Advances in Group Processes 9781785600760, 9781785600777

Citation preview

ADVANCES IN GROUP PROCESSES

ADVANCES IN GROUP PROCESSES Series Editors: Edward J. Lawler and Shane R. Thye Recent Volumes: Volumes 1
17:

Edited by Edward J. Lawler

Volume 18:

Edited by Edward J. Lawler and Shane R. Thye

Volume 19:

Group Cohesion, Trust and Solidarity
Edited by Edward J. Lawler and Shane R. Thye

Volume 20:

Power and Status
Edited by Shane R. Thye and John Skvoretz

Volume 21:

Theory and Research on Human Emotions
Edited by Jonathan H. Turner

Volume 22:

Social Identification in Groups
Edited by Shane R. Thye and Edward J. Lawler

Volume 23:

Social Psychology of the Workplace
Edited by Shane R. Thye and Edward J. Lawler

Volume 24:

Social Psychology of Gender
Edited by Shelley J. Correll

Volume 25:

Justice
Edited by Karen A. Hegtvedt and Jody Clay-Warner

Volume 26:

Altruism and Prosocial Behavior in Groups
Edited by Shane R. Thye and Edward J. Lawler

Volume 27:

Edited by Shane R. Thye and Edward J. Lawler

Volume 28:

Edited by Shane R. Thye and Edward J. Lawler

Volume 29:

Edited by Will Kalkhoff, Shane R. Thye and Edward J. Lawler

Volume 30:

Thirtieth Anniversary Edition
Edited by Shane R. Thye and Edward J. Lawler

Volume 31:

Edited by Shane R. Thye and Edward J. Lawler

ADVANCES IN GROUP PROCESSES VOLUME 32

ADVANCES IN GROUP PROCESSES EDITED BY

SHANE R. THYE Department of Sociology, University of South Carolina, SC, USA

EDWARD J. LAWLER School of Industrial and Labor Relations and Department of Sociology, Cornell University, NY, USA

United Kingdom
North America
Japan India
Malaysia
China

Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2015 Copyright r 2015 Emerald Group Publishing Limited Reprints and permissions service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. Any opinions expressed in the chapters are those of the authors. Whilst Emerald makes every effort to ensure the quality and accuracy of its content, Emerald makes no representation implied or otherwise, as to the chapters’ suitability and application and disclaims any warranties, express or implied, to their use. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-78560-077-7 ISSN: 0882-6145 (Series)

ISOQAR certified Management System, awarded to Emerald for adherence to Environmental standard ISO 14001:2004. Certificate Number 1985 ISO 14001

CONTENTS LIST OF CONTRIBUTORS

vii

EDITORIAL ADVISORY BOARD

ix

PREFACE

xi

BEYOND THREE FACES: TOWARD AN INTEGRATED SOCIAL PSYCHOLOGY OF INEQUALITY Jane D. McLeod, Tim Hallett and Kathryn J. Lively

1

SEQUENCE-NETWORK ANALYSIS: A NEW FRAMEWORK FOR STUDYING ACTION IN GROUPS Benjamin Cornwell and Kate Watkins

31

EXPECTATION STATES, SOCIAL INFLUENCE, AND AFFECT CONTROL: OPINION AND SENTIMENT CHANGE THROUGH SOCIAL INTERACTION Kimberly B. Rogers

65

HOW DOES STATUS AFFECT POWER USE? NEW PERSPECTIVES FROM SOCIAL PSYCHOLOGY Ko Kuwabara

99

RECRUITING SOURCE EFFECTS: A SOCIAL PSYCHOLOGICAL ANALYSIS Richard L. Moreland

123

ON IDENTIFYING HUMAN CAPITAL: FLAWED KNOWLEDGE LEADS TO FAULTY JUDGMENTS OF EXPERTISE BY INDIVIDUALS AND GROUPS David Dunning

149

v

vi

CONTENTS

CONSTRUCTIVE LEADERSHIP ACROSS GROUPS: HOW LEADERS CAN COMBAT PREJUDICE AND CONFLICT BETWEEN SUBGROUPS Michael A. Hogg

177

COLLABORATION AMONG HIGHLY AUTONOMOUS PROFESSIONALS: COSTS, BENEFITS, AND FUTURE RESEARCH DIRECTIONS Heidi K. Gardner and Melissa Valentine

209

LIST OF CONTRIBUTORS Benjamin Cornwell

Department of Sociology, Cornell University, NY, USA

David Dunning

Department of Psychology, Cornell University, NY, USA

Heidi K. Gardner

Harvard Law School, Harvard University, MA, USA

Tim Hallett

Department of Sociology, Indiana University, IN, USA

Michael A. Hogg

Department of Psychology, Claremont Graduate University, CA, USA

Ko Kuwabara

Columbia Business School, Columbia University, NY, USA

Kathryn J. Lively

Department of Sociology, Dartmouth College, NH, USA

Jane D. McLeod

Department of Sociology, Indiana University, IN, USA

Richard L. Moreland

Department of Psychology, University of Pittsburgh, PA, USA

Kimberly B. Rogers

Department of Sociology, Dartmouth College, NH, USA

Melissa Valentine

Department of Management Science, and Engineering, Stanford University, CA, USA

Kate Watkins

Department of Sociology, Cornell University, NY, USA

vii

This page intentionally left blank

EDITORIAL ADVISORY BOARD Stephan Bernard Indiana University, USA

David Melamed University of South Carolina, USA

Karen Hegtvedt Emory University, USA

Jane Sell Texas A&M University, USA

Michael Hogg Claremont Graduate University, USA

Robb Willer Stanford University, USA

Will Kalkhoff Kent State University, USA

ix

This page intentionally left blank

PREFACE Advances in Group Processes publishes theoretical analyses, reviews, and theory based empirical chapters on group phenomena. The series adopts a broad conception of “group processes.” This includes work on groups ranging from the very small to the very large, and on classic and contemporary topics such as status, power, trust, justice, social influence, identity, decision-making, intergroup relations, and social networks. Previous contributors have included scholars from diverse fields including sociology, psychology, political science, economics, business, philosophy, computer science, mathematics, and organizational behavior. The volume opens with a fresh take on Jim House’s famous “three faces” of social psychology article. Jane D. McLeod, Tim Hallett, and Kathryn J. Lively propose a new framework from the social structure and personality face of social psychology in “Beyond Three Faces: Toward an Integrated Social Psychology of Inequality.” This analysis focuses on the micro-to-macro dimensions of social inequality in groups. The approach connects structural and cultural dimensions systems, local contexts and the lived experiences of individuals. Overall, the chapter should help to promote integration and cross-fertilization across the diverse traditions of social psychology. The next three chapters address aspects of networks and structures in producing activity and influence in groups. The first chapter is “SequenceNetwork Analysis: A New Framework for Studying Action in Groups” by Benjamin Cornwell and Kate Watkins. These authors examine the daily activity patterns of both employed and unemployed people. Using over 13,000 24-hour time diaries from the 2010
2013 American Time Use Survey, the authors find that employed and unemployed people participate in significantly different types of activities, and that unemployed individuals engage in much less synchronized behavior. An important methodological advancement of this work is that it uses network-analytic methods to visualize behavior over time. Next, Kimberly B. Rogers examines three complementary theories in “Expectation States, Social Influence, and Affect Control: Opinion and Sentiment Change through Social Interaction.” This chapter uses Expectation States Theory, Affect Control Theory, and Social xi

xii

PREFACE

Influence Network Theory to examine opinion change that is driven by affective impressions and performance expectations. Rogers offers new experimental evidence to bear on hypotheses at the intersection of these theories. An important aspect of this paper is that it examines how behavioral interchange patterns (BIPs) drive opinion change in a novel and creative way. Using these theories together Rogers shows that group members’ opinions were pulled toward the group leader’s opinion irrespective of BIPS. This work adds to the growing body of work that seeks to integrate fundamental theories of group processes. Also, integrating research across normally distinct areas Ko Kuwabara suggests new insights regarding the connection between power and status in “How Does Status Affect Power Use? New Perspectives from Social Psychology.” This chapter examines three variables that might moderate the effect of status on the use of power: the legitimacy of status, achieved versus ascribed status, and the presence of an individualistic versus collectivist culture. Overall, the chapter sheds light on the often contingent connection between power and status will be of interest to sociologists and those in organizational psychology. The remaining chapters in this volume all use facets of social psychology to explore problems and processes common to modern organizations. Richard L. Moreland explores job success in “Recruiting Source Effects: A Social Psychological Analysis.” Specifically, Moreland reviews the theoretical and empirical literature on recruiting sources
that is, when a current employee sponsors a new recruit in the workplace. Traditionally, the two most popular notions regarding the benefits of sponsorship are that (i) new employees have more realistic job expectations and that (ii) sponsored employees are of higher quality. Moreland offers a third explanation − that sponsorship brings with it social pressures to perform such as attempts to repay the sponsor. This is a highly creative social psychological approach to the problem that will certainly inspire new research in social psychology and organizations. Next, David Dunning provides a selective review of theory and research regarding how individuals judge human capital in “On Identifying Human Capital: Flawed Knowledge Leads to Faulty Judgments of Expertise by Individuals and Groups.” The chapter focuses on theory and research surrounding the “Dunning-Kruger” effect, that is, the tendency for incompetent people to not recognize their own incompetence. Dunning applies this to the problem of evaluations in organizations and finds that because evaluators are often flawed they fail to recognize the brilliance of others and their superior ideas. This chapter uses basic social psychological principles to analyze a problem faced by virtually all organizations.

Preface

xiii

Michael A. Hogg presents a new theory of leadership in “Constructive Leadership across Groups: How Leaders Can Combat Prejudice and Conflict between Subgroups.” The chapter summarizes the social psychological research related to reducing intergroup conflict. Hogg then offers a new theory that shows leaders how to reconstruct a common social identity and reduce prejudice and conflict between groups. The result is a highly creative and novel approach to conflict resolution that is grounded firmly in the known principles of social identity theory. This chapter should especially interest anyone who is faced with warring factions − such as any department head, chair, or dean. Finally, the volume closes with “Collaboration among Highly Autonomous Professionals: Costs, Benefits, and Future Research Directions” by Heidi K. Gardner and Melissa Valentine. This chapter explores the tendency to collaborate among powerful and highly autonomous peers. Whereas traditionally teamwork offers many benefits, Gardner and Valentine show that this is not necessarily the case in these sorts of collaborations. Using qualitative data from three professional service firms they find that often times collaborative efforts fail or are never attempted. This chapter offers an emergent theory of the costs and benefits of collaboration that should be of broad appeal to social and organizational scholars alike. Shane R. Thye Edward J. Lawler Series and Volume Co-Editors

This page intentionally left blank

BEYOND THREE FACES: TOWARD AN INTEGRATED SOCIAL PSYCHOLOGY OF INEQUALITY Jane D. McLeod, Tim Hallett and Kathryn J. Lively ABSTRACT Purpose
We propose an elaboration of the social structure and personality framework from sociological social psychology that is intended to promote integration across social psychological traditions and between social psychology and sociology, using the study of inequality as an example. Methodology/approach
We develop a conceptualization of “generic” proximate processes that produce and reproduce inequality in face-toface interaction: status, identity, and justice. Findings
The elaborated framework suggests fundamental questions that analysts can pose about the macro-micro dynamics of inequality. These questions direct attention to the “how” and “why” of macro-micro relations by connecting structural and cultural systems, local contexts, and the lives of individual persons; highlighting implicit processes; making meaning central; and directing our attention to how people act efficaciously in the face of constraint.

Advances in Group Processes, Volume 32, 1
29 Copyright r 2015 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0882-6145/doi:10.1108/S0882-614520150000032001

1

2

JANE D. McLEOD ET AL.

Practical implications
Applying this framework, scholars can use existing theories and generate new ones, and can do so inductively or deductively. Social implications
Research on inequality is enriched by social psychological analyses that draw on the full complement of relevant methods and theories. Originality/value
We make visible the social psychological underpinnings of sociological research on inequality and provide a template for macro-micro analyses that emphasizes the centrality of social psychological processes. Keywords: Inequality; social psychology; status; identity; justice

In his classic 1977 article on the three faces of social psychology, House described and defined three isolated areas of social psychological theory and research. The first, psychological social psychology, focused on psychological processes as they operate in relation to social situations and stimuli, and relied primarily on experimental research. The second, symbolic interactionism, emphasized the processes through which persons interpret situations and construct their actions with respect to each other, especially in the context of participant observation research. The third, psychological sociology, or social structure and personality research, analyzed the relation of macrosocial conditions to psychological attributes and behavior using quantitative modeling approaches, most often applied to survey data. House saw these three faces as having complementary strengths and weaknesses and used those strengths and weaknesses to motivate an argument about the need for, and benefits of, interchange across the three (Jackson, 1988). Specifically, House encouraged scholars who were interested in the relationship between society and the person to pay more attention to the “microsocial interpersonal relations and/or psychological processes through which macrosocial structures come to have” (p. 172) their effects
in essence, to embrace the insights of symbolic interactionism and psychological social psychology. Likewise, he directed scholars within symbolic interactionism and psychological social psychology to acknowledge explicitly that psychological processes and face-to-face

Beyond Three Faces

3

interaction are shaped by, and shape, macro-social conditions. He saw the greatest potential for integration in the application of these distinct traditions to the analysis of specific social phenomena or problems. In essence, House saw social structure and personality research as a broad frame with which to bridge the concepts and tools of macro- and micro-sociologists and to build linkages among distinct social psychological perspectives. Almost 40 years later, the integration House hoped for has not occurred. Symbolic interactionists and social structure and personality researchers are no closer now than they were then, and group processes research has emerged as a distinct (some might say new) face of sociological social psychology (Zelditch, 2013). Nor has the social structure and personality framework achieved greater visibility and recognition within the discipline of sociology. Since 2000, there has been only one explicit mention of the framework in the American Sociological Review (McLeod & Kaiser, 2004), one in the American Journal of Sociology (Schnittker, 2008), and three in Social Forces (Brand & Burgard, 2008; Kohn, Wang, & Yue, 2013; Longest, Hitlin, & Vaisey, 2013).1 Many sociologists study the effects of macro-conditions and processes on individuals but the principles of the social structure and personality framework are not acknowledged regularly or applied consistently. For the discipline of sociology, this represents a missed opportunity. The social structure and personality framework is a useful heuristic for research that spans multiple levels of analysis. As such, the framework has the potential to serve as an important common language for sociologists working across a variety of theoretical and methodological orientations as they examine the micro-meso-macro link.2 As for the arena of interdisciplinary social psychology, the lack of recognition of the SSP framework in existing macro-micro research is one indicator of social psychology’s declining influence in sociology and other social sciences. As House (2008) notes, since 1970, membership in the American Sociological Association section on Social Psychology has declined steadily as have the number of sociology graduate programs with specializations in social psychology and the proportion of sociology faculty listing social psychology as a specialty. Ironically, at the same time, sociologists have become increasingly interested in topics and themes that have long been central to social psychology. In particular, cultural sociologists have laid claim to orienting concepts such as meaning, symbols, and situated practice. Since the Culture Section of the American Sociological Association was founded in 1986 with 96 members, it has grown to 1,209 members in 2013, and it has been the ASA’s largest section since 2008.3 Sociologists clearly want to bring meaning, symbols, and situated practice into their analyses but are seemingly

4

JANE D. McLEOD ET AL.

reluctant to turn to sociological social psychology for guidance (see DiMaggio & Markus, 2010 for a similar argument). In response to this state of affairs, we propose an elaboration of the social structure and personality framework (SSP) that is intended to promote integration across social psychological traditions and between social psychology and sociology. In doing so we have two hopes: first, this elaboration will encourage social psychologists from different traditions to realize their commonalities and work collaboratively on topics of common interest, and second, the heuristic will provide sociologists more generally with a common language to communicate and conceptualize the social psychological processes that are relevant to macro-micro relations. Our elaboration involves a conceptualization of “generic” interactional processes (Schwalbe et al., 2000) through which macro-micro relations are produced, reproduced, and resisted. For the purposes of this chapter, we narrow our conceptualization to relations of inequality and discuss only three generic processes: status, identity, and justice. This makes our task more tractable and, given the centrality of inequality to social psychology and sociology, seems reasonable (McLeod, Schwalbe, & Lawler, 2014). Our decision to discuss only three processes is primarily a function of space limitations but our choices are not random. Status, identity, and justice processes resonate with the interests of cultural sociologists, thereby connecting our discussion to contemporary theoretical developments in sociology (Lamont, Beljean, & Clair, 2014). These processes also have currency among social psychologists from different disciplinary, theoretical, and methodological backgrounds,4 thereby highlighting our common interests. As a starting point, we motivate our discussion with reference to limitations in House’s original SSP conceptualization.

THE SOCIAL STRUCTURE AND PERSONALITY FRAMEWORK Although other scholars had published work that we now think of as SSP research prior to the 1970s (e.g., Inkeles, 1959), House’s writings were most influential in defining the framework and articulating its unique perspective on the analysis of human social life.5 House (1981) defined social structure and personality research as the study of “(t)he relation of macro-social structures (for example, societies, organizations, communities, social classes, racial or ethnic groups, and so forth) or processes (industrialization,

5

Beyond Three Faces

urbanization, social mobility) to individual psychological attributes and behavior” (p. 526). As seen from a SSP perspective, the social world is a set of embedded circles with the individual at the core surrounded by progressively larger and more complex social groupings, including dyads, small groups, communities, organizations and institutions, and the larger social system. SSP researchers attempt to trace how the components of the larger social system affect individuals and through which individuals affect social systems via these intermediary structures (Fig. 1). House articulated three analytical principles that follow from the framework. These principles distinguish it from other social psychological approaches to macro-micro relations and position it as an integrative framework that allows social psychologists to speak to one another and that facilitates dialogue with sociologists more broadly. The first principle, the components principle, involves specifying the components of macro-social conditions that are most relevant to understanding the individual outcome of interest. For example, researchers interested in socioeconomic inequalities in health might ask whether income,

Fig. 1.

The Social Structure and Personality Framework.

6

JANE D. McLEOD ET AL.

education, occupation, or some combination best represents socioeconomic position for that purpose. To address this principle, social psychologists must engage with conceptualizations of structural and cultural arrangements that have been developed by more macro-oriented sociologists. The second principle, the proximity principle, directs attention to the proximate experiences through which macro-social structures impinge on persons and through which persons, in turn, reproduce or resist those structures, in particular, the “smaller structures and patterns of intimate interpersonal interaction or communication” (House, 1981, p. 540) that constitute the day-to-day lives of persons.6 These interactions happen in a range of contexts
families, workplaces, neighborhoods, and small groups
that are the starting points for much social psychological research. Consequently, this principle encourages conversation with the full range of social psychological approaches for analyzing the structure and content of interpersonal interactions and communications. The third principle, the psychological principle, takes the final step toward bridging social arrangements and the person. It involves an examination of the psychological processes that determine “when, how, and to what extent macro-social phenomena” (House, 1981, p. 541) affect the individual. This principle directs us to engage with work that analyzes how interpersonal interactions shape, and are shaped by, individual participants, for example, research on social comparisons and attributions.7 Although early social structure and personality research was concerned primarily with attitudinal and behavioral outcomes, the framework can and has been extended to other individual outcomes such as socioeconomic attainment and health and to other individual-level processes such as physiological processes (e.g., House, 2002; Kerckhoff, 1995; Williams, 1990). The potential of SSP to integrate across social psychological traditions and between social psychology and sociology is evident in its principles. The framework has intuitive sociological appeal, and resonates with calls to study the micro-macro link (Alexander, Giesen, Mu¨nch, & Smelser, 1987). Moreover, it adds rigor to sociological analyses of macro-micro relations by encouraging explicit attention to macro, meso, and microcomponents, and situates sociological social psychology at the center of those analyses. Why, then, do sociological macro-micro analyses rarely acknowledge SSP or relevant theories and concepts from sociological social psychology? Part of the answer is methodological. Most SSP analyses approach macromicro relations by using regression-based models to evaluate potential mediators of the relation of interest. For example, researchers studying

Beyond Three Faces

7

socioeconomic inequalities in health might estimate models in which the quality of medical care, health behaviors, stress exposures, and perceived social status are entered as mediators (e.g., Marmot et al., 1998). The variables in these models are indicators of social psychological processes which themselves go unanalyzed. Or, to say it differently, meditational models allow researchers to test explanations for macro-micro relations without having to think about process. To continue the example, researchers studying socioeconomic inequalities in health do not necessarily acknowledge that the quality of medical care, health behaviors, and perceived social status, have their origins in status, identity processes, and justice processes, respectively. This variable-oriented approach distances SSP research from symbolic interactionists and group processes researchers for whom the analysis of process is a central concern and, thereby, impedes integration across social psychological traditions. Despite these limitations, the SSP framework has strengths that complement ethnographic and experimental research. Group processes researchers, symbolic interactionists, and “microsociologists”8 conduct analyses of proximate process(es) that are attentive to how social contexts shape interaction. However, these analyses often fail to explicitly analyze the influence of specific macro-conditions on proximate interactions or individual outcomes (e.g., Collins, 2004; Katz, 1999; Molm, Melamed, & Whitham, 2013), thereby setting them at odds with the SSP agenda.9 Although there are notable exceptions (e.g., Cancian, 1990; Correll, Benard, & Paik, 2007; Ridgeway, 2011), scholars who study social psychological processes in situ often do not direct their efforts toward understanding specific macro-micro relations. Developing a conceptualization of the processes that connect macro-social conditions to individual experiences would encourage sociologists to engage social psychological research while also integrating work across social psychological traditions.

GENERIC PROXIMATE PROCESSES AS A BRIDGE We borrow the term “generic” from Schwalbe et al.’s (2000) analysis of processes that produce and reproduce inequality in face-to-face interactions. They define generic processes as those that “occur in multiple contexts wherein social actors face similar or analogous problems” (p. 421). In other words, although the specifics of these processes take different forms in different historical, institutional, and interactional contexts, the general

8

JANE D. McLEOD ET AL.

processes are trans-situational and widely observable. The generic processes we discuss
status, identity, and justice
focus on how interpersonal interactions explain, heighten, or dampen the associations of inequitable social arrangements with individual outcomes. By so doing, these processes create space for the systematic analysis of agency, resistance, and social change, and create an imperative to study these dynamics.10 These processes overlap and are mutually reinforcing, giving each analytic power that transcends its associated theories and concepts. Importantly, they occur within a range of proximate contexts
families, schools, workplaces, small groups, etc.
and offer varied approaches to understanding how experiences in those contexts come to matter for persons. Our approach is heavily influenced by previous work on conceptualizing generic processes at the macro-micro interface (Hollander & Howard, 2000; Jackson, 1988; Lamont et al., 2014; Lawler et al., 1993; McLeod, 2013; Schwalbe et al., 2000; Turner, 2006) but diverges from that work in its explicit intent to foster integration across social psychological traditions. It hews toward a more Weberian, and less Marxist, approach to inequality or, in Wallace’s (1983) terms, toward an emphasis on “culture structure” (shared meanings) over “social structure” (patterns of interaction, interdependence, and inequality), consistent with the emphasis of most social psychological theories. There are limitations to that emphasis that we will address later. Although these three processes are our primary focus, before turning to them, we emphasize that each depends on the existence of meaningful, taken-for-granted social distinctions (Bourdieu, 1984; Ridgeway, 2011). Social categorization (or the construction of difference) is a cognitive process that undergirds status, identity, and justice. Research on categorization demonstrates that, when we meet people, we immediately and unconsciously classify them according to a few basic characteristics. In the contemporary U.S., these include gender, race, and age (Fiske, 2002; Wilkins, Mollborn, & Bo´, 2014). These classifications serve the purpose of reducing the cognitive demands of interaction. In essence, they are shorthand that reduces the amount of information we have to process. Social categories have purposes beyond cognitive efficiency; however, they also facilitate interaction by helping us coordinate behavior (Ridgeway, 2011). To coordinate behavior, we have to be able to anticipate the feelings, attitudes, and behaviors of others. Placing people into categories helps us do that. We do not treat social categorization as one of the generic processes, although one could argue for doing so, because for our purposes categories are important primarily inasmuch they carry value, provide identities, and serve as

9

Beyond Three Faces

the basis for evaluating one’s life conditions. In other words, social categorization is important to the reproduction of inequality primarily through these additional processes.11 In what follows, we present a brief description of each process (or set of processes), show how it integrates across social psychological traditions, and demonstrate how it connects disparate lines of research on inequality.

Status Processes Status processes are those through which individuals, groups, or objects are ranked as superior or inferior according to a shared standard of social value (Ridgeway & Nakagawa, 2014) and through which those evaluations are expressed. Status processes reflect and create status beliefs, defined as shared cultural beliefs about the relative superiority or inferiority of different social categories. Status beliefs may arise from initial resource differences, but they are reinforced and reproduced through social interaction (Gusfield, 1963; Ridgeway, 2011; Sauder, 2005); these differences have profound implications for the self. Status processes are an appropriate elaboration of the SSP framework because they connect macro-social conditions to the daily lives of persons and because a full understanding of their relevance requires engagement with diverse social psychological traditions. Group processes research demonstrates that shared cultural beliefs shape performance expectations and, thereby, the opportunities that are available to participants in task-oriented groups (Ridgeway & Nakagawa, 2014). High-status actors set the agenda, talk more, and are more influential than low-status actors; they thereby derive greater power in the interaction and reaffirm their superiority. Beliefs about competence, commitment, and the like shape the distribution of material rewards in not only small experimental task-oriented groups but also in real-world settings, thereby influencing the distribution of consequential outcomes such as employment and wages (e.g., Correll et al., 2007). The construction and expression of status are more general processes that have been studied extensively by symbolic interactionists (Sauder, 2005). Deference rituals are a common feature of interaction (Anderson, 1976; Goffman, 1956). We often express relative status without even thinking about it; for example, we know that we are expected to use honorific forms of address to greet a superior. Behaving in this way signals respect; not behaving in this way signals disrespect and places us at risk of a range of formal and informal sanctions (Anderson, 1999; Schwalbe & Shay,

10

JANE D. McLEOD ET AL.

2014). Because most of us choose to avoid sanctions, we tend to present ourselves in ways that are consistent with our social value and to treat others accordingly, a conservative bias that reinforces status hierarchies. Survey-based social psychological research complements experimental and ethnographic research by showing that membership in devalued social categories diminishes mastery
our assessments of our own competence and effectiveness
by reducing opportunities to exercise power and control (Mirowsky & Ross, 1989). These self-evaluations impede efforts to move up in the status hierarchy by making people feel that their low social status is justified and by depriving them of essential personal skills of resistance (Callero, 2014). Through these processes, status beliefs connect social interactions to individual outcomes. In addition to integrating across social psychological traditions, the concept of status offers a language that can be used to integrate distinct areas of sociological research on inequality. The role of status in the “motherhood penalty” is now well established. Employers judge mothers as less competent and committed than other female workers and, consequently, offer them lower wages (Correll et al., 2007). Status processes are also relevant to socioeconomic health inequalities through their implications for quality of physician care (Fiscella, Franks, Gold, & Clancy, 2000). Lutfey and Freese (2005) find that physicians attribute the noncompliance of diabetes patients from lower socioeconomic groups to lower motivation, lower cognitive ability, and lower ability to maintain a more complex (and also more effective) diabetes regimen. In essence, physicians view those patients as less competent and less committed than socioeconomically-advantaged patients. Building on this work, one might imagine that physicians’ evaluations would influence patients’ beliefs about their own efficacy (although Lutfey and Freese do not consider this question). In these examples, status processes offer a way to understand the “how and why” of inequality. Acknowledging the common origins of wages and health care in status processes directs researchers to important questions about how the processes differ depending on the actors, context, and outcome. The studies we have cited come from different substantive areas, employ a range of different methodologies, and begin from different orientations or social psychological “faces.” They could be seen as quite different, and in some ways they are, but by recognizing their common emphasis on status processes we build a more robust social psychology. Moreover, by incorporating status processes into the SSP framework, we develop a common heuristic for examining micro-macro relations that sociologists from different ilk can use without doing violence to their own theoretical and

11

Beyond Three Faces

methodological proclivities. We build from these points to pose a number of questions for ongoing research regarding status processes in macromicro relations as listed: • What specific status beliefs (e.g., worth, competence, and mastery) are most relevant to understanding the effects of specific dimensions of inequality on the outcome of interest? What components of the macrosociological context are those status beliefs linked to and supported by? How are they deployed, reproduced, and modified in social interactions? • In what proximate environments are those beliefs most likely to be constructed and enacted? How do those beliefs shape the nature and content of interpersonal interactions in those contexts and vice-versa? • How do status beliefs influence people’s views of their own abilities and how do those views influence the outcome of interest and feed back into status beliefs?

Identity Processes The concept of identity offers similar insight into the relevance of social psychology to the study of macro-micro relations and the potential of an expanded SSP framework. The term identity refers to “the socially constructed categories that are used to establish meaningful understandings of persons” (Callero, 2014). Identities can be based on personal attributes, as in “I am athletic” or “I am kind”; social roles, such as “mother” or “corporative executive”; or in social categories, such as “woman” or “African American.” By identity processes, we mean those by which people construct, use, and confirm personal, role-based, and categorical identities in interaction with others. Because identities are grounded in social structure they hold a central position in macro-micro analyses (as one can see in the embracement of various identity concepts, if not the relevant social psychological research, by cultural sociologists, e.g., Armstrong, 2002; Bernstein, 1997; Cerulo, 1997; Lamont, 2002; Nagel, 1994). The study of identity and identity processes also draws from many different social psychological traditions. Some lines of research, primarily based on surveys and experiments, take identities as the starting point and analyze how identities influence behavior and attitudes. In general, people seek consistency between their self-conceptions, their behaviors, and situated meanings. Experimental research in psychological social psychology shows that people seek out

12

JANE D. McLEOD ET AL.

and recall feedback that is consistent with their self-conceptions; for example, people who are depressed prefer to interact with people who give them negative feedback (Swann, Wenzlaff, & Tafarodi, 1992). People also make behavioral choices
for example, whether to attend classes or not
based, in part, on the salience of and levels of commitment to role-based identities, in this case, that of a student (Stryker, 1980). A related but distinct line of research on affect control theory (Heise, 1979) has established that people are motivated to maintain alignment between the fundamental sentiments attached to their role identities, attributes, emotions, and behaviors in the course of their daily, situated interactions. For example, people who hold identities (e.g., a professional) or attributes (e.g., self-disciplined) that are culturally viewed as positive, powerful, and active are more likely to engage in behaviors that are also culturally viewed as positive, powerful, and active (Lively & Heise, 2004). Role-based identities reproduce inequalities in at least two ways: (1) through their association with specific ages, races, classes, and genders (Thoits & Virshup, 1997) and (2) through the construction of marked identities. With respect to the former, Thoits and Virshup (1997) assert that “the hierarchical organization of positions within a system or institution is in part determined by the sociodemographic attributes of conventional incumbents” (p. 123). In other words, “broad social categories penetrate or permeate most social roles” (p. 123). With respect to the latter, Lively and Heise (2014) argue that individual attributes and social roles combine to create marked identities, such as female attorney or Black physician (also see Averett & Heise, 1987). Identities that are marked with a marginalized attribute
such as female or Black
are viewed as less potent than their unmarked counterparts and are experientially and behaviorally distinct.12 In this example, the characteristic emotions and behaviors associated with female attorneys are less potent than those of unmarked attorneys
that is, male attorneys, which results in less powerful actions and emotions, as well as lower status shields (Hochschild, 1983) when interacting with colleagues and clients alike (also see Heise, 2013). Symbolic interactionists extend experimental and survey-based analyses of identity confirmation with ethnographic analyses of identity work, defined as processes through which people create, present, and sustain identities that are congruent with and support the self-concept (Snow & Anderson, 1987). Ethnographic research shows that identity work can be used to both resist and reinforce privilege. For example, upper-level managers strive to cultivate an image as dedicated, trustworthy, and loyal, among other qualities, presenting themselves as having these characteristics

Beyond Three Faces

13

expresses a valued identity and legitimates their claim to that identity (Schwalbe & Shay, 2014). Identity work can also challenge inequality, as when homeless men assert that they are “not like” the men who hang out at the Salvation Army or when they claim the positive characteristics associated with being a “bum” or “tramp” (Snow & Anderson, 1987). Yet even when intended to challenge inequality, identity work can inadvertently confirm stereotypes and reinforce disadvantage. For example, working class men may emphasize masculine traits such as aggression and emotional invulnerability to legitimate their claim to an identity as men (Schwalbe & Mason-Schrock, 1996). Exaggerated masculinity compensates for their exclusion from socially valued ways of expressing masculinity, such as economic success, but also highlights how they are different from middle or upper class men. Categorical identities are also implicated in inequality through processes of boundary maintenance, that is, processes through which people differentiate people “like us” from people “like them” (Lamont & Molna´r, 2002; Tajfel & Turner, 1986). Experimental social psychological research finds that, when people meet, they place each other almost instantly into a small number of categories (Fiske & Taylor, 1991). In-group preferences follow quickly from these distinctions (Tajfel & Turner, 1986). The process of constructing categories is complex and fluid, implicating ideologies, political action, and interpersonal interactions (see Wilkins et al., 2014 for a review). Group identities support the construction of competing interests and the intergroup conflict, opportunity hoarding, and exploitation that follow (Bobo & Hutchings, 1996). Such identities also offer a path through which the people can resist the effects of inequality by supporting self-esteem and mastery among members of disadvantaged groups (Phinney, Cantu, & Kurtz, 1997; Portes & Rumbaut, 2001; Uman˜a-Taylor & Updegraff, 2007) and encouraging collective action (Snow & Owens, 2014). Identity processes connect distinct lines of research on inequality. Research on class variation in health behaviors finds that socioeconomically disadvantaged college students viewed health promotion as a characteristic of the white, middle class and not as part of their in-group identity (Oyserman, Fryberg, & Yoder, 2007); as a consequence, they are less likely to engage in healthy behaviors. Identity processes also contribute to racial disparities in labor market success. Some African American men who hold professional positions struggle to establish authority because other workers resist their attempts to claim professional identities (Winfield, 2010), as do male and female African American professors on predominately white campuses (Harlow, 2003). Although distinct in substance, these lines of

14

JANE D. McLEOD ET AL.

research both show how identities support inequality and suggest new questions about how different types of identities operate across contexts. As with our discussion of status processes, the studies we cite come from different substantive areas in sociology, use different methodologies, and draw from different faces of social psychology. To mitigate these boundaries we emphasize, instead, their common focus on identity processes. Moreover, we envision how an expanded SSP framework that incorporates these processes can promote a more general set of questions that sociologists who are interested in the macro-micro dynamics of inequality can pose: • What role and group identities are associated with the dimension of inequality under consideration? How are those identities used to reproduce and resist macro-imperatives? What components of the macrosociological context do they reflect, and how they buttressed by those components? • What behavioral, affective, or attitudinal expectations are associated with those identities? How are they shaped by proximal environments? How do those expectations influence behavior and how do they become manifest in and evolve through social interaction? • How do categorical memberships shape the identities people are able to claim and their negotiations over the expectations associated with those identities? What identity conflicts arise and how are they reconciled?

