Interpersonal Perception: The Foundationof Social Relationships, 2nd Edition (David A. Kenny)
David A. Kenny
March 12, 2020
Book logo
Interpersonal Perception
The Foundation of Social Relationships
Second Edition

Go to a general Description of the book.
Go to a the Table of Contents of the book.
Order a copy the book!
To view a sample chapter, Chapter 1, click here.
Critical acclaim.

Description

People make judgments about others all the time, often without realizing they are doing so. How are interpersonal impressions formed? How accurate are our perceptions of other people's traits and our own? In this major revision of my often-cited work, I provide a reader-friendly examination of these and other critical questions, identifying key components that shape impressions and their accuracy. Topics include how to estimate perceiver, target, and relationship effects; the extent to which different perceivers see a target in the same way; the impact of group membership and stereotypes; and whether others see us as we see ourselves. Implications for interpersonal relationships and social behavior are highlighted.

This second edition is a new book with hardly any of the text of the first edition included. This edition incorporates 25 years of theoretical, empirical, and methodological advances. Several new and expanded topics, including first impressions, individual differences in accuracy, implicit measures, and narcissism. The book is grounded in a reformulated conceptual model, the 2004 PERSON model. Moreover, I use accessible non-technical language, humor, popular culture, and simplified figures to elucidate complex ideas. Each chapter has "Practical Suggestions" apply the science to real-world social situations and "Philosophical Implications."

Table of Contents, Summary, and Miscellanea

Preface

Chapter 1. The Atomic Structure of Interpersonal Perception

The three SRM components of perceiver, target, and relationship are introduced as the basic elements of interpersonal perception, much like the atom's protons, neutrons, and electrons. Those elements are the perceiver effect (how the perceiver generally views targets), the target effect (how the target is viewed generally by perceiver), and the relationship effect (how the perceiver uniquely views the target. The book examines not only the perception of others but also self-perceptions and metaperceptions or the perception of a target's perceptions. These perceptions are primarily trait perceptions that are organized using the Big Five or Big Two (Dominance and Friendliness), but attraction is also discussed. The context of perception matters, especially the acquaintance between the perceiver and the target. The chapter outlines the nine basic questions of interpersonal perception that are considered in the remainder of the book.

The chapter notes the important contribution of Timothy Leary on the number of underlying traits (1 to 32) and discusses the personalities of Marge and Homer Simpson. To view a sample chapter, Chapter 1, click here.

Chapter 2: They All Look Alike, Assimilation

Assimilation is the tendency to a perceiver to see others as similar, and can be measured by the proportion of total variance due to the perceiver effect. On average, assimilation explains about a quarter of the total variance for trait judgments, but as the chapter describes the level of assimilation varies considerably. For attraction, the level of assimilation is generally lower than it is for traits. For trait judgments, assimilation declines with increasing familiarity, which supports the view that the perceiver effect represents the perceiver's best guess of what targets are like, what is called a personal stereotype. Also consistent with the view that the perceiver effect is a best guess is the finding that assimilation is greater for less observable or visible traits, especially when they are highly evaluative. The largest part of the perceiver effect represents positivity, which is the tendency for some perceivers to see others generally positively on all traits and for other perceivers to see others generally negatively. A smaller part of assimilation is due to acquiescence or the tendency to endorse both positive and negative traits. Generally, perceivers see different sets of targets, for example ingroup and outgroup members, on traits in a similar fashion. However, there are some differences between the average perceptions of members from different groups, which weakens the argument that there is a single view of the other, i.e., the generalized other. Attraction judgments of ingroups and outgroups show weaker levels of consistency than do trait judgments. The perceivers who see others positively are women, the securely attached, and non-narcissists, but none of these effects is very large. Generally, those with the larger perceiver effect enjoy slightly better mental health.

Two of the key papers discussed in this chapter were published in 2010, one by Sanjay Srivastava, Steve Guglielmo, and Jennifer Beer, and another by Dustin Wood, Peter Harms, and Simine Vazire. One finding discussed in this chapter is that how people see other people is how they see dogs, research conducted by Virginia Kwan, Samuel Gosling, and Oliver John. Differences between perceiver effects should be used more often to study prejudice and discrimination towards groups.

