In this spirit, much of the initial empirical research on group entitativity has been devoted to identifying (e.g., Lickel et al., 2000; Harasty, 1996) or manipulating (e.g., Dasgupta, Banaji, & Abelson, 1999; L.Gaertner & Schopler, 1998; S.L.Gaertner et al., 1989; McGarty, Haslam, Hutchinson, & Grace, 1995; Welbourne, 1999) properties of social groups that lead to higher or lower ratings of those groups as “real” social entities. An alternative perspective on Campbell’s criteria for entitativity is to think of them as theories that individual perceivers hold about specific social groups and that endow the groups with social meaning and predictive value (Murphy & Medin, 1985; McGarty et al., 1995). From this perspective, entitativity can be defined as a theory of common origin underlying expected similarities of attitude or behavior on the part of group members. In other words, groups are seen as “real” groups because some aspects of member behavior are believed to arise from some common source. Origin theories provide explanations for group behavior and at the same time afford the boundaries and coherence that make the group a social unit. In this view, entitative social groups are always characterized by homogeneity, similarity, and consistency in some respect, but the question remains, with respect to what? As explanations for behavior, the roots of group entitativity can be further divided into “causes” (distal or proximal) and “reasons” (Malle, 1999). Table 2.1 depicts our taxonomy of origin theories as defined by the nature of the common source underlying group behavior, the type of explanation it represents, and the

20

THE PSYCHOLOGY OF GROUP PERCEPTION

dimension of similarity that would characterize membership in that group. Common attributes are innate, internal dispositions (e.g., genetic characteristics, personality traits, psychiatric disorders) that are shared by all group members and are believed to give rise to similarity and consistency among group members in specific trait-related behaviors or behavioral orientations. When entitativity is based on an implicit theory of common attributes in this sense, intragroup homogeneity is believed to arise from internal, stable causes that are common to all group members. Common history, unlike person attributes, refers to causes of behavior that are external to the individual— past events or experiences that have been shared by all group members, including, for example, a common ancestry, cultural socialization, or life event. Depending on the nature of the shared experience, it may be expected to give rise to similarities in attitudes, values, or physical appearance. In any case, the common history provides a distal (situational) cause for current similarities in (internal) dispositions and associated behaviors.1 Whereas common history refers to past experiences, common fate refers to current or future outcomes that befall all members of the social group by virtue of their group membership. Common fate gives rise to status (or role) similarity among group members that is believed to shape or constrain their behavior. Sociopolitical categories associated with specific social policies (e.g., persons over 65 years old, apartheid categories, labor unions) would be characterized as sharing a common fate, as would groups defined by proximity, such as residents of geographic regions subject to particular meteorological or environmental TABLE 2.1. A Typology of Theories of Group Entitativity Common Origin

Type of Explanation

Domain of Similarity

Common attribute(s) Common history Common fate Common purpose

Person cause Causal history Situational cause Reason

Traits and trait-related behavior Values, attitudes, appearance Status, role Motives, intentions

conditions. Entitativity based on common fate provides an immediate situational cause for commonalities among group members and similarity of behavior. Finally, groups can derive their meaning from common purpose, shared intentions and goals that give rise to collective action. The common purpose provides a reason for group members’ behaviors, which are perceived to be driven by similarity in motives even in the face of diversity of traits, roles, and actions. Common purpose groups are most likely to be characterized by organization and structure (Hamilton, Sherman, & Lickel, 1998), and this is the type of entitativity associated with perceptions of groups as active collective agents with intentions, motives, and goal-directed behavior (Abelson et al., 1998). Implicit theories of common purpose provide a reason (rather than causal explanation) for similarities or consistency in group members’ actions. The different sources of group entitativity depicted in Table 2.1 are by no means mutually exclusive. A specific social group or category can have more than one basis of common origin, and we might expect the different sources to combine additively or multiplicatively to determine the degree of perceived entitativity of the group as a whole. Ethnic groups, for example, are defined primarily in terms of common history/ ancestry, but within many social-political environments they are also characterized by common fate, the perception of innate, common attributes, and the presence of common goals and collective action. Origin theories might also be implicitly correlated. Abelson et al. (1998), for instance, suggest that common attributes may give rise to inferences of common intentions and motives. Nonetheless, although the