Justice Processes The final set of generic processes we discuss are justice processes. By these, we refer to processes through which people develop evaluations of whether inequalities are fair or unfair, and their emotional and behavioral responses to those evaluations. Justice scholars examine three types of injustice: distributional, procedural, and interactional. Respectively, these refer to the distribution of societal benefits and costs, the decision-making practices that produce the distribution, and the interpersonal treatment of persons within groups. Justice processes shape the perceived relevance and import of macro-social conditions and provide a basis for subsequent action. They connect macro to micro by influencing personal well-being and by motivating personal and collective resistance. As such, they are central considerations in the production and maintenance of inequality (Hegtvedt & Isom, 2014; Snow & Owens, 2014).

Beyond Three Faces

15

A key insight of social psychological research is that objective inequalities do not automatically produce perceptions of injustice. Rather, conditions are considered just when the “observed distribution, procedure, or interaction meets the expectations set up by relevant and shared rules” and unjust when it does not (Hegtvedt & Isom, 2014, p. 68). Perceptions of injustice are a function of the rules that apply to a given situation (i.e., which “justice principles” shape people’s expectations and how those are determined), and people’s interpretations of the rules. These interpretations depend, in turn, on ideologies, social comparisons, and cognitive justifications. Justice research considers all of these contingencies, as well as the implications of injustice for individual and collective responses, using a range of theoretical and methodological approaches. Survey research provides basic descriptive evidence about dominant justice principles and related societal ideologies. According to this research, people in western societies generally apply the principle of equity when evaluating distributive justice. According to the principle of equity, a fair distribution is one in which rewards are proportionate to contributions. The evaluation of whether or not proportionality holds depends not only on objective circumstances but also on how those circumstances are explained with reference to accepted ideologies. For example, most Americans evaluate income inequality as fair because they accept a success ideology: that opportunity is widely available and that individuals are responsible for their own economic fates (Kluegel & Smith, 1986). Social comparison processes also play a key role in people’s evaluations of their circumstances as just or unjust. Beginning with Stouffer’s early survey research on soldiers’ experiences in the U.S. Army, scholars have established that people judge their situations as unfair and judge themselves as unfairly disadvantaged if they believe that they have received fewer rewards than those in their comparison groups (Crosby, 1976; Stouffer, Suchman, Devinney, Star, & Williams, 1949). People tend to compare themselves to others who are “close” in some way, by virtue of shared characteristics or physical propinquity (Singer, 1981). For example, using the National Survey of Families and Households, Greenstein (1996) observed that the association of inequality in the household division of labor was not related to perceived inequity for women who held traditional gender role ideologies, presumably because they compared themselves to other women who were in similar circumstances. Yet social comparisons are neither simple nor automatic. Symbolic interactionist research on justice demonstrates that comparators, the meanings of contributions and rewards, and ideologies can be negotiated

16

JANE D. McLEOD ET AL.

and reframed cognitively to bring them into accord. For example, according to Hochschild’s (1989) research on household labor, people change their comparators and redefine the contributions they and others have made in order to maintain consistency with their ideologies. Some women in Hochschild’s study who held egalitarian ideologies compared their husband’s limited household contributions to those of men who contributed even less in order to claim an egalitarian distribution. Justice evaluations motivate emotional, cognitive, and behavioral responses that have consequences for people as individuals and within groups (Hegtvedt & Isom, 2014). Experimental research confirms that persons who perceive that they have been rewarded equitably report positive emotions and persons who are disadvantaged inequitably report negative emotions (e.g., Hegtvedt, 1990). Research on relative deprivation, specifically, reveals that feeling disadvantaged relative to similar others is associated with anger and hostility (Singer, 1981). Yet, perceptions of inequity can also be motivating. Research on social movements show that perceived injustices mobilize collective action and support organized forms of resistance (Snow & Owens, 2014). As Hunt (2014) notes, the questioning of ideologies that legitimate inequality by certain groups can support calls for change. As with status and identity processes, our focus on justice processes connects disparate lines of sociological research on inequality. Studies of health disparities find that unfair treatment is strongly associated with poor health (for a review, see Williams, Neighbors, & Jackson, 2003). Interestingly, whites report rates of unfair treatment as high as or higher than those of racial and ethnic minorities (Kessler, Mickelson, & Williams, 1999). For this reason, studies that use measures of “unfair treatment” often find that those measures do not explain racial health inequalities. However, studies that ask whether the unfair treatment is attributable to race (i.e., whether it is racial discrimination, a group-specific injustice) observe differences in the prevalence of discrimination across groups and find that racial discrimination is a powerful explanatory variable (Krieger, 2012). In another application of justice theories, justice perceptions have been invoked to explain women’s relatively high job satisfaction despite their lower compensation relative to men (e.g., Mueller & Wallace, 1996). These considerations suggest the following questions for researchers interested in the role of justice processes in macro-micro relations: • What type of justice is at issue in the macro-micro relation of interest
distributional, procedural, or interactional? What justice principles shape

Beyond Three Faces

17

people’s expectations? What ideologies influence the interpretation of objective circumstances? • What types of social comparisons are available to people within specific proximate contexts? What justice evaluations are likely to arise from those comparisons, and how do they unfold and develop in social interaction? • What cognitive strategies do people use to negotiate discrepancies between their life circumstances and their beliefs? How do those strategies shape their responses to inequality? In sum, as was true for status and identity, focusing on justice processes allows us to erase unproductive divisions and link existing lines of research to social psychological theories and concepts that suggest interesting research questions (e.g., how do dominant ideologies influence perceptions of discrimination?) while highlighting the importance of social psychology for macro-micro research in sociology more broadly.

GENERIC PROCESSES IN AN INTEGRATED SOCIAL PSYCHOLOGY We set ourselves the goal of developing a conceptualization of generic proximate processes that integrates subfields in sociological social psychology and that reaches out to our parent discipline. Our conceptualization builds on House’s (1981) articulation of the social structure and personality framework by emphasizing processes that apply across specific contexts and that invite the contributions of social psychologists from diverse theoretical and methodological traditions. In addition to identifying generic interests that bridge subfields of sociological social psychology, these processes share four common characteristics that identify key contributions of social psychology to understanding the “how and why” of macro-micro relations. First, they cut across “levels” of analysis. In mainstream sociological parlance, status, identity, and justice processes link macro, meso, and micro levels of analysis. This is true in a strict SSP sense
they come between macro structure and micro experiences. This is also true in the sense that each process can be represented and analyzed at each of those levels. For example, statuses and identities can be studied as sociohistorical phenomena, interactive accomplishments, and personal experiences; justice

18

JANE D. McLEOD ET AL.

evaluations draw on ideologies, interactions, and personal cognitions. And this is true in that these processes are built into the practices of institutions. In this way, they connect structural and cultural systems, local contexts, and the lives of individual persons. Second, they are often implicit. While inequality arises and is reproduced through deliberate, conscious oppressive action, it is also supported by unconscious, implicit processes. The production and reproduction of inequality depend, in part, on processes that exist outside of conscious awareness. These generic processes add nuance to sociological accounts of inequality that rely on brute force, exploitation, and opportunity hoarding by turning our attention to how power is constructed and enacted without intent or awareness. Third, they make meaning central. Meaning is not inherent in objects, persons, or situations; rather, meanings are constructed by social actors engaged in interaction (Blumer, 1969; Strauss, 1978). Because meanings are socially constructed, they are subject to change, as political and economic conditions change, and as groups and individuals negotiate for relative advantage. This implies that status, identities, and justice evaluations must be reaffirmed continually through the construction of difference and supporting ideologies. This happens at the societal level through historical processes that justify differences with reference to resource inequalities and ideologies of superiority/inferiority. These differences are also signified and made meaningful in interaction. Importantly, inequality operates through self-meanings as well as the meanings ascribed by others. Through processes of interaction-based meaning construction, identities and selves are constituted in ways that reproduce social structures (MacKinnon & Heise, 2010). Fourth, they allow for structure and agency, freedom and constraint. People actively participate in the creation and reproduction of inequality, but they do not do so under conditions of their own choosing. Structural conditions shape the values, attitudes, beliefs, self-conceptions, and feelings that form the basis for how people create, experience, and reproduce inequalities. At the same time, social structures provide the tools
language, practices, and resources
that make effective social action possible. In other words, social hierarchies and cultural ideologies channel human agency (McLeod et al., 2014). The study of status, identity, and justice draws our attention to how people make sense of their worlds and act efficaciously in the face of constraint. These characteristics are consistent with work in others areas of sociology that have grown since the time that House (1977) first wrote about

19

Beyond Three Faces

social structure and personality research, such as cultural sociology, microsociology, and the sociology of emotions. Scholars who work in these fields (e.g., Cancian, 1990; Eliasoph & Lichterman, 2003; Hochschild, 1983, 1989) present analyses that, in many ways, conform to the goals of the social structure and personality framework: they identify specific features of the macro-social environment that influence specific dimensions of personal experience, and they analyze interpretive, interactional processes that transmit and translate those influences. Despite clear affinities, most of these scholars do not identify themselves as SSP scholars, or even, in some cases, as social psychologists. While some might argue that affiliations and identities do not matter as long as the analytic goals of macro-micro analysis are fulfilled (see, e.g., Fine, 1993), we take seriously the basic social psychological insight that identities have important behavioral and attitudinal consequences. To reach out to these scholars, we suggest a new banner for SSP research: the structure process and person framework. This revision asserts the importance of process while retaining House’s three foundational principles as touchstones for analysis (components, proximity, and psychological). We replace “personality” with “person” because “personality” connotes trait characteristics that are inconsistent with the focus of sociological social psychology and sociology more broadly. This change also has symbolic importance. When House developed his proposed integration, he identified SSP research both as an original face of sociological social psychology and an integrative model. With the proposed change in the banner, we intend to highlight the connections, rather than distinctions, across social psychological traditions.

Limitations of Our Approach We intend our conceptualization as a starting point for future conversations about how best to characterize the generic processes through which macro-conditions and processes shape, and are shaped, by the lives that people lead in specific interactional contexts. Our conceptualization builds on analyses of the structural characteristics of the specific contexts that are considered part of the proximate environment (e.g., neighborhoods, families, and schools), their major characteristics, and the roles attached to them (McLeod & Lively, 2003) by emphasizing the processes to which those structures give rise. We chose processes that highlight the symbolic aspects of inequality in order to forge the strongest connections with

20

JANE D. McLEOD ET AL.

symbolic interactionist work, the burgeoning field of cultural sociology, and the dialogue that is beginning to develop between social psychology and cultural sociology (Collett & Lizardo, 2014; DiMaggio & Markus, 2010; Rivera, 2012). Yet we realize that our conceptualization can be challenged in the specifics. We hope that other scholars will build on our conceptualization in a collective effort to bridge social psychological traditions and reach out to the sociological mainstream. In any such effort, three important limitations of our conceptualization must be addressed. First, it does not explicitly address the processes through which macro-social arrangements influence allocation to proximate environments and the structure of interactions in those environments
that is, how we become part of specific institutions, organization, and social groups. Although these questions are central to the analysis of macro-micro relations, focusing on what happens in interactions gives us the strongest points of convergence among different traditions in social psychology. Moreover, we believe that, ultimately, allocation into proximate environments depends on these same processes. For example, although people’s decisions about where to live depend in part on their material resources, they also depend on subtle and overt forms of social exclusion that derive from identities and status (Bobo & Zubrinsky, 1996). Allocation processes, as well as other processes, must also be considered in future work. Second, and related, our conceptualization does not incorporate social exchange theories, a major line of work in sociological social psychology, although it is not antagonistic to it (indeed, there are strong affinities between some lines of justice research and research on social exchange). Exchange theories are concerned with “the social structures created by exchange relations and the ways in which such structures constrain and enable actors to exercise power and influence in their daily lives” (Cook & Rice, 2003, p. 53). Research drawing from these theories emphasizes how people’s positions in exchange networks affect their power and influence in the exchange and the inequalities that result from exchanges in which actors have differential access to resources. Given these foci, one could imagine giving exchange processes a central place in our conceptualization. However, social exchange research has not traditionally engaged with sociological research on categorical inequalities. In addition, social exchange typically emphasizes resource allocation over the construction of meaning. Extensions of that work, especially on social capital, do consider how people’s positions in existing social hierarchies influence their access to material and symbolic resources, that is, how those positions influence the networks to which they belong and the types of exchanges in which they

Beyond Three Faces

21

participate, although they give less emphasis to the symbolic components of those processes when compared to status, identity, and justice processes (Cook, 2014). Third, and finally, although the generic processes on which we focus demonstrate how status, identity, and justice evaluations are produced and reproduced in interaction
indirect forms of socialization
our conceptualization does not address direct, or intentional, socialization. We know that parents, peers, and the media intentionally socialize children into cultural belief systems that are the basis for status, identity, and justice. For example, upper-middle class parents may coach their children to advocate for themselves and to believe in their own entitlement, whereas working class parents are more likely to teach their children to respect authority (Calarco, 2014; Lareau, 2011). Parents may teach African American children to be restrained and deferential in interaction with whites (Ferguson, 2000; Froyum, 2010; Tyson, 2003). Neither adults nor children passively accept direct messages about difference, as theories of interpretive reproduction emphasize (Corsaro, 1992). However, the direct messages we receive about what social distinctions are important to observe and what those distinctions mean, nevertheless, have power that may not be adequately recognized by our analysis. Despite these limitations, we believe that our conceptualization has value both to sociology proper and to sociological social psychology. The discipline of sociology cannot achieve its analytical goals with a social psychology that is taken-for-granted and hidden in the shadows. All too often, implicit social psychological theories reverberate through the discipline of sociology without any grounding in the relevant social psychological literatures (e.g., Armstrong, 2002; Bernstein, 1997; Nagel, 1994; Piven, 2008; Tilly, 1998). When scholars fail to recognize the social psychological underpinnings of their work, both sociological social psychology and the discipline lose: sociological social psychology becomes increasingly marginalized, and sociological analyses of the macro-micro interface lose analytic rigor (see also Komarovsky, 1973). Sociological social psychologists have a collective obligation to ensure that our theories and methods achieve the recognition they deserve within sociology. While there may be many paths to that end, we argue here that a framework that articulates the proximate social psychological processes that serve as interfaces between macro-structural arrangements and individual responses is among our strongest potential contributions to the rest of the discipline. In each of our points of focus in this chapter
status, identity, and justice
it is the recognition of process that dissolves boundaries

22

JANE D. McLEOD ET AL.

between the faces of social psychology, creating a more integrated framework that not only provides for a more robust social psychology but also serves as a heuristic for sociologists interested in macro-meso-micro links more generally. In emphasizing that the heuristic is a framework and not a theory, the approach provides a broad starting orientation within which sociologists from multiple theoretical perspectives and methodological proclivities can operate. From within this orientation, scholars can use existing theories, or generate new ones, to specify the possible relationships between the SSP principles (component, proximate, and psychological), the processes that cut across those levels, and the substantive topics of interest. And they can do so deductively, at the start of their research, or inductively, based on their observations. Within this framework, scholars can use a range of methodologies, choosing the most appropriate methods to fit and answer the question of interest. What this yields is not homogeneity, but a common language that we can use to communicate across our diverse interests. In an era of fragmentation, within social psychology as well as sociology, this is vital. To develop the expanded framework, we have used the topic of inequality, but whatever the substantive focus, as scholars embark on a research project they can ask themselves, how do I account for the components principle, both theoretically and methodologically, and why do I account for it in this way? The same questions can be asked regarding the proximate and psychological principles. How do I account for status, identity, and justice processes, both theoretically and methodologically? Should those processes prove irrelevant for the topic or case at hand, what other processes serve as bridges across the component, proximate, and psychological principles, and how can these other processes be integrated into the framework to create a more comprehensive heuristic? If the focus rests primarily on a particular level of analysis and/or on a particular process, how can that choice be justified in the context of the larger framework, and how can I use the framework to create a connection between my particular interest and the broader discipline? In encouraging scholars to ask and answer these types of questions, our goal is not to make all research works the same, but rather, to recognize just how many of us are, in different ways, SSP researchers.

NOTES 1. Our claim is based on an analysis of explicit references to the framework in articles that appeared in these journals between 2000 and 2013. We searched for

Beyond Three Faces

23

“social structure” and “personality” as separate terms and as a phrase using the online within-journal search features in JSTOR for the American Sociological Review and the American Journal of Sociology and Project Muse for Social Forces. 2. In making the distinction between heuristic and theory, we emphasize that the SSP framework provides an orientation for research but it does not articulate specific expectations about relationships, as a theory would. When seen in this way, different theories and theorists can make use of the same general framework, albeit with different specifications, points of emphasis, hunches, hypotheses, and causal foci. 3. In contrast, in 1970 Social Psychology was the largest section, with 756 members, but by 2013 was only the 18th largest section, with 689 members (American Sociological Association, 2013). 4. We address one potential objection to our earlier argument. One could interpret the lack of recognition of social psychology in mainstream sociological research as evidence that social psychological insights have become so widely accepted as to no longer bear mention (a la Fine’s, 1993 “sad demise, mysterious disappearance, and glorious triumph of symbolic interactionism”). If this were true, our concern about the position of social psychology within sociology would be misplaced and our conceptualization unnecessary. To the contrary, however, a recent analysis demonstrates that, not only is social psychology not mentioned explicitly in macromicro research on a regular basis, but relevant social psychological research is also often ignored (McLeod et al., 2014). Far from representing a widespread acceptance of social psychology, the lack of recognition appears to represent a widespread ignorance of social psychology that has serious consequences for the status of social psychology within the discipline and for the quality of sociological research on macro-micro relations (Ridgeway, 2014). 5. House’s response was one among many responses to the “crisis in social psychology” of the late 1960s and 1970s, a crisis characterized by strong criticism from within and without that centered on the limitations of experimental research, the lack of relevance of social psychological insights to real-world problems, and the absence of a core set of theoretical, methodological, and ideological principles (Jackson, 1988). House’s response was, in Jackson’s words, one of the more “thoughtful and magnanimous” (p. 92). 6. These correspond to what Lawler, Ridgeway, and Markovsky (1993) called “microstructures.” 7. This principle corresponds roughly to Lawler et al. (1993) focus on “encounters.” 8. Collins (1981, p. 984) defines microsociology as “the detailed analysis of what people do, say, and think in the actual flow of momentary experience.” This would seem to be social psychological, but microsociology has developed as a somewhat distinct field, and many of those who identify as microsociologists, such as Randall Collins, do not identify as social psychologists. In an effort to bring microsociologists “back into the fold,” during his editorship of Social Psychology Quarterly, Gary Alan Fine went so far as to add the descriptor, The Journal of Microsociologies (Hallett, 2014). However, this was not uncontroversial, as many social psychologists, including those who represent the SSP “face,” do not consider their work “micro.” The descriptor did not endure past Fine’s editorship. The status

24

JANE D. McLEOD ET AL.

of microsociology as perhaps another face of social psychology or as an alternative to social psychology is a worthy topic for another paper. 9. In cultural sociology, Bourdieu derisively terms this the “interactionist fallacy,” the erroneous notion that interactions can be understood without reference to the broader conditions of possibility within which they are embedded (Bourdieu, 1984, pp. 578
579). 10. Some scholars may object to the notion of macro, meso, and micro levels of analysis as distinct and distinguishable. Experiences at the different “levels” do not exist independent of the interactional processes through which macro- and microstructures are constituted. While sympathetic to this position, we retain the language of levels because it is widely recognized in sociology and because it is strongly allied with the SSP framework. The generic processes we describe can all be conceptualized across “levels” as they are traditionally defined. 11. The generic processes on which we focus also depend on socialization, defined by Lutfey and Mortimer (2003) as “the process by which individuals acquire social competence by learning the norms, values, beliefs, attitudes, language characteristics, and roles appropriate to their social groups” (p. 183). Status, identity, and justice processes reproduce cultural belief systems and, in that sense, are socializing processes. We argue that, much like social categorization processes, socialization is foundational to the processes we discuss here, and becomes important to inequality through them. 12. Unsurprisingly, traditionally female roles, such as nurse or baby sitter who is marked as a male is actually viewed as more powerful than an unmarked caregiver.

ACKNOWLEDGMENTS We gratefully acknowledge the many colleagues who gave us feedback on this manuscript over the years, including Celeste Campos-Castillo, David Heise, Neil MacKinnon, Michael Schwalbe, Peggy Thoits, and the participants in the Social Psychology, Health, and the Life Course workshop at Indiana University.

REFERENCES Alexander, J. C., Giesen, B., Mu¨nch, R., & Smelser, N. J. (Eds.). (1987). The micro-macro link. Berkeley, CA: University of California Press. American Sociological Association. (2013). Final 2013 section membership counts. Retrieved from http://www.asanet.org/sections/Final2013Counts.cfm. Accessed on March 3, 2014. Anderson, E. (1976). A place on the corner. Chicago, IL: University of Chicago Press. Anderson, E. (1999). The code of the street. New York, NY: W.W. Norton & Company.

Beyond Three Faces

25

Armstrong, E. A. (2002). Forging gay identities: Organizing sexuality in San Francisco, 1954
1990. Chicago, IL: University of Chicago Press. Averett, C., & Heise, D. R. (1987). Modified social identities: Amalgamations, attributions, and emotions. Journal of Mathematical Sociology, 13, 103
132. Bernstein, M. (1997). Celebration and suppression: The strategic uses of identity by the lesbian and gay movement. American Journal of Sociology, 103, 531
565. Blumer, H. (1969). Symbolic interactionism: Perspective and method. Englewood Cliffs, NJ: Prentice Hall. Bobo, L., & Hutchings, V. L. (1996). Perceptions of racial group competition: Extending Blumer’s theory of group position to a multiracial social context. American Sociological Review, 61, 951
972. Bobo, L., & Zubrinsky, C. L. (1996). Attitudes on residential integration: Perceived status differences, mere in-group preference, or racial prejudice? Social Forces, 74, 883
909. Bourdieu, P. (1984). Distinction: A social critique of the judgment of taste. Cambridge, MA: Harvard University Press. Brand, J. E., & Burgard, S. A. (2008). Job displacement and social participation over the lifecourse: Findings for a cohort of joiners. Social Forces, 87, 211
242. Calarco, J. M. (2014). Coached for the classroom: Parents’ cultural transmission and children’s reproduction of educational inequalities. American Sociological Review, 79, 1015
1037. Callero, P. L. (2014). Self, identity, and social inequality. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 273
294). Dordrecht, the Netherlands: Springer. Cancian, F. M. (1990). Love in America: Gender and self-development. New York, NY: Cambridge University Press. Cerulo, K. A. (1997). Identity construction: New issues, new directions. Annual Review of Sociology, 23, 385
409. Collett, J. L., & Lizardo, O. (2014). Localizing cultural phenomena by specifying social psychological mechanisms: Introduction to the special issue. Social Psychology Quarterly, 77, 95
99. Collins, R. (1981). On the microfoundations of macrosociology. American Journal of Sociology, 86, 984
1014. Collins, R. (2004). Interaction ritual chains. Princeton, NJ: Princeton University Press. Cook, K. S. (2014). Social capital and inequality: The significance of social connections. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 207
228). Dordrecht, the Netherlands: Springer. Cook, K. S., & Rice, E. (2003). Social exchange theory. In J. Delamater (Ed.), Handbook of social psychology (pp. 53
76). New York, NY: Springer. Correll, S. J., Benard, S., & Paik, I. (2007). Getting a job: Is there a motherhood penalty? American Journal of Sociology, 112, 1297
1338. Corsaro, W. A. (1992). Interpretive reproduction in children’s peer cultures. Social Psychology Quarterly, 55, 160
177. Crosby, F. (1976). A model of egoistical relative deprivation. Psychological Review, 83, 85
113. DiMaggio, P., & Markus, H. R. (2010). Culture and social psychology: Converging perspectives. Social Psychology Quarterly, 73, 347
352. Eliasoph, N., & Lichterman, P. (2003). Culture in interaction. American Journal of Sociology, 108, 4735
4794.

26

JANE D. McLEOD ET AL.

Ferguson, A. A. (2000). Bad boys: Public schools in the making of black masculinity. Ann Arbor, MI: University of Michigan Press. Fine, G. A. (1993). The sad demise, mysterious disappearance, and glorious triumph of symbolic interactionism. Annual Review of Sociology, 19, 61
87. Fiscella, K., Franks, P., Gold, M. R., & Clancy, C. M. (2000). Inequality in quality: Addressing socioeconomic, racial, and ethnic disparities in health care. Journal of the American Medical Association, 283, 2579
2584. Fiske, S. T. (2002). What we know now about bias and intergroup conflict: The problem of the century. Current Directions in Psychological Science, 11, 123
128. Fiske, S. T., & Taylor, S. E. (1991). Social cognition. New York, NY: McGraw-Hill. Froyum, C. M. (2010). The reproduction of inequalities through emotional capital: The case of socializing low-income black girls. Qualitative Sociology, 33, 37
54. Goffman, E. (1956). The nature of deference and demeanor. American Anthropologist, 58, 473
502. Greenstein, T. N. (1996). Gender ideology and perceptions of the fairness of the division of household labor: Effects on marital quality. Social Forces, 74, 1029
1042. Gusfield, J. R. (1963). Symbolic crusade: Status politics and the American temperance movement. Urbana, IL: University of Illinois Press. Hallett, T. (2014). Introduction of Gary Alan fine: 2013 recipient of the Cooley-Mead Award. Social Psychology Quarterly, 77, 1
4. Harlow, R. (2003). “Race doesn’t matter, but …”: The effect of race on professors’ experiences and emotion management in the undergraduate college classroom. Social Psychology Quarterly, 66, 348
363. Hegtvedt, K. A. (1990). The effects of relationship structure on emotional responses to inequity. Social Psychology Quarterly, 53, 214
228. Hegtvedt, K. A., & Isom, D. (2014). Inequality: A matter of justice? In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 65
94). Dordrecht, the Netherlands: Springer. Heise, D. R. (1979). Understanding events: Affect and the construction of social action. Cambridge, MA: Cambridge University Press. Heise, D. R. (2013). Modeling interactions in small groups. Social Psychology Quarterly, 76, 52
72. Hochschild, A. R. (1983). The managed heart: The commercialization of human feeling. Berkeley, CA: University of California Press. Hochschild, A. R. (1989). The second shift: Working parents and the revolution at home. New York, NY: Viking Penguin. Hollander, J. A., & Howard, J. A. (2000). Social psychological theories on social inequalities. Social Psychology Quarterly, 63, 338
351. House, J. S. (1977). The three faces of social psychology. Sociometry, 40, 161
177. House, J. S. (1981). Social structure and personality. In M. Rosenberg & R. Turner (Eds.), Social psychology: Sociological perspectives (pp. 525
561). New York, NY: Basic Books. House, J. S. (2002). Understanding social factors and inequalities in health: 20th century progress and 21st century prospects. Journal of Health and Social Behavior, 43, 25
142. House, J. S. (2008). Social psychology, social science, and economics: Twentieth century progress and problems, twenty-first century prospects. Social Psychology Quarterly, 71, 232
256.

Beyond Three Faces

27

Hunt, M. O. (2014). Ideologies. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 325
352). Dordrecht, the Netherlands: Springer. Inkeles, A. (1959). Personality and social structure. In R. K. Merton, L. Broom, & L. S. Cottrell, Jr. (Eds.), Sociology today (pp. 249
274). New York, NY: Basic Books. Jackson, J. M. (1988). Social psychology, past and present: An integrative orientation. Hillsdale, NJ: Lawrence Erlbaum. Katz, J. (1999). How emotions work. Chicago, IL: University of Chicago Press. Kerckhoff, A. C. (1995). Institutional arrangements and stratification processes in industrial societies. Annual Review of Sociology, 21, 323
347. Kessler, R. C., Mickelson, K. D., & Williams, D. R. (1999). The prevalence, distribution, and mental health correlates of perceived discrimination in the United States. Journal of Health and Social Behavior, 40, 208
230. Kluegel, J. R., & Smith, E. R. (1986). Beliefs about inequality: Americans’ views of what is and what ought to be. New York, NY: Aldine de Gruyter. Kohn, M. L., Wang, W., & Yue, Y. (2013). Social structure and personality during the transformation of urban China: A comparison to transitional Poland and Ukraine. Social Forces, 91, 347
374. Komarovsky, M. (1973). Presidential address: Some problems in role analysis. American Sociological Review, 38, 649
662. Krieger, N. (2012). Methods for the scientific study of discrimination and health: An ecosocial approach. American Journal of Public Health, 102, 936
944. Lamont, M. (2002). Culture and identity. In J. Turner (Ed.), Handbook of sociological theory (pp. 171
186). New York, NY: Kluwer Academic. Lamont, M., Beljean, S., & Clair, M. (2014). What is missing? Cultural processes and causal pathways to inequality. Socio-Economic Review, 12, 573
608. Lamont, M., & Molna´r, V. (2002). The study of boundaries in the social sciences. Annual Review of Sociology, 28, 167
195. Lareau, A. (2011). Unequal childhoods: Class, race, and family life. Berkeley, CA: University of California Press. Lawler, E. J., Ridgeway, C., & Markovsky, B. (1993). Structural social psychology and the micro-macro problem. Sociological Theory, 11, 268
290. Lively, K. J., & Heise, D. R. (2004). Sociological realms of emotional experience. American Journal of Sociology, 109, 1109
1136. Lively, K. J., & Heise, D. R. (2014). Emotions in affect control theory. In J. E. Stets & J. H. Turner (Eds.), Handbook of the sociology of emotions (Vol. 2, pp. 51
75). Dordrecht, the Netherlands: Springer. Longest, K. C., Hitlin, S., & Vaisey, S. (2013). Position and disposition: The contextual development of human values. Social Forces, 92, 1499
1528. Lutfey, K., & Freese, J. (2005). Toward some fundamentals of fundamental causality: Socioeconomic status and health in the routine clinic visit for diabetes. American Journal of Sociology, 110, 1326
1372. Lutfey, K., & Mortimer, J. (2003). Development and socialization through the adult life course. In J. Delamater (Ed.), Handbook of Social Psychology (pp. 183
202). New York, NY: Kluwer/Plenum. MacKinnon, N. J., & Heise, D. R. (2010). Self, identity, and social institutions. New York, NY: Palgrave Macmillan.

28

JANE D. McLEOD ET AL.

Marmot, M., Fuhrer, R., Ettner, S. L., Marks, N. F., Bumpass, L. L., & Ryff, C. D. (1998). Contribution of psychosocial factors to socioeconomic differences in health. The Milbank Quarterly, 76, 403
448. McLeod, J. D. (2013). Micro-processes of inequality: What are they? Why do they matter? Plenary panel. Paper presented at American Sociological Association, New York, NY. McLeod, J. D., & Kaiser, K. (2004). Childhood emotional and behavioral problems and educational attainment. American Sociological Review, 69, 636
658. McLeod, J. D., Schwalbe, M., & Lawler, E. J. (2014). Introduction. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality. New York, NY: Springer. McLeod, J. D., & Lively, K. J. (2003). Social structure and personality. In J. Delamater (Ed.), Handbook of social psychology (pp. 77
102). New York, NY: Springer. Mirowsky, J., & Ross, C. E. (1989). Social causes of psychological distress. Hawthorne, NY: Aldine de Gruyter. Molm, L. D., Melamed, D., & Whitham, M. M. (2013). Behavioral consequences of embeddedness: Effects of the underlying forms of exchange. Social Psychology Quarterly, 76, 73
97. Mueller, C. W., & Wallace, J. E. (1996). Justice and the paradox of the contented female worker. Social Psychology Quarterly, 59, 338
349. Nagel, J. (1994). Creating and recreating ethnic identity and culture. Social Problems, 41, 152
176. Oyserman, D., Fryberg, S. A., & Yoder, N. (2007). Identity-based motivation and health. Journal of Personality and Social Psychology, 93, 1011
1027. Phinney, J. S., Cantu, C. L., & Kurtz, D. A. (1997). Ethnic and American identity as predictors of self-esteem among African American, Latino, and white adolescents. Journal of Youth and Adolescence, 26, 165
185. Piven, F. F. (2008). Can power from below change the world? American Sociological Review, 73, 1
14. Portes, A., & Rumbaut, R. G. (2001). Legacies: The story of the immigrant second generation. Berkeley, CA: University of California Press. Ridgeway, C. L. (2011). Framed by gender. New York, NY: Oxford University Press. Ridgeway, C. L. (2014). Why status matters for inequality. American Sociological Review, 79, 1
16. Ridgeway, C. L., & Nakagawa, S. (2014). Status. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 3
25). Dordrecht, the Netherlands: Springer. Rivera, L. A. (2012). Hiring as cultural matching: The case of elite professional service firms. American Sociological Review, 77, 999
1022. Sauder, M. (2005). Symbols and contexts: An interactionist approach to the study of social status. The Sociological Quarterly, 46, 279
298. Schnittker, J. (2008). Happiness and success: Genes, families, and the psychological effects of socioeconomic position and social support. American Journal of Sociology, 114, S233
S259. Schwalbe, M., Godwin, S., Holden, D., Schrock, D., Thompson, S., & Wolkomir, M. (2000). Generic processes in the reproduction of inequality: An interactionist approach. Social Forces, 79, 419
452. Schwalbe, M., & Mason-Schrock, D. (1996). Identity work as group process. Advances in Group Processes, 13, 113
147.