Chapter 3. Perception at First Sight

Research dating back to 1924, has shown that level of agreement or consensus is about .20, i.e., 20 percent of the total variance is target variance. Alternatively, the .20 value can be viewed as the correlation between perceivers' judgments of two different targets. Much of the work on consensus in first impressions is based on the "zero-acquaintance" paradigm, pioneered by Linda Albright, Thomas Malloy, and myself, which was based on earlier work by Norman and Goldberg. Weaker, though non-zero, consensus is found for the judgment of faces. Consensus is greatest for Extraversion, followed by Conscientiousness. Viewing people's bedrooms or offices also shows evidence for consensus, though consensus is greater for Openness, followed by Conscientiousness.

Evidence also supports the view that first impressions are surprisingly accurate, with an estimate an accuracy value of .35 or the average judgment a .16 value for an individual's judgment. In large part, the acceptance of this conclusion is due to Nalini Ambady and Robert Rosenthal's pioneering work on thin slices. It is important to realize that first impressions are not 100% accurate. In fact, a .16 correlation for individual implies that roughly 42% of the time the judgment would be mistaken.

The chapter examined the question of whether a first impression is a lasting impression, and the evidence is partially consistent with the naive view that indeed "a first impression is a lasting impression." Interestingly, the consensual piece appears to be relatively permanent, but the relational part is more mutable.

The research reported in this chapter should lead readers to reappraise their view of stereotypes. Agreement and accuracy of perceivers' first impressions is largely based on stereotypes. Evidently, there is a kernel of truth due stereotypes, a topic extensively revisited in Chapter 5.

The chapter considers the claim by several researcher that first impressions are inaccurate. While I reject these arguments, I nonetheless present them so that the reader can make his or her own determination. It is interesting to note that the initial title of Jane Austin's book Pride and Prejudice was First Impressions.

Chapter 4: The Content of Our Apparent Character, Consensus

Consensus represents the agreement between perceivers in their judgment of a common target. It is indexed by the proportion of the total variance due to the target effect and can alternatively be viewed as the correlation between different perceivers in judging a common target. The overall level of consensus for long-term acquaintances is around .39, as measured by the ratio of target variance to the total variance. Studies that measured consensus when perceivers were relatively unacquainted with the target show consensus, about .23, not nearly the same level as with long-term acquaintance.

Target effects are highly consistent over time. In fact, moving beyond zero acquaintance, the finding is exceptionally large level of stability of the target effect. There is also a fair amount of consistency across different target groups, with a reasonable estimate for that consistency correlation being .70. Different groups of perceivers see the target similarly but not exactly the same. Lastly, in terms of the Big Two, Dominance and Friendliness are relatively uncorrelated. In terms of the Big Five, there appears to be some weak correlations.

In terms of moderators, there is evidence for greater consensus for Extraversion followed by Openness, with the other three Big Five factors trailing. Both visibility and evaluativeness moderate consensus, with visibility having a positive effect on consensus and evaluativeness a negative effect, and the effect of visibility is larger than the effect of evaluativeness. Positive and negative traits do not appear to differ in terms of consensus. Outcome dependency is also an important factor and may be implicated in cultural differences in consensus. Finally, it is found that there is usually more consensus in judging ingroup members than outgroup members. Interestingly and surprisingly, across 15 longitudinal studies and 64 measures, consensus does not increase with increased acquaintance. This result is in contradiction to most of the cross-sectional studies, reviewed above, that show an increased consensus as a function of acquaintance. The best resolution is that there is a large increase in consensus as a function of initial acquaintance, but once there is some acquaintance, the rate of increase is very low.

The PERSON model presumes that a perceiver's impression of a target is a function of the perception of the target's appearance and the appraisal of the target's actions. (The PERSON model is very complicated, though the details of those complications are presented in Appendix C and so many readers may wish to skip reading those complicated details.) It is a complex model with six different variables that create the acronym PERSON and parameters for the number of acts perceived, the overlap of the behavioral information perceived by two perceivers, two kernel-of-truth parameters, and the communication between perceivers. The PERSON model can explain some of the results in this chapter. It does predict an initial large increase in consensus as a function of increasing acquaintance and then a slow, gradual increase. It also predicts the surprisingly small effect of overlap on consensus, as well as the incredibly large levels of stability of the target effect over time. The model needs to be altered to explain the less-than-perfect cross-context consensus.