DYNAMIC ENTITATIVITY

21

different origins of entitativity may combine and interact, our focus in this chapter will be on the differences among the various types as distinct theories of group entitivity. ESSENCE VS AGENCY: TWO VERSIONS OF ENTITATIVITY The four bases of common origin listed in Table 2.1 have been arrayed along a dimension of “essence” versus “agency” as theories about groups. The distinction between essence and agency is related to the distinction between “entity theory” and “incremental theory” as implicit theories about individual abilities and personal character (Chiu, Hong, & Dweck, 1997; Dweck, Chiu, & Hong, 1995). Entity theorists believe that abilities and personality reflect innate characteristics that are fixed and immutable, whereas incremental theorists believe that the underlying bases of abilities and behavior are malleable and change and develop over time and circumstance. At the group level, essentialist models assume that entitative groups are characterized by a shared “essence” that is fixed, inherent, and immutable (Haslam, Rothschild, & Ernst, 2000; Yzerbyt, Rocher, & Schadron, 1997). The extreme form of essentialism is represented by common attribute categories that are perceived as “natural kinds” (Rothbart & Taylor, 1992). At the other end of the spectrum are theories about entitative groups that emphasize their status as actor-agents. From this perspective, aggregates “become” groups (rather than “are” groups) by virtue of recognizing shared goals and purposes and organizing or banding together for collective action to achieve those goals. Because motives, organization, and action are temporal and malleable (rather than innate and fixed), agency theories of group entitativity correspond to incremental implicit theories at the individual level. Table 2.2 contrasts essence and agency as theories of group entitativity. Drawing on parallels with entityincremental theories of social cognition, essence and agency as theories of group entitativity are expected to set up different processing goals and modes of analysis. Specifically, within the framework of essence theory, perceivers would be motivated to diagnose and evaluate fixed characteristics as the basis for making judgments and decisions about the group and predicting member behavior. In contrast, within the framework of agency theory, perceivers would be motivated to understand the temporal dynamics of group behavior and to predict behavior based on perceived goals in relation to the current situation. With such distinct processing orientations, essence vs. agency theory might orient perceivers to analyze groups differently. Table 2.2 contrasts the analytical processes of essence vs. agency theory on four dimensions. First, in terms of unit of analysis, an essence theory would orient perceivers to focus on personality traits and trait-related units. That is, within this theory framework, groups are perceived and analyzed in terms of the innate, fixed personality traits of their members (e.g., “Blacks are athletic”). Essentialistic theories underlie the formation and use of social stereotypes (Yzerbyt, Rocher, & Schadron, 1997; Yzerbyt, Rogier, & Fiske, 1998). In contrast, an agency theory would orient perceivers to focus on the psychological processes (e.g., goals and values) of group members when they are acting on behalf of the group, which are sensitive to situational changes and account for dynamic group processes. TABLE 2.2. Contrasts between Essence Theory versus Agency Theory Essence Theory Unit of Analysis Personality traits; stereotypes Intragroup similarity Ingroup consistency; look for similarities among group members and consistency of behavior

Agency Theory Psychological processes (e.g., goals, values) Ingroup heterogeneity: look for variations and heterogeneity among group members

22

THE PSYCHOLOGY OF GROUP PERCEPTION

Essence Theory Intergroup similarity Look for clear intergroup distinctions, definition of the boundary of the target group Temporal dimension Look for temporal consistency: e.g., focus on a group’s historical background or traditions

Agency Theory Look for relationships with outgroups; ways to locate the target group in a larger social structure Look for dynamic change: e.g., focus on a group’s future and development

In terms of intragroup similarity, essence theory would have perceivers look for similarities and consistency in behavior among group members as diagnostic of their common underlying attributes. In contrast, agency theory would have perceivers look for variations in behavior and variability among group members as diagnostic of underlying goals and responsiveness to changing situations and functional needs. Coordination, rather than similarity, of behavior is the hallmark of group agency. In terms of intergroup similarity, an essence theory would have perceivers accentuate the differences among groups so as to establish clear group boundaries. With an agency theory, perceivers would be more likely seek to understand the relations among groups and to locate specific groups in a larger social structure rather than as distinct individual units. Finally, in terms of the temporal dimension, essence theory would have perceivers conduct a search into the group’s history so as to locate its common, shared ancestry, history, or traditions. In contrast, agency theory would lead to a search into the future of the group, what it will become (i.e., the development of the group). It should be made clear that we think of essence and agency as two different bases for judgments of group entitativity (even though any one individual may use both theories). This is in contrast to Rothbart and Park (this volume), who posit a single continuum of entitativity (group boundedness) in which common goals are seen as intermediate between common attributes and “essence” in terms of degree of entitivity. In our model, essence and agency reflect two different meanings of the entitativity concept, one based on perceived common attributes, the other based on perceived common goals and intentions. A particular group can be perceived as high or low in entitativity on either basis. Attribute similarity alone does not correspond to high entitativity, unless the attributes have been essentialized. Similarity of goals and purposes alone does not correspond to high entitativity, unless the group is perceived as acting in concert on their common goals. These distinctions between essentialist and agency notions of group entitativity help explain why entitativity judgments are associated only imperfectly with any specific group properties such as similarity or longevity (Lickel et al., 2000). Which group characteristics will be diagnostic will depend on what the underlying basis for judged entitativity is presumed to be. The postulated differences in processing orientations associated with different meanings of group entitativity underscore the role of implicit theories in determining what information about groups is attended to and how group membership is used to explain individual group members’ behaviors and to predict future behavior. The distinction between essentialist and agency theories of group entitativity crosscuts the typological distinctions identified by Lickel et al. (2000) among groups of different size and function (intimacy groups, task groups, social categories, and loose associations). Groups of any type may be viewed in terms of static common properties and similarities or in terms of their dynamic goals and purposes. The judged entitativity of any particular group will depend on which lens is being used. Members of a family, for instance, may be seen as sharing essential features by virtue of heredity and common experience, but if they rarely interact as a group or coordinate their activities, that family will not be seen as a unit in the actor-agency sense. Large

