Social Relations Modeling of Behavior in Dyads and Groups

By Thomas E. Malloy

Social Relations Modeling of Behavior in Dyads and Groups covers software, interpersonal perception (adult and children), the SRM with roles (e.g. in families), and applications to non-human research. Written in an accessible way, and for advanced undergraduates, graduate students and researchers, author Thomas E. Malloy strives to make inherently abstract material and unusual statistics understandable. As the social relations model provides a straightforward conceptual model of the components that make up behaviors in dyads and groups, this book will provide a powerful conceptual and methodological toolbox to analyze behaviors in dyads and groups across the sciences.

This book is specifically designed to make this toolbox accessible - beyond interpersonal perception phenomena. It helps identify the relevant phenomena and dynamics surrounding behaviors in dyads and groups, and goes on to assess and analyze them empirically.

• Captures essential conceptual and methodological topics around the scientific analyses of behaviors in groups and dyads
• Situates the SRM in the history of dyadic research
• Offers detailed guidance on research design and measurement operations
• Organizes models and empirical results into easily read figures and tables
• Demonstrates how SRM variances and covariances can be used as dependent measures in experiments
• Conceptualizes novel phenomena in personality psychology using the SRM

• Language: English
• Publisher: Academic Press
• Release date: Aug 10, 2018
• ISBN: 9780128119662

Thomas E. Malloy

Thomas E. Malloy Professor of Psychology Mary Tucker Thorpe Professor Department of Psychology Rhode Island College Providence, Rhode Island 02908 [phone] Office [e-mail] Thomas E. Malloy has conducted research on interpersonal perception, peer perceptions in classrooms, intergroup relations, and reconciliation, individual differences and behavior, cross-cultural psychology, research methodology, and healthy psychology. He is currently funded by RI-INBRE and the National Institute of General Medical Sciences, a component of the National Institutes of Health (NIH) to study Visual Attention to Faces of in-group and out-group members. Professor Malloy directs the Intergroup Relations Laboratory at Rhode Island College. He works with researchers at the Finnish National Institute of Health and Welfare on the 1987 Finnish Birth Cohort Study, a 25 year longitudinal study of all those born in Finland in 1987. Professor Malloy is collaborating with researchers at The Hebrew University of Jerusalem on the quality of listening in dyadic interactions. He is also collaborating with researchers at the University of Ulster in Ireland on face-to-face dyadic interaction. He has offered methodological workshops at annual meetings of the Association for Psychological Science on Social Relations Modeling of Dyadic Data.

Book preview

Social Relations Modeling of Behavior in Dyads and Groups

First Edition

Thomas E. Malloy

Mary Tucker Thorp Professor of Psychology, Rhode Island College, Providence, RI, United States

Table of Contents

Cover image

Title page

Copyright

Dedication

Foreword 1

Foreword 2

Chapter 1

Chapter 2

Chapter 3

Chapter 4

Chapter 5

Chapter 6

Chapter 7

Chapter 8

Chapter 9

Chapter 10

Chapter 11

Chapter 12

Chapter 13

Chapter 14

Final appreciation

Preface

1: Dyads and groups

Abstract

The nature of dyads

Dyads within groups

The adaptive function of dyads and groups

The dyad and group in social psychology

Representative design in dyadic research

The dyad awaited a solution to the dependence problem

The heuristic value of the social relations model

General aims of this book

2: The logic and mathematics of social relations modeling

Abstract

Conceptual and analytic concerns with dyadic data

The componential structure of dyadic data

The components of dyadic scores specified by the social relations model

Conceptual examples of actor, partner, and relationship components

Validity criteria for SRM components

Computation of SRM effect estimates

Computing effect estimates in the half-block design

Computing effect estimates in the round-robin design

SRM variance components

Covariances of SRM effect estimates

Self-actor correlations

Self-partner correlations

Splitting generalized and dyadic interpersonal phenomena

Generalized reciprocity correlations

Dyadic interpersonal reciprocity

Bivariate reciprocity

Dyadic intrapersonal and interpersonal reciprocity of metaperspectives

Multivariate models with effect estimates

Summary

3: Research designs for social relations analysis

Abstract

The group in designs for social relations analysis

Distinguishable and indistinguishable dyad members

The nature of multiple interaction designs

The round-robin design

Half-block design

Half-block with a nested structure

The symmetric block design

The asymmetric block design

One-with-many design

Key person design: Nomothetic and idiographic

Generation design

Integrating traditional experimental and multiple interaction designs

Summary

4: Planning research for social relations modeling

Abstract

Measurement operations

Research contexts and design implications

Summary

5: Interpersonalism: Personality processes in dyads

Abstract

From interactionism to interpersonalism

Goals of this chapter

Interpersonalism: Behavior in the dyadic context

Interim summary

Personality processes: Variance components and covariances

An empirical example: Individual differences and cross-situational consistency of mice (Mus musculus) behavior in opposite sex interactions

Summary of the prescriptions of interpersonalism

6: The psychophysics of trait perception: Accurately detecting minimal differences between people