The chapter relies heavily on the extensive meta-analysis conducted by Brian Connelly and Deniz Ones in 2010. I suggest that there might exist a Mercurialness factor in the target effect. That is, someone might exhibit highly agreeable traits (e.g., warm) as well as highly disagreeable traits (e.g., cold).


Chapter 5: You Are Very Unique, Uniqueness and Reciprocity

This chapter shows that the relationship effect dominates perceptions of others' personalities, as well as how much people like one another. For trait ratings of the Big Five, about 40% of variance is due to relationship effects, and for interpersonal attraction, that percentage rises to about 60%. Person perception is fundamentally idiosyncratic, a phenomenon that is not well appreciated by the research or practitioner communities. Although not the primary focus of this chapter, results for assimilation and consensus in perceptions for both trait ratings, 26 studies, and attraction studies, 18 studies, are also reviewed. For trait ratings at zero acquaintance, there is 33% assimilation, whereas well-acquainted people show only a 22% assimilation. Consensus appears to increase with acquaintance, starting at 20% at zero acquaintance and ending up at 39% for long-term acquaintance. For interpersonal attraction, consensus accounts for about 20% of the variance, regardless of the level of acquaintance. However, for romantic attraction, the variance partitioning is somewhat different than it is for nonromantic attraction in that there is about 75% more target variance. It appears that the reason for this increase in consensus is the importance of physical attractiveness in determining initial romantic attraction. Assimilation is strongest at brief acquaintance, accounting for about one-third of the total variance. Likely, the result is reflecting how much the person likes being a member of the group.

Attraction and trait judgments are highly related for all three SRM components. However, attraction does not explain all of the variation in trait judgments. In fact, there is good evidence that all three SRM components for traits are multidimensional. Consequently, it is impossible for one factor, attraction, to explain all of their variance. The question about causal direction remains uncertain, and it is not known whether attraction leads to trait judgments or vice versa.

Within the SRM, there are two different reciprocity correlations: One is the correlation between perceiver and target effects, called generalized reciprocity. The other is the correlation between relationship effects, called dyadic reciprocity. Little or no reciprocity for traits was found. However, there is some evidence of positive generalized reciprocity for Friendliness. For attraction, especially when people are well acquainted, there is a great deal of dyadic reciprocity, with the average correlation being around .52. Romantic attraction shows negative generalized reciprocity (people desired by others are very choosy), and that effect appears to be due in part to the effects of physical attractiveness (attractive people are both desirable and selective). Given the strong dyadic reciprocity for attraction, it seems unlikely that liking is a strong cause of trait judgments, as trait judgments do not show dyadic reciprocity.

Understanding the meaning of uniqueness and the relationship effect is a much more difficult task than describing it. The most commonly accepted theoretical approach to uniqueness is matching -- characteristics of the perceiver and target are matched -- but operationally it is difficult to test. Other approaches should be considered: person models proposed by Bernadette Park and agent-based models. The view that relationship effects are emergent within the dynamics of interpersonal interactions also has possibilities, but a strategy for testing the model needs to be developed.

It is suggested that the low over-time stability of the relationship effect for attraction may be part of the reason why societies make it difficult for married couples to divorce.

Chapter 6: Peeling Away the Onion, Target Accuracy

Determining whether trait perceptions are valid is not so each a question to answer. One issue is what can be taken as the truth. Typically, self-judgments, knowledgeable informants, and behavioral measures are used to measure the truth. Each has advantages and disadvantages. The second issue is that there is not a single accuracy, but rather many. Based a framework the I initially developed with Linda Albright, the focus is on four different SRM measures or accuracy: generalized accuracy (If a target is generally seen by perceivers in a certain way, does the target act that way?), perceiver accuracy (If a perceiver generally sees targets in a certain way, do those targets act in that way with the perceiver?), dyadic accuracy (If a perceiver sees a target in a particular, does that target act that way with that perceiver?), and constant bias(Do people over or estimate a particular behavior in others?). It is noted that generalized and dyadic accuracy approximately correspond to William Swann's global and circumscribed accuracy.