Beyond Three Faces

29

Schwalbe, M., & Shay, H. (2014). Dramaturgy and dominance. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 155
180). Dordrecht, the Netherlands: Springer. Singer, E. (1981). Reference groups of social evaluations. In M. Rosenberg & R. H. Turner (Eds.), Social psychology: Sociological perspectives (pp. 66
93). New York, NY: Basic Books. Snow, D. A., & Anderson, L. (1987). Identity work among the homeless: The verbal construction and avowal of personal identities. American Journal of Sociology, 92, 1336
1371. Snow, D. A., & Owens, P. B. (2014). Social movements and social inequality: Toward a more balanced assessment of the relationship. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 657
682). Dordrecht, the Netherlands: Springer. Stouffer, S. A., Suchman, E. A., Devinney, L. C., Star, S. A., & Williams, R. M., Jr. (1949). The American soldier: Adjustment during Army life. Oxford: Princeton University Press. Strauss, A. (1978). Negotiations: Varieties, contexts, processes, and social order. San Francisco, CA: Jossey-Bass. Stryker, S. (1980). Symbolic interactionism: A social structural version. Menlo Park, CA: Benjamin Cummings. Swann, W. B., Jr., Wenzlaff, R. M., & Tafarodi, R. W. (1992). Depression and the search for negative evaluations: More evidence of the role of self-verification strivings. Journal of Abnormal Psychology, 101, 314
317. Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (2nd ed., pp. 7
24). Chicago, IL: Nelson-Hall. Thoits, P. A., & Virshup, L. K. (1997). Me’s and we’s: Forms and functions of social identities. In R. D. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental issues (pp. 106
133). New York, NY: Oxford University Press. Tilly, C. (1998). Durable inequality. Berkeley, CA: University of California Press. Turner, J. H. (2006). The state of theorizing in sociological social psychology: A grand theorist’s view. In P. J. Burke (Ed.), Contemporary social psychological theories (pp. 353
374). Stanford, CA: Stanford University Press. Tyson, K. (2003). Notes from the back of the room: Problems and paradoxes in the schooling of young black students. Sociology of Education, 76, 326
343. Uman˜a-Taylor, A., & Updegraff, K. A. (2007). Latino adolescents’ mental health: Exploring the interrelations among discrimination, ethnic identity, cultural orientation, selfesteem, and depressive symptoms. Journal of Adolescence, 30, 549
567. Wallace, W. L. (1983). Principles of scientific sociology. Hawthorne, NY: Aldine. Wilkins, A. C., Mollborn, S., & Bo´, B. (2014). Constructing difference. In J. D. McLeod, E. J. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 125
154). Dordrecht, the Netherlands: Springer. Williams, D. R. (1990). Socioeconomic differentials in health: A review and redirection. Social Psychology Quarterly, 53, 81
99. Williams, D. R., Neighbors, H. W., & Jackson, J. S. (2003). Racial/ethnic discrimination and health: Findings from community studies. American Journal of Public Health, 93, 200
208. Winfield, A. H. (2010). Are some emotions marked “whites only”? Racialized feeling rules in professional workplaces. Social Problems, 57, 251
268. Zelditch, M. (2013). Thirty years of advances in group processes: A review essay. In S. R. Thye & E. J. Lawler (Eds.), Advances in group processes (Vol. 30, pp. 1
19). Advances in Group Processes. Bingley, UK: Emerald Group Publishing Limited.

This page intentionally left blank

SEQUENCE-NETWORK ANALYSIS: A NEW FRAMEWORK FOR STUDYING ACTION IN GROUPS Benjamin Cornwell and Kate Watkins ABSTRACT Purpose
The ability to analyze social action as it unfolds on micro time scales
particularly the 24-hour day
is central to understanding group processes. This chapter describes a new approach to this undertaking, which treats individuals’ involvement in specific activities at specific times as bases for: (1) sequential linkages between activities; as well as (2) connections to others who engage in similar action sequences. This makes it possible to examine the emergence and internal functioning of groups using existing network analysis techniques. Methodology/approach
We illustrate this approach with a specific application
a quantitative and visual comparison of the daily activity patterns of employed and unemployed people. We use data from 13,310 24-hour time diaries from the 2010
2013 American Time Use Surveys.

Advances in Group Processes, Volume 32, 31
63 Copyright r 2015 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0882-6145/doi:10.1108/S0882-614520150000032002

31

32

BENJAMIN CORNWELL AND KATE WATKINS

Findings
Employed and unemployed people engage in significantly different types of activities and at different times. Beyond this, network analyses reveal that unemployed individuals experience much lower levels of synchrony with each other than do employed individuals and have much less organized action sequences. In short, there is a chronic lack of prevailing norms regarding how unemployed people organize the 24-hour day. Research implications
Future research that uses time-stamped data can employ network methods to analyze and visualize how group members sequence and synchronize social action. These methods make it possible to study how the structure of social action shapes group and individual-level outcomes. Keywords: Social network analysis; sequence analysis; small groups; microsociology; employment; group cohesion

The study of social action that unfolds on the order of minutes or hours is central in the scholarly effort to understand fundamental group processes. This is evident in numerous related analytic frameworks that examine the action sequences that constitute group activity. Bales (1951), for example, developed interaction process analysis
a method that records detailed sequences of behaviors (e.g., utterances, gestures, and other acts) that occur within group settings
to show how group outcomes like problem solving emerge (e.g., Bales & Strodtbeck, 1951). Similarly, the Time-by-Event-byMember Pattern Observation system includes a coding schema for tracking larger task-related group processes (e.g., Lehmann-Willenbrock, Allen, & Kauffield, 2013). These and other frameworks emphasize that how action among group members progresses over time shapes key social phenomena, including the emergence of social hierarchy (e.g., Ridgeway & Diekema, 1989; Skvoretz & Fararo, 1996), decision making (e.g., Gibson, 2012; Stasser & Davis, 1981), the development of divisions of labor (e.g., Bianchi & Milkie, 2010), social exchange (e.g., Bearman, 1997), and social cohesion (e.g., Friedkin, 2004; Tuckman, 1965). Because these social processes involve complex minutiae of micro-time action, however, scholars have struggled to map and analyze the web of activities that link group members to each other in a coordinated system.

Sequence-Networks and Group Action

33

The purpose of this chapter is to describe a new approach to analyzing micro-time data, called sequence-network analysis (Cornwell, in press). The starting point for this approach is the idea that the things that a given group member does on a given day
for example, “I drove in to work, I caught up on email in my cubicle, I talked with coworkers by the water cooler, and then I attended a meeting”
are temporally situated within larger chains or sequences of action. The temporal position of a given activity within one’s schedule (i.e., when that activity occurs) provides a basis for linking that person to other group members who follow similar sequences. This insight makes it possible to construct generic affiliation networks (Baldassarri, 2011; Cornwell & Harrison, 2004; Light, 2015; Simmel, 1922/1955) that reflect which activities tend to precede and follow which other activities within a group, giving rise to dominant activity pathways that reflect common experiences or functions among its members. Our approach is suited to the study of action as it unfolds within groups of people who are consciously oriented to each other. But it can also be used to study generic affiliations among people who are in the same general population but who may not know each other. Thus, we employ a very general definition of “group”
not necessarily considering closed task performance systems nor systems that structure actual social interaction (McGrath & Kravitz, 1982). Rather, groups in this chapter refer to sets of individuals who possess similar attributes, occupy similar positions, or play similar roles in a larger social system (Simmel, 1922/1955). We illustrate this approach using data on activity patterns reported by members of two generic social groups
employed people and unemployed people. Employment is widely seen as one of the key factors that regulate everyday action, as it demands routine, synchrony, and coordination (e.g., Durkheim, 1893/1997; Giddens, 1984; Zerubavel, 1981). Thus, it provides an ideal case for demonstrating our method of analyzing common patterns of social action that link group members together. We analyze time diaries provided by 12,160 employed people and 1,150 unemployed people in the 2010
2013 American Time Use Surveys (ATUSs). By comparing the activity-sequence networks of employed and unemployed individuals, we demonstrate the capacity for this method to uncover the temporal organization of social action both within explicit groups and in the general population. Below, we briefly revisit key theories regarding how everyday activities are structured, theoretical foundations for the idea that group structure can be analyzed using a network framework, and existing research on the role of employment specifically in shaping everyday activity.

34

BENJAMIN CORNWELL AND KATE WATKINS

SOCIAL STRUCTURE AND EVERYDAY ACTION One thing that has hampered the development of a method for illustrating interconnected action sequences within groups is that these sequences are very complex. Group processes involve countless interconnected acts across the numerous time points that compose a given period of time (e.g., a workday). Such detailed data were not always widely available when some group process theories were first developed. More importantly, the capacity to analyze such complex data, even within small groups, was severely limited until relatively recently. Thanks to major improvements in data collection and computer technology, the analysis and depiction of a sequential structure is more feasible.

Using Networks to Analyze Group Action Our theoretical starting point is the idea that social action is made up of series of sequentially connected acts and that group processes depend on these acts being interconnected. This means not only that certain acts are necessarily sequenced in a certain way, but also that certain group members are linked (e.g., by position or role) by virtue of the acts they execute at certain time points within a sequence. Simmel (1922/1955) argued that people develop affinities, or a sense of group membership, by virtue of the fact that they share similar attributes, have similar experiences and backgrounds, and express similar interests. These affinities, in turn, provide a basis for deeper social connections
an insight that is born out in research on “homophily” (Laumann, 1973; McPherson, Smith-Lovin, & Cook, 2001). Homogenous social ties, in turn, provide opportunities for greater social influence between actors, mutual understanding and intersubjectivity, the capacity for collective action, social support, relationship stability, diffusion, and other social phenomena (e.g., see McPherson, Popielarz, & Drobnic, 1992; Rogers, 2010; Suitor & Keeton, 1997). In outlining the elements that link individuals to each other, Simmel (1922/1955) did not restrict his attention to any specific set of elements or criteria that serve as bases of group membership, like having face-to-face contact or sharing organizational goals. Rather, he mentioned numerous sources of generalized affiliation, including inadvertent or involuntary bases such as shared geography and ascribed status (e.g., age and sex), similarities in personal interests (p. 128), similar occupation (p. 128),

Sequence-Networks and Group Action

35

mutual religious, political, and intellectual pursuits (p. 137), as well as citizenship, social class, and involvement in social clubs (p. 138). Simmel emphasized the fact that physical interaction is not necessarily foundational to this type of group affiliation. Rather, actors’ affiliations come from their awareness or consciousness of their social conditions, not from their awareness of each other. Affiliations stem from shared activities, similar role performances, mutual sentiments, experiences with particular events, similar life-course transitions, involvement in particular developmental stages, preferences, and just about anything else that unfolds in a sequential fashion. This provides the theoretical basis for our treating temporally parallel activity sequences as a basis of generic group affiliations between individuals. Whenever two people engage in the same activity at the same time, there is evidence of a degree of similarity, mutual orientation, or shared social role or position between them. For example, two people who are at work at three o’clock in the morning are linked through that mutual experience, through the fact that they are oriented in a similar way to society’s temporal structure (e.g., institutional scheduling constraints). Whether this similar experience ever serves as a basis for actual social interaction is beside the point
it links them together in social structure, and thus constitutes one of many coordinates that defines their positions within it. Though it does not evince a direct social tie, it does provide evidence of an indirect social connection, or affinity, due to occupying a similar position or social role within a larger social system (Simmel, 1922/1955). Thus, a key assumption of this chapter is that while different individuals’ sequences may be reported in isolation, they are not unrelated to each other. Instead, they reflect meaningful group affiliations between individuals (which may or may not also manifest as actual face-to-face contact or cooperation) and, simultaneously, systematic relationships between activities. Every sequence is a reflection of a larger complex of socialstructural forces that combine to give rise to regular patterns of behavior within groups. These forces include norms, social institutions, and networks of obligations (Durkheim, 1893/1997; Giddens, 1984; Merton, 1957; Nadel, 1957; Parsons, 1951; Weber, 1922/1978; Zerubavel, 1981). Accordingly, which patterns of behavior are common for members of a given social group will depend on that group’s particular set of social statuses and roles, the institutional constraints (e.g., business hours) and contexts in which they engage in action, and their embeddedness in a system of social obligations and expectations. Given these observations, one strategy for demonstrating the value of an analytic approach to showing how a given group’s repertoire of social

36

BENJAMIN CORNWELL AND KATE WATKINS

action is sequentially organized is to compare the action sequences of that group’s members to the action sequences of individuals in another group. We attempt to do that in this chapter. To make our illustration as straightforward as possible, we choose one of the most highly regulatory social statuses in everyday life
employment
and attempt to show the vastly different sequential structuring of activities within these generic groups. Before doing so, we first cover existing research that provides some insight into what the temporal structure of action within employed and unemployed groups might look like.

EMPLOYMENT AS A SOURCE OF STRUCTURE FOR ACTION We are most concerned here with the role paid work plays in structuring individuals’ everyday action. Much research shows that paid work and nonwork activity alike
including family, social, and leisure time
is structured by the routine and constraints of paid labor. Employment increases coordination and synchrony among working individuals, partly by dictating coordinated divisions of labor and partly by making social action so routine (Giddens, 1984; Zerubavel, 1981). Thus, it is not surprising that research reveals widespread (but slowly changing) norms with respect to how much time workers dedicate to paid work relative to other activities (see Gershuny, 2000; Gimenez-Nadal & Sevilla, 2012). Likewise, and more directly related to the issue of sequencing, there are widespread norms regarding the time at which paid work occurs. The prevailing, stereotypical norm suggests a 40-hour workweek, during which people work from 9 to 5, Monday through Friday. It is estimated that 82.6% of employed persons over the age of 15 do average 7.9 hours of work on weekdays, and the vast majority of these hours fall between 9 a.m. and 5 p.m. (Bureau of Labor Statistics, 2014a; see also Lesnard, 2010). This leads to high levels of synchrony and predictability among paid workers.1 There is some evidence of a fragmenting influence of paid work on the temporal structure of everyday activity. For one, there is the rise of nonstandard work schedules, including part-time work, shift work, temporary work, weekend work, and rotating and flexible work schedules (Kalleberg, 2000, 2009; Presser, 2003; Wight, Raley, & Bianchi, 2008). Beers (2000) reported that in 1997, 17% of all full-time workers worked outside of the 6-to-6 time window. According to estimates from special supplements of

Sequence-Networks and Group Action

37

the U.S. Current Population Survey, flexible work arrangements rose from 13.6% in 1985 to 29.6% in 2004 (McMenamin, 2007). By the late 1990s, 71% of American workers either worked a majority of their hours in the evening or at night, worked more or less than 35
40 hours per week, or worked on weekends (Presser, 1999). We also consider how unemployment shapes everyday action. The loss of work may have an even more pronounced impact on the structure of individuals’ daily activities than do nonstandard work arrangements. This is a timely question. The severity of the economic downturn during the 2007
2009 recession resulted in unemployment rates not experienced since the early 1980s and prior to that the Great Depression (Bureau of Labor Statistics, 2014b). Unemployment has been linked to an array of adverse outcomes for workers and their families, not limited to foregone income and material deprivation. Several studies provide evidence of psychological strain resulting from unemployment (e.g., Creed & Evans, 2002; Stro¨m, 2002), as well as substance abuse, poorer health, poorer sleep, and reduced recreational activity (e.g., Eliason, 2014). Negative effects also extend beyond unemployed individuals to their families, friends and communities
including relationship strain between unemployed individuals and their spouses and children, divorce, and child abuse (Stro¨m, 2003; Jones, 1990; Vinokur, Price, & Caplan, 1996).2 One of the motivating ideas behind the analysis we present here is that the loss of structure in everyday life that employment otherwise supplies
for example, synchrony with others and greater predictability in everyday life
has additional negative consequences for many unemployed individuals. Unemployment changes the nature of the activities people do. Research shows that macroeconomic cycles influence the share of childcare between mothers and fathers, where men dedicate more time to childcare during downturns (Casper & O’Connell, 1998; Stro¨m, 2002). Aguiar, Hurst, and Karabarbounis (2013) report that during the recent U.S. recession, foregone work hours were reallocated as follows: roughly 30
35% went toward increased nonmarket work (e.g., home production); 50% were allocated to leisure (especially sleeping and television watching); and the remaining time was allocated to other activities (e.g., civic involvement, health care, and education). Only a limited amount of time was reallocated to job search. Krueger and Mueller (2010) find that, on average, Americans spend 41 minutes of job search on weekdays, compared to 12 minutes among Europeans (see also DeLoach & Kurt, 2013). The general impression one gets from this line of work is that the lack of employment results in a drastic loss of temporal regulation of action. But

38

BENJAMIN CORNWELL AND KATE WATKINS

this disorganization has not been measured directly, nor depicted visually. To show how this can be done, we turn to detailed time diary data provided by employed and unemployed individuals. The goal is to see how everyday action is sequenced within each of these groups. Our core hypothesis is that, despite the increasing prevalence of nonstandard work arrangements, employment remains a key force in temporally structuring everyday activities. To demonstrate this empirically, we need to document not just that people in these groups spend different amounts of time on certain activities, but also that the sequential structure of action within the employed group is different from that within the unemployed group. To do this, we describe and present: (1) measures of the extent to which action is synchronized among the individuals within each group; (2) indicators of the roles specific activities play within the action sequences of each group; and (3) diagrams depicting the structure of action within each group.

DATA To conduct a study like this, one needs time-stamped data on the social activities of multiple group members that occur within a relatively short, bounded period of time. One could do this, for example, using data on the behaviors of the children who attend a given school on a given day, or the activities of all of the residents of a given household or neighborhood. These would be more explicit group applications of our approach. But to demonstrate the broader generalizability of our method, we examine weekday time diary data from presumably unacquainted respondents in a nationally representative survey. Specifically, we use data from the ATUSs, which are annual nationally representative studies of adults from a wide variety of backgrounds throughout the United States. To obtain ATUS samples, the Bureau of Labor Statistics begins by drawing a random sample of households from those leaving the Current Population Survey (CPS) rotation each month. An eligible person from the household (a civilian who is at least 15 years old) is randomly selected from the household to be interviewed by phone. The ATUS is administered to roughly 10,000 people each year, resulting in a dataset that contains time diary records for over 148,000 Americans from 2003 to 2013.3 The annual response rates for the ATUS average 55%. The main attraction of the ATUS for our present purposes is the detailed information it provides about what individuals were doing, and

Sequence-Networks and Group Action

39

when, during the course of the previous 24-hour period. The ATUS collects 24-hour recall diaries from individuals over the telephone. ATUS interviewers start by asking respondents about the beginning of the previous day: “So, let’s begin. Yesterday [e.g., Thursday], at 4:00 AM. What were you doing?” They then work forward through the day, collecting information about what the respondent was doing, the times when each activity began and ended, where each activity occurred, and whom the respondent was with during that activity. The shortest unit of time reported for any given activity is 5 minutes, allowing for a maximum 288 distinct activity reports for a given day. The ATUS employs a three-tiered activity classification schema, including noncollapsed and collapsed version of over 400 specific activity codes. In the present analysis, we use the highest-tier activity codes, which code each 5-minute increment into one of 17 general types of activities. We are interested in the extent to which employed and unemployed people engage in different activities and, more importantly, the extent to which their activity sequences evince typical patterns that provide the basis of generic affinities among members within these groups. We focus on the most recent four years of the ATUS
2010 through 2013
a period that saw high but steadily declining unemployment in the United States. In the wake of the “Great Recession,” unemployment rates peaked toward the end of 2009 at 10% nationally. At the end of 2013, the national unemployment rate declined to 6.7% (Bureau of Labor Statistics, 2014b). To highlight how work activity itself shapes activity patterns, we restrict our analysis to individuals who reported weekday (Monday
Friday) diaries. Because we are most interested in how work structures everyday activity
and the contrast of that with everyday activity among nonworking individuals
we restrict our analysis to activity that occurs in the 9-to-5 time period. Our final dataset includes 12,160 employed individuals and 1,150 unemployed individuals, yielding a total of 13,310 cases. Within these diaries, we focus only on activity patterns that unfold during the typical 9-to-5 workday period. These restrictions are put in place to provide the clearest possible contrast between the weekdays of employed and unemployed people.

SEQUENCE-NETWORK ANALYSIS We now turn to our description of how one can map the connections that exist among group members by virtue of how they sequence their social

40

BENJAMIN CORNWELL AND KATE WATKINS

activities. We are interested in the activities
that are reported  by a given set of g individuals, contained in the set N = n1 ; n2 ; :::; ng . We have timestamped activity data from each of these g individuals that describe which of the k elements (here, activities) in the set E = fe1 ; e2 ; :::; ek g each individual was doing at
each of t time  points during a given period of time, contained in set P = p1 ; p2 ; :::; pt :

Sequence-Network Construction This data setup lends itself to network analysis. All social networks involve two social elements: Nodes, or vertices
which may represent individuals, events, or other social entities
and the links, or ties, that associate nodes with each other. The sequence-network method (Bearman, Faris, & Moody, 1999; Bearman & Stovel, 2000; Bison, 2014; Cornwell, in press) treats the activities that occur at specific times as the nodes in the network. These activities are connected to each other via their adjacency in a temporal sequence of action. The network is composed of k × t = q possible time-activity
 combinations, which are captured in the set C = c1 ; c2 ; …; cq : For example, a node in a given activity-sequence network may be “Emailing at 10:45 a.m.” That time-activity will be followed by another time-activity, such as “Standing by the water cooler at 11:00 a. m.” To denote the time-activity combinations, we use two subscripts for c, the first is a number representing the time period in P and the second a letter representing the activity in E. The ordered activity data that are contained in a typical time diary dataset can be recorded in a g-by-q rectangular matrix, which we will call A. In matrix A, the g individuals are arranged down the rows of the matrix and the q activity-time combinations are arranged along the matrix’s columns. For example, let us assume that we have time diary data that indicates which activities each of 10 individuals were doing between 7 a.m. and 7:40 a.m. on a given day
a period of time when many people transition between home and work. Information on what these individuals were doing is available for each of the four 10-minute time periods that comprise this time window. Let us also assume, for the sake of simplicity, that analysts recorded very generic activity data for each period, employing only four activity codes: “Eating” (=1), “Household activity” (=2), “Commuting” (=3), and “Working” (=4). Table 1 puts this information for the 10 individuals (n1 through n10) into the format of matrix A. Each of the g × q cells of matrix A indicates

41

Sequence-Networks and Group Action

Table 1.

Layout of Matrix A for a Hypothetical Set of 10 Individuals. Time-Activity

Subject

n1 n2 n3 n4 n5 n6 n7 n8 n9 n10

c1A

c1B

c1C

c1D

c2A

c2B

c2C

c2D

1 1 1 1 1 1 0 0 0 0

0 0 0 0 0 0 1 1 1 1

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

1 0 0 0 0 0 0 0 0 0

0 1 1 1 1 0 0 0 0 0

0 0 0 0 0 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0

Time-Activity

Subject

n1 n2 n3 n4 n5 n6 n7 n8 n9 n10

c3A

c3B

c3C

c3D

c4A

c4B

c4C

c4D

1 0 0 0 0 0 0 0 0 0

0 1 1 0 0 0 0 0 0 0

0 0 0 1 1 1 0 0 0 0

0 0 0 0 0 0 1 1 1 1

0 0 0 0 0 0 0 0 0 0

1 1 1 0 0 0 0 0 0 0

0 0 0 1 1 0 0 0 0 0

0 0 0 0 0 1 1 1 1 1

Note: Cells that contain a “1” indicate that the individual in the corresponding row engaged in a given activity at a given time as displayed in the corresponding column. Cells that contain a “0” indicate that the individuals did not engage in that activity at that time. Vertical lines are used to separate the four time intervals.

whether a given individual was doing the corresponding activity at the corresponding time point. For example, because person n1 was eating at time t1, the cell corresponding to row n1 and column c1A contains a “1.” Likewise, the cells corresponding to row n1 and columns c1B, c1C, and c1D each contain a “0” to indicate that person n1 was not doing those other three activities (“B,” “C,” or “D”) at time t1. Organizing the activity data using this matrix is the first step in creating an ordered sequence network. Matrix A does not convey how the q time-activity elements are related to each other in a larger sequence chain of activities, based on these individuals’ reports. In network analysis, the idea of sequential order can be

42

BENJAMIN CORNWELL AND KATE WATKINS

captured in what is referred to as a “directed” network. One can transform matrix A into a directed matrix that reflects the ordered relationships among the activities. This can be done through matrix multiplication. First, one multiplies the transpose of A (denoted AT) by A to obtain a new matrix, B. In matrix algebra terms: B = AT(A). Unlike matrix A, the timeactivity elements are the only elements in B. Individuals are not explicitly included in this matrix. Rather, B is a q-by-q square matrix. For the uninitiated, this might look like a complex operation, but the cells in the resulting matrix reflect simply how often a given pair of time-activities
for example, c1B and c2C
co-occur in the same sequences. The new matrix that results from this procedure is shown in Table 2. The cells of this matrix are valued, so that relationships between pairs of successive activities
like “commuting from 7:10 to 7:20 a.m.” (or c2C) and “working from 7:20 to 7:30 a.m.” (or c3D)
are expressed in terms of the number of individuals who reported this particular transition. The cell corresponding to row c2C and column c2C contains a “4,” which indicates that this sequence of activities occurred at this particular time in four individuals’ sequences. This suggests that this particular subsequence is common, given that so few other activity sequences occur as often. (Notice that three sub-matrices within this larger matrix are demarcated using shading and borders. The significance of this will be explained in a moment.) Matrix B can be further examined to show exactly how specific activities are sequentially connected. For the purposes of constructing a directed sequence network, we can null the diagonal, the cells below the diagonal, and all remaining cells that link nonadjacent time periods. This leaves us with a partial version of matrix B which contains only those three shaded sub-matrices in Table 2. One can keep the matrix as a whole but simply null out (i.e., set to “0”) all of the values outside of these three bounded sub-matrices. Manipulating the matrix in this fashion will facilitate both the analysis of the structural properties of the sequence-network and the visual representation of its structure. The data in these cells is all one needs to construct a sequence network from time-stamped activity data. We will illustrate the value of visualizing matrix B as a network diagram in the Results section.

Measuring Connections among Group Members How does this network framework help us to study how group processes are structured? One goal is to document the extent to which a given set

43

Sequence-Networks and Group Action

Table 2. Affiliation Matrix B Based on the Hypothetical Matrix A in Table 2, Showing the Number of Individuals Who Reported Specific Pairs of Time-Activities within Their Activity Sequences. Time-Activity

Time-Activity

c1A c1B c1C c1D c2A c2B c2C c2D c3A c3B c3C c3D c4A c4B c4C c4D

c1A

c1B

c1C

c1D

c2A

c2B

c2C

c2D

6 0 0 0 1 4 1 0 1 2 3 0 0 3 2 1

0 4 0 0 0 0 4 0 0 0 0 4 0 0 0 4

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 0 0 0 1 0 0 0 1 0 0 0 0 1 0 0

4 0 0 0 0 4 0 0 0 2 2 0 0 2 2 0

1 4 0 0 0 0 5 0 0 0 1 4 0 0 0 5

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Time-Activity

Time-Activity

c1A c1B c1C c1D c2A c2B c2C c2D c3A c3B c3C c3D c4A c4B c4C c4D

c3A

c3B

c3C

c3D

c4A

c4B

c4C

c4D

1 0 0 0 1 0 0 0 1 0 0 0 0 1 0 0

2 0 0 0 0 2 0 0 0 2 0 0 0 2 0 0

3 0 0 0 0 2 1 0 0 0 3 0 0 0 2 1

0 4 0 0 0 0 4 0 0 0 0 4 0 0 0 4

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 0 0 0 1 2 0 0 1 2 0 0 0 3 0 0

2 0 0 0 0 2 0 0 0 0 2 0 0 0 2 0

1 4 0 0 0 0 5 0 0 0 1 4 0 0 0 5

Note: The shaded areas represent sub-matrices that contain the information that is needed to determine which time-activities were sequentially adjacent in this hypothetical set of 10 individuals’ activity sequences.

44

BENJAMIN CORNWELL AND KATE WATKINS

of individuals’ activity sequences link them to a group by virtue of the fact that they are doing similar things at certain times. As Simmel (1922/ 1955) pointed out, much information can be gleaned about individuals’ “affinities” with each other based on the extent to which they possess similar attributes, occupy similar positions, engage in similar activities, or otherwise share social characteristics. This is our definition of generic groups. This also provides the basis of the social network analysts’ practice of linking otherwise unconnected individuals by virtue of their similar beliefs, attributes, activities, or other social characteristics via generic affiliation networks (e.g., Baldassarri, 2011; Cornwell & Harrison, 2004; Light, 2015). In the context of everyday action, this kind of affiliation between individuals can be measured by comparing the activity sequences of each pair of individuals within the group to each other, quantifying the degree of similarity in each pair, and then averaging the extent of similarity across all pairs in the group. This can be thought of as grouplevel “synchrony” (e.g., see Kingston & Nock, 1987; Lesnard, 2008; Wight et al., 2008). A first step is to quantify the degree of this synchrony between a given pair of group members with respect to what they were doing during each of a set of fixed time intervals throughout the day. A pair’s sequences intersect at any time point when they report doing the same thing at the same time. Synchrony, or s, can be measured by quantifying the amount, or proportion, of time that the two group members in question were doing the same thing during the specified time period. This calculation for a given pair of group members, ni and nj , can be expressed as follows: sðni ; nj Þ =

t X

up

p=1

where u represents the number of time units (e.g., minutes) that members ni and nj were doing the same thing during a given time period, p represents a given time period in the sequence, and t represents the total number of time intervals. Thus, s(ni, nj) can range from 0 to the total number of time units covered by the sequences (e.g., 1,440 minutes). This measure may also be expressed as a proportion of the total amount of synchrony possible by dividing s(ni, nj) by t. The final step is to combine the measures of synchrony for all of the pairs of group members by averaging them. The resulting measure provides a sense of the extent to which two randomly chosen members of the group tend to do the same things at the same time.

Sequence-Networks and Group Action

45

Assessing the Structure of Action Sequences within a Group The measure of synchrony provides a sense of how similar the members of a given group are in terms of how they structure their activities. But we are also interested in helping researchers provide quantifiable and visible representations of what those action sequences look like. There are a couple of ways one can do this. One is the simple approach of showing how much time group members spend doing certain things, and when those activities tend to occur within the course of the time period in question. Thus, we describe some basic analyses showing the amount of time employed and unemployed people spend doing each of the activities in the time diary, and what proportion of group members do those activities at given times of the day. One can also use networks to take the analysis a step further by exploring how specific activities gain social significance not just via their popularity at particular times, but also in terms of the extent to which they serve as intersections, or junctions, in longer sequence pathways. Some activities play a more pivotal role in linking different activities together in time, thus providing the contexts by which individuals can transition from one activity to another.4 The simplest way to quantify this linking function of specific activities is to measure the number of different types of activities that lead into a given activity at a given time (“in-degree”) and the number of different activities that follow from that activity-time (“out-degree”). These quantities provide basic insight into the tendency for a given activity to be stochastically linked to both antecedent and subsequent activities (Wasserman & Faust, 1994), and thus help us identify activities that function as a pathway, or bridge, between other activities.

Comparing Groups’ Action Sequences As just described, we measure aggregate time-use patterns, levels of synchrony, and the structural roles of different activities in the sequence networks of employed and unemployed people. Our goal is to compare whether these aspects of activity patterns differ between groups. An important wrinkle is that network measures (e.g., each activity’s average in- and out-degree) will automatically vary with group size. For example, the more people in the group, the more likely it is by random chance that a given activity at a given time will be connected to all 16 other activities at the preceding time point through at least one group member’s sequence.