DYNAMIC ENTITATIVITY

23

social categories are usually defined in terms of similarity of characteristics, but nation-states as political units are social actors in a very real sense. In terms of consequences of perceived entitativity, it is our contention that static, essentialist theories of groupness have implications primarily for the perception and evaluation of group members as individuals and of the group prototype (Brewer & Harasty, 1996). These are the processes characteristically studied in social cognition research on stereotypes and stereotyping (e.g., Macrae, Stangor, & Hewstone, 1996). For those who hold essentialist theories of group entitativity, knowledge of group membership provides significant information about the character of individual group members and justifies treating individuals in accord with group stereotypes. For those who hold dynamic agency theories of group entitativity, on the other hand, the nature of groups and group boundaries should be viewed as malleable and changeable over time. A group is a “group” only so long as the collective intent and purpose is salient and shared, and groups can dissolve and re-form as purposes and interests change. Attributions are made to the group as a collective actor, but group membership is not necessarily a meaningful cue to individual characteristics outside of the group context. In this framework, attitudes about the group as a whole may or may not generalize to feelings and evaluations of individual group members. Further, the goals and intents attributed to the group may or may not be linked to essential traits or characteristics of its members. It is possible to attribute aggressive intent to a group without believing that aggressiveness is an essential trait shared by individual group members. Actor-agency theories of groups have more implications for emotional and behavioral responses to the group as a whole (M.G.Alexander, Brewer, & Herrmann, 1999; Brewer & Alexander, 2002). Attributing motives, intent, and purposive behavior to a group as a unit leads to responding to the group as if it were a single individual, but on a larger scale. Given the larger-than-life property of a group acting in concert, it is not surprising that attributions and emotional reactions are magnified in response to perceived entitativity in this sense (Abelson et al., 1998; Dasgupta, Banaji, & Abelson, 1999, Thakkar, 2000). Further, the consequences of perceiving a group as a dynamic entity will vary, depending on whether the group is an ingroup or an outgroup. Agency perception of one’s own ingroup enhances both security and efficacy and hence should be associated with high ingroup identification and positivity (see Castano, this volume). The consequences of outgroup agency, on the other hand, depend on the nature of the relationship between ingroup-outgroup goals, and their relative status and power (Alexander et al., 1999). A trusted outgroup would be more valuable as an ally if it is also perceived as an entitative actor-agent. However, a distrusted or disliked outgroup becomes more threatening and dangerous to the extent that it is perceived as a dynamic entity or collective actor (Dasgupta et al., 1999). Static traits or essences that are attributed to the outgroup may contribute to the initial evaluation of the group as a potential friend or enemy, but it is the dynamic attributes of the group-as-actor that determine whether the outgroup can benefit or harm the ingroup. Thus, the study of the relationship between group cognition and intergroup relations first requires an understanding of whether an essentialist or agency theory of group existence has been activated in any given case. Essentialist notions of group entitativity are covered extensively in other chapters in this volume. The remainder of this chapter will be devoted to further discussion of defining entitativity in terms of agency theories of groups. PERCEIVING GROUPS AS ACTORS: WHEN AND WHY Activating dynamic agency theories of group entitativity is a joint function of permanent and situational properties of the group and of chronic and temporary beliefs and motives of the individual perceiver.