In Chapter 3, the accuracy of an individual's first impressions was estimated to be .16. What is the level of accuracy when perceivers know the target well? Using the 1994 fraternity study and the 2007 aggression study, the average generalized accuracy correlation for individual is .32. Converting these correlations to rough probabilities, the .16 zero-acquaintance accuracy correlation implies being right about 58% of the time, whereas a correlation of .32 between well-acquainted individuals implies about a 66% chance of being correct. Thus, about half of what perceivers would ultimately know about targets is learned in the first few minutes, if not seconds. Moreover, even when perceivers are well acquainted with the target, they are still wrong about one time out of three.

The implications of the PERSON model for target accuracy are explored. That model predicts that most of the increase in accuracy happens very early on in the acquaintance process. Moreover, the presence of a kernel of truth can lead to interesting relationships of acquaintance with accuracy. A kernel of truth tends to flatten the relationship whereas a "grain of falsehood" results an initial negative accuracy that eventually becomes positive.

Arie Kruglanski quite incisively observed that in measuring the truth in person perception, the measure of the truth is itself a person perception. He went on to argue that all measures of accuracy are actually a form of consensus.

Chapter 7: Mirror, Mirror, on the Wall: Self-Other Agreement and Assumed Similarity

Self-other agreement is often used in accuracy research. A person does see him- or herself in the same way as others, self-other agreement. However, the level of self-other agreement is lower than consensus; that is, two peers agree more than that self does the self with one peer. Self-other agreement increases with increased acquaintance, visibility of the trait, and likely the evaluativeness of the trait.

Assumed similarity has a long history in psychology and perhaps most familiar to readers as the false consensus effect. Perceivers assume that others are similar to themselves, assumed similarity. However, the level of assumed similarity is lower than assimilation; that is a person sees two peers as more similar than does the self views him or herself as similar to a peer. Assumed similarity is lowered by visibility and increased by ingroup membership.

Generally, overall most people see themselves as better than they see others. However, this effect is weakened and even reversed if self and other are close. Following the analysis of Virginia Kwan, self-enhancement is viewed the self-perception minus how the person sees others (the perceiver effect) and minus how others see the person (the target effect). This measure of self-enhancement can be interpreted as the relationship effect for self-perceptions. An extension of the Kwan approach, the "k and q" method is presented. In terms of consistency, self-relationship effects are highly differentiated across the Big Five. If a person self-enhances on one Big Five trait, he or she may not do so on another trait. However, there is strong consistency over time in self-enhancement. There is some evidence that relational self-perceptions are more varied than the relational impressions of others. There is also some evidence that the perceptions of the self vary across interaction partners but no consistent evidence that certain persons elicit a particular self-perception.

Are others more accurate perceivers than the self? The evidence is somewhat mixed, but certainly there are cases in which self-perceptions are less accurate than the perceptions of others. One advantage of other-perceptions is that they can be aggregated across perceivers to increase reliability.

Self-perceptions are integrated into the PERSON model and its implications for self-other agreement, accuracy, and communication are discussed. Finally, the empirical results are related to Samuel McAbee and Brian Connelly's Reputation-Trait-Identity model and Simine Vazire's Self-Other Knowledge Asymmetry model.

The chapter discusses the work of a Robins, a Robbins, and a Robyn. I also pose the question: Who sees himself as more humble: Donald Trump or Pope Francis?

Chapter 8: Putting Yourself in Someone Else's Tevas, Assumed Reciprocity and Meta-Accuracy

People believe that they make very consistent impressions on others. Some think they make very good impressions, whereas others think they make very poor impressions. This consistency in metaperceptions is stronger for traits than for attraction, and it is somewhat greater earlier in the acquaintance process. Although evidence for perceiver variance is overwhelming, there is little evidence for variance in the target effect: People do not agree about which targets are lenient judges of others and which are harsh. People do recognize that others see them differently, but this recognition is stronger for attraction than for traits and stronger for long-term than for short-term interactions. The consistency of metaperception is examined in various ways.