46

BENJAMIN CORNWELL AND KATE WATKINS

Therefore, we cannot directly compare the whole sequence network of the 1,150 unemployed individuals to the whole sequence network of the 12,160 employed individuals. We use a bootstrap approach to sidestep this problem. Because each group is a different size, we compare the network measures that are calculated from the whole unemployed group to network measures that are calculated from randomly chosen, equally sized subsets of 1,150 employed individuals. Of course, just one randomly chosen subset of employed individuals is not enough to make a comparison, however, because that subset might be statistically different from the larger set of employed individuals. Therefore, we draw 100 subsets of employed persons, each containing 1,150 randomly chosen members of that group. We calculate network measures separately for each of these 100 networks that contain employed individuals. We then average those network measures to obtain values that can be compared to the values that were calculated for the unemployed network. We use the percentile method (Mooney & Duval, 1993) to obtain a 95% confidence interval around the averaged estimates obtained from the 100 employed-person samples. If an observed network measure for the unemployed group falls outside of the confidence interval around the corresponding employed group measure, this provides evidence that the unemployed group’s activity network is structurally different from the employed group’s activity network (Cornwell & Dokshin, 2014).

RESULTS A common approach to characterizing the activity patterns within a given group is to analyze their aggregate time-use patterns. A useful way to begin is to examine the overall average amount of time spent by members of a given group doing certain activities during a given period of time. To this end, Table 3 presents estimates of the average amount of time spent by members of these two groups on each of the 17 major activities in the ATUS coding lexicon between the hours of 9 a.m. and 5 p.m. This table reveals clear differences between the two groups in terms of how their time is spent. Workers’ time is dominated by work-related activity, which accounts for 293.3 minutes (or 61.1%) of this 8-hour period. The second-most common set of activities are much less prevalent, and include socializing/relaxing/leisure (38.9 minutes), travel/commute (30.6 minutes), personal care (25.0 minutes), and household activities (61.5 minutes). This

47

Sequence-Networks and Group Action

Table 3. Average Overall Amount of Time Spent in Certain Activities during the Course of the Day Among Employed and Unemployed Respondents in the 2010
2013 ATUS Studies. Activity

Personal care (e.g., sleeping and dressing) Household activities (e.g., cleaning) Helping household members (e.g., childcare) Helping nonhousehold members Work-related activities Education Consumer purchases (e.g., shopping) Professional/personal care services (e.g., banking) Household services (e.g., maintenance/ repair) Government services/civic obligations Eating and drinking Socializing, relaxing, and leisure Sports, exercise, and recreation Religious and spiritual activities Volunteer activities Telephone calls Travel/commute a

Employeda

Unemployedb

Mean

s.d.

Mean

s.d.

25.10 (22.23, 27.99) 29.39 (26.03, 32.49) 8.18 (6.93, 9.46) 2.37 (1.59, 3.39) 289.01 (281.45, 297.30) 6.75 (4.97, 8.82) 8.61 (7.38, 9.97) 2.70 (1.99, 3.41) .48 (.22, .84) .20 (.05, .37) 26.11 (25.04, 27.34) 38.19 (35.15, 41.27) 5.12 (3.73, 6.43) .66 (.25, 1.18) 1.92 (1.11, 2.83) 1.09 (.82, 1.41) 29.11 (27.52, 31.05)

67.75

44.90*

76.28

69.22

85.13*

106.74

33.69

20.97*

56.88

20.67

8.92*

41.61

168.14

37.30*

87.56

43.65

52.32*

119.06

28.54

16.87*

39.82

16.89

5.66*

26.06

5.83

.80

11.82

3.77

1.57*

16.63

26.72

23.04*

27.19

75.65

114.04*

120.80

26.17

12.87*

46.39

9.71

1.97*

16.84

18.71

6.77*

43.11

7.60

3.35*

16.83

41.10

38.53*

45.93

Estimates are based on values averaged across 100 1,150-person samples drawn randomly from among all 12,160 employed ATUS respondents. 95% confidence intervals are shown within parentheses. b Estimates are based on values from all 1,150 unemployed respondents. A “*” means that the value for unemployed persons is significantly different from that of the employed (at the p .05). Assertive leaders elicited positive change in activity (Δa = .4), while deferential leaders elicited slight negative change (Δa = −.15). The agreement of the third group member did not have a significant effect on meaning change. Interestingly, respondents’ average ratings of the receptionist after the interaction most closely fit with ratings of concepts like “tease” and “hussy” from previous work (Francis & Heise, 2006). Meanings for the “manager” identity changed only slightly as a function of group interaction. As expected, managers were viewed as less powerful after the group leader expressed support for the receptionist. Contrary to predictions, they were also viewed as slightly more good. Neither of these differences reached significance (p > .05). As for the receptionist identity, sentiment change on the evaluation and potency dimensions was stronger for assertive (Δe = .15, Δp = −.18) than for deferential group leaders (Δe = .04, Δp = −.15). Change in identity meanings was stronger when the third group member agreed with the leader (Δe = .13, Δp = −.21) than when they disagreed (Δe = .05, Δp = −.11). Change on the activity dimension was small and nonsignificant (p > .05). Ratings of the manager after the group deliberation most closely fit with ratings of concepts like “novice” and “stranger” from previous work (Francis & Heise, 2006).

88

KIMBERLY B. ROGERS

For conditions in which the group leader endorsed the manager, the definition of the situation followed the form “receptionist accuses manager.” Simulations predict that this interpretation of events will result in an impression that the receptionist is substantially less good and substantially more powerful and active than anticipated (−1.79, 1.13, 1.45), and an impression that the manager is substantially less good and powerful and somewhat less active than anticipated (−1.53, .36, 1.02). Notably, the power relations between them are largely equalized by this definition of the situation, and both are seen as “rivals”. The receptionist might feel eager and the manager indignant. Simulations predict that the manager might restore the status order by confronting the receptionist or penalizing her. Alternatively, the receptionist might attempt to resolve the situation with nice, deferent behaviors like obedience. Once again, the simulation predictions were largely upheld by the data. When the group leader expressed support for the manager, the “receptionist” identity was viewed as significantly less good (Δe = −.36, p < .001) than before the task. This identity was also seen as more powerful (Δp = .39, p > .05) and active (Δa = .28, p > .05), although these differences did not reach significance. Sentiment change on the evaluation and potency dimensions was stronger when the third group member agreed with the leader’s opinion (Δe = −.49, Δp = .45) than when they disagreed (Δe = −.23, Δp = .33). The receptionist was viewed as less good when the leader was assertive (Δe = −.71) rather than deferential (Δe = .07), but this had no effect on perceived potency. There was a positive change in activity overall, and this change was strongest when leaders were assertive (Δa = .44) rather than deferent (Δa = .07). Counter to other findings, change in activity was less pronounced when the third group member agreed with the leader (Δa = .19) than when they disagreed (Δa = .36). Respondents’ average ratings of the receptionist after the interaction most closely fit with ratings of concepts like “daughter” and “little sister” from previous work (Francis & Heise, 2006). As predicted, the “manager” identity was viewed as significantly less powerful (Δp = −.15, p < .05) given this definition of the situation. This identity was also seen as less good (Δe = −.29, p > .05) and more active (Δa = .27, p > .05) than before the task, though these differences did not reach significance. Sentiment change on the evaluation and potency dimensions was stronger for assertive (Δe = −.43, Δp = −.2) than for deferential group leaders (Δe = −.11, Δp = −.09), and when the third group member agreed with the leader (Δe = −.37, Δp = −.45) than when they disagreed (Δe = −.2, Δp = .15). Change on the activity dimension followed the

Expectations, Social Influence, and Affect Control

89

opposite pattern; less change was observed when the leader was assertive (Δa = .14) than when they were deferential (Δa = .44), and when the third group member agreed with the leader (Δa = −.07) than when they disagreed (Δa = .61). Another way of viewing this result is to say that activity ratings most closely fit with the simulation predictions when leaders were assertive and the third group member agreed. Ratings of the manager after the group deliberation most closely fit with ratings of concepts like “supervisor” and “authority” from previous work (Francis & Heise, 2006). In sum, results largely supported Hypotheses 3a and 3b for the dimensions of evaluation and potency, while results were mixed for the activity dimension. Identity meanings for the “receptionist” and “manager” changed as a function of the group leader’s expressed opinion during the deliberation task, in ways that tended to fit with the predictions of ACT’s models of impression change. Meaning change was often more pronounced following interaction with a leader that displayed assertive rather than deferential behavior. It is also noteworthy that the behavioral predictions of ACT’s impression formation models roughly seem to follow the results presented previously regarding the resolution of the claim. Respondents were more likely to side with the receptionist and support the settlement of the claim out of court when the leader showed support for the receptionist, fitting with the model’s prediction that compromise would be a likely next step given the event interpretation “manager victimizes receptionist.” Respondents were more likely to side with the manager and support fighting the charges in court when the leader supported the manager, fitting with the model’s prediction that confrontation would be more likely given the event interpretation “receptionist accuses manager.” Sentiments and Relative Influence Friedkin and Johnsen (2003) proposed that relative sentiments for self and other may translate between the status order, as conceptualized by EST, and hierarchies of interpersonal salience, which feature prominently in SINT. In other words, comparisons of identity meanings for self and other are likely a precursor of inferences about the influence structure of the group. Dippong and Kalkhoff (2015) have shown that affective impressions of a group member along the dimensions of evaluation, potency, and activity are significantly and positively related to performance expectations for that group member. Those who are seen as good, powerful, and active are also expected to contribute more to the group task. In the present research, two identity labels were used to assign respondents (and their purported interaction partners) to roles in the task

90

KIMBERLY B. ROGERS

group: leader and employee. Respondents rated these concepts prior to reading about the workplace issue or participating in the group deliberation task and again upon its completion. Prior to participation in the group task, leaders were seen as quite good, powerful, and active (1.78, 2.7, 1.87); employees (i.e., the respondent and the third group member) were typically seen as good but much less powerful and active than leaders (1.1, −.02, .44). Using the regression parameters from Dippong and Kalkhoff’s (2015) study, initial performance expectations were calculated from respondents’ affective impressions of the leader and employee roles prior to participation in the group discussion. As expected based on the design of the study, initial expectations were significantly higher for leaders (7.81) than for employees (6.79, p < .05). Following the guidelines published in Kalkhoff et al. (2010), predicted expectations were used to calculate relative influence, assuming initial consensus in expectation standing. Self-weights were estimated to be higher for the group leader (.6) than for the respondent (.53), and respondents were expected to attribute greater influence to the leader (.26) than the third group member (.21). Thus, mathematical predictions corresponded with the hypotheses outlined previously regarding the relationship between sentiments and influence. Following the group interaction, both sentiments and predicted expectations had shifted. Leaders were seen as substantially less powerful and active after the interaction than before (1.66, 1.15, .88), while employees were seen as much more good, powerful, and active (2.04, .79, 1.09). The gap in predicted performance expectations was also reduced substantially across conditions. Behavioral interchange patterns were the primary determinant of post-interaction expectation standing. In conditions where the leader had behaved assertively, predicted performance expectations suggested a slight status advantage to the leader (7.23) over the respondent (7.17). In conditions where the leader was deferential, this pattern was reversed, and the respondent (7.24) was predicted to have a slight status advantage over the leader (7.19). To test hypotheses about the relationship between sentiments and influence, seven variables were created. Six of these variables capture relative sentiments, using a ratio of respondents’ evaluation, potency, activity ratings of self and other both before and after the group task (“self as employee”/“partner as leader”). The seventh captures relative influence, using a ratio of respondents’ ratings of the perceived influence of self and other at the end of the study (own influence/leader influence). Two linear regressions (N = 122) were run to test the effects of relative sentiments at pre-test and post-test on relative influence.

Expectations, Social Influence, and Affect Control

91

Results offer only limited support for Hypothesis 3c. Relative sentiments for the employee and leader identities prior to participation in the group task did not have a significant effect on the relative influence attributed to self and other (p > .05). The relative potency of identity meanings for self and other following the interaction was a significant predictor relative influence (p < .05), while evaluation and activity were not significant (p > .05). Thus, emergent perceptions of relative power seem to be most essential in shaping the influence composition of the group. It is possible that the scope conditions of EST, which call for both collective orientation and task orientation, necessitate maintaining a view of all group members, including the self, as generally good and active. Moreover, Dippong and Kalkhoff (2015) found that affective impressions on the potency dimension were the strongest predictor of performance expectations following group interaction, and suggest that “interaction renders power differences more affectively and cognitively salient” (p. 10). Indeed, potency ratings have been linked with perceived competence in earlier research (Rogers et al., 2013), suggesting an important role for this dimension in communicating relative performance expectations for members of the group. However, Dippong and Kalkhoff also found that affective impressions at pre-test on all three dimensions were significant predictors of performance expectations, with particular importance placed on judgments of group members’ goodness.

Emergent Gender Attributions Logistic regression (N = 182) was used to examine both conditional manipulations and respondents’ emergent perceptions of social influence as predictors of gender attributions for the group leader. In support of Hypothesis 4a, the opinion expressed by the group leader had a significant effect on attributions about their gender (p < .05). Pro-receptionist leaders were more often viewed as female, and pro-manager leaders as male. Hypothesis 4b was not supported; behavioral interchange patterns did not have a significant effect on gender attributions, nor did group consensus (p > .05). However, the descriptive results reveal an interesting pattern. Respondents were more likely to believe that assertive leaders were female when they endorsed the receptionist and male when they endorsed the manager. Weak leaders were likely to be perceived as female regardless of their expressed opinions. Respondents were also more likely to believe the group leader was male when the third group member agreed with their opinion and female when the third group member disagreed.

92

KIMBERLY B. ROGERS

The perceived influence of the group leader did not have a significant effect on gender attributions (p > .05) contrary to Hypothesis 4c. Unexpectedly, the perceived influence of the third group member was a significant predictor of gender attributions (p < .05). The group leader was more likely to be seen as male when the third group member was seen as highly influential and female when they were seen as low in influence. This effect may be a result of the network structure used in all conditions of the study, as the group leader was relaying this person’s opinions to the respondent.

DISCUSSION The present research sought to assess the relative contributions of experimentally manipulated expectations for group members, affective impressions of group members, and emergent perceptions of their influence to opinion and sentiment change. When status-bearing roles were assigned prior to the task, group leaders carried sufficient social influence to elicit change in respondent opinions and identity sentiments in keeping with the opinions expressed by the group leader. Following earlier research on expectation states, those who occupied high-status social positions (i.e., leaders) and whose contributions were supported by others had the largest effect on respondent opinions. Leaders who behaved assertively during group interaction, however, had no greater effect on opinions than those who were more submissive. This finding suggests that assigned leaders’ status garnered sufficient influence to shape respondent opinions, and did not require further substantiation through their behavior during the interaction. This lends support to the contention of expectation states theorists that initial performance expectations for the group relate directly to the group’s influence structure, so long as leaders are perceived as legitimate. All three manipulations employed during the group interaction had implications for change in identity sentiments. While the group leader’s opinion in many cases determined the direction of sentiment change, the assertiveness of their behavior and support of the third group member often determined the magnitude of change. In keeping with SINT, respondents’ emergent perceptions of the group’s social influence composition were important predictors of opinion and sentiment change, above and beyond the effects of the conditional manipulations. Generally speaking, high-influence group leaders tended to produce

Expectations, Social Influence, and Affect Control

93

greater change in opinions than low-influence leaders, in the direction of the leader’s expressed opinion. The perceived potency of self compared to leader was also an important determinant of judgments of relative influence associated with the task. Thus, results offer support for key intersections between the three theories proposed by Friedkin and Johnsen (2003). Results also suggest that inferences about the relative status of other group members (e.g., gender attributions) may legitimize and reinforce the influence order of the group, when information about status characteristics is unavailable. Just as performance expectations and identity meanings shape the relative influence of group members, influence structures may shape identity sentiments and inferences about status characteristics, supporting the equilibrium of the influence network. Some limitations of the present research are worth noting. As mentioned previously, the experimental design had explanatory tradeoffs. It allowed for a test of the interplay between performance expectations, affective impressions, and social influence in small groups, through direct manipulation of group consensus, opinions, and behavioral interchange patterns. This was useful in testing the propositions of earlier research, as well as some new hypotheses that were previously untested. It did not, however, allow for formal testing of the relationship between the full matrices of expectation standing, affective impressions, and social influence for the groups under study. Such a test would have required all group members to be respondents in the study and to have completed measures on each metric. This is an important area of study for future research, as it would allow for more rigorous testing of the mathematical propositions forwarded by earlier research, particularly in circumstances where the group does not agree about the initial expectation standing of group members. The findings presented here suggest that interpersonal influences shape the semantic content of group discussion as well as definitions of the ongoing interaction. Future studies should seek to determine whether estimates of relative influence and expectation standing based on the ratings of all group members act upon each in the same manner as shown here. In addition, all groups under study were three-person task groups with an identical network structure, wherein the respondent only interacted directly with the purported group leader. Future work should explore groups of different sizes and structures to determine whether these findings can be generalized across network structures in the manner suggested by earlier research. The study is also unable to conclude with certainty that a causal relationship exists between respondents’ social influence ratings and opinion change. The design follows Friedkin’s (1999) earlier work in having

94

KIMBERLY B. ROGERS

respondents rate perceived influence after the group interaction is complete and final opinions have been reported. As a result, opinion change may have shaped perceived influence rather than the reverse. Future research should pursue a causal test of these effects to ensure their directionality corresponds to that specified in the network theory of social influence. Friedkin (1999) predicts that influence ratings following an initial interaction will predict opinion change in a second, topically related interaction; such a design may be useful for pursuing questions of causality. Social influence network theorists have shown that patterns of social influence are importantly related to network cohesion (Friedkin, 1993), as well as actors’ proximity (Friedkin, 2009), centrality (Friedkin, 1991, 1993), and structural equivalence within networks (Friedkin & Johnsen, 1997). Future studies should further examine how features of network structure and position act upon individual and interpersonal processes to shape opinions and sentiments. These issues were beyond the scope of the present research, which focused on a single network structure in the interest of exploring the interaction between numerous concepts and predictors of both social influence and opinion change. In addition, status characteristics and behavioral interchange patterns constitute only two of the three established predictors of performance expectations in EST (Correll & Ridgeway, 2003). Unequal distributions of resources are an important means by which social characteristics take on status value and also contribute important, independent effects on task-related behavior. More work is needed to understand how resource distributions contribute to perceived influence and opinion change, in dialogue with the two other theories discussed here. In sum, expectations states theory, affect control theory, and social influence network theory provide complimentary views of the multi-level processes operating on the stability and change of opinions through social interaction. EST offers insight into the micro-level processes that define and shape the status order and, thus, relative influence within group interaction. ACT links identity meanings representing status, power, and agency with inequalities of behavior and emotion in interaction, and models dynamic changes in meanings in response to interaction dynamics. SINT extends our understanding of influence processes to networks of arbitrary size and explains the implications of dynamic changes in influence for opinion change and norm formation. The present research sought to advance the integration of these theories by testing the relationship between key predictors of opinion change associated with each. Future work should continue to pursue this effort, with particular focus on the areas of study proposed herein. Together, the three theories help further our understanding of the relationship between the mental

Expectations, Social Influence, and Affect Control

95

representations of individual group members and the creation, maintenance, and transformation of opinions within social networks.

NOTES 1. The following identities were rated prior to the group deliberation task: receptionist, manager, flirt, sexist, victim, leader, employee, subordinate, myself as I really am, and myself as others see me. 2. There is some debate about the appropriateness of using conditional change models versus employing difference scores as a dependent measure in analyses of change (Allison, 1990). To ensure that the results presented here are robust, opinion change hypotheses were also tested with conditional change models (Finkel, 1995), which yielded identical results. 3. Notably, many respondents had a strong and stable preference for the settlement of the claim, with 50.6% in favor of settlement both before and after the group interaction. In contrast, only 14.8% of respondents had a stable preference for fighting the charges in court.

ACKNOWLEDGMENTS This chapter was made possible by a dissertation improvement grant awarded to Kimberly B. Rogers by the National Science Foundation (#1003419) and funding support from the Graduate School and the Department of Sociology at Duke University. Many thanks to Lynn Smith-Lovin, Linda K. George, Miller McPherson, Kenneth Spenner, Mark Leary, and several anonymous reviewers for their helpful comments on earlier versions of the manuscript.

REFERENCES Allison, P. (1990). Change scores as dependent variables in regression analysis. Sociological Methodology, 20, 93
114. Anderson, C., Brion, S., Moore, D. A., & Kennedy, J. A. (2012). A status-enhancement account of overconfidence. Journal of Personality and Social Psychology, 103, 718
735. Anderson, C., Srivastava, S., Beer, J. S., Spataro, S. E., & Chatman, J. A. (2006). Knowing your place: Self-perceptions of status in face-to-face groups. Journal of Personality and Social Psychology, 91, 1094
1110. Asch, S. E. (1955). Opinions and social pressure. Scientific American, 193, 31
35.

96

KIMBERLY B. ROGERS

Bennis, W., & Goldsmith, J. (2010). Learning to lead: A workbook on becoming a leader (4th ed.). New York, NY: Basic Books. Berger, J., Cohen, B. P., & Zelditch, M., Jr. (1972). Status characteristics and social interaction. American Sociological Review, 37, 241
255. Berger, J., Conner, T. L., & Fisek, M. H. (1974). Expectation states theory: A theoretical research program. Cambridge, MA: Winthrop. Berger, J., Fisek, M. H., Norman, R. Z., & Wagner, D. G. (1985). Formation of reward expectations in status situations. In J. Berger & M. Zelditch, Jr. (Eds.), Status, rewards, and influence (pp. 215
261). San Francisco, CA: Jossey-Bass. Berger, J., Fisek, M. H., Norman, R. Z., & Zelditch, M., Jr. (1977). Status characteristics and social interaction: An expectation states approach. New York, NY: Elsevier. Berger, J., Rozenholtz, S. J., & Zelditch, M., Jr. (1980). Status organizing processes. Annual Review of Sociology, 6, 470
508. Cast, A. D. (2003). Power and the ability to define the situation. Social Psychology Quarterly, 66, 185
201. Cook, K. S. (1975). Expectations, evaluations, and equity. American Sociological Review, 40, 372
388. Correll, S. J., & Ridgeway, C. L. (2003). Expectation states theory. In J. Delamater (Ed.), Handbook of social psychology (pp. 29
51). New York, NY: Plenum. Crano, W. D. (2000). Milestones in the psychological analysis of social influence. Group Dynamics: Theory, Research, and Practice, 4, 68
80. Dippong, J., & Kalkhoff, W. (2015). Predicting performance expectations from affective impressions: Linking affect control theory and status characteristics theory. Social Science Research, 50, 1
14. Dovidio, J. E., Brown, C. E., Heltman, K., Ellyson, S. L., & Keating, C. F. (1988). Power displays between women and men in discussions of gender-linked tasks: A multichannel study. Journal of Personality and Social Psychology, 55, 580
587. Driskell, J. E., & Mullen, B. (1990). Status, expectations, and behavior: A meta-analytic review and test of the theory. Personality and Social Psychology Bulletin, 16, 541
553. Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum. Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice towards female leaders. Psychological Review, 109, 537
598. Eaton, A. A., Majka, E. A., & Visser, P. S. (2008). Emerging perspectives on the structure and function of opinion strength. European Review of Social Psychology, 19, 165
201. Festinger, L., Schachter, S., & Back, K. W. (1950). Social pressures in informal groups: A study of human factors in housing. New York, NY: Harper and Brothers. Finkel, S. E. (1995). Causal analysis with panel data. Thousand Oaks, CA: Sage. Fisek, M. H., Berger, J., & Norman, R. Z. (1991). Participation in heterogeneous and homogeneous groups: A theoretical integration. American Journal of Sociology, 97, 114
142. Fontaine, J. R. J., Scherer, K. R., Roesch, E. B., & Ellsworth, P. C. (2007). The world of emotions is not two-dimensional. Psychological Science, 18, 1050
1057. Foschi, M. (1992). Gender and double standards for competence. In C. L. Ridgeway (Ed.), Gender, interaction, and inequality (pp. 181
207). New York, NY: Springer-Verlag. Foy, S., Freeland, R. E., Miles, A., Rogers, K. B., & Smith-Lovin, L. (2014). Emotions and affect as source, outcome and resistance to inequality. In J. McLeod, E. Lawler, & M. Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 295
324). New York, NY: Springer.

Expectations, Social Influence, and Affect Control

97

Francis, C. A., & Heise, D. R. (2006). Mean affective ratings of 1,500 concepts by Indiana University undergraduates in 2002
2003 [computer file]. Retrieved from http://www. indiana.edu/~socpsy/ACT/interact/JavaInteract.html Friedkin, N. E. (1991). Theoretical foundations for centrality measures. American Journal of Sociology, 96, 1478
1504. Friedkin, N. E. (1993). Structural bases of interpersonal influence in groups: A longitudinal case study. American Sociological Review, 58, 861
872. Friedkin, N. E. (1999). Choice shift and group polarization. American Sociological Review, 64, 856
875. Friedkin, N. E. (2001). Norm formation in social influence networks. Social Networks, 23, 167
189. Friedkin, N. E. (2005). A structural theory of social influence. New York, NY: Cambridge University Press. Friedkin, N. E. (2009). Jaccard-Spline index of structural proximity in contact networks. Social Networks, 31, 76
84. Friedkin, N. E., & Johnsen, E. C. (1997). Social positions in influence networks. Social Networks, 19, 209
222. Friedkin, N. E., & Johnsen, E. C. (1999). Social influence networks and opinion change. Advances in Group Processes, 16, 1
29. Friedkin, N. E., & Johnsen, E. C. (2003). Opinion change, affect control, and expectation states in the formation of influence networks. Advances in Group Processes, 20, 1
29. Gould, R. V. (2002). The origins of status hierarchies: A formal theory and empirical test. American Journal of Sociology, 107, 1143
1178. Heise, D. R. (1997). INTERACT: Introduction and software. Retrieved from http://www. indiana.edu/∼socpsy/ACT/interact.htm Heise, D. R. (1999). Controlling affective experience interpersonally. Social Psychology Quarterly, 62, 4
16. Heise, D. R. (2007). Expressive order: Confirming sentiments in social action. New York, NY: Springer. Kalkhoff, W., Friedkin, N. E., & Johnsen, E. C. (2010). Status, networks, and opinions: A modular integration of two theories. Advances in Group Processes, 27, 1
38. Koch, S. C., Mueller, B., Kruse, L., & Zumbach, J. (2005). Constructing gender in chat groups. Sex Roles, 53, 29
41. MacKinnon, N. J., & Langford, T. (1994). The meaning of occupational prestige scores: A social psychological analysis and interpretation. The Sociological Quarterly, 35, 215
245. Nowak, A., Szamrej, J., & Latane´, B. (1990). From private opinion to public opinion: A dynamic theory of social impact. Psychological Bulletin, 97, 362
376. Osgood, C. E., May, W. H., & Miron, M. S. (1975). Cross-cultural universals of affective meaning. Urbana, IL: University of Illinois Press. Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. Urbana, IL: University of Illinois Press. Ridgeway, C. L. (2001). Gender, status, and leadership. Journal of Social Issues, 57, 611
655. Ridgeway, C. L. (2014). Why status matters for inequality. American Sociological Review, 70, 1
16. Ridgeway, C. L., & Berger, J. (1986). Expectations, legitimation, and dominance behavior in groups. American Sociological Review, 51, 603
617.

98

KIMBERLY B. ROGERS

Ridgeway, C. L., Berger, J., & Smith, L. (1985). Nonverbal cues and status: An expectation states approach. American Journal of Sociology, 90, 955
978. Ridgeway, C. L., Boyle, E. H., Kuipers, K., & Robinson, D. (1998). How do status beliefs develop? The role of resources and interaction. American Sociological Review, 63, 331
350. Ridgeway, C. L., & Erickson, K. G. (2000). Creating and spreading status beliefs. American Journal of Sociology, 106, 579
615. Ridgeway, C. L., & Smith-Lovin, L. (1994). Structure, culture, and interaction: Comparing two generative theories. Advances in Group Processes, 11, 213
239. Ridgeway, C. L., & Smith-Lovin, L. (1999). The gender system and interaction. Annual Review of Sociology, 25, 191
216. Robinson, D. T., & Smith-Lovin, L. (2006). Affect control theory. In P. Burke (Ed.), Contemporary social psychological theories (pp. 137
164). Stanford, CA: Stanford University Press. Rogers, K. B., Schro¨der, T., & Scholl, W. (2013). The affective structure of stereotype content: Behavior and emotion in intergroup context. Social Psychology Quarterly, 76, 125
150. Scholl, W. (2013). The socio-emotional basis of human interaction and communication: How we construct our social world. Social Science Information, 52, 3
33. Shelly, R. K. (1997). Sequences and cycles in social interaction. Small Group Research, 28, 333
356. Sherif, M. (1936). The psychology of social norms. New York, NY: Harper. Smith-Lovin, L. (1987). Impressions from events. Journal of Mathematical Sociology, 13, 35
70. Smith-Lovin, L., & Heise, D. R. (1988). Analyzing social interaction: Research advances in affect control theory. New York, NY: Gordon and Breach. Stewart, P. A., & Moore, J. C. (1992). Wage disparities and performance expectations. Social Psychology Quarterly, 55, 78
85. Visser, P. S., & Mirabile, R. R. (2004). Opinions in the social context: The impact of social network composition on individual-level opinion strength. Journal of Personality and Social Psychology, 87, 779.

HOW DOES STATUS AFFECT POWER USE? NEW PERSPECTIVES FROM SOCIAL PSYCHOLOGY Ko Kuwabara ABSTRACT Purpose
Are people more or less likely to use their power if they have high social status? This chapter discusses how having status affects the use of power by those in positions of power in exchange relations or small groups. Although status and power are typically assumed to be mutually reinforcing, there is growing recognition that having status may actually inhibit the use of power under certain conditions. Methodology/approach
I review relevant research findings and consider three variables in particular that may moderate the effects of status on the use of power: legitimacy of status, achieved versus ascribed status, and individualist versus collectivist cultures. Research implications
While status and power are close correlates, there is growing recognition
particularly in organizational psychology
that, under certain conditions, having status may inhibit the use of power or that lacking status increases power use. These studies shed new light on how status interacts with power in hierarchical groups and challenge the pervasive view of power and status as mutually

Advances in Group Processes, Volume 32, 99
121 Copyright r 2015 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0882-6145/doi:10.1108/S0882-614520150000032004

99

100

KO KUWABARA

reinforcing forces that perpetuate inequalities. Understanding more precisely when and why status and power have convergent or divergent effects on power use is an important task for scholars of group processes. Originality/value
The possibility that status and power can have distinct consequences, let alone opposite effects, presents an intriguing opportunity for scholars of group processes to rethink and extend our understanding of social hierarchies in a new light. Keywords: Status; power; social hierarchy; culture; legitimacy

Questions of power and status have long inspired sociological inquiry, informing our understanding of bureaucracy and work (Abbott, 1981; Blau & Duncan, 1967; Dornbusch & Scott, 1975; Weber, 1978), organizations and markets (Pfeffer & Salancik, 1978; Podolny, 1993), stratification and inequality (Bendix & Lipset, 1966; Ridgeway, 2011), conflict and social change (Dahrendorf, 1959; Marx, 1888), as well as social exchange and group dynamics (Berger & Zelditch, 1998; Cook & Emerson, 1978). Although early theorists proposed various bases of power (Blau, 1964; French & Raven, 1959), contemporary scholars generally refer to power as the structural potential to promote or create unequal redistribution of resources and benefits through actual use of power (Molm, 1988; Thye, 2000). Power therefore varies with one’s access to or control over resources valued by others, including the allocation of rewards and punishment. In contrast, social status refers to one’s position in a social system based on recognition, respect, or deference given to certain individuals by others (Berger, Cohen, & Zelditch, 1972). It is one’s relative rank in a social hierarchy based on how actors view or regard each other. While status and power are close correlates, they are nevertheless distinct attributes of social hierarchies
much like the weight and height of a person
with potentially distinct consequences on social action and dynamics. The purpose of this chapter is to explore how having high status affects the use of power by those in positions of power in exchange relations or small groups. Does status increase power use? Are people more likely to use their power if they feel respected or disrespected? For instance, research from psychology has long shown that power corrupts (Kipnis, 1972; Milgram, 1963; Zimbardo, Maslach, & Haney, 1999).

How Does Status Affect Power Use?

101

Assigning people to positions of power, even randomly, is enough to trigger various coercive, aggressive, or even antisocial tendencies of dominance. But does this effect hold across different levels of status, from low to high? Some people hold high power by virtue of, or in conjunction with, high status, such as those in positions of organizational leadership. Others hold high power while occupying relatively low ranks, such as secretaries, nightclub bouncers, or IT support staff. Who is more likely to use the power given to them? Although sociologists have been careful to distinguish power and status as distinct concepts (Berger & Zelditch, 1998; Thye, 2000; Weber, 1978), much remains unknown about whether, when, and why status might have distinct effects from power. Conceptually, power and status are both properties of social relations
power based on relative dependence on one another for valued resources and status based on respect or recognition from others
at once differentiating and constraining people into social hierarchies. In effect, power and status are often highly correlated and difficult to isolate in natural settings. Even small differences in the initial distribution of resources or respect can create mutually reinforcing cycles that reproduce inequality, as power begets status and status reinforces power (DiPrete & Eirich, 2006; Merton, 1968; Weber, 1978). Laboratory studies designed to tease apart power and status under careful experimental control have also shown by and large that power confers status (Lovaglia, 1995; Ridgeway, 1991), and status creates power differences (Thye, 2000). Nevertheless, there is growing recognition
particularly in organizational psychology
that, under certain conditions, having status may actually inhibit the use of power (Blader & Chen, 2012) or that lacking status increases power use (Fast & Chen, 2009; Fast, Halevy, & Galinsky, 2012; Wiltermuth & Flynn, 2012). These studies shed new light on how status interacts with power in hierarchical groups and challenge the pervasive view of power and status as mutually reinforcing forces that perpetuate inequalities (Gould, 2002; Magee & Galinsky, 2008). In turn, understanding more precisely when and why status and power have convergent or divergent effects on power use is an important task for scholars of group processes. The purpose of this essay is to review relevant research findings and consider three variables in particular that may moderate the effects of status on the use of power: legitimacy of status, achieved versus ascribed status, and individualist versus collectivist cultures. These variables, I argue, can interact with status to affect how people use their power in given exchange tasks when they feel respected. This review is not meant to be

102

KO KUWABARA

comprehensive, but to draw attention to some novel results from recent studies, particularly those outside of sociology.