Abstract

Acknowledgment

Social relations modeling of consensus and accuracy

Psychophysics of consensual and accurate trait perception

Variance component analysis and difference detection functions

General experimental procedures and analyses

Procedures: Between-subjects studies

Variance component analyses: Between-subjects studies

Integration of results from four between-subjects studies

Sequential processing of trait information

Variance component analyses: Within-subjects study

Idiographic accuracy in the detection of minimal target differences

Summary

7: The generalized and dyadic interpersonal self

Abstract

Self in the context of others

Generalized and dyadic models of the self

An empirical example: The dyadic self in listening and intimacy

Generalized and dyadic self-other models

Summary and conclusion

8: Interpersonal perception

Abstract

Componential approaches to interpersonal perception

Interpersonal perception at the individual level: Assimilation and consensus

Origins of perceivers’ agreement and disagreement

Dyadic interpersonal perception: Uniqueness

The second Cronbach critique of interpersonal perception

Methodological implications of the second Cronbach critique

9: ARRMA: assumed reciprocity, reciprocity, and metaperception accuracy

Abstract

The ARRMA model

ARRMA is a multivariate componential model

Specification of ARRMA parameters at the individual level

Theoretical predictions: Individual level of analysis

Specification of ARRMA parameters at the dyadic level

Theoretical predictions: Dyadic level of analysis

Organization of SRM effect estimates for ARRMA analysis

Model fit and comparison: Individual and dyadic levels

Summary

10: Interpersonal similarity in dyads

Abstract

Self-referenced interpersonal perception

Self-referenced perceived interpersonal similarity

Profile analysis

Summary

11: Interpersonal attraction in dyads and groups

Abstract

Aims of the chapter

Determinants of interpersonal attraction in dyads

Measurement operations in interpersonal attraction research

Social relations analysis of interpersonal attraction

Variance components in interpersonal attraction

Variance components in metaperception of interpersonal attraction

Consistency of interpersonal attraction across groups

Components of interpersonal attraction across groups

Consistency of attraction to key persons across groups

Consistency of the key persons’ attraction to others across groups

Undecomposed scores should not be used to estimate consistency across groups

ARRMA model of interpersonal attraction

Interpersonal attraction hypotheses derived using the SRM

ARRMA predictions

Summary of the Malloy (2018) results

Interpersonal attraction among the well-acquainted

Splitting the similarity-attraction correlation: Individual and dyadic

Implications of componential analysis of interpersonal attraction

12: The componential structure of social vision: Face processing

Abstract

Visual attention, categorization, and differentiation of faces (ACD)

Perceiver, facial feature, and perceiver by facial feature effects on social vision

Hedonic relevance and face processing

Facial features and stereotypes

Target status and perceiver visual attention

Unmediated and mediated effects of facial features

Measuring visual attention

The componential structure of visual attention to faces

Using SRM effect estimates and variance components to test the ACD model

Empirical applications of the componential model of face processing

Face centricity, visual attention, and ability judgments

Facial attractiveness and visual attention

Interim summary

Attractiveness and face recognition

Facial attractiveness in human relations

Empirical findings: Facial attractiveness and recognition memory

Research methods: Facial attractiveness and recognition accuracy

Variance components: Perceiver, generalized, and idiosyncratic distinctiveness

Generalized distinctiveness and perceived attractiveness of faces

Recognition accuracy: Signal and noise faces

Summary and conclusions

13: Social relations modeling in groups

Abstract

The intergroup relations model

Social relations modeling of intragroup and intergroup phenomena

Design considerations for social relations modeling in groups

Social relations modeling of in-group and out-group responses

Social relations modeling of archival intergroup relations data

Variance component analysis of out-group covariation bias

Social relations modeling of longitudinal intergroup processes

Simultaneous social relations modeling of intragroup and intergroup processes

Social relations modeling with experimental and quasiexperimental groups

Intercultural processes

A cautionary note: Simulating ratios of SRM individual level variance components

Summary

14: Social relations analysis of dyadic data structures: The general case

Abstract

The problem

Notation

Sums of cross-products

Variance-covariance matrix among observed scores

The coefficient matrix

Exact and estimated standard errors

An example

Monte Carlo simulations

Alternative approaches

Index

Copyright

Academic Press is an imprint of Elsevier

125 London Wall, London EC2Y 5AS, United Kingdom

525 B Street, Suite 1650, San Diego, CA 92101, United States

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom

© 2018 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher's permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-0-12-811967-9

For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Nikki Levy

Acquisition Editor: Emily Ekle

Editorial Project Manager: Barbara Makinster

Production Project Manager: Mohana Natarajan

Designer: Mark Rogers

Typeset by SPi Global, India

Dedication

This book is dedicated to my father, Thomas E. Malloy, Sr. and my mother, Margaret M. Malloy.

Foreword 1

Charles F. Bond , Nashville, Tennessee

Bob punches Carl. Debbie stares at Heather. George believes that Julie likes him. Each of these examples involves a pair of individuals, that is, a dyad. The book Social Relations Modeling of Behavior in Dyads and Groups considers examples like these. It seeks a rigorous explanation for dyadic phenomena. To explain why Bob punched Carl, it would consider Bob's aggressiveness, Carl's tendency to be the target of aggression, and the relationship between Bob and Carl.

It was in 1979 that Warner, Kenny, and Stoto devised a statistical analysis of dyadic phenomena. It decomposed those phenomena into actor effects (individual differences in the tendency to engage in a certain behavior), partner effects (individual differences in the tendency to elicit that behavior from others), and relationship effects (tendencies that are specific to particular pairs of individuals). Since 1979, these Social Relations techniques have been widely applied, and our understanding of social psychology has been revolutionized.

