Racial Pain Beyond Black and White

Identity Maintenance Assignment

Background

Identity Maintenance pertains to the activities—cognitive, affective, actional—that help us to maintain the sense of identity stability: What makes me feel “myself” yesterday, today, and tomorrow.  Of course, such activity is made necessary because, despite ideas of a stable, unchanging identity, we do change—and often!

In class, we discussed the role of selective perception—seeing somethings but not others—in this identity maintenance project.  Two assigned articles discuss the interplay under focus:

Shively, Joellen. 1992. “Perceptions of Western Film Among American Indians and Anglos.” American Sociological Review 57:725-734.

Robinson, Dawn T. and Smith-Lovin, Lynn.  1992 (March).  “Selective Interaction as a        Strategy for Identity Maintenance: An Affect Control Model.”  Social Psychology Quarterly, 55(1):12-28.

The Task

Read these two essay, then read Chapter 2 in America’s Atonement, on Racial Pain in the 21^st Century.  Focus on the section “Racial Pain Beyond Black and White,” pages 42-48.

Drawing on the essays, discuss the various relational conflicts illustrated in these pages; that is, the Jewish-Anglo episode in Dr. Phil’s excerpt, etc.

Selective Interaction as a Strategy for Identity Maintenance: An Affect Control Model
Author(s): Dawn T. Robinson and Lynn Smith-Lovin
Source: Social Psychology Quarterly , Mar., 1992, Vol. 55, No. 1 (Mar., 1992), pp. 12-28
Published by: American Sociological Association
Stable URL: https://www.jstor.org/stable/2786683
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