POWER, POWER USE, AND STATUS From Power to Power Use How power is allocated and used has long been a topic of considerable interest across the social sciences and beyond. In sociology, a significant body of work has explored how the allocation of power shapes or converts to the use of power in exchange relations (Emerson, 1976; Molm, 1990). Drawing on various traditions, including power-dependence theory (Cook & Emerson, 1978), elementary theory (Willer, Markovsky, & Patton, 1989), and network exchange theory (Willer, 1999), this line of research broadly defines power use as behavioral patterns that create asymmetric exchange outcomes in favor of the power holder, typically through direct bargaining over the division of benefits, assigning rewards and punishment to each other, and withholding exchange opportunities from certain others. Thus, power is distinct from the actual, behavioral use of power although power typically promotes power use, exactly how it seems to depend on the nature of power and the social structure in which it is embedded (Molm, 1988, 1989). While sociologists have focused almost exclusively on structural power, research in organizational psychology has examined subjective power, that is, feeling powerful (by virtue of having power). One increasingly common way to induce feelings of power (Galinsky, Gruenfeld, & Magee, 2003) is a recall task in which experimental participants are asked to spend a few minutes describing a time when they had power over another person (high power), or vice versa (low power). Using this or other tasks to manipulate feelings of power, research has found that simply feeling powerful can increase a variety of disinhibited behaviors (Hirsh, Galinsky, & Zhong, 2011; Keltner, Gruenfeld, & Anderson, 2003), leading people to make more selfish decisions (Blader & Chen, 2012; Chen, Lee-Chai, & Bargh, 2001), show less empathy for others (Galinsky, Magee, Inesi, & Gruenfeld, 2006; Van Kleef, De Dreu, Pietroni, & Manstead, 2006), devalue or objectify others (Gruenfeld, Inesi, Magee, & Galinsky, 2008; Kipnis, 1972), and endorse harsher punishment (Wiltermuth & Flynn, 2012).

How Does Status Affect Power Use?

103

How Status Promotes or Inhibits Power Use More recent studies have shown, however, that having high status can counteract these dominance tendencies among the powerful. People given status and made to feel respected are less likely to engage in aggressive bargaining (Blader & Chen, 2012), assign demeaning tasks to others (Fast et al., 2012), and mete out punishment to enforce group norms (Kuwabara, 2013). There are several explanations for why status might restrain use of power. First, status sensitizes people to how they might be viewed by others for resorting to coercive power (Blader & Chen, 2012). Second, having high status promotes a sense of self-worth, reducing the need to display dominance to earn others’ respect (Fast et al., 2012). Third, status legitimates social influence, reducing the need for coercive means to steer others (Berger, Ridgeway, Fisek, & Norman, 1998; Tyler, 2005). Finally, clear hierarchical differences can discourage group members from resorting to competitive displays of dominance (Anderson, Ames, & Gosling, 2008; Gould, 2003; Halevy, Chou, Galinsky, & Murnighan, 2012). The idea that power and status have distinct
indeed, opposite
effects is surprising. It is typically assumed that power and status are mutually reinforcing (Magee & Galinsky, 2008). Power reinforces status and status confers power, increasing power use. According to Thye’s (2000) status-value theory of exchange, for instance, the social status of an actor imbues exchangeable resources with status value, creating bargaining-power advantage by making the resources of the high-status holder subjectively more valuable. To demonstrate, he conducted laboratory studies using simple bargaining tasks over the division of fixed resources. To create status differences, participants (undergraduates) were told that they were paired against an older graduate student with a superior GPA or a younger highschool student with a low GPA, inducing relatively low versus high status compared to the partner. The dependent measure was the number of resources earned through bargaining. Participants who were assigned highstatus positions (paired with a lower status, high-school student) earned significantly more than those in low-status positions (paired with a higher status, graduate student), controlling for structural power (i.e., number of alternative partners or resources) in their exchange networks. More recently, however, Blader and Chen (2012) demonstrated in laboratory studies that high-status actors can be less aggressive in bargaining tasks. Their experiments examined the likelihood of making the first

104

KO KUWABARA

move in dyadic negotiations, often considered a display of power and assertiveness, by people assigned to high-status, high-power, or baseline conditions. The experimental treatment was based on text included in the role materials for the negotiation, which characterized their particular role as one of high power or status; the baseline condition lacked the text manipulation. They found that those given high power were more likely, but those given high status were less likely, than the baseline condition, to make the first move. Thus, assigning high power versus status had the opposite effects. These studies differ from each in a number of ways to allow direct inferences about the effects of power versus status. For instance, the dependent variable for Blader and Chen was the likelihood of making the first offer rather than the final agreement. Second, their negotiations were face-toface rather than computer mediated. Such differences in methodology as well as the interpretation of results suggest that an important task for scholars of group processes and exchange is to better understand when and why status promotes or inhibits power use. Below, I consider three variables that may affect how status affects the use of power: legitimacy of hierarchies, achieved versus ascribed status, and individualist versus collectivist culture. The rationale for considering these possible moderators is the idea that status effects are often context specific. Whereas power is typically grounded in objective, exogenous allocation of resources or value, status stems from collective perceptions of social worth, which can depend in critical ways on how people understanding the nature and normative significance of their status differences.

POSSIBLE MODERATORS OF STATUS EFFECTS ON POWER USE Legitimacy of Hierarchies To what extent status inhibits acts of dominance may depend on the extent to which status invokes legitimacy concerns. Legitimacy refers to the degree of perceived consensus about what is recognized or respected by peers as apparently or presumably proper (fair and appropriate) and valid (widely accepted) in a given situation or role (Dornbusch & Scott, 1975; Johnson, Dowd, & Ridgeway, 2006). Legitimate people are “taken for granted” while those lacking legitimacy (illegitimate or nonlegitimate ones) may be

How Does Status Affect Power Use?

105

met with disapproval and resistance. Legitimacy is an important element of whether a status hierarchy is accepted or rejected. This definition of legitimacy precludes the idea of illegitimate status. Although a person’s low status may be deemed legitimate
that is, a person may be assigned to a low-status position for perfectly legitimate reasons
one cannot enjoy high status (respect or recognition) that lacks claim to collective acceptance or endorsement, because having status depends critically on how one is viewed by others as not only worth of respect but also valid and proper. It is possible, however, for individuals to have high status in a hierarchy that lacks legitimacy, because legitimacy is a property of individuals and their acts as well as systems. For instance, some people enjoy respect or deference from others for values or traits that are largely irrelevant to a given situation because of hierarchies that impose status orders on arbitrary bases. Many diffuse status characteristics like gender, race, or physical attractiveness can accord high status to members of one group (e.g., white men) over others, even when such status differences carry no logical or apparent merit. Thus, it is possible to confer someone high status for a particular trait or quality, even while questioning the validity and propriety of allocating privileges or resources based on such bases. I suggest that the effect of status on power use hinges directly on legitimacy, because an important effect of having high status is to sensitize people to how they are viewed or accepted by others, including whether they are fair, proper, and hence legitimate (Blader & Chen, 2012). Having high status makes people attentive to others, because status derives from how people regard them. This is in effect the opposite of power, which alters how people view others; rather than sensitizing people to the perspectives and feelings of others, power in fact desensitizes people, making them less empathetic and more self-centric (Galinsky et al., 2006; Gruenfeld et al., 2008; van Kleef et al., 2008). How should the legitimacy of a status hierarchy affect power use? If the effect of status is to sensitize people to the concern that acts of power and dominance maybe seen by others as selfish, unjust, and inappropriate, then delegitimizing one’s status might reduce such effects. People with illegitimate status should therefore care less about appearing fair and appropriate. In this view, lacking legitimate status is functionally equivalent to lacking status. Another possibility, however, is that having high but nonlegitimate status reduces power use even more than high legitimate status. The reason is that people who lack legitimacy might feel less entitled to use their resources and privileges. Nonlegitimate high status makes people feel

106

KO KUWABARA

respected but not necessarily accepted because their status lacks the confidence of peer acceptance and hence feels precarious. Compared to people in positions of legitimate high status, those in positions of high but nonlegitimate status might therefore care even more about how they will be viewed by others for using power against them. Using power, they feel, is not only unbefitting of their status but also inappropriate given their lack of claim to legitimacy. Examining Status Effects on Punishment in Legitimate versus Nonlegitimate Hierarchies To explore the effect of legitimate versus nonlegitimate hierarchies on power use by high-status actors, I conducted a public goods experiment with punishment (Kuwabara, 2013). The public goods game models situations in which members of a group must independently decide how much to contribute to the production of a public good that benefits everyone in the group. Because no person can be excluded from consuming the public good, each person is better off free riding on others to contribute, but nothing will be produced if everyone decides to shirk their share of the burden. Understanding how to enforce cooperation and solve such social dilemmas is a critical issue for the success and survival of many teams, communities, and institutions (Ostrom, 1990). To examine such issues, public goods games with “peer punishment” (Fehr & Ga¨chter, 2002) alters the standard paradigm by allowing group members (“peers”) to punish each other. Each round consists of two stages. In the contribution stage, each member is given an endowment of 20 points, of which they could contribute any amount to a team project or keep. Each point contributed to the team project yields .5 point for each member, and thus 1.5 points for the whole team, whereas keeping 1 point yielded 1 point for that member only. Thus, the earning πi,t for member i in the first stage of period t is π i;t = 20 − ci;t þ :5

3 X

cm;t

m=1

where ci,t is i’s contribution at t and m denotes the three members of i’s team. After each round, each member’s contribution is shown to the other members of the team on the computer screen. In the punishment stage, each punisher i is given an opportunity to assign deduction points pim (0
10 points) out of i’s earnings to each other

107

How Does Status Affect Power Use?

member m. Each punishment point costs the punisher 1 point and the target 3 points. Thus, the final payoff in each period t for player i in peer punishment is 3 3 X X π^ i;t = π i;t − pim;t − 3 pmi;t i≠m

i≠m

In my study, I made another important modification to this design by designating one member to be the “solo P P3 punisher” in each group, such that 3 3i ≠ m pmi;t = 0 for punishers and i ≠ m pim;t = 0 for nonpunishers in the final payoff. In effect, designating a solo punisher created within-group differences in power, that is, who could versus who could not punish. I created three experimental conditions based on the solo punishment paradigm. In the baseline condition, the solo punisher was chosen randomly in each group, creating power difference without introducing status differences a priori. In the legitimate high-status condition, the solo punisher was chosen based on a “leadership aptitude” questionnaire that was purported to correlate with effectiveness in leadership, thus creating differences in power and status; similar procedures have been used to manipulate status in past research (Anderson & Berdahl, 2002; Lount & Pettit, 2012). In the nonlegitimate high-status condition, the solo punisher was chosen based on a trivia quiz about explicitly task-irrelevant topics (e.g., art, botany, and geography), thus creating status based on recognition of high scores but low legitimacy. Unbeknownst to participants, all punishers were assigned randomly. However, the manipulation checks (t-tests of preexperimental control questions) confirmed that solo punishers in the legitimate and illegitimate conditions were perceived as higher status than in the baseline conditions by both punishers and nonpunishers. At the same time, using the leadership questionnaire to assign punishers was perceived as more legitimate (fair) than using the trivia quizzes. Assigning punishers randomly was also perceived as more legitimate than using the trivia quizzes; random assignment was thus perceived as more legitimate than “arbitrary” assignment. After the assignment to condition, teams, and roles, participants played eight rounds of public goods games with punishment. To see the effects of legitimate and illegitimate status against the baseline condition, I examined the amount of punishment in each condition. As Fig. 1 shows, punishers with legitimate status punished less than punishers without a priori status and punishers with illegitimate status. Because punishment depends on the level of contributions, I verified the results in Tobit regression predicting

108

KO KUWABARA

Deduction points

.8

.6

.4

.2

0 No status

Fig. 1.

High status

Nonlegitimate status

Average Punishment Points Assigned per Target Per Round. Source: Kuwabara (2013).

punishment per round per target, controlling for the contribution from each group member (the punisher, target, and the third member), results from the previous round, as well as clustered errors at the individual level for repeated measures. The analytic results confirmed the descriptive patterns. To summarize, this study showed that assigning high status to solo punishers reduced punishment as a form of power use, whether status was viewed as legitimate or nonlegitimate. In fact, solo punishers with high but nonlegitimate status were even less likely to use punishment than those with high legitimate status as well as those with no a priori status. Thus, lacking legitimacy is functionally distinct from lacking status. More generally, the legitimacy of the hierarchy seems to be an important moderator of how status affects power use. Costly Punishment as a Basis of Legitimacy In the study above, legitimacy derived from assigning solo punishers based on task-relevant or irrelevant status. To see if similar patterns occur under other forms of legitimate versus nonlegitimate punishment, we designed another study (Kuwabara & Yu, 2014) to manipulate the legitimacy of hierarchies by making punishment costly or costless to punishers, such that assigning each deduction point reduced the punisher’s earning by 1 (costly) or 0 (costless) point. The rationale is that punishment should be perceived as more legitimate when it is costly rather than costless, because enforcing

How Does Status Affect Power Use?

109

group cooperation at one’s own expense is perceived as more selfless and fair. Such acts of self-sacrifice for collective interests are often rewarded with peer approval and compliance (Barclay, 2006). In this view, costly punishment is a social signal that conveys prosocial motives and claims legitimacy as an in-group member. Conversely, using punishment that is costless is more likely to be seen as abusive or selfish. In addition to the effect of costly versus costless punishment, we also predicted that solo punishers may be more sensitive than peer punishers (i.e., members of peer-punishment groups in which everyone has the power to punish) to legitimacy concerns
appearing fair and just to others
that may curtail such antisocial tendencies, because their unique status and power attracts greater attention and scrutiny (Devlin-Foltz & Lim, 2008). In comparison, legitimacy is less salient for peer punishers because each person has the same role; their power to punish is therefore more likely to be taken for granted, whether it is costly or costless. To test this prediction, we first ran a pilot study in which 238 volunteers were asked to read a written description of the public goods game with either costly or costless punishment and elicit their impressions of the enforcement system (“how fair/appropriate is this enforcement system?”) on 7-point scales. Solo punishment was perceived as more fair and appropriate when costly (4.16) than costless (3.48), p < .01. However, peer punishment was perceived as equally fair and appropriate when costly (4.01) versus costless (4.16), p = .54. These results provided initial support for the idea that making punishment costly versus costless for punishers changes the perceived legitimacy of punishment, but only by solo punishers. To examine the behavioral consequences of this idea, we modified the public goods game with punishment to manipulate the cost of punishment to punishers (holding constant the cost to targets) in groups with peer versus solo punishment, creating four conditions: peer-costless, peer-costly, solo-costless, solo-costly. In all conditions, participants played six rounds of public goods games in groups of three. In each round, each member was given an endowment of 20 points, of which they could contribute any amount to the group or keep. In the peer conditions, each member could review individual contributions from the round and assign 0
10 punishment points to each other; each deduction point reduced the target’s total earning by 3 points and the punisher’s earning by 0 (costless) or 1 (costly) point. In the solo conditions, one member was randomly chosen to be the solo punisher. The results showed that, as expected, peer punishers punished less when punishment was costly than costless. This was not the case for solo

110

KO KUWABARA

punishers, however. As Fig. 2 suggests, there was almost no difference in total punishment between the solo-costless and solo-costly conditions. Although this alone is contrary to the rational prediction (as well as what we observe in the peer punishment conditions), we also find distinct patterns when we examine punishment toward relative defectors versus cooperators, that is, prosocial punishment (punishing a target who contributed fewer points than the punisher) versus antisocial punishment (punishing a target who contributed equal or more points than the punisher). While costly punishment decreased antisocial punishment by both peer and solo punishers, it decreased prosocial punishment among peer punishers but increased it among solo punishers. These results, confirmed in two-way (costly × solo punishment) ANOVAs on total points spent per punisher on anti- and prosocial punishment, suggest that peer punishers engaged less in both pro- and antisocial punishment when punishment was costly rather than costless, presumably because they are concerned primarily with the economics of punishment rather than legitimacy. In contrast, solo punishers under costly punishment punished more prosocially and less antisocially because they were concerned about legitimacy. To confirm these patterns, we used Tobit regression with standard errors clustered at the level of individual punishers and teams, controlling for contributions from the punisher, the target, and the team total in each round

Deduction points

15

10

5

0

P0

P1

S0

S1

Fig. 2. Total Punishment Points Assigned per Punisher in Kuwabara and Yu (2014). Note: P0 = peer costless, P1 = peer costly, S0 = solo costless, and S1 = solo costly. Dark and light bars indicate the total number of punishment points that were assigned prosocially versus antisocially.

How Does Status Affect Power Use?

111

and the fixed effects of rounds. The results converged with the descriptive patterns. First, we found significant effects of costly × solo punishment, suggesting that the cost of punishment had different effects on peer versus solo punishers; this effect held, whether antisocial punishment was included or excluded. Second, looking at antisocial punishment only, both peer and solo punishers were less likely to punish antisocially under costly punishment. In the exit survey after the public goods games, participants were asked how legitimate (fair and respected) they felt while playing their assigned roles on 7-point scales. Solo punishers in S1 reported feeling more legitimate than in S0, 4.88 > 3.55, p = .012, in a t-test. Nonpunishers also felt more legitimate in S1 (4.61) than S0 (3.57), p = .008. No such differences were found between P0 (3.78) and P1 (4.14), p = .20. In addition, in the designated punishment conditions only, nonpunishers were asked how fair the punisher in their team was. They evaluated their solo punisher as more fair in S1 (4.53) than S0 (3.43), p = .013. Finally, an indirect but consequential measure of legitimacy is compliance, that is, how targets of punishment respond (Baldassarri & Grossman, 2011). To examine compliance, we submitted changes in contributions from targets immediately after punishment (in round t + 1) to Tobit regression against punishment in round t × costly punishment, controlling for lagged effects of contributions. Results revealed a significant positive effect for targets of solo punishers (p = .03) but not peer punishers (p = .28), indicating that contributions increased more after receiving costly (vs. costless) punishment from solo punishers, but not from peer punishers. Summary While high status reduced power use, compared to hierarchies of power with no a priori status differentiation, this constraining effect of status was even stronger for punishers with nonlegitimate status (Kuwabara, 2013). In Kuwabara and Yu (2014), we found a similar effect of nonlegitimacy to reduce punishment by solo punishers (but not peer punishers in groups with not status differentiation). Thus, across two studies using public goods games, I found evidence that how status affects power use depends on the legitimacy of hierarchies. It should be noted, however, that lacking legitimacy is different from illegitimacy. Although punishers in the nonlegitimate conditions
whether chosen based on trivia quizzes or giving power to punish costlessly
were not seen as particularly fair or proper, neither were they rejected as unfair or improper. More research is needed to elucidate how illegitimate status affects power use.

112

KO KUWABARA

Achieved versus Ascribed Status Another potential moderator of how status affects power use is whether status is achieved or ascribed. Consider the following two studies by Blader and Chen (2012) and Visser and Roelofs (2011). Both studies used the Dictator Game, one of the simplest tasks for studying power in the division of resources between two people. One person plays the role of the proposer, who is asked to determine the division of certain endowment, typically money or points, between him- or herself and another person. The proposer may keep any amount of the endowment; the remainder is given to the second player who is entirely passive, given no way to reject the offer or make a counter offer, creating explicit inequality in the distribution of power. Blader and Chen (2012) examined how social status affects allocation decisions in the Dictator Game by assigning participants to high or default status (control). They found that high-status participants gave greater portions of the endowment (42.71%) than those in the control group (29.67%). Thus, they still kept more for themselves than gave away in both conditions, they gave away more generously when they had high status. This result is consistent with their primary argument that having status, that is, feeling respected, sensitizes people to concerns about how they are viewed by others. However, Visser and Roelofs (2011) find more mixed results. In their study, participants who had been assigned high status gave either less, or no less and no more, to their partners than did those with low status. What explains the different patterns? While these studies, despite their ostensible similarity, are different in several respects, one possibility is that status was earned or achieved in one study but given or ascribed in the other. In Blader and Chen (2012), participants were randomly assigned to a high-status position and told to assume the role of someone “regarded with greater esteem, respect, and admiration from others.” Hence, status was simply assigned or ascribed to participants. In comparison, Visser and Roelofs followed Ball, Eckel, Grossman, and Zame’s (2001) procedure in which participants are first given trivia quizzes (on obscure economics questions unrelated to the experimental task) and assigned to different status groups based on their actual answers. Although the questions were scored more or less randomly (and unbeknownst to participants), participants were led to belief that they earned their status. To reinforce the status assignment, those in high status received a golden star to wear and a round of applause from lower status participants.

How Does Status Affect Power Use?

113

Why does it matter whether status is earned or given? One possibility is legitimacy. Assigning status randomly may seem fair, but it is unclear whether it satisfies the “propriety” component of legitimacy, that is, it desirable or appropriate? In this sense, randomly ascribing status may seem illegitimate, compared to achieving status based on some external measure. Nevertheless, achieved status may also lack legitimacy if how it is earned is questionable. Using trivia quizzes on task-unrelated topics may be viewed as unfair and inappropriate, as was the case in my aforementioned study (Kuwabara, 2013). In other words, both ascribed and achieved status can be legitimate or nonlegitimate, depending on whether they are recognized by peers or not. Hence, legitimacy may not be the root issue of contradictory results from Blader and Chen (2012) and Visser and Roelofs (2011). Another possibility is that earning status through a competitive process (e.g., outscoring others on trivia quizzes) could prime people to assert their status against others by framing subsequent tasks or interactions in more competitive terms as well. If outdoing others is recognized with higher status, participants may feel encouraged to compete against lower status counterparts and affirm their status. In comparison, when status is ascribed randomly or matter-of-factly based on taken-for-granted societal conventions (e.g., diffuse status characteristics), high-status actors may feel less competitive and less compelled to assert their status because they take their status more for granted. The distinction between achieved and ascribed status is not simply a methodological matter. They have long been a topic of core sociological inquiry (Foladare, 1969; Lipset & Bendix, 1991). As such, examining specific questions about how achieved versus ascribed status affects power use to reinforce or mitigate power dynamics and inequalities suggests an important avenue for future research.

Culture: Individualism versus Collectivism Culture has received relatively little empirical attention by scholars of the psychology of social status (Brashears, 2008; Torelli, Leslie, Stoner, & Puente, 2014). This is surprising given that the idea of power distance, that is, how people allocate and express power and status differences, is one of the core dimensions of cultural differences (Triandis, 1995). Moreover, it is widely acknowledged that status plays a more integral role in more collectivist cultures (Hofstede, 1980; Triandis, 1995). For instance, one characteristic of vertical collectivism that prevails in Asian cultures is the normative

114

KO KUWABARA

acceptance of inequalities within one’s group (Triandis & Gelfand, 1998), including conformity to role expectations, even at the expense of individual rights, interests, or identity (Kitayama, Markus, Matsumoto, & Norasakkunkit, 1997). In contrast to the Western notion of social status as a component of individual identity and self-worth (Blader & Chen, 2012), status is tied more strongly in Asia to one’s place in a collective group that prescribes one’s expected role and social standing within a hierarchy. Viewing social status in terms of individual versus collectivist values has novel implications for understanding how status might affect power use across cultures. As already discussed, research in the United States has shown status tempers acts of power and dominance by sensitizing people to the concern that acting aggressively may be viewed as unjust and disrespectable (Blader & Chen, 2012; Cheng, Tracy, Foulsham, Kingstone, & Henrich, 2012). Having high status also promotes a sense of self-worth, reducing the need to display dominance to earn others’ respect (Fast et al., 2012). These patterns may be particularly salient in individualist cultures where power use is typically viewed in terms of individual needs or self-interest (Torelli & Shavitt, 2010). In contrast, status may actually promote expressions of dominance by the powerful in collectivist cultures because the “ostensible” purpose of power among collectivists is to maintain the existing social order. For instance, sinologists have argued that the idea of social harmony in the Confucian ideology does not imply simply avoiding conflict between people (Leung, Koch, & Lin, 2002). Rather, “the Confucian emphasis on harmony is better understood as the need to maintain relations in a hierarchical social structure, rather than for seeking smooth relations with others” (Lun, 2012, p. 474). To this end, Confucian ethics actually mandates acts of dominance, not for the pursuit of personal ends per se, but to maintain and reinforce the status quo of hierarchical relations by asserting one’s relative status (Ip, 2009; Pye & Pye, 1985; Weatherley, 2002). As Pye and Pye observe in Asian Power and Politics (1985, chapter 2), “power … was an end value, not to be debased for utilitarian purposes,” and “any claims to power which fell outside [of formal hierarchies] were seen as illegitimate.” In the collectivist cultures of Asia, having status therefore requires using power while exercising power without the benefit of high status is seen as disruptive and counter normative. Experimental Tests of Status Effects on Power Use across Cultures My study with Siyu Yu and Adam Galinsky (Kuwabara, Yu, & Galinsky, 2014) illustrates these ideas. In three experiments, we examined

How Does Status Affect Power Use?

115

how social status affects punishment, as a form of power to enforce cooperation, in individualist versus collectivist cultures. We predicted that centralized punishers would punish less in the United States when they have high status, but more in Asian collectivist cultures such as India and China. Experiment 1 had participants from the United States and India (N = 75 per country, recruited from Amazon Mechanical Turk) engaged in “thirdparty punishment” (Fehr & Fischbacher, 2004) in which punishers monitored and punished teammates playing two-person public goods games over 19 rounds. Each round consisted of two stages. In the contribution stage, the workers separately decided how many hours (0, 10, or 20) to contribute to team projects that benefit the whole team versus personal projects that would only benefit themselves. In the punishment stage, the monitor evaluated each worker’s effort and productivity and provided feedback by assigning 0 to 8 punishment points. Each punishment point costs the target 3 points and the monitor 1 point. The monitor earned 50% of the joint earnings by the workers, minus any punishment points assigned to them. We simulated the workers in order to hold constant the level of cooperation that the punisher observed. Our experimental manipulation varied the status of punishers by assigning them to the punisher role “randomly” (equal-status condition) or based on a preexperimental leadership aptitude questionnaire (high-status condition). Manipulation checks confirmed that the high-status treatment successfully induced stronger feelings of status and respect, compared to equal status. Consistent with our prediction, Americans assigned more punishment in the equal-status (1.52 points) than the high-status condition (.99 point) per round. In contrast, Indians punished more in the high-status (2.15 points) than the equal-status condition (1.20 points). The culture × status interaction effect was significant in an ANOVA. Furthermore, we found that this interaction effect was mediated by “feeling respected” (measured as a manipulation check of the status manipulation), suggesting that feeling respected on the basis of having status had different effects on punishment in the United States and India (Fig. 3). Experiment 2 replicated these patterns using participants from the United States (N = 159) and China (N = 162) playing public goods games with punishment in which each punisher was a in-group member of a three-person team instead of a third-party outsider, and teammates were not simulated. The procedures were similar to my aforementioned study (Kuwabara, 2013). Each group played eight rounds of public goods games.

116

KO KUWABARA

3

Deduction points

US

India

2

1

0 Equal Status High Status Equal Status High Status

Fig. 3.

Results from Study 1 in Kuwabara et al. (2014).

The results dovetailed with Study 1. In the United States, equal-status punishers (1.11 points per round) punished more than high-status punishers (.75 point) and baseline punishers (.31 point). In China, higher status punishers (.86 point) punished more than equal-status punishers (.33 point) and baseline punishers (.52 point). Because these patterns do not control for contribution levels, we submitted our data to Tobit regression, controlling for contributions by each team member, fixed effects of rounds, and clustered errors at the individual level. Using this model, the effect of high status × culture (the United States) was significant, p =
.012, indicating that status had different effects across cultures again. The negative coefficient suggests that having high status reduced power use in the United States but increased it in China. Experiment 3 examined the causal role of culture by recruiting 200 Asian-Americans and activating either their Asian or American cultural identity. In the survey experiment, participants were first asked to list two to three ways in which they identify with their American or Asian culture. Next, they were asked to read a vignette about a team of four employees of an organization working on team versus personal projects, based on Experiments 1 and 2. Participants were assumed to take the role of the monitor, chosen randomly (equal status) or by their boss based on their “technical expertise and managerial competence” (high status). Finally, for the dependent variable, they were asked how much they agreed with the

How Does Status Affect Power Use?

117

statement: “As manager of this team, I am willing to punish teammates if they are lazy or uncooperative” (1 = disagree completely and 7 = agree completely). The results showed that respondents primed to identify with Asian culture reported greater willingness to punish in the high-status than equalstatus condition. In contrast, those primed with American culture did not differ in their willingness to punish between the equal-status and highstatus conditions. However, the status × q cultural identity interaction was significant and positive in the pooled sample, suggesting that having high status had different effects between American and Asian cultures. Altogether, three studies found that that having high status tends to decrease punishment by solo punishers in the United States but to increase punishment in collectivist cultures. Although the idea that status matters more in collectivist cultures is not new (Triandis, 1995), our research demonstrates that status can have qualitatively different effects on punishment among Asians versus Americans.

CONCLUSION Research on the psychology of power suggests a pernicious dilemma: leaders need power to achieve personal and organizational goals, yet having power also corrupts (Kipnis, 1972). Even inducing temporary, subjective feelings of power can make people less empathic (Galinsky et al., 2006). According to more recent studies (Blader & Chen, 2012; Fast et al., 2012), one simple way to tame the powerful may be to give them status and make them feel respected in the eyes of others. Yet, this solution may depend critically on a number of contextual factors that affect how status is perceived, namely the legitimacy of a hierarchy, whether status is achieved or ascribed, and cultures of collectivism versus individualism. Using public goods games with punishment, my studies have found that status reduces punishment by designated solo punishers regardless of whether the hierarchy has legitimacy or not, but somewhat surprisingly, the effect seems to be stronger in nonlegitimate hierarchies. By contrast, the effect of status on power use seems to flip when it is achieved versus ascribed. Having achieved high status through a competitive process seems make people feel more entitled to claim their share in subsequent episodes of social exchange. Similarly, the effect of status also reverses in collectivist cultures in which having high status mandates the use of power to maintain existing hierarchical relations and reinforce the social order.

118

KO KUWABARA

These ideas are still largely speculative. Much more work is needed to examine their causal logic, scope conditions, and effect sizes under different theoretical and methodological conditions. Nevertheless, the possibility that status and power can have distinct consequences, let alone opposite effects, presents an intriguing opportunity for scholars of group processes to rethink and extend our understanding of social hierarchies in a new light.

REFERENCES Abbott, A. (1981). Status and status strain in the professions. American Journal of Sociology, 86, 819
835. Anderson, C., Ames, D. R., & Gosling, S. D. (2008). Punishing hubris: The perils of overestimating one’s status in a group. Personality and Social Psychology Bulletin, 34, 90
101. Anderson, C., & Berdahl, J. L. (2002). The experience of power: Examining the effects of power on approach and inhibition tendencies. Journal of Personality and Social Psychology, 83, 1362. Baldassarri, D., & Grossman, G. (2011). Centralized sanctioning and legitimate authority promote cooperation in humans. Proceedings of the National Academy of Sciences, 108, 11023
11027. Ball, S., Eckel, C., Grossman, P. J., & Zame, W. (2001). Status in markets. Quarterly Journal of Economics, 116, 161
188. Barclay, P. (2006). Reputational benefits for altruistic punishment. Evolution and Human Behavior, 27, 325
344. Bendix, R., & Lipset, S. M. (1966). Class, status, and power: Social stratification in comparative perspective. London: Taylor & Francis. Berger, J., Cohen, B. P., & Zelditch, M. (1972). Status characteristics and social interaction. American Sociological Review, 37, 241
255. Berger, J., Ridgeway, C. L., Fisek, M. H., & Norman, R. Z. (1998). The legitimation and delegitimation of power and prestige orders. American Sociological Review, 63, 379
405. Berger, J., & Zelditch, M. (1998). Status, power, and legitimacy: Strategies and theories. New Brunswick, NJ: Transaction Press. Blader, S. L., & Chen, Y.-R. (2012). Differentiating the effects of status and power: A justice perspective. Journal of Personality and Social Psychology, 102, 994
1014. Blau, P. M. (1964). Exchange and power in social life. New Brunswick, NJ: Transaction Press. Blau, P. M., & Duncan, O. D. (1967). The American occupational structure. New York, NY: Wiley. Brashears, M. E. (2008). Sex, society, and association: A cross-national examination of status construction theory. Social Psychology Quarterly, 71, 72
85. Chen, S., Lee-Chai, A. Y., & Bargh, J. A. (2001). Relationship orientation as a moderator of the effects of social power. Journal of Personality and Social Psychology, 80, 173
187. Cheng, J. T., Tracy, J. L., Foulsham, T., Kingstone, A., & Henrich, J. (2012). Two ways to the top: Evidence that dominance and prestige are distinct yet viable avenues to social rank and influence. Journal of Personality and Social Psychology, 104, 103
125.

How Does Status Affect Power Use?