Malloy earned a doctoral degree in counseling psychology and, in the late 1970s, held a tenure-track position in that field. Though well positioned to earn tenure, Malloy resigned his counseling psychology position to pursue a second doctorate—the latter in social psychology with David Kenny. It was under Kenny's tutelage that Malloy commenced a 35-year immersion into Social Relations research. No one has done more of it; no one has thought more about it; no one has taken Social Relations in more directions than Thomas E. Malloy. Yes, he has investigated Social Relations among college sophomores. But he has also studied mice. Yes, much of his research has been done with American subjects, but he has also used the Social Relations Model in cross-cultural, archival, and statistical investigations. Yes, Malloy has taught social psychology to undergraduates at his home institution Rhode Island College, but he has also offered Social Relations workshops to PhDs at locales around the world. No one is better qualified to write a book on Social Relations Modeling than Thomas E. Malloy.

This book is not for the casual reader. Math phobics may cringe at the author's equations. Dabblers may miss the most crucial details. Malloy's book targets those who seek a deep understanding of Social Relations Modeling, people who are willing to commit themselves to the close reading this Model merits. Study the book, and your industry will be rewarded.

Foreword 2

Avraham N. Kluger , The Hebrew University of Jerusalem

I have learned about social relations modeling (SRM) and started using it—about 2 years ago. Thus I could be considered an advanced novice. I wrote my foreword to this fine book from my personal point of view with the hope to share with the reader my excitement about the various gems that can be found in every chapter both for the novice and for more seasoned users of SRM.

Chapter 1

In 2004, I experienced a professional identity crisis, being an organizational psychologist working on individual responses to feedback (e.g., Kluger & DeNisi, 1996). The crisis was sparked when I was introduced to Appreciative Inquiry (Cooperrider & Srivastva, 1987) and was interviewed by a partner, and interviewed that partner, during a small workshop. In that workshop, I was requested to tell a story about a moment at work during which I felt full of life. Not only did I find the experience mesmerizing, I recognized that the experience was created somehow due to the interaction, such that neither my partner nor I could have had similar personal insights about our own work without the presence of the other. That is, the experience was dyadic. This led to my disillusionment with all the statistical and conceptual tools that were available to me. Furthermore, this was also coupled with disillusionment regarding the relevance of my research, as human resource managers and educators seemed to be oblivious to my research findings regarding the potential damage of feedback, even positive feedback, to performance (Van Dijk & Kluger, 2011).

I first overcame the relevance crisis when I switched my attention from feedback to listening, that is, from asking what happens, for example, when a supervisor talks about performance, versus when a supervisor listens to subordinate's ideas about performance, or about anything else. Next, through a chance communication with Thomas E. Malloy in 2015, I became aware of SRM—the topic of this book. When I understood the potential of SRM to address variety of questions about listening, and social behavior in general, I felt as if I were an astronomer of the early 17th century being given for the first time in one's life a telescope. Suddenly, I could see phenomena that I could not imagine beforehand, such as the possibility that good listening is much more a function of one's partner than one's trait. I felt as if I were Galileo watching the moons of Jupiter.

Yet, reading Chapter 1 of this book gave me another surprise. While, I experienced SRM as a tool that allows the discovery of new social phenomenon, Malloy amassed in this chapter a collection of theoretical claims raised by luminaries of social psychology before they had the tools of the SRM. They all point to one social dictum: social life is dyadic to a large degree. Thus my take home message for the reader is whether you start from theory or from applying SRM as a tool, you are bound to discover new territories.

Chapter 2

Chapter 2 led me to grasp the complexity of SRM in a gradual manner, starting with a half-block design and only then moving to explain the round-robin design. That is, showing that SRM is a general approach that could be applied to multiple designs (with different purposes). Moreover, the chapter helped me better understand the meaning of actor versus partner effects as the consistency of behavioral responses of one to many (i.e., actor) and many to one (i.e., partner). This chapter also laid a foundation for longitudinal SRM research. If dyadic interactions occur on multiple occasions, to establish truly unique interpersonal responses there should be temporal stability (i.e., convergence) of unique responses to specific partners. I will definitely apply this insight in my future research.

Upon reflecting on this chapter, I realized that it offers many insights and points to new research possibilities both for the industrial/organizational psychologist and for the social psychologist. Perhaps, the deepest insight borrowed from Tagiuri (1958) is the idea that successful leaders annul (or perhaps change) reciprocities among their followers. SRM would allow testing such hypotheses, exposing leadership effects that are otherwise not visible. The idea that round-robin data that are not decomposed are hiding useful information could be a boon to employee selection experts who have data from assessment centers, especially when each participant rates the behavior of all others in the group (e.g., panel interview). It could be that estimating partner scores, with the adjustment to actor scores (random variation among raters in leniency) and relationship score would produce much more reliable scores as to increase validity, while at the same time alerting selection specialists that the criteria for the selection system (e.g., performance) may as well have SRM components. Finally, this chapter demonstrated how the SRM can address questions regarding the construction of metaperception, or how a person perceived being perceived by others. Thus my view of the potential of SRM, as fantastic as it is, is probably capturing a fraction of the possibilities, and experts in different domains of human behavior may be able to see infinite new possibilities afforded by the SRM.