24

THE PSYCHOLOGY OF GROUP PERCEPTION

Certainly many of the properties identified by D.T.Campbell (1958) as cues that perceivers might use in assessing whether an aggregate of individuals constitutes a meaningful group could be considered indicative of common purpose and intent. In particular, proximity and covariation of movement in time and space are dynamic properties that elicit perceptions of the group as a coordinated, acting unit (Dasgupta et al., 1999). However, outside of the behavior of sports teams and angry mobs, coordination is not always so visibly evident as a property of groups. Groups can share purposes and organize action even at a distance, with or without formal communication networks. Thus, the dynamic properties of shared intention and coordination may be psychological states and relationships among group members rather than formal organizational structures. As such, the degree of dynamic entitativity perceived may be as much a product of the beliefs and motives of the perceiver as of the reality of the group itself (Brewer & Harasty, 1996). In the political domain, the portrayal of outgroup nations or political parties as monolithic actors with malevolent intent and organized purpose is often used strategically to mobilize ingroup cohesion, loyalty, and collective action. There is experimental evidence that this strategy is effective in that manipulation of the perceived entitativity of an outgroup is sufficient to enhance perceptions of the outgroup’s power and resources and to increase perceived ingroup entitativity and the level of ingroup identification (Castano, Sacchi, & Gries, in press). The use of this strategy in intergroup contexts may have self-fulfilling effects. If outgroup entitativity increases perceived cohesion and identification with the ingroup, it creates the conditions under which members of the ingroup are most likely to behave as a collective actor with a common purpose. Apart from strategic motives, implicit theories about the world (Dweck, Chiu, & Hong, 1995; Chiu, Dweck, Tong, & Fu, 1997) may increase the likelihood that social groups are perceived as actor-agents with intents and purposes. Specifically, to the extent that individuals perceive the social world and its institutions as malleable, they are likely to focus on how social groups change and develop and so imbue social groups with the properties of active agents. In contrast, to the extent that perceivers hold the view of a fixed social world, they may focus on identifying the static attributes that define the groups (such as the common attributes, common history, and common appearance among members) rather than the dynamic aspects (such as common purpose). This tendency to endorse the essentialist theory of group entitativity should be especially intense if the fixed worldview is coupled with a belief in low efficacy of social groups in causing important outcomes (i.e., low group efficacy belief). In this case, the fixed world may be seen as consisting mostly of social groups that are formed according to traditions or common practices, but seldom function as effective actor-agents. The role of the individual in this system might be to navigate a way through these fixed structures and get the best out of them. Previous research has also shown evidence of cross-cultural differences in the endorsement of the fixed vs. malleable theories of the world. Specifically, Chiu and Hong in several studies have consistently found that the Chinese samples hold a more fixed (vs. malleable) belief about the world than did the American samples (Chiu & Hong, 1999; see review in Levy, Plaks, Hong, Chiu, & Dweck, 2001). These crosscultural differences in implicit theories of the world have been shown to link to American people’s tendency to shape social structures to accommodate individuals’ needs and Chinese people’s tendency to accommodate to the social structure (Su, Chiu, Hong, Leung, Peng, & Morris, 1999). In perceiving social groups, American participants might also endorse the actor-agent theory of group entitativity. In contrast, Chinese participants might endorse the essentialist theory of group entitativity, especially those who also think that the social groups have low efficacy in causing important outcomes (i.e., low group efficacy belief). These ideas were tested in a study that we will discuss next.