119

Cook, K. S., & Emerson, R. M. (1978). Power, equity and commitment in exchange networks. American Sociological Review, 43, 721
739. Dahrendorf, R. (1959). Class and class conflict in industrial society. London: Routledge. Devlin-Foltz, Z., & Lim, K. (2008). Responsibility to punish: Discouraging free-riders in public goods games. Atlantic Economic Journal, 36, 505
518. DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advantage as a mechanism for inequality: A review of theoretical and empirical developments. Annual Review of Sociology, 32, 271
297. Dornbusch, S. M., & Scott, W. R. (1975). Evaluation and the exercise of authority. San Francisco, CA: Jossey-Bass. Emerson, R. M. (1976). Social exchange theory. Annual Review of Sociology, 2, 335
362. Fast, N. J., & Chen, S. (2009). When the boss feels inadequate: Power, incompetence, and aggression. Psychological Science, 20, 1406
1413. Fast, N. J., Halevy, N., & Galinsky, A. D. (2012). The destructive nature of power without status. Journal of Experimental Social Psychology, 48, 391
394. Fehr, E., & Fischbacher, U. (2004). Third-party punishment and social norms. Evolution and Human Behavior, 25, 63
87. Fehr, E., & Ga¨chter, S. (2002). Altruistic punishment in humans. Nature, 415, 137
140. Foladare, I. S. (1969). A clarification of “ascribed status” and “achieved status”. The Sociological Quarterly, 10, 53
61. French, J. R. P., Jr., & Raven, B. (1959). The bases of social power. In Studies in social power (pp. 150
167). Ann Arbor, MI: University of Michigan. Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to action. Journal of Personality and Social Psychology, 85, 453
466. Galinsky, A. D., Magee, J. C., Inesi, M. E., & Gruenfeld, D. H. (2006). Power and perspectives not taken. Psychological Science, 17, 1068
1074. Gould, R. V. (2002). The origins of status hierarchy: A formal theory and empirical test. American Journal of Sociology, 107, 1143
1178. Gould, R. V. (2003). Collision of wills: How ambiguity about social rank breeds conflict. Chicago, IL: University of Chicago Press. Gruenfeld, D. H., Inesi, M. E., Magee, J. C., & Galinsky, A. D. (2008). Power and the objectification of social targets. Journal of Personality and Social Psychology, 95, 111
127. Halevy, N., Chou, E. Y., Galinsky, A. D., & Murnighan, J. K. (2012). When hierarchy wins: Evidence from the National Basketball Association. Social Psychological and Personality Science, 3(4), 398
406. Hirsh, J. B., Galinsky, A. D., & Zhong, C.-B. (2011). Drunk, powerful, and in the dark how general processes of disinhibition produce both prosocial and antisocial behavior. Perspectives on Psychological Science, 6, 415
427. Hofstede, G. H. (1980). Culture’s consequences: International differences in work-related values. Beverly Hills, CA: Sage. Ip, P. K. (2009). Is confucianism good for business ethics in China? Journal of Business Ethics, 88, 463
476. Johnson, C., Dowd, T. J., & Ridgeway, C. L. (2006). Legitimacy as a social process. Annual Review of Sociology, 32, 53
78. Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach, and inhibition. Psychological Review, 110, 265
284. Kipnis, D. (1972). Does power corrupt? Journal of Personality and Social Psychology, 24, 33
41.

120

KO KUWABARA

Kitayama, S., Markus, H. R., Matsumoto, H., & Norasakkunkit, V. (1997). Individual and collective processes in the construction of the self: Self-enhancement in the United States and self-criticism in Japan. Journal of Personality and Social Psychology, 72, 1245. Kuwabara, K. (2013). When status restrains power use: Divergent effects of power and status in public goods games with solitary punishment. Unpublished manuscript. Kuwabara, K., & Yu, S. (2014). Costly punishment makes punishment by centralized punishers more pro-social. Working Paper. Kuwabara, K., Yu, S., & Galinsky, A. D. (2014). Status decreases punishment in the west but increases punishment in the east. Working Paper. Leung, K., Koch, P. T., & Lin, L. (2002). A dualistic model of harmony and its implications for conflict management in Asia. Asia Pacific Journal of Management, 19, 201
220. Lipset, S. M., & Bendix, R. (1991). Social mobility in industrial society. New Brunswick, NJ: Transaction Publishers. Lount, R. B., Jr., & Pettit, N. C. (2012). The social context of trust: The role of status. Organizational Behavior and Human Decision Processes, 117, 15
23. Lovaglia, M. J. (1995). Power and status: Exchange, attribution, and expectation states. Small Group Research, 26, 400
426. Lun, V. M.-C. (2012). Harmonizing conflicting views about harmony in Chinese culture. In X. Huang & M. H. Bond (Eds.), Handbook of Chinese organizational behavior: Integrating theory, research and practice (pp. 467
479). Northampton, MA: Edward Elgar. Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of power and status. Academy of Management Annals, 2, 351
398. Marx, K. (1888). Communist manifesto. Waiheke Island: The Floating Press. Merton, R. K. (1968). The Matthew effect in science. Science, 159, 56
63. Milgram, S. (1963). Behavioral study of obedience. Journal of Abnormal and Social Psychology, 67, 371
378. Molm, L. (1990). Structure, action, and outcomes: The dynamics of power in social exchange. American Sociological Review, 55, 427
447. Molm, L. D. (1988). The structure and use of power: A comparison of reward and punishment power. Social Psychology Quarterly, 51, 108
122. Molm, L. D. (1989). Punishment power: A balancing process in power-dependence relations. American Journal of Sociology, 94, 1392
1418. Ostrom, E. (1990). Governing the commons: The evolution of institutions for collective action. New York, NY: Cambridge University Press. Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations: A resource dependence perspective. New York, NY: Harper & Row. Podolny, J. M. (1993). A status-based model of market competition. American Journal of Sociology, 98, 829
872. Pye, L. W., & Pye, M. W. (1985). Asian power and politics: The cultural dimensions of authority. Cambridge, MA: Belknap Press. Ridgeway, C. (1991). The social construction of status value: Gender and other nominal characteristics. Social Forces, 70, 367
386. Ridgeway, C. L. (2011). Framed by gender: How gender inequality persists in the modern world. New York, NY: Oxford University Press. Thye, S. R. (2000). A status value theory of power in exchange relations. American Sociological Review, 65, 407
432.

How Does Status Affect Power Use?

121

Torelli, C. J., Leslie, L. M., Stoner, J. L., & Puente, R. (2014). Cultural determinants of status: Implications for workplace evaluations and behaviors. Organizational Behavior and Human Decision Processes, 123, 34
48. Torelli, C. J., & Shavitt, S. (2010). Culture and concepts of power. Journal of Personality and Social Psychology, 99, 703. Triandis, H. (1995). Individualism and collectivism. Boulder, CO: Westview Press. Triandis, H. C., & Gelfand, M. J. (1998). Converging measurement of horizontal and vertical individualism and collectivism. Journal of Personality and Social Psychology, 74, 118
128. Tyler, T. R. (2005). Psychological perspectives on legitimacy and legitimation. Annual Review of Psychology, 57, 375
400. Van Kleef, G. A., De Dreu, C. K. W., Pietroni, D., & Manstead, A. S. R. (2006). Power and emotion in negotiation: Power moderates the interpersonal effects of anger and happiness on concession making. European Journal of Social Psychology, 36, 557
581. Van Kleef, G. A., Oveis, C., van der Lo¨we, I., LuoKogan, A., Goetz, J., & Keltner, D. (2008). Power, distress, and compassion: Turning a blind eye to the suffering of others. Psychological Science, 19, 1315
1322. Visser, M. S., & Roelofs, M. R. (2011). Heterogeneous preferences for Altruism: Gender and personality, social status, giving and taking. Experimental Economics, 14, 490
506. Weatherley, R. (2002). Harmony, hierarchy and duty based morality: The Confucian antipathy towards rights. Journal of Asian Pacific Communication, 12, 245
267. Weber, M. (1978). Economy and society: An outline of interpretive sociology. New York, NY: Bedminster Press. Willer, D. (1999). Network exchange theory. Westport, CT: Praeger. Willer, D., Markovsky, B., & Patton, T. (1989). Power structures: Derivations and applications of elementary theory. In J. Berger, M. Zelditch, & B. Anderson (Eds.), Sociological theories in progress (pp. 313
353). Boston, MA: Houghton-Mifflin. Wiltermuth, S., & Flynn, F. (2012). Power, moral clarity, and punishment in the workplace. Academy of Management Journal, 56, 1002
1023. Zimbardo, P. G., Maslach, C., & Haney, C. (1999). Reflections on the Stanford prison experiment: Genesis, transformations, consequences. In T. Blass (Ed.), Obedience to authority: Current perspectives on the Milgram paradigm (pp. 193
237). New York, NY: Psychology Press.

This page intentionally left blank

RECRUITING SOURCE EFFECTS: A SOCIAL PSYCHOLOGICAL ANALYSIS Richard L. Moreland ABSTRACT Purpose
I present and evaluate various explanations for why new workers who were sponsored by oldtimers tend to have better job outcomes (better performance, more satisfaction, and less turnover) than do new workers who were not sponsored. Methodology/approach
My evaluations involve searching for evidence that fits (or does not fit) each of the explanations. Findings
The two most popular explanations argue that the job benefits of sponsorship arise because (a) sponsored newcomers have more realistic job expectations than do unsponsored newcomers, or (b) the quality of sponsored newcomers is greater than that of unsponsored newcomers. Unfortunately, these explanations have weak empirical support. A third explanation, largely untested as yet, attributes the performance benefits of sponsorship to social pressures that can arise when someone is sponsored for a job. These pressures include efforts by newcomers to repay the people who sponsored them, efforts by sponsors to assist the newcomers they sponsored after those persons have been hired, and

Advances in Group Processes, Volume 32, 123
148 Copyright r 2015 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0882-6145/doi:10.1108/S0882-614520150000032005

123

124

RICHARD L. MORELAND

stereotypes among coworkers about the kinds of people who get jobs through sponsors. Although limited as yet, the evidence regarding this new explanation seems promising. Research implications
More research on this third explanation for sponsorship effects should be done. Suggestions for how to do such research are reviewed and a relevant experiment is presented. Social implications
The ideas and evidence presented here could help employers who want to improve the job outcomes of their new workers. Poor outcomes among such persons are a major problem in many settings. Originality/value
Although some of my ideas have been mentioned by others, they were not been described in much detail, nor were they tested. My hope is that this chapter will promote new theory and research on the performance benefits of sponsorship, a topic that has been largely ignored in recent years. Keywords: Recruiting sources; sponsorship; job outcomes

How do qualified workers get suitable jobs? The process through which workers and jobs are matched is complex and imperfect. Some workers are left unemployed, although suitable jobs are available, and some jobs go unfilled, despite the availability of qualified workers. And even when workers are hired, their jobs may not be suitable or the workers themselves may be unqualified. Perhaps these problems could be solved, or at least ameliorated, if more were known about the matching process. I will focus on one aspect of that process, namely the help that people often get from “sponsors” (relatives, friends, or acquaintances) during a job search. This help can take two forms. First, sponsors can help workers to discover suitable jobs and/or help companies to discover qualified workers. Second, sponsors can sometimes influence companies to make job offers to workers and/or influence workers to accept those offers. Relatives, friends, and acquaintances can thus help workers to both locate and obtain jobs. I will begin this chapter by evaluating how often jobs are arranged through sponsors and analyzing the advantages and disadvantages of this

Recruiting Source Effects: A Social Psychological Analysis

125

practice for both workers and companies. I will then review research on the job outcomes (satisfaction, performance, and turnover) of workers who get jobs in this way. That research suggests that such workers often experience better outcomes. Two broad explanations for that finding have been proposed.1 One explanation focuses on workers’ expectations about their jobs and the other focuses on the kinds of workers who are likely to be helped. I will evaluate the evidence for both explanations and then propose a third explanation, one that focuses on the social pressures that can arise when people help one another to get jobs.

USING SPONSORS Workers can locate and obtain jobs in several ways (see Schwab, Rynes, & Aldag, 1987). Many reviewers have classified these methods as “formal” versus “informal” although other classification schemes are certainly possible. Formal methods include using placement services and employment agencies, responding to advertisements in newspapers or trade journals, and scanning relevant websites. Informal methods include applying for jobs directly and relying on help from sponsor (relatives, friends, acquaintances). Several researchers have studied how often these methods are actually used by workers. Reviews of their research can be found in Schwab (1982) and Wiley (2011). Much of this research involves surveys of workers, who are simply asked how they got their jobs, but company hiring records are often analyzed as well (many companies routinely ask job applicants how they heard about the jobs they are seeking). What proportion of workers say that they got their jobs through sponsors? Most studies (see Rees, 1966) suggest that half or more of all workers get their jobs in this way, and that sponsorship is used more often than any other method.2 Why are sponsors used so often? Their use may be advantageous for both workers and companies. One advantage is that sponsors are relatively cheap and easy to use. When a worker is searching for a job, it is generally simpler and less expensive to ask for help from relatives, friends, and acquaintances than it is to pay for the services of an employment agency. In fact, it may not even be necessary to ask sponsors for help, which is often offered spontaneously, and once sponsors have begun to help somebody get a job, little further effort from that person may be required. Similar considerations can influence a company that is searching for qualified workers to fill its jobs. Many companies encourage current workers to

126

RICHARD L. MORELAND

identify people they know who might make good employees, sometimes offering special incentives if those persons are later hired (Martin, 1987). These inducements may cost some money, of course, but they are generally less expensive than more formal recruiting methods. And once a company’s workers are mobilized to find new employees, further assistance from those workers may require relatively little coordination or guidance by the company. Finally, the use of sponsors by a company can strengthen its social capital (Fernandez & Castilla, 2001). Another advantage of sponsors is their effectiveness. Computer simulations of job markets (e.g., Boorman, 1975; Delany, 1988; see also Saloner, 1985) have shown that sponsors can be quite good at matching qualified workers with suitable jobs. Several researchers (e.g., Azavedo, 1974; Holzer, 1988; Kirnan, Farley, & Geisinger, 1989; Reid, 1972) have even found that sponsors are more effective than other methods are in this regard. Holzer, for example, surveyed young American workers and found that those who sought help from friends and relatives, rather than using more formal search methods, received more job offers and accepted more of the offers that they received. And Reid, who surveyed older British workers, found that seeking help from friends and relatives not only led to more jobs, but also produced jobs faster than did other, more formal search methods. Several factors may help to explain such results. Sponsors probably find more and better jobs for workers because they are more familiar with those persons and more concerned about maintaining good relationships with them. And workers probably evaluate such jobs more favorably, not only because they are appealing, but also because the people who found them are liked and trusted (cf. Fisher, Ilgen, & Hoyer, 1979; Rynes, 1991). Workers may also fear offending sponsors by rejecting (too quickly or too readily or at all) whatever jobs those persons find (Rosen, Mickler, & Collins, 1987). Similar factors may operate in companies. Sponsors usually find jobs for other people in companies where they work themselves (Corcoran, Datcher, & Duncan, 1980; Jones & Azrin, 1973; Lin, Ensel, & Vaughn, 1981; Marsden & Hurlbert, 1988). Because they are familiar with those companies and concerned about protecting their careers, sponsors probably find qualified and appealing workers for those jobs. And companies may evaluate such workers more favorably, not only because the workers are more appealing, but also because their sponsors are liked and trusted. Companies may also be susceptible to influence by sponsors, who try to promote their relatives, friends, and acquaintances in various ways (see Jones & Azrin, 1973; Knouse, 1989; Marsden & Hurlbert, 1988; Pfeffer, 1989).

Recruiting Source Effects: A Social Psychological Analysis

127

A final advantage of sponsors is that they can produce better outcomes, both for workers seeking jobs and for companies seeking workers. Most workers, for example, have at least some relatives, friends, or acquaintances who are more successful than themselves. By asking such people for help, these workers can locate and obtain better jobs than would be possible otherwise (Lin & Dumin, 1986; Lin et al., 1981; Marsden & Hurlbert, 1988). In fact, several studies have shown that as the status of a sponsor rises, so does the quality of whatever jobs that person arranges for workers (DeGraaf & Flapp, 1988; Lin & Dumin, 1986; Lin et al., 1981; Lin, Vaughn, & Ensel, 1981; Marsden & Hurlbert, 1988; Wegener, 1991). This effect not only transcends workers’ actual job qualifications, but also helps to explain why more qualified workers get better jobs (see DeGraaf & Flapp, 1988; Lin et al., 1981). Better educated workers tend to know more successful people, and when that advantage is taken into account, the ability of such workers to help someone get better jobs is weakened. Do companies get better workers when they recruit them through sponsors? Surveys of personnel managers (see Schwab, 1982) suggest that this practice is popular primarily because of the widespread belief that it yields workers who are more skilled and motivated. The available evidence, which I will discuss shortly, supports that belief. Although the use of sponsors clearly has some advantages for both workers and companies, a few potential disadvantages should also be noted. These include problems experienced by the workers themselves. For example, the jobs that sponsors find for them are sometimes unsuitable, yet workers may feel obliged to take those jobs anyway, because they don’t want to offend their sponsors. Another problem involves the feelings of indebtedness (see Greenberg, 1980), that may arise in workers who get their jobs through sponsors. These feelings are aversive, especially when sponsors ask for reciprocal favors that workers are unwilling or unable to provide. Finally, workers who need help from sponsors to get jobs may begin to see themselves in more negative ways (Fisher, Nadler, & WhitcherAlagna, 1982; Nadler & Fisher, 1986). For example, the self-esteem of such workers may decline, along with their sense of personal competence or control. These changes could induce feelings of dependence on others, which might impair the ability of those workers to get jobs on their own later on. Other disadvantages involve problems experienced by companies who use personal contacts to recruit workers. For example, people may resent coworkers who are hired through sponsors, because that practice seems like a form of nepotism. And if many people are hired in this way, then everyone may come to believe that success in the company depends on

128

RICHARD L. MORELAND

relationships, rather than knowledge or skills. Such beliefs can weaken workers’ motivation and commitment, thereby reducing their productivity (see Beehr & Taber, 1993; Beehr, Taber, & Walsh, 1980). Another problem involves the challenges associated with balancing personal and work relationships (Bridge & Baxter, 1992). When many of the workers in a company are linked by family or friendship ties, issues of discipline and discretion may arise and be difficult to resolve. Similarly, the tendency for sponsors to recruit workers who are similar to themselves (see Jackson et al., 1991; Moreland & Levine, 1982; Schneider, 1987) can be a problem. For example, if the quality of a company’s current workers is poor, then hiring more people of the same sort is clearly unwise (see Franzen, 1994). But even when the quality of workers (new and old) is good, recruiting through sponsors can limit the diversity of a company’s workforce (Elliott, 2001; Jackson, 1992; Moreland & Levine, 1992; Pfeffer, 1989). This can have several consequences, some helpful (e.g., increased cohesion and improved communication among workers) but others harmful. As workers become less diverse, for example, a company may have more trouble developing innovative products or procedures (Bantel & Jackson, 1989; Nemeth & Staw, 1989). And when certain kinds of people are excluded from a company simply because they lack sponsors, ethical and legal issues may arise (see Braddock & McPartland, 1987; Mier & Giloth, 1985; Peters, 1993). Some companies try to cope with these issues by adopting special policies to regulate hiring practices, but such policies are not always effective, and their implementation can create problems of other sorts (Ford & McLaughlin, 1986).

SPONSORS AND JOB OUTCOMES As noted earlier, companies often use sponsors for recruiting because they believe that this practice produces better workers. Many researchers have actually compared the job outcomes of workers who were hired in different ways. Reviews of this work, some narrative and others meta-analytic, can be found in Rynes (1991), Wanous (1992), Wanous and Colella (1989), and Zottoli and Wanous (2000). Turnover is the job outcome that has been studied most often, and although the evidence is somewhat mixed, it seems to favor the use of sponsors. Workers hired through sponsors are generally less likely to leave their jobs (voluntarily or involuntarily) than are workers hired through other

Recruiting Source Effects: A Social Psychological Analysis

129

methods (Breaugh & Mann, 1984; Caldwell & Spivey, 1983; Decker & Cornelius, 1979; Gannon, 1971; Kirnan et al., 1989; Reid, 1972; Ullman, 1966). There are a few studies in which the use of sponsors had no special effects on turnover (e.g., Swaroff, Barclay, & Bass, 1985; Taylor & Schmidt, 1983; Williams, Labig, & Stone, 1993), but methodological factors may account for these anomalous results. Another outcome that many researchers have studied is job performance. The evidence from these studies is less clear. Workers hired through sponsors often perform no better than do those hired through other methods (Hill, 1970; Kirnan et al., 1989; Swaroff et al., 1985; Williams et al., 1993) and sometimes their performance is worse. For example, workers who are rehired by companies (Taylor & Schmidt, 1983), or who apply for jobs on their own (Blau, 1990; Breaugh & Mann, 1984; Caldwell & Spivey, 1983), seem to outperform workers hired through personal contacts. A final outcome, studied by just a few researchers, is job satisfaction. The evidence from these studies is meager, but does favor the use of sponsors. Sponsored workers tend to enjoy their work more than do workers hired through other methods (Latham & Leddy, 1987; Reid, 1972; but see also Taylor & Schmidt, 1983). The available evidence thus shows that some job outcomes are indeed better when workers are sponsored. Several limitations of that evidence, however, deserve to be noted. First, the total number of studies is modest and has not grown much in recent years. Stronger conclusions about the effects of sponsors on workers’ job outcomes could be drawn if more evidence were available. Second, the methods that workers use to locate and obtain jobs are often measured rather crudely. Most researchers rely on self-reports by workers about those methods, ignoring the risk of possible response biases associated with poor memory, beliefs about which methods are expected to be effective, and self-presentation concerns. And despite the fact that many workers use more than one method to get a job (see Azavedo, 1974; Bortnick & Ports, 1992; Ellis & Taylor, 1983; Reid, 1972; Williams et al., 1993), researchers usually ask each person to attribute his or her employment to a single hiring method. This may obscure the true effects of sponsors on job outcomes, because those outcomes could differ as a function of how many and what other hiring methods are used (Bortnick & Ports, 1992; Williams et al., 1993), or the order in which different methods are used (Reid, 1972). Third, job outcomes are often measured poorly. Researchers who study the effects of sponsorship on job performance, for example, often rely on evaluations by supervisors, whose beliefs about workers could be biased by knowledge about how they were hired (see Dipboye, 1986; Bretz, Milkovich, & Read, 1992). Single measures of

130

RICHARD L. MORELAND

job performance are common (Breaugh & Mann, 1984; Caldwell & Spivey, 1983; Hill, 1970), despite their uncertain reliability, and even when multiple performance measures are gathered, they are sometimes weakly correlated with one another (Blau, 1990; Swaroff et al., 1985). When job performance is measured a long time after workers were hired, it may be confounded with turnover (see Blau, 1990; Caldwell & Spivey, 1983; Kirnan et al., 1989; Swaroff et al., 1985; Williams et al., 1993). Perhaps this explains why workers hired through sponsors do not always outperform workers hired through other methods. The latter workers are more likely to leave their jobs, and if (as seems likely) the worst of those workers leave first, then only the best workers may still remain by the time job performance is assessed. Finally, few researchers consider contextual factors, such as job, company, and industry characteristic (see Rynes & Barber, 1990). These factors could moderate the effects of sponsors on job outcomes. For example, turnover might be lower among workers hired through personal contacts because the jobs those workers perform are more appealing. All of these limitations could be overcome by conducting more and better research, but that research should be guided by theory. This raises a critical issue, namely why job outcomes are affected by how workers are hired.

REALISTIC EXPECTATIONS OR INDIVIDUAL DIFFERENCES? There are several reasons why workers hired in different ways might experience different job outcomes, but most research has examined just two broad explanations. One explanation involves the kinds of expectations that workers develop about their jobs, whereas the other explanation involves the kinds of people who use different job search methods. Neither explanation focuses specifically on the use of sponsors, which is regarded as just one of the several informal methods for locating and obtaining jobs. Yet both explanations are relevant to workers who get their jobs in this way, and both may clarify why such workers often experience better job outcomes.

Expectations about Jobs The importance of workers’ job expectations has been noted by many theorists (e.g., Ashforth, 1989; Louis, 1980; Nelson & Sutton, 1991), especially

Recruiting Source Effects: A Social Psychological Analysis

131

Wanous and his colleagues (see Premack & Wanous, 1985; Wanous, 1992; Wanous & Colella, 1989; Wanous, Poland, Premack, & Davis, 1992). While workers are searching for jobs, they often develop unrealistic, overly optimistic expectations about what those jobs are like (Wanous, 1992; see also Brinthaupt, Moreland, & Levine, 1991). These expectations can cause trouble later on, after the workers are hired. If more accurate information were available earlier, then workers might develop more realistic expectations, which could help to minimize that trouble. This hope underlies efforts by some companies to provide workers with realistic job previews, which do indeed appear to be beneficial (Breaugh & Billings, 1988; Wanous, 1992). Maybe relatives, friends, or acquaintances serve a similar function when they provide workers with information about jobs. Insofar as that information is accurate and complete, it could improve workers’ job outcomes in several ways. For example, information from sponsors could help workers to choose more suitable jobs. And when problems arise on their jobs, workers who were warned about those problems by their sponsors might be less surprised or disappointed, and better prepared to cope effectively. [Note that sponsors could also provide valuable information to companies about the new workers that they are planning to hire]. This explanation for the effects of sponsors on workers’ job outcomes suggests the need or three kinds for research evidence. First, do sponsors indeed provide workers with information that helps them to develop more realistic expectations about jobs? Second, do workers with more realistic expectations actually experience better job outcomes? Finally, when the realism of workers’ expectations about their jobs is taken into account, are the effects of sponsorship on job outcomes weakened? Put another way, do realistic job expectations mediate the effects of sponsors on job outcomes? What does the literature show? The general level of support for the realistic expectations explanation of recruiting source effects is weak. Many researchers, while attempting to interpret their findings, merely speculate about workers’ job expectations without ever actually measuring them (e.g., Breaugh, 1981; Caldwell & Spivey, 1983; Decker & Cornelius, 1979; Latham & Leddy, 1987; Reid, 1972; Taylor & Schmidt, 1983). And even when job expectations are measured, the findings are often clouded by conceptual, methodological, or analytical problems. So, the available evidence does not provide very strong support for the claim that sponsored workers experience better job outcomes because they have more realistic expectations about their jobs. And although several studies (e.g., Blau, 1990; Breaugh & Mann, 1984; Quaglieri, 1982) have indeed shown that workers hired through personal contacts have more realistic job

132

RICHARD L. MORELAND

expectations than workers hired through other methods, only one study (Williams et al., 1993) suggests that workers with more realistic job expectations experience better job outcomes, and in that study, job expectations were related to turnover, but not to performance (see also Blau, 1990; Hill, 1970). There is no evidence that realistic job expectations mediate the effects of hiring methods on job outcomes. Both Blau (1990) and Williams et al. (1993) found that those effects did not change when the job expectations of workers were taken into account. This lack of support suggests that other explanations should be explored for the relatively positive job outcomes experienced by sponsored workers. But before realistic expectations are discounted in this regard, two important points are worth noting. First, only researchers have yet studied how the job expectations of workers hired through various methods can affect their job outcomes. Second, conceptual, methodological, and analytical problems cloud the results from many of those studies. At the conceptual level, for example, several important issues have been ignored. For example, what kinds of job expectations have the greatest impact on a worker’s job outcomes (Wanous & Colella, 1989)? Are some sponsors more willing or able than others to provide accurate information about jobs (Levine & Moreland, 1991)? And how are job outcomes affected by the information about workers that sponsors provide to companies about them (Knouse, 1989; see also Moreland & Levine, 1989)? Other problems, at the methodological level, are also apparent. For example, job expectations have often been measured after workers are hired, rather than before. Can workers really recall, weeks or months later, what they once expected their jobs to be like? Memories about such matters can be faulty and could be biased by job outcomes. If job expectations are measured before workers are hired, then their realism can be assessed in better ways, through comparisons with a variety of standards (e.g., objective information about jobs, perceptions of jobs among workers already employed by the company, and workers’ own job perceptions after they are hired). And regardless of when or how workers’ job expectations are measured, more effort should be made to study the process by which those expectations influence workers’ job outcomes. For example, several factors that could moderate the effects of realistic job expectations on workers’ outcomes deserve to be studied. These factors include individual (e.g., intelligence, locus of control, and tolerance for ambiguity), organizational (e.g., training or socialization practices), and job (e.g., complexity, stability, and familiarity) characteristics. Evidence involving the actual use of job information by workers, before and/or after they are hired, would also be valuable. For example, do

Recruiting Source Effects: A Social Psychological Analysis

133

workers with better information choose more suitable jobs (cf. Allen & Keaveny, 1980)? Once they are hired, are those workers less surprised or disappointed by any problems that arise? And are they better than other workers at coping with such problems? Finally, analytical problems weaken much of the research on realistic job expectations, especially when it comes to mediation. For example, some researchers (e.g., Breaugh & Mann, 1984) collect the data needed to explore mediation, but fail to perform relevant analyses. Other researchers (e.g., Blau, 1990; Williams et al., 1993) analyze their data in clumsy or misleading ways. There are several data analytical techniques (see Baron & Kenny, 1986; Coovert, Penner, & MacCallum, 1990) available for testing mediation, but they have not been used to explore how realistic job expectations shape the job outcomes of workers hired through personal contacts.

Individual Differences among Workers As noted earlier, there is another way in which the effects of sponsorship on workers’ job outcomes might be explained. This second explanation involves possible individual differences among workers hired through various methods (see Schwab, 1982). Maybe there is something special about workers who use personal contacts to get jobs. These workers may differ from others on a variety of characteristics, including age, sex, or race; education or experience; attractiveness; abilities (general or specific); motivation; attitudes or beliefs; and dispositions. Some of these characteristics might lead to better job outcomes on their own, producing an apparent effect of personal contacts on those outcomes. Yet that effect would be artifactual in nature, reflecting a tendency for better workers to be hired more often through sponsors. In other words, sponsored workers may have better job outcomes because of their personal characteristics, rather than their job search methods. These workers thus succeed because they have the “right stuff”
the fact that relatives, friends, or acquaintances helped them get their jobs is irrelevant. Three kinds of research evidence could provide support for this second explanation. First, sponsored workers must possess distinctive characteristics. Second, workers with those characteristics must experience better job outcomes. Finally, when those characteristics are taken into account, the effects of sponsorship on job outcomes must be weakened. That is, individual differences among workers should mediate the effects of sponsorship on job outcomes.

134

RICHARD L. MORELAND

The general level of support for this second explanation is also rather weak. Some researchers speculate about the kinds of workers who are hired through sponsorship, but do not measure their actual characteristics (e.g., Gannon, 1971; Hill, 1970; Ullman, 1966). Other researchers compare the characteristics of workers hired through various methods, but fail to measure their job outcomes (see Corcoran et al., 1980; Holzer, 1988; Lipset, Reinhard, & Malm, 1955; Marsden & Campbell, 1990; Rees & Gray, 1982). For example, sponsors are used more often by workers who are young, white, and male, but after such workers are hired, how do those characteristics affect their turnover, job performance, or job satisfaction? Few researchers have measured both the personal characteristics and the job outcomes of sponsored workers. Sadly, many of their findings are also clouded by conceptual, methodological, and analytical problems that failed to resolve the issue of mediation (see Breaugh & Mann, 1984; Caldwell & Spivey, 1983; Kirnan et al., 1989; Swaroff et al., 1985), or suggested that such mediation did not occur (see Blau, 1990; Breaugh, Greising, Taggart, & Chen, 2003; Williams et al., 1993). So, the available research evidence does not provide strong support for the claim that sponsored workers experience better job outcomes because of their distinctive characteristics. Despite several attempts to identify such characteristics, workers hired though sponsorship seldom differ from those hired through other methods. And when distinctive characteristics are found, they are just as likely to be negative (Blau, 1990; Taylor & Schmidt, 1983) as positive (Kirnan et al., 1989; Williams et al., 1993). Many of the characteristics that distinguish sponsored workers are not especially positive or negative (see Blau, 1990; Breaugh & Mann, 1984; Kirnan et al., 1989; Swaroff et al., 1985). In fact, when their effects on job outcomes are actually tested, those characteristics seldom turn out to be important (see Kirnan et al., 1989; Swaroff et al., 1985; Williams et al., 1993). Moreover, job outcomes are often related to characteristics that do not distinguish workers hired though sponsorship (Blau, 1990; Caldwell & Spivey, 1983; Ellis & Taylor, 1983; Kirnan et al., 1989; Swaroff et al., 1985; Williams et al., 1993). This lack of support suggests again that other explanations must be found for the relatively positive job outcomes that are experienced by workers hired through sponsors. But the same points that I made earlier, regarding studies of job expectations, are also relevant to studies of individual differences. Few researchers have studied how the characteristics of workers hired through various methods actually influence their job outcomes, and other problems can cloud the results of those studies. For

Recruiting Source Effects: A Social Psychological Analysis

135

example, several conceptual issues are often ignored. Workers obviously vary in many ways, but which characteristics are likely to have the greatest impact on their job outcomes (Rynes, 1991; Wanous & Colella, 1989)? And why should workers hired though sponsors have any distinctive characteristics at all? Maybe some workers, because they are more intelligent, motivated, or socially skilled, are more likely than others to ask for (or receive) help from others while searching for jobs. These workers may also have larger and more diverse social networks, networks whose members are better able to provide help (Campbell, Marsden, & Hurlbert, 1986). Maybe relatives, friends, and acquaintances only offer help to certain kinds of workers, such as those who seem physically attractive (Benson, Karabenick, & Lerner, 1976), competent (Saloner, 1985; Ullman, 1966), similar to themselves (Dovidio, 1984; Jackson et al., 1991; Rushton, 1989), or likely to “fit” into a company (Ullman, 1966). Sponsors may even encourage workers to develop characteristics (e.g., education, experience, and motivation) that seem related to job outcomes. Other problems are apparent at the methodological level as well. For example, many researchers measure workers’ characteristics only after workers are hired, when some selection has probably already occurred (Rynes, 1991). If a characteristic is indeed related (positively or negatively) to job outcomes, then it is likely to shape a company’s hiring decisions. As a result, workers who are hired may show little variability on that characteristic, weakening its apparent relationships with both search methods and job outcomes (cf. Dunbar & Linn, 1991). Under these circumstances, researchers may mistakenly conclude that the characteristic is unimportant. Workers’ characteristics should thus be measured before the workers are hired. Better measures of workers’ characteristics would also be helpful, and a wider variety of characteristics should be measured. Too many researchers focus on characteristics (e.g., age, race, and sex) that are measured fairly easily, while ignoring more subtle characteristic, such as opinions and dispositions, whose measurement is more challenging. Finally, many studies of individual differences are also plagued by analytical problems, especially when it comes to the issue of mediation. Once again, some researchers (e.g., Breaugh & Mann, 1984; Kirnan et al., 1989) collect the data needed to explore that issue, but fail to perform relevant analyses. Other researchers (e.g., Blau, 1990; Taylor & Schmidt, 1983) analyze their data in clumsy or misleading ways. For example, none of the special analytical techniques for testing mediation has yet been used to explore how individual differences shape the job outcomes of workers hired through sponsors.