Chapter 3

This chapter introduced me to two designs not commonly found in applications of SRM. First, Malloy demonstrates how SRM could be used for one-with-many designs that are more familiar to researchers studying phenomenon like leadership and know how to use hierarchical linear modeling (HLM). The SRM could become useful when the nesting becomes more complex. For example, some workers have both an administrative and a professional supervisor, studying the unique effects of different supervisors could be accommodated by SRM. Second, the key-person design combined with measurement of metaperception shed light on the gap between how people perceive their behavior and how it is perceived by others: People around the globe believe that members of different groups judge their traits similarly when, mostly, they do not. That is, people seem not to be aware of how they change their behavior as they traverse between one sphere of life (e.g., family) to another (e.g., coworker). This tool of the key person design, when applied to large numbers of key persons, may allow in the future calculating discrepancy matrices, as to identify people who excel at understanding the malleability of their own behavior across spheres of life.

Chapter 4

This chapter challenged me to consider design choice when planning a study. Among the challenges are When self is measured in a dyadic study, one should consider explicitly the referent to be used when rating the other and should be guided theoretically. and One should consider the possibility that the minimal symmetric block (and asymmetric block) may increase relationship variance because of the structure of the design. In addition, this chapter raises interesting questions about what people think that other people think about them (metaperception): Partner variance in metaperceptions is often near zero with most stable variance determined by actor and relationship effects; and Much remains to be done to understand when metaperception is a function of self-perception, and when it is determined by one's behavior with another. Finally, this chapter intrigued me to consider analyzing archival data such as commercial exchanges between nations with the question of whether SRM structure will be found in data of interest to economists.

Chapter 5

Chapter 5 elucidated for me the role of partner effect as an ignored aspect of personality: the reliable individual difference in the tendency to elicit or suppress behaviors in others. The chapter also shows how to get out from a trait/state debate and integrate both perspectives into a single paradigm. Moreover, Malloy proposes predictors of various components of personality-related parameters such as cognitive development that may determine the amount of variance accounted by the situation created by a specific partner.

Chapter 6

Presents a surprising application of the SRM logic to the question of detection of differences between traits of other people. This insight and calculation tools may prove useful in domains unimagined by the author. For example, in panel interviews where issues of reliability (among judges) and validity (prediction of actual job performance) have serious implications, SRM could be useful in detecting differences among candidates that are consensual versus differences that are random. This has far-reaching implication for fairness in interviews.

Chapter 7

Introduces yet another novel application of SRM with the idea that self-perception of behavior changes with particular dyadic partners, such that I may perceive myself as a funny person in the company of X, but not in the company of Y. While theories regarding the dialogical self predict such behaviors, the use of bivariate SRM, where one of the SRM variables is self-perception, opens the door for many new inquiries regarding the self, dyadic accuracy, and sources of metaperception.

Chapter 8

Describes the application of SRM to analyze perceptions, where the terms actor and partner are designated as perceiver and target. The perceiver effect measures raters’ tendency to perceive others as consistently high or low on a trait. This variance can be studied across time and it is shown how it is reduced along with acquaintance and maturation. As such, the changes in perceiver effects opens a window for understanding the development of cognitive representation of others. Perceiver effect may also reflect low motivation of study participants to differentiate targets, either due to fear of lack of anonymity, disinterest in the study, and more. Thus modeling the effect of motivation on perceiver effect may have both theoretical and methodological benefits.

The target/perceiver ratio is an interesting measure that can reveal behavior of different variables and changes across time. For example, Malloy demonstrates how the rise in T/P ratio in measures of popularity between grade 1 and grade 6 reveal potential implication for the developmental trajectories of cognitive representation of social hierarchy. I found this as, yet another, exciting insight afforded by SRM.

Another fascinating implication is that target effects (consensus) regarding observable traits (e.g., cognitive ability in the classroom) can explain outcome variance much better than typical aggregation of judges score, because the SRM decompose the variance and only the relevant variance is used in the prediction.

Comparison of meta-analyses of consensus regarding the Big 5 reveals that in short-term one-to-one interactions the level of consensus is very low, whereas in short-term group interaction consensus is markedly higher. This may hint that one-to-one interaction is a special kind of activity that affects perception and knowledge. Given the emphasis of the book on the importance of dyads as the building block of the self and of society, it also raises questions about the different roles that dyads versus groups have in shaping the self and society.

This chapter also introduces an example of SRM with roles in families and demonstrates a high degree of dyadic variance and dyadic reciprocities. Thus it seems that relationships in families are highly dyadic. That is, unique relationships are formed between mothers, fathers, and various children. These pairings open the door to many new questions about formation of coalitions within families and exposing their predictors and functions.

Chapter 9

Shows how bivariate SRM applied to one variable and metaperception of the same variable allows studying assumed reciprocity, actual reciprocity, and accuracy (ARRMA model). This sheds light on classical theories in social psychology (e.g., balance theory) and shows how this logic could be extended from dyads to triads as to study people watching. Reviewed evidence showing how people assume that the way they behave toward others is similar to the way others behave toward them, but that actual reciprocity and accuracy are often low and seem to depend on the type of variable measured and the type of people involved in the rating. Without a doubt, ARRMA opens yet another door to understanding the fabric of social life, including all the illusions that allow people to navigate social life. Fortunately, Malloy walks the reader through the technical details needed for these insights and properly warns the reader about places where confusion in data preparation may occur.