DYNAMIC ENTITATIVITY

25

DYNAMIC ENTITATIVITY ACROSS CULTURES: A PRELIMINARY INVESTIGATION As an initial exploration of possible cultural differences in the perception of groups as entities—particularly as actor-agents—we undertook a survey study among college students in Hong Kong and in the United States. During the same time period at the beginning of the fall 2000 academic term, 101 students at the Hong Kong University of Science and Technology and 246 students at the Ohio State University completed questionnaires assessing their perceptions of specific social groups to which they belonged.2 The target groups included in the survey questionnaires were selected to represent a range of group size and type. Specifically, the groups chosen for investigation were (1) immediate family (intimacy group), (2) a social club or other small group affiliation (intimacy or task group), (3) university (task group or social category), and (4) nation (social category or loose association). Individual respondents completed the full questionnaire for two of these groups (either family and university or small-group and nation), with order counterbalanced. The questionnaires were written in Chinese for the Hong Kong sample and in English for the U.S. sample. Entitativity Measures Sets of items in the questionnaire were written to assess three different entitativity-related aspects of group perceptions. One set of items assessed dynamic properties; a second set assessed relatively static properties and essentialist beliefs; and the third measure was a judgment of the general “groupness” of the group as a whole. Static Characteristics and Essence. The perception that members of a group have “common attributes” is one basis for group entitativity (see Table 2.1). However, high attribute similarity does not necessarily imply high perceived entitativity in and of itself, but only when these attributes are seen as fixed, innate characteristics that reflect the “core essence” of the group. In order to tap this essentialist belief directly, two items in the questionnaire assessed the extent to which the group was perceived as having such fixed characteristics and a core essence. These two items constituted our measure of perceived essence ( =.68). Dynamic (Agency) Characteristics. Each group was also rated by selected dynamic properties, including common fate (future), shared purpose, group cooperation, and effectiveness as a group. These four items were intended to represent the bases of group entitativity associated with an actor-agent theory of entitativity. Ratings of each group on these five items were averaged together to form our measure of perception of the group-as-actor ( =.82). Entitativity Rating. Finally, two items were included in the questionnaire to create a general measure of perceived entitativity of the group. These items were (1) the extent to which the group should be “thought of as a whole” (rather than a collection of individual members), and (2) the extent to which the group is a coherent unit ( =.82). All ratings were made on an 11-point scale, ranging from 0 to 10, with 5 as the midpoint. Implicit Theories of the Social World and Group Efficacy Beliefs In addition to the measures of target groups, the questionnaire included a measure of the implicit theories about the social world developed by Dweck et al. (1995). A revised version of the measure consisted of four items that depict a fixed view about the world (e.g., “Though we can change some social phenomena, it is unlikely that we can alter the core characteristics of our social world.” “Our social world has its basic or ingrained characteristics, and you really can’t do much to change them”), and four items that depict a

26

THE PSYCHOLOGY OF GROUP PERCEPTION

malleable view about the world (e.g., “We can change some social phenomena, and even the core characteristics of our social world.” “Just as societal trends can change, so can the fundamental nature of our social world”). Participants were asked to rate the extent of their agreement with the statements from 1, not at all, to 6, very much. The internal reliability of the scale derived from the eight items (with malleability items reverse scored) was =.83. The efficacy scale (developed by Hong and Wong, 2000) consisted of 12 items asking the respondents’ agreement with statements such as “The will of individuals is the most powerful force in society,” “Social groups and organizations influence what happens in an individual’s life,” “What happens in an individual’s life is of his or her own making,” and “Social groups and organizations take control of the situations around them and exercise free will.” Six of the items referred to individual will and influence, and six of the items referred to group will and influence, creating individual efficacy and group efficacy subscales, respectively. These measures were intended to tap chronic individual beliefs that were hypothesized to be related to perception of groups as static or dynamic entitities. Cultural Differences in Perceiving Groups as Entities: Some Findings Our first analyses of the survey data took a look at mean differences on the three entitativity-related measures as a function of respondent culture and target group, and aggregated across all four target groups within each culture. The relevant means are reported in Table 2.3. As is evident from inspection of the table. Hong Kong respondents consistently rated the target groups lower in overall entitativity than did U.S. respondents. Across all target groups, the main effect of culture on entitativity ratings was statistically significant. However, given issues of comparability due to translation and possible differences in the use of response scales, interpreting mean differences across cultures in an absolute sense is tenuous at best.3 Of more interest are differences between cultures in the relative ratings on the different measures, as well as differences within cultures in the perceived entitativity of the various target groups. In both cultures, large social categories (university and nation) were rated as lower in entitativity than the family group. However, Hong Kong respondents made a sharp distinction between the family as a rela TABLE 2.3. Mean Entitativity Ratings by Culture and Target Group Hong Kong Family Small Group University Nation Aggregated United States Family Small Group University Nation Aggregated

Entitivity

Dynamic

Essence

6.93 6.24 6.06 6.04 6.32

6.00 6.08 5.65 5.78 5.88

6.94 5.95 6.24 6.31 6.36

7.69 7.70 6.68 6.67 7.18

7.36 7.41 6.72 6.20 6.92

7.53 6.90 6.63 6.78 6.96

DYNAMIC ENTITATIVITY

27

tively high entitative group and all other types of groups, including small social clubs. U.S. respondents, on the other hand, rated the two types of small, interactive groups essentially the same, and relatively high on entitativity and dynamic properties especially. Of more interest to our purposes than mean ratings of entitativity were the relationships among the entitativity measures within and between cultures. More specifically, we were interested in the extent to which perceptions of a group in terms of essence or actor would predict the overall entitativity rating of that group. For each target group within each culture, multiple regression was used to assess the relative weight of dynamic (actor) versus static (essence) group TABLE 2.4. Multiple Regressions: Predictors of Entitativity Ratings Hong Kong Family Small Group University Nation Aggregated United States Family Small Group University Nation Aggregated Note: *=p