First, there is considerable stability of the perceiver effect was found, although, as with impressions, the consistency is lower at the level of relationship. Second, there is limited research on the consistency of measures, but the little evidence supports consistency. Third, the one study that examined the consistency of metaperceptions across ingroups and outgroups does show some consistency. The perceiver effect has substantial positive correlations with a constellation of variables, the primary one being social anxiety. Additionally, people think that others see them very much the same way that they see themselves.

Are these metaperceptions correct? Alternatively, is it the case that metaperceivers have no clue as to how they are seen? The inclination for metaperceivers to think others see them favorably (or unfavorably) turns out to be generally accurate, but these tendencies are exaggerated: Metaperceivers who think that they make good impressions are right in that targets do see them favorably, but not nearly as favorably as they think. Conversely, people who perceive others as not viewing them positively are also correct; however, they are not thought of as negatively as they think. How is such accuracy achieved? The evidence for traits generally supports the view that accuracy is achieved by inferring others' perceptions by using one's self-concept. When given the opportunity, metaperceivers also observe their own behavior and use it to infer target perceptions. People do try to read the other person's behavior, but it appears that people are not very successful at this task. Metaperception enhancement, or the extent to which metaperceivers think that others see them more favorably than others actually do, is examined. In general, the finding is for relatively weak levels of enhancement, certainly weaker levels than for self-enhancement.

People metaperceive not only how targets see them, but also how targets see others, so s that much of the variance is relational and that there is both accuracy and bias in these perceptions. Metaperceivers guess the perceptions of individual targets' perceptions of not only individuals but also social groups, something called meta-stereotypes. The evidence suggests what is called pessimism: People believe that outgroup members see their group more negatively than they actually do. The concept of metaperception can be broadened to include not only the perceptions of perception but also the perceptions of expectations or meta-expectations.

Much of the content of the chapter is based on my prior collaborations with Bella DePaulo and Erika Carlson.


Chapter 9: Good, Bad, and Ugly Judge, Target, and Pair

The chapters considers the extent to which there are individual differences between perceivers in the ability to judge targets' personalities, to recognize emotion in targets, and to determine whether the target is lying or not. Besides individual differences between perceivers, also considered are individual difference between targets and relationships. Generally, individual differences between perceivers are relatively small, especially for judging personality and detecting lies. For judging emotion, the individual differences are considerably larger, but still not very large, being much smaller than the individual differences found in most other areas of psychology. Consistently, target differences in accuracy are stronger than perceiver differences. This result was found for all three types of individual differences investigated in this chapter. At the very least, given the large differences between targets, to maximize power, a study needs to include more targets than perceivers. Relationship or dyad variance has not been a topic that has attracted much interest but is shown to be present in judgment of emotion.

The chapter features Jeremy Biesanz's Social Accuracy Model, or SAM, which provides the most complete analysis of individual differences in interpersonal accuracy. Moreover, besides measuring accuracy, SAM can offer insights into how perceptions of others are biased. SAM is complicated, and it can be difficult to conduct the analysis and to understand the meaning of the analysis. Likely, SAM or some variant of SAM will be increasingly used by researchers who are interested in the accuracy of social perception.

The chapter includes a tribute to Robert Rosenthal, who pioneered research in individual differences in the accuracy in the perception of emotions. It also extensively relies on a meta-analysis conducted by Katja Schlegel, Thomas Boone, and Judith Hall, another meta-analysis conducted by Charles Bond and Bella DePaulo, and several studies conducted by Hillary Anger Elfenbein.

Chapter 10: Finis

The major point of this book is that the study of interpersonal perception requires a partitioning of perceptions into perceiver, target, and relationship components. Such a partitioning reveals the fundamentally different processes operating for each of the components. For these components, three different types of consistency can be examined. The book provides answers to basic questions of interpersonal perception, and often those answers are at different levels. The PERSON model can be used to integrate many of these disparate findings. Considered first is an additional term in the SRM: the group effect. Then I broaden the discussion beyond personality judgments and interpersonal attraction to consider emotions, leadership, trust, and behavior, both naughty and nice.