Chapter 10

Tackles a seemingly simple question of similarity, both perceived and actual, between people. Understanding similarity is obviously of interest for psychologists because it predicts attraction. Yet, as the chapter shows, similarity could be calculated relative to many referents such as oneself, specific others, or a generalized other. Two findings, one empirical and one methodological, captured my eyes. First, using the key-person design, Malloy reports that people systematically differ from each other in the degree of uniqueness (dissimilarity) they experience across three life domains (family, work, and friends). People also tend to believe that their uniqueness is similarly assessed by their partners in these life domains. Yet, they are largely wrong, as there is very little agreement among partners from different life domains regarding the uniqueness of the focal person. This exposes that people navigate, somehow, social life with mostly a fixed idea about their uniqueness that they carry from one domain to another, although their partners perceive their uniqueness in different ways as a function of the life domain in which they get to know the focal person. This raises unexplored questions regarding potential individual differences in these phenomena. For example, who are the people who are aware that they are perceived uniquely (or not) in one domain, but not in another. Are these people better adjusted? Or, if objective data were available, in which social domain do judges perceive the focal person more accurately (taking into account the likely fluctuation in objective behavior across domains)?

The second fascinating discussion is methodological and pertains to profile similarity between two people (e.g., preferences of husband and wife for similar recreation activities). The discussion shows that [r]emoving the wrong base rate (the stimulus mean across dyads and persons) can drastically alter any assessment of the components of the data. When properly modeled, such data could be analyzed to assess actor effect (the smaller it is the more similar are dyad members), stimulus effect (shared attractiveness of a given recreational activity), and actor by stimulus effect (the unique preferences of one dyad member of various activities).

Chapter 11

Builds on previous models and research, and asks questions about attraction, and about the similarity-attraction hypothesis. First, it reveals that some people are generally attracted to others while some others are not, and that there are some people that are more attractive than others. This led Malloy to conclude, Attention to the factors that determine one's actor and partner effects is a promising direction for future research. I suggest that Bowlby's attachment theory may offer a bridge. It is possible to add personality variables, or even view attachment style (Mikulincer & Shaver, 2016) with SRM components to understand these phenomena.

Another fascinating finding is that people assume that attraction is reciprocal but largely it is not. This led Malloy to suggest paying attention to the science of people watching and trying to understand how A and B believe that C sees them, and hence, Malloy concludes, Triadic analysis of interpersonal attraction holds considerable promise.

The final finding that I found very interesting pertains to similarity in how one is perceived by family members, coworkers, and friends. In fact, of the nine estimates of the consistency of key persons’ partner effects in attraction across groups, only one was reliably different from zero. This led to the conclusion, affective responses to an individual by members of different groups are context specific. The meaning of these findings is that although people tend to think about themselves as a single person, in every social situation they are perceived as different people.

Chapter 12

Presents yet one more unexpected application of SRM, this time to social vision. Malloy reports multiple studies of social vision based on eye tracking (attention) to different faces (varying on race prototypically and on attractiveness). The overarching finding is that most of the variance stems from the perceiver and from the perceiver × face interaction, but not from differences in the faces. Malloy notes, Counter-intuitively, face effects on social vision are weaker than the other effects. This raises several questions. First, what predicts high versus low level of attention to faces among different people? Second, what is the meaning of research on human attention that failed to decompose the SRM components (as different faces explain very small amount of variance in visual attention)? Third, what is it about different people that pay attention to different faces (some people look more at race-typical faces and some at race-untypical faces)?

Chapter 13

This chapter expanded my understanding of SRM in two ways. First, it taught me that SRM components could drastically change when different variables are involved. Second, it taught me several extensions of SRM by combining designs and by considering the relationships among SRM components. Specifically, whereas in most research reviewed in the book, partner (consensus) effects are small or nil, when perceived leadership was measured in zero-acquaintance teams the partner effect was substantial. Next, I describe several methodological extensions that I found fascinating.

Among the extensions, this chapter shows how to combine the experimental approach with SRM to manipulate SRM parameters. This chapter also shows how to use repeated measurement to obtain insight into social behavior. For example, in mice, actor effects in sniffing are stable across time, but relationship effects that are substantial in each time are not stable. This raises questions about relationship stability and reciprocities across time in humans. Another twist is in using archival data. For example, Malloy shows, using archival data from the 1960 UN General Assembly meeting that those world leaders with low actor score on approaching others and high partner score on being approached by others are those that have high relative power. This demonstrates how Partner score minus Actor score can have psychological meaning (power). Another extension is combining the power of round-robin design with block design into one study a Block Round Robin design that permitted estimation of SRM actor and partner effect estimates and their variance components, both within and between genders. For example, with this design Malloy demonstrates that there is more consensus among both males and females regarding male leadership emergence than regarding female leadership emergence. SRM offers a powerful method when intragroup and intergroup processes are of simultaneous interest.

Another twist is the use of target/perceiver (T/P) ratio to model developmental processes. For example, the ratio increases from grade 1 to grade 6 for some, but not all traits. Finally, SRM could be compared within and between groups, and when SRM is used across cultural groups and across groups within the culture, it can reveal subtleties. For example, one study compared consensus regarding traits among family members and friends. In Mexico, consensus about a person's traits was similar among family members and friends on 80% of the trait judgments. In China, there was no transfer of consensus among family members and friends. In short, when I thought I got the gist of the book, I learned many new things from this chapter.