136

RICHARD L. MORELAND

Other Explanations Why should sponsored workers experience better job outcomes than those who are hired in other ways? Two broad explanations have been proposed, but neither one is strongly supported by the available research evidence. Moreover, both explanations seem narrow in scope. For example, their temporal perspectives are limited to the time before employment begins
little has been said about what happens to a sponsored worker after he or she is hired, and how events during that period might affect his or her later job outcomes. Maybe sponsored workers have distinctive experiences, after they are hired, that help them to succeed. The social perspectives of the two explanations seem limited as well, focusing almost exclusively on the thoughts and feelings of new workers. Little has been said about how other people might affect the job outcomes of such workers. Maybe sponsored workers are treated in special ways by others (e.g., their sponsors, coworkers, or other stakeholders) and are more or less successful as a result. A few analysts have argued that broader temporal and/or social perspectives on sponsorship and job outcomes are indeed needed (see Breaugh, 1981; Castilla, 2005; Elliott, 2001; Skolnik, 1987), but little progress has been made so far toward developing those perspectives.

A SOCIAL PSYCHOLOGICAL ANALYSIS My analysis focuses on the period after someone is hired for a job, examining the effects of sponsorship on three key actors during that period. Those actors are the new worker, that person’s sponsor(s), and anyone who works with the newcomer. Each of these actors experiences social pressures related to sponsorship, pressures that could have an impact (usually positive, but sometimes negative) on the job outcomes of the new worker. Without considering those pressures, it seems unlikely that the impact of sponsorship on workers’ job outcomes can be fully understood.

Newcomer Gratitude for the Sponsor’s Help When one person helps another to get a job, the new worker is likely to feel some gratitude and may try to repay the favor (Tesser, Gatewood, & Driver, 1968; see also Gouldner, 1960). One obvious strategy in that regard

Recruiting Source Effects: A Social Psychological Analysis

137

is to perform the job well, focus on its positive features, and keep the job for awhile. This point becomes clearer if we imagine a sponsored worker who acts in the opposite ways
performing a job poorly, complaining about it to others, and leaving the job soon after taking it. That person’s sponsor would probably view these acts as signs of ingratitude and feel angry as a result. Such anger could damage the relationship between the sponsor and the new worker, perhaps leading to other problems later on. Someone’s gratitude to a sponsor is likely to be greater when (a) sponsorship arose from a genuine concern for the worker, rather than serving a more selfish motive; (b) it was costly for the sponsor to provide help, and (c) the job was attractive. Thus, a sponsored worker might not feel much gratitude if the organization paid the sponsor a “bounty” for helping to hire him or her, it was relatively easy for the sponsor to arrange the new job, or that job was difficult to perform, yet paid poorly (Tesser et al., 1968). There are even conditions when the new worker might feel little or no gratitude to his or her sponsor. For example, if the sponsor and the new worker had a communal rather than exchange relationship (Clark & Mills, 2011), then sponsorship might seems like a normal aspect of that relationship, rather than something special. Imagine, for example, a person who got a new job through his or her parents. That person might expect such help and even feel resentful if it were not offered. Finally, there are situations where sponsorship could actually harm the relationship between a sponsor and the person he or she helps, because the new worker feels indebted to the sponsor (see Greenberg, 1980). Feelings of indebtedness are likely to arise if the favor seems too large (e.g., the sponsor has arranged an unusually attractive job); the sponsor seems to have an ulterior motive (e.g., a parent finds a job for a grown child who still lives at home, so that the child can use his or her salary to pay rent); or the sponsor explicitly asks to be repaid in some way for the help that he or she provided. Although there has been some research on both gratitude and indebtedness, few researchers have yet studied those phenomena in the context of sponsorship and job seeking. One piece of potentially relevant evidence was reported by Castilla (2005), who studied the role of social networks in the performance of people working in a call center. Castilla found that the departure of someone’s sponsor from the organization had a negative effect on the performance of that worker, even dropping it below the performance of unsponsored workers. Maybe repayment of a sponsor is no longer an issue after the sponsor leaves. In my opinion, more research of this sort would be very helpful.

138

RICHARD L. MORELAND

Sponsor Concerns about Newcomer Performance A sponsor’s efforts to help someone are unlikely to end when that person is hired. In most cases, the sponsor wants the person to succeed at the new job and will thus provide various forms of social support to ensure that the newcomer does well. In fact, the newcomer might even ask his or her sponsor for such support. Social support from a sponsor can be indirect at times. For example, if the sponsor finds the newcomer a job at his or her own workplace, then the sponsor might simply serve as a role model for the newcomer, showing him or her, for example, how to dress and act properly. More direct forms of social support are also possible. For example, the sponsor could (a) spread positive information (accurate or not) about the new worker, so that he or she is treated better by coworkers and managers; (b) intervene to improve the newcomer’s working conditions (e.g., arrange for better task assignments); or (c) encourage the newcomer, answer his or her questions, and offer advice. Some evidence of social support for workers by their sponsors was reported by Schoorman (1988), who studied job performance evaluations within a large organization. Schoorman found that when managers were uninvolved with someone’s hiring, they later gave that person lower performance evaluations than they gave to someone whose hiring they had favored, and higher performance evaluations than they gave to someone whose hiring they had opposed. In other words, performance evaluations were apparently used by sponsors to help the people they brought into the organization to succeed. Note that a sponsor might have ulterior motives for helping a new worker to succeed. For example, a sponsor might hope that the people he or she helps will later support the sponsor’s ideas and initiatives within the company. And if a newcomer performs well, then that reflects well on his or her sponsor (see Cialdini, 2013; Elliott, 2001), to the extent that the sponsorship is known to have occurred. This could benefit the sponsor in several ways, such as improving his or her reputation in the organization. In contrast, a newcomer who performs poorly will have the opposite kinds of consequences for his or her sponsor. Thus, when someone is known to have sponsored a new worker, he or she would be wise to help that person succeed. Biases toward Sponsored Newcomers among Coworkers Some newcomers (and sponsors) do not care if others know that sponsorship has occurred, and so they talk about it openly, or at least do not try to hide it. In fact, they might even publicize the sponsorship, hoping that favorable reactions toward the new worker will bias reactions toward the sponsor, or vice versa (Schlenker, 1990). Even if newcomers and their

Recruiting Source Effects: A Social Psychological Analysis

139

sponsors prefer to keep sponsorship hidden, that may not be possible. And, of course, coworkers may believe that sponsorship has occurred, whether it actually did or not. Beliefs among coworkers about sponsorship can have important consequences for how a new worker is treated, and this can influence the newcomer’s job outcomes. On the positive side, coworkers might treat sponsored newcomers better than they would otherwise, in order to please the sponsors of those persons. This is especially likely if those sponsors are liked, feared, or respected. On the negative side, sponsorship might be a sigma that causes coworkers to treat newcomers worse than they would otherwise. In either case, it would be interesting to learn how long these biases last, whether and how they can be changed, and the extent to which a newcomer actually internalizes the negative qualities attributed to him or her by others. Finally, coworkers’ reactions toward sponsored workers may depend in part on how common sponsorship is within the organization. In some organizations, rules or norms may restrict the use of sponsorship, whereas in other organizations, sponsorship may be commonplace. In the latter organizations, sponsorship might have little impact on coworkers’ reactions toward a newcomer, perhaps because the coworkers themselves were sponsored. No one has studied the reactions of coworkers toward sponsored versus unsponsored workers, but it is not difficult to imagine how such research could be done. Some of the paradigms used to study prejudice in the workplace could easily be adapted for this purpose. Consider, for example, Heilman’s research (see Heilman & Haynes, 2006) on how knowing (or even suspecting) that a woman was hired through an affirmative action program can damage her reputation among coworkers. Sponsorship, like affirmative action, could be viewed by coworkers as an unfair advantage.

RESEARCH REFLECTING A SOCIAL PSYCHOLOGICAL ANALYSIS OF SPONSORSHIP EFFECTS To support my analysis, I have mentioned relevant research whenever possible, but the sad fact is that not much research of this sort can be found. More such work is needed. With that goal in mind, I have been engaged for several years now in a broad program of research meant to explore several aspects of sponsorship. A variety of studies have been performed. Two of the largest studies were surveys. In one survey, I identified workers who said they got their jobs through sponsorship. I asked these workers where

140

RICHARD L. MORELAND

they worked, what kind of work they did, and how much they liked their jobs. Questions were also asked about the sponsors and the workers’ relationships with those sponsors. Several questions arising from my analysis were also asked (e.g., did the workers feel any obligation to “repay” the sponsors for their help, and if so, then how was that repayment made?). In the other survey, I identified workers who said they had once sponsored someone for a job. I asked these workers questions about the people they sponsored and their relationships with those persons. And once again, several questions specifically suggested by my analysis were asked (e.g., did the respondents worry about the job performance of the people they sponsored, and if so, then what did they do to improve that performance?). I have also done several scenario studies that explored the reactions of “managers” and “coworkers” to a “job applicant” who was either sponsored or unsponsored. These studies investigated such issues as whether sponsorship of job applicants has much impact on managers’ hiring decisions, and if so, then whether that impact is generally favorable or unfavorable. I have also studied variables that might moderate the benefits of sponsorship for job applicants. For example, are the benefits of sponsorship different for job applicants whose qualifications are weaker versus stronger? And are they different for sponsors located in or out of the organization? I have also explored possible stereotypes that coworkers may have about someone who was sponsored or unsponsored, using methods developed by McCauley and Stitt (1978). Do such stereotypes exist, and if so, then what is their nature and how powerful are their effects? One broad finding from this research is that sponsored workers are believed to have stronger social skills, but weaker task skills (cf. Fiske, Cuddy, Glick, & Xu, 1999) than those who were not sponsored. Finally, let me describe briefly an experiment that I conducted (see Moreland, 2003) on some of the ways in which sponsorship might bias the reactions of coworkers toward a new work group member. The participants in this experiment were college students that were randomly assigned to three-person, same-sex groups, with the restriction that each group’s members had to be unacquainted. I told students that in order to take part in our research, they had to first name and provide contact information for a friend who could be invited to participate with them. Two confederates (one male and one female) were also trained to help with the experiment by pretending to be group members. Every group played a computerized restaurant simulation called QSC Pizza Shoppe (Ness, 1987). In that simulation, decisions must be made

Recruiting Source Effects: A Social Psychological Analysis

141

about how to operate a small take-out pizza restaurant, with the goal of maximizing the restaurant’s profits. These decisions involve such matters as how much to charge for each pizza, what kinds of pizza toppings to use, whether to make coupons available to customers, how to advertise the restaurant, and so on. After these decisions are entered into the computer, an algorithm computes how well the restaurant is doing and provides feedback about that performance to the restaurant’s owner(s). The algorithm gives a large weight to publicity as a factor in the restaurant’s success. The participants were unaware of that weighting; all they were told was how to enter their decisions into the computer and how to interpret the feedback that they received afterward Each experimental session lasted for 2 hours. After operating their restaurant for 1 hour, each group’s members evaluated one another on a series of task (e.g., “How much does this person know about operating a business?”) and social (e.g., “How much do you like this person?”) measures. One person, whom I chose at random, was then removed from the group and replaced by a new person (a confederate). Half of the groups were told nothing about this new person, but the other half were told that he or she was a friend of the person who had left. The group then operated its restaurant for another hour. After about 15 minutes, the newcomer suggested that the restaurant’s advertising expenditures be tripled. This was an unusual suggestion; most of the groups spent little money on advertising earlier in the session. Aside from his or her suggestion about advertising, the newcomer was (trained to be) quiet and passive. When the restaurant’s second hour of operation was over, group members again evaluated one another, using the same task and social measures. The results of this experiment yielded several findings. First, sponsored newcomers (who were believed to be the friends of group members) were evaluated more positively than unsponsored newcomers (whose provenance was unclear). This effect was especially strong for social evaluations, but there was some evidence for it in task evaluations as well. And the effect was stronger when the newcomer’s sponsor was evaluated more favorably himself/herself (on either set of measures). Sponsored newcomers were also more influential, in the sense that their suggestion to increase the restaurant’s advertising expenditures had more impact than did the same suggestion by unsponsored newcomers. Once again, this effect was stronger when the sponsored newcomers were evaluated favorably by others, especially on task measures. And the impact of the sponsored newcomers’ suggestion was stronger when

142

RICHARD L. MORELAND

the sponsors were themselves evaluated more favorably, on either task or social measures. All of this suggests that sponsorship can be valuable for new group members. Future research might explore the limits of this phenomenon, using the same basic paradigm. For example, a researcher could vary (a) the perceived nature and/or strength of the relationship between a newcomer and his or her sponsor; (b) what changes the newcomer suggests in the group’s operating procedures; or (c) the physical presence or absence of the sponsor when the newcomer makes his or her suggestions. I chose to describe this experiment, rather than some of the other studies that I have done, for two reasons. First, there are really many ways in which the effects of sponsorship on the job outcomes of new workers can be studied. Although surveys of newcomers, their sponsors, and coworkers are obvious choices in this regard, someone could also (for example) analyze archival data, do observational studies, and even carry out experiments (in laboratory or field settings). Each method has its own advantages and disadvantages; so using more than one method would allow the disadvantages of one method to be balanced against the advantages of another. Second, although sponsorship effects are nearly always analyzed and studied in business contexts, they probably operate more widely. It can be argued that people enter groups of all kinds through social ties involving their relatives, friends, and acquaintances. If that is so, then a social psychological analysis of sponsorship and its effects may be relevant to nearly all groups.

NOTES 1. Several variations on each of these explanations have been offered. In fact, one could view each as a “family” of explanations whose members all derive from the same basic idea. To save space, I will focus on those basic ideas. 2. My guess is that the actual percentage is higher; many workers may by shy about admitting that they received help from others during the job search process, fearing that it makes them seem passive or dependent.

ACKNOWLEDGMENTS Many thanks are owed to Ms. Ranjani Krishnan, who worked closely with me to find, read, and discuss the literature on recruiting source effects. I

Recruiting Source Effects: A Social Psychological Analysis

143

would also like to thank several undergraduates, including Sheri Booth, Rebekah Greenebaum, Morris Isaacson, Jennifer Leiter, Theodore Requa, and Tara Shetye, who helped me to carry out research exploring some of the ideas presented in this chapter.

REFERENCES Allen, R. E., & Keaveny, T. J. (1980). The relative effectiveness of alternate job sources. Journal of Vocational Behavior, 16, 18
32. Ashforth, B. E. (1989). The experience of powerlessness in organizations. Organizational Behavior and Human Decision Processes, 43, 207
242. Azavedo, R. E. (1974). Scientists, engineers, and the job search process. California Management Review, 17, 40
49. Bantel, K. A., & Jackson, S. E. (1989). Top management and innovations in banking: Does the composition of the top team make a difference? Strategic Management Journal, 10, 107
124. Baron, & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173
1182. Beehr, & Taber, T. D. (1993). Perceived intra-organizational mobility: Reliable versus exceptional performance as means to getting ahead. Journal of Organizational Behavior, 14, 579
594. Beehr, Taber, & Walsh, J. T. (1980). Perceived mobility channels: Criteria for intraorganizational job mobility. Organizational Behavior and Human Performance, 26, 250
264. Benson, P. L., Karabenick, S. A., & Lerner, R. M. (1976). Pretty pleases: The effects of physical attractiveness, race, and sex on receiving help. Journal of Experimental Social Psychology, 12, 409
416. Blau, G. (1990). Exploring the mediating mechanisms affecting the relationship of recruitment source to employee performance. Journal of Vocational Behavior, 37, 303
320. Boorman, S. A. (1975). A combinatorial optimization model for transmission of job information through contact networks. The Bell Journal of Economics, 6, 216
249. Bortnick, S. M., & Ports, M. H. (1992). Job search methods and results: Tracking the unemployed. Monthly Labor Review, 12, 29
35. Braddock, J. H., & McPartland, J. M. (1987). How minorities continue to be excluded from equal employment opportunities: Research on labor market and institutional barriers. Journal of Social Issues, 43, 5
39. Breaugh, J. A. (1981). Relationships between recruiting sources and employee performance, absenteeism, and work attitudes. Academy of Management Journal, 24, 142
247. Breaugh, J. A., & Billings, E. S. (1988). The realistic job preview: Five key elements and their importance for research and practice. Journal of Business and Psychology, 2, 291
305. Breaugh, J. A., Greising, L. A., Taggart, J. W., & Chen, H. (2003). The relationship of recruiting sources and pre-hire outcomes: Examination of yield ratios and applicant quality. Journal of Applied Social Psychology, 33, 2267
2287.

144

RICHARD L. MORELAND

Breaugh, J. A., & Mann, R. B. (1984). Recruiting source effects: A test of two alternative explanations. Journal of Occupational Psychology, 57, 261
267. Bretz, R. D., Milkovich, G. T., & Read, W. (1992). The current state of performance appraisal research and practice: Concerns, directions, and implications. Journal of Management, 18, 321
352. Bridge, K., & Baxter, L. A. (1992). Blended relationships: Friends as work associates. Western Journal of Communication, 56, 200
226. Brinthaupt, T., Moreland, R. L., & Levine, J. M. (1991). Sources of optimism among prospective group members. Personality and Social Psychology Bulletin, 17, 36
43. Caldwell, D. F., & Spivey, W. A. (1983). The relationship between recruiting sources and success: An analysis by race. Personnel Psychology, 36, 67
72. Campbell, K. E., Marsden, P. V., & Hurlbert, J. S. (1986). Social resources and socioeconomic status. Social Networks, 8, 97
117. Castilla, E. J. (2005). Social networks and employee performance in a call center. American Journal of Sociology, 110, 1243
1283. Cialdini, R. N. (2013). Indirect tactics of image management: Beyond basking. In R. A. Giacalone & P. Rosenfeld (Eds.), Impression management in the organization (pp. 45
56). Hillsdale, NJ: Erlbaum. Clark, M. S., & Mills, J. R. (2011). A theory of communal (and exchange) relationships. In P. A. M. Van Lange, A. W. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology (Vol. 2, pp. 232
250). Thousand Oaks, CA: Sage. Coovert, M. D. E., Penner, L. A., & MacCallum, R. (1990). Covariance structure modeling in personality and social psychology: An introduction. In C. Hendrick & M. S. Clark (Eds.), Research methods in personality and social psychology (pp. 186
216). Newbury Park, CA: Sage. Corcoran, M., Datcher, L., & Duncan, G. J. (1980). Most workers find jobs through word of mouth. Monthly Labor Review, 103, 33
35. Decker, P. J., & Cornelius, E. T. (1979). A note on recruiting sources and job survival rates. Journal of Applied Psychology, 64, 463
464. DeGraaf, N. D., & Flapp, H. D. (1988). With a little help from my friends: Social resources as an explanation of occupational status and income in West Germany, the Netherlands, and the United States. Social Forces, 67, 452
472. Delany, J. (1988). Social networks and efficient resource allocation: Computer models of job vacancy allocation through contacts. In W. Wellman & S. D. Berkowitz (Eds.), Social structures: A network approach (Vol. 2, pp. 430
451). Cambridge: Cambridge University Press. Dipboye, R. L. (1986). Some neglected variables in research on discrimination in appraisals. Academy of Management Review, 10, 119
127. Dovidio, J. F. (1984). Helping behavior and altruism: An empirical and conceptual overview. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 17, pp. 361
427). Orlando, FL: Academic Press. Dunbar, S. B., & Linn, R. L. (1991). Range restriction adjustments in the prediction of military job performance. In A. K. Wigdor & B. F. Green (Eds.), Performance assessment for the workplace (pp. 127
157). Washington, DC: National Academy Press. Elliott, J. R. (2001). Referral hiring and ethnically homogeneous jobs: How prevalent is the connection and for whom? Social Science Research, 30, 401
425.

Recruiting Source Effects: A Social Psychological Analysis

145

Ellis, R. A. S., & Taylor, M. S. (1983). The role of self-esteem within the job search process. Journal of Applied Psychology, 68, 632
640. Fernandez, R. M., & Castilla, E. J. (2001). How much is that network worth? Social capital in employee referral networks. In N. Lin, K. Cook, & R. S. Burt (Eds.), Social capital: Theory and research (pp. 85
104). New York, NY: Aldine de Gruyter. Fisher, C. D., Ilgen, D. R. E., & Hoyer, W. D. (1979). Source credibility, information favorability, and job offer acceptance. Academy of Management Journal, 22, 94
103. Fisher, J. D., Nadler, A., & Whitcher-Alagna, S. (1982). Recipient reactions to aid. Psychological Bulletin, 91, 27
54. Fiske, S. D., Cuddy, A. J. C., Glick, P., & Xu, G. (1999). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878
902. Ford, R., & McLaughlin, F. (1986). Nepotism: Boon or bane? Personnel Administrator, 31, 80
89. Franzen, J. (1994). Lost in the mail. New Yorker, 70, 82
77. Gannon, M. J. (1971). Sources of referral and employee turnover. Journal of Applied Psychology, 55, 226
228. Gouldner, A. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161
178. Greenberg, M. S. (1980). A theory of indebtedness. In K. J. Gergen, M. S. Greenberg, & R. H. Willis (Eds.), Social exchange: Advances in theory and research (pp. 3
26). New York, NY: Plenum Press. Heilman, M. E., & Haynes, M. C. (2006). Affirmative action: Unintended adverse effects. In M. F. Karsten (Ed.), Gender, race, and ethnicity in the workplace (Vol. 2, pp. 1
24). Westport, CN: Greenwood Publishing. Hill, R. E. (1970). New look at employee referrals. Personnel Journal, 49, 144
148. Holzer, H. J. (1988). Search method use by unemployed youth. Journal of Labor Economics, 6, 1
20. Jackson, S. E. (1992). Team composition in organizational settings: Issues in managing an increasingly diverse workforce. In S. Worchel, W. Wood, & J. Simpson (Eds.), Group process and productivity (pp. 138
173). Newbury Park, CA: Sage. Jackson, S. E., Brett, J. F., Sessa, V. L., Cooper, D. M., Julin, J. A., & Peyronnin, K. (1991). Some differences make a difference: Individual dissimilarity and group heterogeneity as correlates of recruitment, promotions, and turnover. Journal of Applied Psychology, 76, 675
689. Jones, R. J., & Azrin, H. H. (1973). An experimental application of a social reinforcement approach to the problem of job finding. Journal of Applied Behavioral Analysis, 6, 345
353. Kirnan, J. P., Farley, J. A., & Geisinger, K. F. (1989). The relationship between recruiting source, applicant quality, and hire performance: An analysis by sex, ethnicity, and age. Personnel Psychology, 42, 293
308. Knouse, S. B. (1989). Impression management and the letter of recommendation. In R. A. Giacalone & P. Rosenfeld (Eds.), Impression management in the organization (pp. 45
56Z). Hillsdale, NJ: Erlbaum. Latham, V. M., & Leddy, P. M. (1987). Source of recruitment and employee attitudes: An analysis of job involvement, organizational commitment, and job satisfaction. Journal of Business and Psychology, 1, 230
235.

146

RICHARD L. MORELAND

Levine, J. M., & Moreland, R. L. (1991). Culture and socialization in work groups. In L. B. Resnick, J. M. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 257
279). Washington, D C: APA Press. Lin, N., & Dumin, M. (1986). Access to occupations through social ties. Social Networks, 8, 365
385. Lin, N., Ensel, W. M., & Vaughn, J. C. (1981). Social resources and strength of ties: Structural factors in occupational status attainment. American Sociological Review, 48, 393
406. Lin, N., Vaughn, J. C., & Ensel, W. M. (1981). Social resources and occupational status attainment. Social Forces, 59, 1163
1181. Lipset, S. M., Reinhard, B., & Malm, T. (1955). Job plans and entry into the job market. Social Forces, 33, 224
232. Louis, M. (1980). Surprise and sense making: What newcomers experience in entering unfamiliar organizational settings. Administrative Science Quarterly, 25, 226
251. Marsden, P. V., & Campbell, K. E. (1990). Recruitment and selection processes: The organizational side of job searches. In R. L. Breiger (Ed.), Social mobility and social structure (pp. 59
79). Cambridge: Cambridge University Press. Marsden, P. V., & Hurlbert, J. S. (1988). Social resources and mobility outcomes: A replication and extension. Social Forces, 66, 1058
1059. Martin, B. (1987). Recruitment ad ventures. Personnal Journal, 66, 45
63. McCauley, C., & Stitt, C. (1978). An individual and quantitative measure of stereotypes. Journal of Personality and Social Psychology, 36, 929
940. Mier, R., & Giloth, R. (1985). Hispanic employment opportunities: A case of internal labor markets and weak-tied social networks. Social Science Quarterly, 66, 296
309. Moreland, R. L. (2003, August). The role of sponsorship in attempts by newcomers to influence small work groups. Paper presented at the annual meeting of the Academy of Management, Seattle. Moreland, R. L., & Levine, J. M. (1982). Socialization in small groups. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 15, pp. 137
192). New York, NY: Academic Press. Moreland, R. L., & Levine, J. M. (1989). Newcomers and oldtimers in small groups. In P. Paulus (Ed.), Psychology of group influence (2nd ed., pp. 143
186). Hillsdale, NJ: Erlbaum. Moreland, R. L., & Levine, J. M. (1992). The composition of small groups. In E. J. Lawler, B. Markovsky, C. Ridgeway, & H. A. Walker (Eds.), Advances in group processes (Vol. 9, pp. 237
280). Greenwich, CT: JAI Press. Nadler, A., & Fisher, J. D. (1986). The role of threat to self-esteem and perceived control in recipient reaction to help: Theory development and empirical validation. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 81
122). New York, NY: Academic Press. Nelson, D. L., & Sutton, C. D. (1991). The relationship between newcomer expectations of job stressors and adjustment to the new job. Work & Stress, 5, 241
251. Nemeth, C. S., & Staw, B. M. (1989). The tradeoffs of social control and innovation in groups and organizations. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 22, pp. 176
210). San Diego, CA: Academic Press. Ness, T. E. (1987). QSC pizza shoppe. Chicago, IL: Dryden Press. Peters, S. (1993). Word-of-mouth recruitment can be risky. Personnel Journal, 72, 92
93. .

Recruiting Source Effects: A Social Psychological Analysis

147

Pfeffer, J. (1989). Organizational demography. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (pp. 299
357). Greenwich, CT: JAI Press. Premack, S. L., & Wanous, J. P. (1985). A meta-analysis of realistic job preview experiments. Journal of Applied Psychology, 70, 708
719. Quaglieri, P. L. (1982). A note on variations in recruiting information obtained through different sources. Journal of Occupational Psychology, 55, 53
55. Rees, A. (1966). Information networks in labor markets. American Economic Review, 56, 559
586. Rees, A., & Gray, W. (1982). Family effects in youth employment. In X. Freeman & X. Wise (Eds.), The youth labor market: Its nature, causes, and consequences (pp. 453
473). Chicago, IL: University of Chicago Press. Reid, G. L. (1972). Job search and the effectiveness of job-finding methods. Industrial and Labor Relations Review, 25, 479
495. Rosen, S., Mickler, S. E., & Collins, J. E. (1987). Reactions of would-be helpers whose offer of help is spurned. Journal of Personality and Social Psychology, 53, 288
297. Rushton, J. P. (1989). Genetic similarity, human altruism, and group selection. Behavioral and Brain Sciences, 12, 503
509. Rynes, S. L. (1991). Recruitment, job choice, and post-hire consequences. A call for new research directions. In M. D. Dunnette & M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 2, pp. 399
444). Palo Alto, CA: Consulting Psychologists Press. Rynes, S. L., & Barber, A. E. (1990). Application attraction strategies: An organizational perspective. Academy of Management Review, 15, 285
310. Saloner, G. (1985). Old boy networks as screening mechanisms. Journal of Labor Economics, 3, 255
267. Schlenker, B. R. (1990). Impression management: The self-concept, social identity, and interpersonal relations. Monterey, CA: Brooks/Cole. Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437
453. Schoorman, F. D. (1988). Escalation bias in performance appraisals: An unintended consequence of supervisor participation in hiring decisions. Journal of Applied Psychology, 75, 58
62. Schwab, D. P. (1982). Recruiting and organizational participation. In K. M. Rowland & G. R. Ferris (Eds.), Personnel management (pp. 103
128). Boston, MA: Allyn & Bacon. Schwab, D. P., Rynes, S. L., & Aldag, R. J. (1987). Theories and research on job search and choice. In K. Rowland & G. Ferris (Eds.), Research in personnel and human resources management (Vol. 5, pp. 129
166). Greenwich, CT: JAI Press. Skolnik, Y. (1987). Recruitment source effects: A socialization approach. Unpublished doctoral dissertation, New York University, New York, NY. Swaroff, P. G., Barclay, L. A., & Bass, A. R. (1985). Recruiting sources: Another look. Journal of Applied Psychology, 70, 720
728. Taylor, M. S., & Schmidt, D. W. (1983). A process-oriented investigation of recruitment source effectiveness. Personnel Psychology, 36, 343
354. Tesser, A., Gatewood, R., & Driver, M. (1968). Some determinants of gratitude. Journal of Personality and Social Psychology, 9, 233
236. Ullman, J. C. (1966). Employee referrals: Prime tool for recruiting workers. Personnel, 43, 30
35.

148

RICHARD L. MORELAND

Wanous, J. P. (1992). Organizational entry: Recruitment, selection, and socialization of newcomers. Reading, MA: Addison-Wesley. Wanous, J. P., & Colella, A. (1989). Organizational entry research: Current status and future directions. In G. R. Ferris & K. M. Rowland (Eds.), Research in personnel and human resources management (Vol. 7, pp. 59
120). Greenwich, CT: JAI Press. Wanous, J. P., Poland, R. D., Premack, S. L., & Davis, K. S. (1992). The effects of met expectations on newcomer attitudes and behaviors: A review and meta-analysis. Journal of Applied Psychology, 77, 288
297. Wegener, B. (1991). Job mobility and social ties: Social resources, prior jobs, and status attainment. American Sociological Review, 56, 60
71. Wiley, C. (2011). Recruitment research revisited: Effective recruiting methods according to employment outcomes. Journal of Applied Business Research, 8(2), 74
79. Williams, C. R., Labig, C. E., & Stone, T. H. (1993). Recruitment sources and post hire outcomes for job applicants and new hires: A test of two hypotheses. Journal of Applied Psychology, 79, 163
172. Zottoli, M. A., & Wanous, J. P. (2000). Recruitment source research: Current status and future direction. Human Resource Management Review, 10, 353
382.

ON IDENTIFYING HUMAN CAPITAL: FLAWED KNOWLEDGE LEADS TO FAULTY JUDGMENTS OF EXPERTISE BY INDIVIDUALS AND GROUPS David Dunning ABSTRACT Purpose
To thrive, any individual, organization, or society needs to separate true from false expertise. This chapter provides a selective review of research examining self and social judgments of human capital
that is, expertise, knowledge, and skill. In particular, it focuses on the problem of the “flawed evaluator”: most people judging expertise often have flawed expertise themselves, and thus their assessments of self and others are imperfect in profound and systematic ways. Methodology/approach
The review focuses mostly on empirical work specifically building on the “Dunning
Kruger effect” in self-perceptions of expertise (Kruger & Dunning, 1999). This selective review, thus, focuses on patterns of error in such judgments.