A new typology of the correspondence between dyadic measures, called the Revised Laing Typology, is proposed. This typology considers two types of perceptions: perceptions of the other and perceptions of objects. There are nine measures of correspondence for perception of the other and four for perceptions of the other. A common vocabulary for measures of dyadic correspondence should advance the field of interpersonal perception.

The chapter and the book concludes with a brief history of interpersonal perception research. Two key developments are the interest of personality reseachers in the topic and intense interest of European researchers, most notatbly the PERSOC group and family researchers. Europeans have been especially creative in impoving the estimation of the SRM. Moreover, individual reseachers have made key contributions, including Hillary Anger Elfenbein, David Marcus, Paul Eastwick, Eli Finkel, and Brian Connelly. Interpersonal perception researchers are in most areas of psychology - social, personality, clinical, developmental, and industrial-organizational, human development and family studies, communications, and business have made important discoveries. The future looks very promising and exciting.

The material on the Revised Laing Typology was co-written with Tessa West.

Glossary

Definitions of the major terms are given here.

Appendix A: Details of the Social Relations Model

The appendix begins with a formal description of the model. The estimates of the Social Relations Model (SRM) effects and of the SRM variances and correlations are presented. Described also is how correlations between two different variables can be computed. Also discussed how to measure the simultaneous effect of two or more variables on an outcome. The section that follows contains a description of methods of estimating the SRM. The final section contains details about the Social Accuracy Model (SAM), which is an extension of the SRM to estimate individual differences in judgmental accuracy.

The appendix features the ongoing contributions of Steffen Nestler, Oliver Ludtke, and many others.

Appendix B: Details on the Reported Results

For each of the hundreds of numeric results presented in the text, the source from which the result was from is described.

Appendix C: Details of the PERSON Model

The appendix presents the new standard values that are used in this book, as well as their empirical and theoretical justification. The basic PERSON formulas for consensus and accuracy are presented for both, as well as formulas for how consensus and accuracy are affected by the number of acts, as well as the kernel of truth. It is shown how the consensus and accuracy values are altered as a function of averaging over multiple perceivers (i.e., aggregation) and communication. The effects of stability or consistency over time are presented. In the final section, the assumptions of the model are critically examined.

References

Ordering Information

Online

To order a copy of this book online. click here .

Mailing Address

Guilford Publications
370 Seventh Avenue
Suite 1200
New York, NY USA 10001-1020

Telephone: +1 (800) 365-7006

Email Address

info@guilford.com

Sample Chapter

To view a sample chapter, Chapter 1, click here.

Critical Acclaim

Eli Finkel

Kenny - a towering genius and social/personality psychology's true north - has written a book for the ages. In engaging and nontechnical prose, this book orients readers to the foundational elements of social interaction, offering deep insight into topics like consensus, accuracy, stereotyping, reciprocity, metaperception, self-perception, self-enhancement, and group dynamics. This book is essential reading for anybody seeking to understand human sociality.

Tessa West

With the methods revolution, this book provides a 'must read' of the most solid and replicable social scientific research on person perception. Very few authors could weave a narrative of how decades of person perception research unfolded in the social sciences as astutely-and entertainingly-as Dave Kenny.

William Swann

This volume provides a comprehensive, up-to-date overview of person perception through the lens of Kenny's Social Relations Model-arguably the field's most innovative and creative tool for gaining insights into how people view themselves and others. The years since the first edition have given the author far more material to work with, and he has been generous in sharing his wisdom. Readers will delight that Kenny has made the content much more accessible. This volume is an indispensable reference for all social scientists interested in person perception, and will also make an excellent graduate-level textbook.

David Funder (from the forward)

Readers of this book are in for a rare treat. The book usefully updates and expands Kenny's landmark exposition published in 1994 (a mere 25 years ago), but it is much more than a mere "second edition." The topics are expanded to include, for just one important example, a thorough and even profound exposition of self-perception, as well as fascinating philosophical musings and useful applications at the end of every chapter. Thanks to David Kenny for doing the work, providing the insight, and organizing the field of interpersonal judgment at this crucial point in its history, when the field can be considered mature but is also confronting a number of possible directions as to where to go next.

Return to the top of this page.
Book logo

Go back to the interpersonal perception gateway page.

Go back to homepage.