Chapter 14

Offers a technical solution for bivariate SRM with missing data. In demonstrating the solution, it brings yet another interesting application of SRM to perception of truthfulness in a game played in class. Among the interesting findings, replicating past findings, it shows that people who appear most honest are most likely to perceive others as dishonest.

Final appreciation

By presenting model firsts and demonstration with data later, Malloy created a tension that made me feel as if I was reading a detective novel. I thank him for sharing with me much of his career toil. I hope that the reader will join me in this appreciation.

References

Cooperrider D.L., Srivastva S. Appreciative inquiry in organizational life. In: Pasmore W., Woodman R., eds. Research in organization change and development. Greenwich, CT: JAI Press; 1987:129–169.

Tagiuri R. Social preference and its perception. In: Tagiuri R., Petrullo L., eds. Person perception and interpersonal behavior. Stanford, CA: Stanford University Press; 1958:313–336.

Kluger A.N., DeNisi A. The effects of feedback interventions on performance: a historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin. 1996;119:254–284.

Mikulincer M., Shaver P.R. Attachment in adulthood : Structure, dynamics, and change. 2nd ed. New York: Guilford; 2016.

Nestler S. Restricted maximum likelihood estimation for parameters of the social relations model. Psychometrika. 2016;81(4):1098–1117. doi:10.1007/s11336-015-9474-9.

Van Dijk D., Kluger A.N. Task type as a moderator of positive/negative feedback effects on motivation and performance: a regulatory focus perspective. Journal of Organizational Behavior. 2011;32(8):1084–1105. doi:10.1002/job.725.

Preface

Thomas E. Malloy , The Mill House, Woodstock, CT, United States

Six decades have passed since Lee J. Cronbach first recognized the componential structure of interpersonal behavior, and nearly four decades of next generation social relations modeling of dyadic behavior has ensued. The seminal insight that the behavior of interacting people is a function of their individual psychology, an effect of their interaction partner, and unique effects occurring in specific dyadic arrangements is among the most important principles of 21st century social psychology. The social relations model (SRM) formalized these insights, and this volume is the first to consider componential modeling of a broad range of dyadic phenomena. This book has four basic goals. One is tutorial and aims to introduce the logic, research designs, and the formal structure of the SRM. A second is to provide examples of how the model can be applied to address novel theoretical questions in social and personality psychology. A third goal is to encourage second stage process modeling of phenomena following an initial social relations analysis, and illustrate how the SRM components become variables in multivariate, process models of dyadic behavior. A majority of the models proposed in this volume are theoretical specifications awaiting empirical assessment. Most of the dyadic process models specified assume that dyad members are distinguishable; that is, they can be distinguished on a variable that impacts the outcome of interest. A fourth goal is to introduce general algorithms that produce estimates of the parameters of the social relations model with both complete and incomplete round robin data. This estimation procedure solves the very thorny problem of missing data in round robin designs that can lead to substantial loss of information. These goals make this book relevant to scientists with no knowledge of the model, and equally relevant to those who have used the model, and wish to extend their focus to specialized topics and dyadic process modeling.

The material covered in the book reflects these basic goals. Chapter 1 considers the functional significance of group formation and dyadic relationships in the contexts where people lead their lives and establish fundamental human relationships. Chapters 2–4 address the logic of the SRM and the nuts and bolts of this endeavor; these chapters will be particularly useful for novitiates. Chapter 5 introduces a theoretical approach to individual differences called interpersonalism. That approach presumes that personality phenomena emerge primarily in the context of interacting people, although the logic of interpersonalism is elaborated with an empirical example of the social interaction of male and female mice. This demonstrates interpersonalism's generality. Chapter 6 addresses the psychophysics of trait judgment and illustrates a method for estimating the minimal differences in trait information necessary for the differentiation of two targets’ traits. In Chapter 7, self is conceptualized as a dyadic phenomenon that can change in different interactions, and provides a novel perspective on one of the oldest constructs in psychology. The determinants of interpersonal perception are considered in Chapter 8, and the Second Cronbach Critique of this area culled from personal communication with Professor Cronbach in 1996 is presented. A new process model called ARRMA is specified in Chapter 9 that integrates assumed reciprocity, reciprocity and metaperception accuracy in a single multivariate model at the individual level, and the minimal dyadic ARRMA with distinguishable members. Traditionally, these phenomena have been conceptualized and estimated independently. Chapters 10 and 11 introduce theoretical analyses of interpersonal attraction and perceived interpersonal similarity guided by the heuristic utility of the SRM. These chapters show that each is a set of phenomena that should not be reified as single constructs. Social vision is an emerging area of inquiry in social psychology, and Chapter 12 provides the first evidence documenting the componential structure of visual attention to social stimuli. A model of visual attention, categorization, and differentiation (ACD) is introduced and methods to test the model are elaborated. In addition, methodological prescriptions are offered to enhance the validity of research on face processing and memory. Social relations modeling of intragroup and intergroup phenomena are considered in Chapter 13. This chapter offers a new theoretical assessment of the meaning of responses to in-group and out-group members that emphasizes the categorization of individuals to social units (e.g., personally relevant or irrelevant), and the extent to which they are differentiated based on that placement. And finally, in Chapter 14 algorithms derived for the estimation of the parameters of the general social relations model with both complete and incomplete round robin data are presented. Derivation of the social relations model for arbitrary (i.e., incomplete) round robin structures (ARBSRM) offers a solution to the problem of missing data. This estimation procedure uses available information to produce parameter estimates, and departs from previous attempts to address this nontrivial problem using imputation. Monte Carlo simulations conducted by Charles F. Bond, Jr. addressing the method's ability to produce unbiased estimates of SRM variances and covariances under different patterns of missing data can be found at www.thomasemalloy.org. Also available is code (arbcodeR) and documentation written by Dr. Bond to accomplish the estimation within R. Because the topics addressed vary substantially, I anticipated that specific chapters would be of primary interest for different readers. For this reason, some key concepts are discussed in multiple chapters so that each can stand alone without necessarily consulting other chapters. My concern with clarity guided the writing more than my concern with redundancy.