Advances in Group Processes, Volume 32, 149
176 Copyright r 2015 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0882-6145/doi:10.1108/S0882-614520150000032006

149

150

DAVID DUNNING

Findings
Because judges of expertise have flawed expertise themselves, they fail to recognize incompetence in themselves. Because of their flaws, most people also fail to recognize genius in other people and superior ideas. Practical implications
The review suggests that organizations have trouble recognizing those exhibiting the highest levels of expertise in their midst. People in organizations also fail to identify the best advice and correct flawed ideas. Organizations may also rely on the “wisdom of crowds” strategy in situations in which that strategy actually misleads because too few people identify the best idea available. Keywords: Human capital; self-assessment; peer-assessment; Dunning
Kruger effect; advice; wisdom of crowds

The total wealth of an individual, organization, or nation does not include solely financial assets owned or the material resources that can be sold. Since the mid-twentieth century, economists have recognized that the skills, talents, and knowledge that individuals and groups possess comprise an important form of wealth or capital as well (Becker, 1964). For example, a licensed surgeon dropped into the middle of a city has skills and expertise that he or she can convert to greater prosperity than someone whose only skill is typing 25 words a minute. This human capital, which comprises a person’s intellectual skills and technical knowledge, has increasingly become a focus of economists trying to explain the impact of schooling on the economy (Denison, 1962; Schultz, 1963), how the acquisition of skill influences wages (Mincer, 1974), and how countries develop advantages in trade (Findlay & Kierzkowski, 1983). The importance of human capital can be summed up by the estimate of economist Becker (2002) that 75% of the capital within the United States lies in the skills and knowledge of its citizens. As such, for people, organizations, and societies to gain wealth and bolster well-being, they do well to develop their own human capital and learn how to exploit it. Central to this task is the capacity to recognize human capital
where it is present and where it is absent. The individual thrives if he or she can identify personal skills that can be utilized to best navigate the modern world, as well as weaknesses that should or must be improved upon. Organizations, from small groups to entire nations,

Faulty Judgments of Human Capital

151

succeed to the extent that they can identify or develop true experts within their midst who provide the best recommendations and leadership, while dismissing false authorities who mislead. But herein lies the problem. How does one succeed in recognizing human capital? In economics, research suggests that people depend on such characteristics as years in schooling as a signal that one has acquired human capital (Denison, 1962; Schultz, 1963). But such indicators are not failsafe. A student may obtain a college degree but still suffer severe holes and deficits in important intellectual skills. After all, in handling a question about the central economic concept of opportunity cost, fewer than 22% (slightly less than chance) of nearly 200 professional economists attending an annual conference of the American Economic Association got the question right (Ferraro & Taylor, 2005). In a survey of roughly 2,300 students from 24 different universities in the United States, Arum and Roksa (2011) discovered that 36% of students displayed no improvement in their writing, critical thinking, and complex reasoning skills after four years of university study (Arum & Roksa, 2014). Often, people or organizations go beyond social signals to construct their own assessment of skills and know-how, relying on “informal” methods of assessment based on intuition, commonsense theories, and homespun deliberation of what skill should look like. For example, Google famously asked such questions as How much would you charge to wash all the windows in Seattle? or Why are manhole covers round? (Moss, 2014) to gauge skills and brainpower in potential employees. However, it abandoned such questions in 2013 after empirical investigation showed the questions were completely uninformative of future performance (Moore, 2013). In this chapter, I focus on informal, everyday assessments of knowledge and expertise, outlining some systematic problems that arise in such assessments, whether those evaluations be of self or other people. I argue that the recognition of human capital is not a straightforward task. If left to their own devices, constrained only to their own wits or erudition, both individuals and groups will make systematic errors in their judgments about who has know-how versus who knows not.

THE PROBLEM OF THE INFORMAL EVALUATOR People will suffer difficulty when informally evaluating human capital because of one central observation. That contention is that the person judging expertise is usually, by definition, doing so under the shadow of his or

152

DAVID DUNNING

her own inevitable incompetence. Except for the most trivial of tasks, each individual has gaps and flaws in his or her skills and knowledge (Caputo & Dunning, 2005). To be sure, some people possess gaps that are far wider than other people, as well as mistaken ideas that damage their judgments in occasionally flamboyant ways (Dunning, 2011; Dunning, Johnson, Ehrlinger, & Kruger, 2003; Ehrlinger, Johnson, Banner, Dunning, & Kruger, 2008; Kruger & Dunning, 1999; Williams, Dunning, & Kruger, 2013). Others in contrast have knowledge and proficiency that far outstrip anything their peers might possess. But the central contention herein is that each individual sooner or later hits a ceiling of competence that lies lower than that describing total competence or perfection. As painter Salvador Dali was fond of saying, “Have no fear of perfection
you’ll never reach it.” The inevitability of imperfections in one’s own competence leads to one general implication in many areas of life. Imperfection not only impairs performance, it also impairs a person’s ability to accurately judge performance. To the extent that people have gaps or defects in their expertise, they will not only perform imperfectly but will also make mistakes in their informal judgments of performance, whether it be by self or others. Imagine, for example, that you were asked to grade student performance on a classroom exam, but are asked to generate your own answer sheet to use to grade. To the extent that your personal answer sheet contains flaws and omissions, it will lead to you to giving grades that are flawed. In many areas of life, such imperfection or incompetence thus leads to the individual to suffer a double curse: that person will not only produce flawed performance but will produce flawed judgments of performance as well. People will suffer this double curse because the skills or knowledge they need to produce a correct response are often the very same ones they need to judge the quality of that response. For example, the expertise needed to produce logically sound arguments is exactly the same knowledge needed to recognize whether a person has just made a logically sound argument. Generating a valid physics proof requires the same math skills needed to check the validity of the proof. Thus, the “skill set” needed to attain adequate performance on some intellectual or cognitive task is the exactly the same needed to accurately perform the metacognitive task of evaluating that performance, where metacognition refers to the evaluation of one’s knowledge, reasoning, and learning (Metcalfe & Shimamura, 1994). In these cases, imperfection in tackling the cognitive task implies similar deficiencies in executing the metacognition one.

Faulty Judgments of Human Capital

153

To be sure, there are some tasks that do not share this property. For example, running fast requires good legs and quick feet. But the skill set required to assess how quickly a person runs does not involve those legs and feet. Instead, it entails a stopwatch and enough attention to start and stop it appropriately. In this case, assessing running speed presents none of the problems suggested earlier. However, to the extent that performance and judgment of that performance converge to require the same expertise, people will be double cursed by their shortcomings and defects. Many tasks in life, falling in intellectual, social, or technical realms, present this doubly cursed challenge.

IMPLICATIONS FOR SELF-ASSESSMENT Close to 20 years of research on self-assessment has shown that this doublecurse analysis of expertise carries many important implications. Despite the Delphic admonition to “know thyself,” contemporary psychological research suggests that people tend to have, at best, tenuous to meager insight into their competence, character, and personality (Dunning, 2005; Dunning, Heath, & Suls, 2004). To be sure, the perceptions people have of themselves tend to be correlated with the reality of their performance and behavior, but the relationship tends is far from perfect (Freund & Kasten, 2012; Hansford & Hattie, 1982; Harris & Schaubroeck, 1988; Zell & Krizan, 2014). When it comes to judgments of expertise, a substantial portion of the disconnect between what people think about their skill and the reality of that skill when tested can be attributed to the “double curse.” Consider the classic meta-analysis on self-evaluation published by Mabe and West (1982). When considering skills likely not inflicted with Dunning
Kruger issues, in that the competencies needed to judge skill are different from those necessary to produce good performance, one sees a higher correspondence between selfperception and reality. For example, judgments about physical skills tend to correlate (.47) with objective performance. However, when looking at skills more likely to be inflicted with the double curse, one finds smaller correspondence between self-perception and reality. Judgments of intelligence (.34) and of technical (.33), mechanical (.20), and medical-related (.17) skills, all tend to be only modestly related to actual skill. Judgments of social skills show little correspondence with reality, as evidenced in managerial (.04), job interview (.28), and interpersonal (.17) assessments.

154

DAVID DUNNING

Failing to Recognize Poor Self-Performance But more direct tests of the double curse came from a comparison of those with low levels of skill to those with higher levels. Less skill in performance should also mean less accuracy in judgment. Thus, performers with low levels of competence should provide less accurate opinions of their own or anyone else’s competence. Ultimately, this would mean that low performers would have little insight into just how bad their performances are. Because they, like everyone else, choose what they think is the most reasonable course of action, they should show confidence in the choices they make. They, unfortunately, do not have the expertise to recognize the errors they make, although other people can. They also cannot recognize the superiority of other people’s decisions when presented with them. As a consequence, they remain blithely unaware of the depths of their incompetence. This state of affairs among the incompetent has since become known in popular culture as the Dunning
Kruger effect (e.g., Morris, 2010), and many studies have demonstrated its existence. An example of data from a typical Dunning
Kruger effect study can be found in Fig. 1 (Dunning et al., 2003). This figure represents the relationship between college student perceptions of how well they have performed on a course exam they have just completed and how well they have actually performed. The responses are real; these data are taken from a large-lecture psychology class at Cornell University. In the figure, students have been sorted into four groups based on their actual test performance, from the bottom 25% of performers to the top 25%. Students’ perceptions of “their mastery of course material” and their “performance on this specific test” are tracked as a function of their performance group. Students are asked to report the percentile they think their performance falls in. The figure shows three (actually, four) major findings. First, students on average overrate their performance. By definition, the average performance falls on the 50th percentile, the point at which as many peers outperform the student as the student outperforms. All groups, even the bottom performers, believe they outperform a majority of their peers, with the average self-rating in each group falling at least in the 60th percentile or above. In a sense, this is not a surprise. Psychological research consistently shows that people overestimate their performance, skill, status in the world, and their future prospects (for reviews, see Dunning, 2005; Dunning et al., 2004). Second, there is a significant but shallow relationship between perception of performance and reality, with top performers rating their achievement slightly above bottom performers. Again, this pattern echoes past research

155

Faulty Judgments of Human Capital

Percentile

(a) 100 90 80 70 60 50 40 30 20 10 0 Bottom

Actual

Raw Score

(b)

Third Second Objective Performance Quartile Perceived Mastery

Top

Perceived Performance

45

35

25

15 Bottom

Third Second Objective Performance Quartile Actual

Top

Perceived

Fig. 1. Perceived Performance as a Function of Objective Performance on a Class Exam. (a) Presents Percentile Ratings for Perceived Mastery of Course Material and Performance on the Exam. (b) Presents Perceived Raw Score on the Exam (Out of 45 points). Source: From Dunning et al. (2003). Adapted with permission.

on self-insight, or the lack thereof, into skill and expertise (Dunning, 2005; Dunning et al., 2004; Mabe & West, 1982). It is the third finding that is the most important. Focusing on bottom performers, one sees that they rate their knowledge and performance rather highly, in the 60th percentile, although their actual performance lies much

156

DAVID DUNNING

lower, at the 12th percentile. Similar analyses focusing on estimates of raw test scores show a similar pattern, with bottom performers overestimating their raw performance by 20
40%. Not only found in the classroom, we have found this pattern among students taking logical reasoning or grammar quizzes, or assessing their sense of humor (Kruger & Dunning, 1999). We have seen it among debate teams taking part in college tournaments (Ehrlinger et al., 2008), and other researchers have seen the same among participants in chess tournaments and weekly bridge competitions (Park & Santos-Pinto, 2010; Simons, 2013), as well as among participants at a trap-and-skeet competition doing badly on a quiz concerning firearm care and safety (Ehrlinger et al., 2008). It has been observed among medical lab technicians queried about aspects of their job (Haun, Zeringue, Leach, & Foley, 2000), international students seeking pharmacy licensure in Canada, and obstetrics/gynecology interns completing their rotations (Austin, Gregory, & Galli, 2008). In each case, poor performers show little insight into how poorly they are really doing (for a recent review, see Dunning, 2011). This lack of insight appears to be an honest one. Offering to pay people up to $100 for accurate self-assessments does nothing to enhance the quality of their self-judgments or to make poor performers more negative about their performer (Ehrlinger et al., 2008). Being exposed to how other people approach a task or test
thus, seeing how other people approach the same tasks differently
does nothing to make poor performers revise their flattering and inappropriate self-estimates of skill (Hodges, Regehr, & Martin, 2001; Kruger & Dunning, 1999)

Implications for Behavior These misperceptions also influence behavioral choices. In a clever study, Ferraro (2006) presented 17 college students with two answer sheets to the 10 hardest questions students had faced in a recent economics exam. One sheet contained the answers the student had provided to those questions on the exam. The other contained answers that were all correct. Students were asked to choose the answer sheet they thought represented the best performance on those questions. For every question answered correctly on the sheet they selected, they would receive one extra credit point toward their final exam grade. Even though students themselves in reality got more than half the questions wrong, 13 of 17 (over 75%) chose the answer sheet that represented their own responses over the one with only correct responses.

Faulty Judgments of Human Capital

157

In another ingenious follow-up, Ferraro (2010) presented students with “insurance policies” that hedged against any bad performance they might encounter in a classroom exam. Policy A cost 10 exam points (which would be really deducted from the exam), but it would add 20 points to student scores if they finished in the bottom half of the class. Policy B cost only 2 exam points, but would pay only 4 points if the student’s score fell between the 50th and 75th percentile. Given these terms, 50% of students would have “bought” Policy A and 25% Policy B had they had perfect insight into their exam performance. However, only 33% of students bought a policy, suggesting that many expected to perform better than the 75th percentile. And of the two policies, Policy B turned out to be more popular than A, which again suggested that students expected their performance to be more likely to fall between the 50th and 75th percentile than to fall in the bottom 50%. In fact, of 19 students who had fallen in the 50% on every previous exam in the class, only four bought Policy A, while six bought Policy B.

CRITICS OF THE FRAMEWORK To be sure, these findings and our analysis of them are not without critics. Other researchers have asserted that the Dunning
Kruger pattern of selferror is mere statistical artifact. For example, some researchers have argued that the pattern is simply a regression-to-the-mean effect (Ackerman, Beier, & Bowen, 2002; Burson, Larrick, & Klayman, 2006; Krueger & Mueller, 2002). Simply because of measurement error, perceptions of performance will fail to correlate perfectly with actual performance. This dissociation due to measurement error will cause poor performers to overestimate their performance and top performers to underestimate theirs, the pattern found, for example, in Fig. 1. In response, we have conducted studies in which we estimate and correct for measurement error, asking what the perception/ reality link would look like if we had perfectly reliable instruments assessing performance and perception. We find that such a procedure reduces our pattern of self-judgment errors only trivially (Ehrlinger et al., 2008; Kruger & Dunning, 2002). Other researchers, working from assumptions about the performance distribution underlying our data, have claimed that poor performers face a more difficult job estimating their percentile ranking simply because there are more poor performers than top performers at the rarefied level of

158

DAVID DUNNING

expert achievement (Krajcˇ & Ortmann, 2008). However, when we look, we fail to find evidence for the critical assumptions these critics make to support their analysis, nor much adjustment to our findings if we go ahead and correct for these assumptions anyway (Schlo¨sser, Dunning, Johnson, & Kruger, 2013).

MANAGING THE INCOMPETENT The implications of self-errors inspired by the Dunning
Kruger effect carry obvious implications for the individual making them. But these errors also produce challenges for others or organizations who must deal with the individuals making these errors (Dunning, 2014c).

The Paradox of Advice Some of these implications will be obvious to anyone involved in management. To the extent that a manager must counsel and mentor others, his or her success may very well depend on the competence and expertise of the underling receiving the counsel. All academic advisors know this: often, it is the students who need advice who are the ones least likely to know it and to show up at their academic advisor’s office. Even if they do, they often are the ones who resist or discard any advice they receive. In short, incompetence leads to many paradoxes when it comes to advice. Incompetence can make people unaware that they need advice. In addition, to weigh advice appropriately, one must already have enough expertise to assess the worth of the advice we receive. If one does not have that expertise, then one runs into trouble separating good advice from bad. We have recently demonstrated this in the advice paradigm commonly used in the organizational psychology literature (Bonaccio & Dalal, 2006). In that literature, participants are asked to estimate some value, such as how tall Mt. Everest is or what year the Wright Brothers flew their first airplane at Kitty Hawk. After giving their estimate, participants are provided the estimate of another individual as advice and asked if they wish to revise their initial estimate. Usually, people revise their estimate somewhat toward the one provided, giving their own initial estimate more weight in their final conclusion (Bonaccio & Dalal, 2006; Yaniv & Kleinberger, 2000). This turns out to be a mistake in the long run: if they moved to a point equally

159

Faulty Judgments of Human Capital

between their own initial estimate and the “advice” they were given, they would be significantly more accurate. How does weight given to advice vary according to a person’s own expertise? Do people know when they are right versus grossly wrong? And, more important, can they differentiate good advice from bad? I conducted a study (Dunning, 2014b) in which American participants were asked to estimate the years in which 12 different historical events took place (e.g., the founding of CNN, the end of the Spanish-American war, and Hawaii became a state). Then, for six of the events, participants were given a previous respondent’s advice
which happened to be pretty good, namely the actual years those events took place. For the other six, respondents were given bad advice representing some of the worst estimates given in a previous group. I also assessed how much expertise respondents revealed in their initial estimates. Were they, on average, close to the actual years the events took place or were they further away. How well participants weighed the advice they were given depended on their own expertise, as shown in Fig. 2. For good advice, both knowledgeable (+1 SD from the mean in accuracy) and unknowledgeable (
1 SD from the mean of accuracy) gave equal and substantive weight to the advice, moving over halfway from their initial estimate toward the provided estimate in their final answer (51% and 59%, respectively). However, when it came to bad advice, knowledgeable and unknowledgeable respondents differed. Knowledgeable respondents hardly budged from their initial estimates, moving only 6% toward the bad advice. Unknowledgeable respondents, in contrast, gave the bad advice much weight, moving 58%

0.7 Weight Given

0.6 0.5 0.4 0.3 0.2

Good

0.1

Bad

0 Low

High Knowledge Level

Fig. 2.

Weight Given to Good and Bad Advice as a Function of Participant Knowledge Level. Source: From Dunning (2014b).

160

DAVID DUNNING

toward it from their original estimate, nearly just as much as they moved toward good advice.

The Fate of Feedback In addition, feedback per se is not necessarily an effective means to rid people of their inability to see their incompetence. Students taking multiple exams in a class do not become more accurate in predicting their exam scores as they gain experience (Ferraro, 2010; Schlo¨sser et al., 2013). If anyone does gain accuracy, it tends to be the more competent students. Students performing the worst do not alter their predictions about exam performance to become more accurate (Hacker, Bol, Horgan, & Rakow, 2000). The reasons for this resistance to gaining insight into incompetence may be many. When predicting future performance, people give less weight to their past performance than they do to how well they aspire to perform (Helzer & Dunning, 2012)
even though they give heavy weight to past performance in their predictions of other people. For example, in one study, we gave students an opportunity to predict the upcoming exam performance of one of their peers. They could win up to $5 for accurate predictions. Given the opportunity, participants wanted to learn about their peer’s past performance in the class rather than about the performance they aspired to. However, we then played a second prediction game with these same participants. Some peer was to predict how the participant himself or herself was to perform on an upcoming exam, with accurate predictions earning the participant up to $5. What information did the participant now want to give the peer making the prediction? Most participants chose to give information about the performance they aspired to in the future rather than information about their past performances (Helzer & Dunning, 2012). In addition, people resist negative feedback, sometimes with paradoxical results. In a telling study, we gave business students a chance to buy a selfimprovement book on emotional intelligence at a discount after learning about the concept in their class and taking a test of their own emotional intelligence. The feedback we gave participants had a strong impact on whether they wanted the book, but not necessarily in the way one would expect. Over 65% of top performers bought the book, but only 20% of poor performers did likewise. Ironically, giving performance feedback inspired self-improvement only among the best performers and not among those who arguably needed it the most (Sheldon, Dunning, & Ames, 2014).

Faulty Judgments of Human Capital

161

Additionally, people resist negative feedback because they can be quite creative and flexible about finding routes that allow them to dismiss negative feedback. In another emotional intelligence study, we asked one participant group about whether emotional intelligence was relevant to their lives before giving them feedback about their performance, thus “fixing” their impressions of relevance before receiving any feedback. Students performing badly in this group reacted by disparaging the accuracy of the test, which was then connected to a reluctance to engage in self-improvement. However, with another participant group, we asked them whether the test was accurate before we gave them feedback, thus fixing their opinions of test accuracy. Those in this group performing badly reacted by deriding the relevance of emotional intelligence for their lives, which was then connected, again, to a resistance to self-improvement emotional intelligence skills (Sheldon et al., 2014). Only one intervention seems to work to make poor performers see the error of their ways, and it is a paradox: training poor performers to become competent. Now skilled, they recognize their previous responses as errors, and so become more modest in rating their skill after having, if anything, been trained them to be more proficient at that skill (Kruger & Dunning, 1999). The irony of competence is to be more skilled at seeing examples of one’s own past incompetence; one can see more accurately the unskilled individual that one no longer is.

THE BURDENS OF TOP PERFORMERS At the level of the collective or group, some of the burden imposed by the Dunning
Kruger framework fall disproportionately on some of its members more than others. In particular, top performers do not escape being affected by the imperfect expertise possessed by more typical members of the group.

Failing to See One’s Uniqueness Some of this burden is revealed by looking back at Fig. 1. Much like bottom performers, top performers also tend to misjudge the knowledge
but they do so in a way quite different from their less knowledgeable peers. In the figure, one sees that top performers tend to underestimate their

162

DAVID DUNNING

performances relative to their peers
not seeing how unique or special their performance level is. They do so because top performers suffer from a “curse of knowledge” (Camerer, Loewenstein, & Weber, 1989; Nickerson, Baddeley, & Freeman, 1987). Namely, they infer that other people must have the same level or similar levels of knowledge as themselves. If they know it, others must, too. As a consequence, their errors in the figure come about via a very different psychology from that of poor performers. Top performers, unlike poor performers, assess their own work accurately in an objective sense. They have a more correct sense of what their correct responses are and what they likely get wrong (see Ehrlinger et al., 2008). Where they err is in their assessments of other people, overestimating what their peers know. We have documented this dynamic by collecting information of what top performers think of their peers and then correcting for it statistically. Such a procedure removes a good deal of the overall judgmental errors that top performers make, although it does nothing for poor performers (Ehrlinger et al., 2008), who are mainly wrong about themselves. Additionally, exposing top performers to how their peers approach a task tends to “clue them in” about low levels of knowledge among their peers
and thus leads them to recognize just how unique or “special” their own performances and achievements are (Kruger & Dunning, 1999; see also Hodges et al., 2001).

Genius as Unrecognized However, failing to recognize their own specialness is not the only burden that top performers suffer. As we have moved our research on the Dunning
Kruger framework from the individual to the collective level, we soon came to an important insight about how top performers are likely regarded by others. Consider a group assessing the expertise of the individuals within its midst. The group, as a whole, likely has the skill level to identify poor performers and their deviant mistakes. But consider what happens when the target of judgment becomes more competent. At each rise in competence, fewer in the group have the expertise to spot the true level of competence displayed by the target. Ultimately, at the highest level of competence, the expertise of the target outstrips that of most in the group. Those individuals in the group likely fail to have the expertise to understand just how competent this high performer is. Thus, when his or her judgments deviate from these less

163

Faulty Judgments of Human Capital

competent peers, his or his judgments will be seen as potential errors rather than as genius. It is as Jonathan Swift once observed, “When a true genius appears, you can know him by his sign: that all the dunces are in a confederacy against him.” In short, genius will hide in plain sight. The group on average may have the expertise to spot any poor performers within it. However, the group’s expertise on average will be imperfect, and thus it will largely fail to recognize individuals with competence closer to perfection. This assertion that genius hides in plain sight fits well with studies that showing that groups often have trouble identifying those performing the best among its midst (Littlepage, Schmidt, Whisler, & Frost, 1995; Miner, 1984). It is also consistent with more recent data we have collected expressly to test the hypothesis. In those studies, we asked respondents to take tests of logical reasoning, financial literacy, or intuitive physics. We then give them tests filled out by previous participants, representing quite a range from poor to excellent performance, asking respondents to estimate how well each of these previous participants has done. Fig. 3 presents results from a typical study, depicting the average judgments of 37 respondents examining targets who scored 4, 8, 12, 16, or a perfect 20 on a test of logical reasoning (with average performance for the respondents hovering around 12). Half of the participants were given financial incentives (up to US$50) for accurate judgments; the other half heard no mention of incentives (Dunning & Cone, 2014). 20 18 Estimated Score

16 14 12 10 8 No Incentive

6 4

Financial Incentive

2

Actual Performance

0 0

Fig. 3.

5

10 15 Target Exam Score

20

Average Estimate of Target Performance of that Performance. Source: From Dunning and Cone (2014).

164

DAVID DUNNING

As seen in the figure, the group is largely accurate in its judgments of poor-performing targets. However, as target performance increases, estimates and their accuracy begins to fall off. In fact, the performance of the top performer is underestimated by some 35%
with the group seeing this performance as barely above average. And, as seen in the figure, financial incentives had no impact on the group’s accuracy. In addition, although not shown in the figure, higher performing respondents came closer to recognizing the truly excellent performance of the top performer relative to those doing more poorly (Dunning & Cone, 2014). We have replicated this finding of genius hiding in plain sight in multiple ways. In one study, we ask participants which peer they would go to for financial advice after seeing a set of peer performances on a financial literacy test. A greater proportion of participants spotted the person they should most avoid (i.e., the bottom performer; 43.2%) than identified the person they should most seek out (i.e., the top performer; 29.6%), a difference reaching statistical significance (Dunning & Cone, 2014). In addition, in another study, we asked participants to compare themselves to each individual target and bet whether their score on a logical reasoning test beat was beaten by or tied the target’s score. They would win an additional $1 if their prediction was right. With the bottom performer, participants were largely accurate in their bets. Of the 101 respondents, 84 thought their score would beat that of the target, with additional 7 participants claiming ties. These predictions came quite close to reality (i.e., 89 wins, 6 ties). However, when comparing themselves to the top performer, respondents grossly overestimated themselves. Of the 101, 27 thought their score would beat that of the top performer, with an additional 27 claiming a tie. In reality, only one participant’s performance managed to tie that of the top performer, with the rest losing (Dunning, 2014d).

Leading Back to Self-Flattering Social Comparisons In a sense, the fact that genius hides in plain sight provides an intriguing explanation for an oft-documented phenomenon in self-psychology. That finding is that people on average think of themselves as anything but average. Among college students, 70% believe themselves to be “above average” in leadership ability, but only 2% see themselves as below average (College Board, 1976
1977). A full 94% of university professors state that they do above-average work relative to their peers (Cross, 1977). In one software development firm, a full 32% of engineers thought their skill level

Faulty Judgments of Human Capital

165

put them in the top 5% of employees; in another firm, the figure was 42% (Zenger, 1992). This above-average bias in self-assessment is ubiquitous (for a review, see Dunning et al., 2004). Ironically, people even claim above-average abilities in providing accurate self-evaluations untarnished by bias and distortion (Friedrich, 1996; Pronin, Lin, & Ross, 2002). The inability to recognize talent among top-performing peers, coupled with a true capacity to spot poor performers, would lead to these inflated self-estimates, honestly believed. If accuracy is lop-sided, with people accurately spotting those they outperform but underestimating those who outperform them, then people would have sincere but mistaken evidence about how their expertise compares against that of their peers. Again, we have recent evidence of this dynamic (Dunning, 2014d). In it, I asked participants to complete a reasoning test in which participants, on average, thought they scored 6 out of 10 items right, a significant but slight overestimate. Subsequently, for half of participants, I showed them tests as filled out by five poor performers who averaged a score of 3.4. For the other half, I showed them tests filled out by five superior performers who attained an 8.6 score on average. I asked all participants to estimate how well these other performers had done. Participants moderately overestimated the performance of the poor group by 1.2 items on average, but solidly underestimated the performance of the superior group by 2.0 items. In short, participants accurately saw themselves as outperforming the poor performers, even though they moderately overestimated those performers. However, due to a mixture of overestimating themselves and underestimating superior performers, they failed to see just how much they came up short in comparison to those top-line performers. This was evident in assessments in which they directly compared their own performance to the set of targets they had been given. Participants rated themselves superior on average to the poor-performing group
an accurate result. However, they rated their own performance roughly equal to that of the superior performers, a grossly mistaken impression, but one that followed directly from their misestimates of self- and peer performances. In sum, blindness to performances better than one’s own can lead to self-flattering social comparisons not supported by actual evidence.

Resistance to Correct Beliefs in the Marketplace of Ideas What is true of top-performing people may also be true of very smart ideas. They, too, suffer burdens due to the Dunning
Kruger framework

166

DAVID DUNNING

and might prove too smart for their own good. The collective may not share the expertise or “intellectual scaffolding” to understand their worth. As computer pioneer Howard Aiken put it, “don’t worry about people stealing your ideas; if your ideas are any good, you’ll have to ram them down people’s throats.” Or, as English writer Aldous Huxley observed, “the vast majority of human beings dislike and even dread all notions with which they are not familiar; hence it comes about that at their first appearance innovators have always been derided as fools and madmen.” Smart ideas, especially those of the highest IQ, run the risk of being doubted, unheeded, or discarded in favor of notions that still carry some worth, but not that of the “high IQ” idea. We have begun to conduct research to see if ideas might be too smart to be truly recognized by the collective. Central to this idea is the notion, taken from the literature on the marketplace of ideas, of the “habitat” in which ideas must live and survive. Some ideas have a friendlier habitat, in that people possess beliefs or knowledge that “hook into” those ideas more easily and validate them, thus supporting people’s adoption of them. For example, many more recent stories about witches have arisen than stories about trolls because people already have a rich set of associations and stereotypes concerning the former then they do the latter. People are cognitively prepared to understand and remember any story about witches more than they about trolls, and thus stories about witches “win” in the marketplace of ideas (Berger & Heath, 2005). High IQ ideas, by definition, may exist in harsher habitats, in that people typically fail to have the expertise or intellectual scaffolding necessary to recognize their worth or even remember them accurately, relative to ideas understood by a larger proportion of the population. I have already demonstrated this in an initial study (Dunning, 2014a), asking people how likely they are to pass along various pieces of information they have heard before to a friend who is interested. For example, participants are asked how likely it is that they will pass along the answer “the Sahara” if their friend wants to know what is the largest desert in the world, or “Pluto” if they friend wants to know which planet was discovered last. There is one trick to the study, however. For some participants, they are asked if they will pass along answers that turn out to be popular, but which actually turn out to be wrong, such as the Sahara and Pluto. The correct answers to those two questions are the Antarctic and Neptune (remember, Pluto has been decommissioned as a planet as of 2005). However, when asked the likelihood of passing along the answers, participants reported being more willing, and more disposed to vouch as true,

167

Faulty Judgments of Human Capital

popular but wrong answers over unpopular but right ones. In a more telling example, participants stated they were more likely to pass along information to a friend worried about chronic high blood pressure that the disease producing symptoms of dizziness and headaches (which is inaccurate) over information that the condition had no external symptoms (which is correct) (Dunning, 2014a).

Limits to Correcting Belief Correct ideas may also suffer resistance in another way if they fall too outside the box of people’s intuitions. Given advice about these ideas, people may give them weight, but with limits. This was demonstrated in another study about a counterintuitive statistic
that the top 20% wealthiest of Americans possess roughly 89% of the total wealth in the country. In an experiment in which people were asked to estimate the percentage of wealth owned by that top 20%, respondents on average initially estimated that the figure was roughly 67%, as shown in Table 1. We then exposed participants to an estimate that had been given by a previous respondent, from 30% to 90% (the correct answer) by intervals of 15%, and asked if they wished to revise their estimates. Participants showed some sensitivity to the accuracy of the estimates they were provided. Those exposed to lower, but inaccurate, estimates of 45% and 30% declined to lower their estimates. In the groups given 75% and 90% as an advisor’s estimate, figures closer to the truth, both groups revised their estimates upward to a statistically significant degree. But there Table 1. Impact of Other Person’s Estimate on One’s Own Estimate of Proportion of Wealth Held by Top 20% of the Wealthiest in the United States. Other’s Estimate (%)

a

90 75 60 45 30 a

Correct figure.

Own Estimate (%)

p

Initial

Final

64.1 67.9 68.3 66.7 66.5

73.2 73.3 68.5 67.1 63.9

100 300 Varied

39% 52% 64% Mean = 58%

16 13 10 29

For confidentiality, only general size ranges can be given.

219

Collaboration among Highly Autonomous Professionals

Within each firm we then used purposive sampling (Patton, 1990) to identify interviewees who would help up to balance our sample on several theory-driven dimensions such as organizational tenure, geography, career track, and professional specialization (Gardner, 2012a; Hitt, Bierman, Shimizu, & Kochhar, 2001; Von Nordenflycht, 2010). We began each interview by defining collaboration (many people recalled the definition from the survey), and stated that our purpose was to understand more about the benefits and barriers in their firm as they personally had experienced them. We used probing techniques to ask interviewees to clarify their responses, especially to distinguish their general beliefs about collaboration versus those they had personally encountered. We also relied extensively on a technique known as informant feedback (Miles & Huberman, 1994) or member checks (Lincoln & Guba, 1985). We confirmed our understanding of key concepts, both during interviews and sometimes in follow-up phone calls or emails. In addition, we tested our emerging interpretations by presenting and discussing summaries of our coding and analysis with larger groups of professionals during follow-up workshops. Table 2 summarizes these formal verification events. During the workshops, we presented our findings and encouraged participants to challenge their accuracy, validity, and completeness.

Reported Advantages of Professional Peer Collaboration Knowledge and Learning In the survey responses, the most frequently mentioned benefit of collaboration (34 percent of responses) related to participants’ ability to learn from their peers during collaborative work. Respondents reported gaining “Knowledge about what other parts of the Firm are up to, as well as Table 2.

Formal Verification Events to Interpret Findings.

Location

Number of Attendees

London Boston

125 65

Chicago Naples, FL

40