Throughout this book the collective efforts of my colleagues and students are readily apparent. From 1979 to 1983, I was an assistant professor at New Mexico State University and was interested in dyadic behavior, but found limited analytic guidance. I collaborated on dyadic research with Stephen Clifford, Juan Franco, and Dolores Ludwig at NMSU. Eventually I discovered the SRM, and after correspondence with Dave Kenny about Kramer-Jacklin equations, I joined the social psychology graduate program at the University of Connecticut in 1984. The social relations model was very new and a group of collaborators were pursuing this new way of thinking about social behavior. Linda Albright, Dave Kenny, and I focused on interpersonal perception research among unacquainted and well-acquainted people in different cultural contexts. The National Science Foundation supported summer workshops we offered at the University of Connecticut to train scientists to use the SRM. While at UConn I also had the opportunity to collaborate with Reuben Baron, Jeffrey Fisher, William Fisher, Hilik Klar, Arie Nadler, and Stan Scarpati on projects centered in Storrs.

Most of my career has been at Rhode Island College, and I have been fortunate to have talented colleagues and students with whom I have collaborated; particularly, Fredric Agatstein, Beverly Goldfield, Robin Montvilo, and David Sugarman. I owe the greatest debt to my undergraduate and graduate students. Their curiosity and dogged hard work has been the foundation of research projects, and without them, many of the findings reported in this book would not exist. I direct the Social Relations Laboratory at Rhode Island College (thomasemalloy.org) and since 2008 the lab has been supported by funds from RI-INBRE (Rhode Island Institutional Development Award (IDeA) Network of Biomedical Research Excellence) and the National Institutes of Health. INBRE awards supported the establishment of an eye-tracking laboratory for social neuroscience research, and summer fellowships for students. Many of our findings are reported in Chapter 12, and I am very grateful for this support. I also want to thank the students with whom I have worked most closely at Rhode Island College and elsewhere: Elise Aruda, Suzy Barcelos, Avi Ben-Zeev, Rosalie Berrios-Candeleria, Gregg Bromgard, John Capman, Sathiarith Chau, Jason Dollard, Mike DeRosa, Brandon DeSimone, Carissa DiPietro, Jessica Hunter, Claire Janowski (Trinity College, CT), Lorin Kinney, Irina Kushid, Jennifer LaFountain, Johanna Martin, Scott Miller, Peter Murphy, Stephen Peters (Lincoln University, PA), Tiina Ristikari, Keri Silva, Angela Viphakone, Lyn Winquist, and Aaron Yarlas.

Two colleagues and friends deserve special mention for their contributions to this project. Charles F. Bond, Jr. of Nashville, Tennessee and Avraham N. Kluger of The Hebrew University of Jerusalem read every word of the manuscript and provided detailed comments. Their reactions shaped the content, and their honest critiques helped me avoid conceptual and statistical blunders. I own any that remain. I met Charlie in a causal modeling course taught by Dave Kenny in 1984 when he was an assistant professor at Connecticut College. Charlie is among the most talented quantitative social psychologists I have met, and we are authors of the only co-written chapter in the book that presents the general SRM. I first met Avi remotely and then face to face in Jerusalem in August 2016 when I offered a workshop on the SRM at The Hebrew University of Jerusalem. We initiated a program of research on the quality of listening in dyads and the consequences for interpersonal behavior. Avi's conceptual and statistical sophistication has motivated me to think in new ways about the application of the SRM. I am forever grateful to Avi and Charlie. I also want to thank Dave Kenny for his support and friendship since the summer day in 1984 when I arrived at his office in the Weston A. Bousfield Psychology Building while he was preparing for a softball game in a UConn summer league. As my advisor, Dave shaped the way I think about, and study, social behavior. Reuben Baron influenced my theoretical perspective on social behavior and his impact is apparent throughout this book. I also want to thank David Funder and Daniel Ozer for their comments on Chapters 5 and 6, respectively.

In May 2016, I received an email from Emily Ekle, Senior Acquisitions Editor for Psychology at Elsevier/Academic Press, suggesting that we meet in Chicago at the annual conference of the Association for Psychological Science. I was offering a methodology workshop on the SRM, and the plan was to discuss the possibility of a book focused on dyads. On January 1, 2018 the manuscript was delivered to Barbara Makinster, Senior Editorial Project Manager with Elsevier/Academic Press, who was helpful throughout the project and particularly patient as we explored cover art possibilities.

After signing the contract with Elsevier/Academic Press in July, 2016 this project was center stage in my life. My wife, Gina Malloy, tolerated my preoccupation and was always available, interested, and supportive. She understands the SRM and on many evenings patiently listened as I tried out new ideas for componential process models. I thank Gina for her love and support. Jeffrey, Stephen, Grace, Madeline, and Abigail diligently queried about progress on the book at family gatherings, and I appreciated their interest as they pursued their own academic and professional careers. And finally, I thank you, the reader. This writer's goal was to provide you with useful ideas, and to the extent that our minds have met, that goal has been realized. But beware; once the logic of social relations modeling becomes automatic, it is impossible to think about social behavior noncomponentially.

1

Dyads and groups

Abstract

Dyads and groups occupied the theoretical and empirical attention of social psychologists in the early and mid-20th century (e.g., Becker & Useem, 1942; Murphy, Murphy & Newcomb, 1937; Simmel, 1955), but as the cognitive revolution ensued (Miller, Galanter, & Pribram, 1960; Neisser, 1967) interest was redirected to social cognition. While that revolution freed psychology from a myopic focus on connections between observable stimuli and responses, many social psychologists abandoned the textured interactions of people and directed attention to social information processing. The theoretical and empirical yield of that revolution has been substantial (Fiske & Taylor, 1991) and paved the way for a recent focus on social neuroscience (Harmon-Jones & Inzlicht, 2016). Coupled with a waning interest in dyads was the inherent complexity of methods for dyadic research that dampened the enthusiasm of many investigators (Cronbach, 1955). The net effect was that topics once central to social psychology occupied small tranquil eddies on the edge of the social cognitive mainstream. In the post-Cronbach era, scientists interested in dyads faced an absence of guiding principles, adequate research designs, mathematically sufficient models, and software to ease the computational burden. That changed with the derivation of a random effect ANOVA model for round-robin data (Warner, Kenny & Stoto, 1979) and the specification of the social relations model (SRM; Kenny & La Voie, 1984). These elegant ideas were novel, intuitively appealing, empirically useful, and reignited interest in dyads. Advances in science are propelled by the development of methods for the observation of phenomena, and the SRM offered a glimpse into dyadic processes that had been invisible or shrouded in statistical confounds.

Keywords

Dyads and groups; Social context hypothesis; Social relations model; Dyadic research; Variance component analysis

Dyads and groups occupied the theoretical and empirical attention of social psychologists in the early and mid-20th century (e.g., Becker & Useem, 1942; Murphy, Murphy, & Newcomb, 1937; Simmel, 1955), but as the cognitive revolution ensued (Miller, Galanter, & Pribram, 1960; Neisser, 1967) interest was redirected to social cognition. While that revolution freed psychology from a myopic focus on connections between observable stimuli and responses, many social psychologists abandoned the textured interactions of people and directed attention to social information processing. The theoretical and empirical yield of that revolution has been substantial (Fiske & Taylor, 1991) and paved the way for a recent focus on social neuroscience (Harmon-Jones & Inzlicht, 2016). Coupled with a waning interest in dyads was the inherent complexity of methods for dyadic research that dampened the enthusiasm of many investigators (Cronbach, 1955). The net effect was that topics once central to social psychology occupied small tranquil eddies on the edge of the social cognitive mainstream. In the post-Cronbach era, scientists interested in dyads faced an absence of guiding principles, adequate research designs, mathematically sufficient models, and software to ease the computational burden. That changed with the derivation of a random effect ANOVA model for round-robin data (Warner, Kenny, & Stoto, 1979) and the specification of the social relations model (SRM; Kenny & La Voie, 1984). These elegant ideas were novel, intuitively appealing, empirically useful, and reignited interest in dyads. Advances in science are propelled by the development of methods for the observation of phenomena, and the SRM offered a glimpse into dyadic processes that had been invisible or shrouded in statistical confounds.

The focus of this book is on random effect variance component analysis (Searle, Casella, & McCulloch, 1992) of dyadic data using the SRM. This approach estimates dyadic phenomena primarily with variances and covariances rather than point estimates, such as the mean. A new way of thinking about estimation is required, as will be seen throughout this book. The goal is to introduce the logic of the SRM for studying dyadic and group phenomena, and to illustrate its application in research on classic and novel theoretical problems in social and personality psychology. A foundational assumption is that social behavior in the context of interacting people should be among the core problems addressed in contemporary social psychology. If one were to critique the social cognitive revolution, a primary concern would be the asocial nature of many paradigms. The SRM offers logic and methods that have revolutionized dyadic research making it inherently social, authentic, and interpersonally relevant while enhancing the validity and ecological scope of social psychological research (Albright & Malloy, 2000).

The nature of dyads

One of the first comprehensive analyses of dyads and their specific functions was provided by Becker and Useem (1942). Dyads were organized in two superordinate categories: "comprehensive dyads and segmentalized dyads." Within each were dyads that serve different social functions. Comprehensive dyads included friendships, sexual, and intergenerational pairs; whereas segmentalized dyads included aider-aided (e.g., therapist-client), teacher-pupil, status pairs (e.g., employer-employee), common interest (e.g., hobbyists), and patterned contact (e.g., two commuters). Respectively, they represent dyads with emotional and formal relationships.

A dyad is two people whose behavior is dependent because of the function of their interaction. Two randomly selected people may be treated as a dyad in statistical analysis (i.e., pseudo-dyads) as a control for real dyads, but they are not dyads psychologically. Dyads originate in biological mandates (e.g., parent-child), interpersonal