Terms and Conditions of Use

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend
access to Social Psychology Quarterly
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
Social Psychology Quarterly 1992, Vol. 55, No. 1, 12-28 Selective Interaction as a Strategy for Identity Maintenance: An Affect Control Model* DAWN T. ROBINSON Stanford University LYNN SMITH-LOVIN University of Arizona In this paper we answer the question: Do people select interaction partners to enhance their self-image, or do they strive to maintain a stable view of self? Affect control theory, a quantified version of symbolic interactionism, predicts that individuals adopt strategies which maintain their identities in order to secure a stable definition of the situation. When individuals have low self-esteem, they select interactions that maintain this low self-esteem, even when these interactions cause negative emotions. Two experiments examined the cognitive, affective, and behavioral effects of identity-confirming and -disconfirming information. Study 1, which assessed the cognitive and emotional outcomes of identity-relevant feedback, revealed that people with both high and low self-esteem feel good when praised and bad when they receive negative feedback on their performance. People with low self-esteem, however, think that the criticism is accurate, and like the critic more than do people with high self-esteem. Study 2 examined the implications of these responses for social interaction. Consistent with affect control predictions, subjects selected interaction partners who provided identity-consistent feedback. People with low self-esteem preferred to interact with a critic rather than with someone who said flattering things about their performance, even though the negative feedback made them feel bad. We discuss our results in terms of their implications for role stability and for the behavior of people with chronically negative self-identities. Suppose you have just completed a presen- tation, and two people expressed opinions about your performance. One was a harsh critic, who described you as nervous, uncom- fortable, awkward, and a poor speaker. The other, who was quite appreciative, said that you were a poised and graceful speaker who communicated effectively and was at ease in social situations. Whom would you choose to talk with for the next half hour? This paper asks whether people select their interaction partners to enhance their self- image or to maintain a consistent, stable view of themselves. Affect control theory, based on symbolic interactionism principles, makes the counterintuitive prediction that some people indeed will choose to talk to the harsh critic as an interaction partner. If a person has * The authors thank Susan Dessert, Randi Freidman, and Kelly Hyatt Wrolsen for assistance in collecting the data. Ronald L. Breiger, David R. Heise, Alice M. Isen, Miller McPherson, and Johnmarshall Reeve provided valuable comments on earlier versions of the manuscript. Some of these data were presented at the 1989 meeting of the Society for the Study of Symbolic Interactionism, held in Tampa. This project was supported by National Science Foundation Grant SES 9001078 to Dawn T. Robinson and National Science Foundation Grant SES 9008951 to Lynn Smith-Lovin. a fundamentally negative self-image in regard to public speaking, interacting with such a critic will support that self-image. Even though negative information is emotionally distressing, it is consistent with prior self- conceptions and thus confirms one’s identity. We argue that the self-concept guides choices among interaction partners in ways that act to confirm and maintain self-image. In people with high self-esteem, this choice results in positive, self-enhancing interactions; in peo- ple with low self-esteem, it may result in emotionally distressing, negative interactions that maintain a sense of failure. THE SELF CONCEPT AS A SOCIAL CONSTRUCTION The self-concept has been hailed as the ultimate intersection between the individual and society (Rosenberg 1988). According to Rosenberg, the capacity to abide by social rules and to play social roles depends heavily on what he calls self-objectification, the ability to “involve the self as object to itself” (Rosenberg 1988, p. 549). By providing a sense of both differentiation and continuity across time, self-identities also help us to 12
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 13 make choices, provide us with goals, and give us the strength and resilience to pursue them (Baumeister 1986). By accomplishing these functions, self-identities make possible our relationships with other people. At the same time, relationships with other people make possible the formation of identities. Historically, sociologists have stressed the process of coming to know one’s self through the observation of one’s own interactions with others (Mead 1934). The ability to view one’s own actions in relation to other people makes the self inherently social. This paper focuses on the sociological importance of stable self-identities in shaping social interaction. We believe that the impact of social interactions on one’s self-image transforms choices among interaction partners into powerful tools for identity negotiation.1 Many theories embrace the notion that human beings rely on information obtained through social interaction to formulate concepts of self, while the self simultaneously provides the motivation and the agency to shape social interaction. Examples of such theories are symbolic interactionism (Hewitt 1979; Mead 1934; Stryker 1980), cognitive dissonance theory (Festinger 1957), self-consistency the- ory (Backman and Secord 1962; Lecky, 1945; Swann 1983, 1985, 1987), social identity theory (Tajfel 1981), self-perception theory (Bem 1967), self-esteem theory (Jones 1973; Tesser 1988), and affect control theory (Heise 1977, 1979; Smith-Lovin and Heise 1988). Although each of these theories acknowledges the importance of self-perception in determin- ing social behavior, they debate the underly- ing motivation for this process. These per- spectives can be grouped roughly into two motivational orientations: those which pro- pose that people seek and use self-information to enhance their self-images (self-enhance- ment) and those which propose that people seek and use self-information in order to maintain consistent self-views (identity main- tenance). 1 This is not to say that we view selection of interaction partners as an entirely purposive enterprise. We assert, however, that there is some voluntary element, and we choose to explore this element here. SELF-ENHANCEMENT VERSUS IDENTITY MAINTENANCE Self-Enhancement A prevalent view of the self’s role in motivating social behaviors and perceptions is that people perceive, seek out, and create events in such a way as to enhance their views of themselves with respect to their social worlds. Self-enhancement theories are predi- cated on the belief that people are motivated to think of themselves “as favorably as they can get away with” (Smith 1968, p. 368). This desire is thought to predominate over the aspiration for objective self-knowledge, and has been labeled a “law of human develop- ment” (Becker 1968) and a “master senti- ment” (McDougall 1932). An abundance of empirical literature sug- gests that most people have overwhelmingly positive views of self and that they adopt attributional and recall strategies which pro- mote such self-views (see review in Taylor and Brown 1988). For instance, individuals tend to remember past performances as more successful than they were (Langer and Roth 1975), to recall successes more clearly than failures (Silverman 1964), and to overesti- mate their abilities at even obviously random tasks such as predicting the outcome of a coin toss (Langer and Roth 1975). Who should exhibit self-serving biases? Theoretically this self-enhancement tendency applies to those with both positive and negative self-images. Tesser and Campbell (1983) proposed a theory of self-evaluation maintenance (SEM) in which individuals are motivated to maintain positive self-evalua- tions regardless of self-esteem. In fact, Tesser and Campbell state that “persons low in self-esteem tend to engage in these [SEM] behaviors more than persons high in self- esteem” (1983, p. 10). Taylor and Brown (1988), however, observe that persons who are low in self-esteem or moderately de- pressed may be immune to the self-serving bias. These authors interpret this immunity as consistent with the maladjusted state of persons in these groups. Identity Maintenance Another interpretation of the finding that people with low self-esteem do not demon- strate self-enhancing behaviors is that these people are acting to maintain a consistent (and
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
14 SOCIAL PSYCHOLOGY QUARTERLY negative) view of themselves. An early view of the self-consistency position is contained in symbolic interactionist theory. Symbolic in- teractionism (Cooley 1964; Mead 1934) is based on the proposition that people act towards objects and events on the basis of the meanings those objects and events hold for them. Once an object becomes a symbolic representation of meaning for a person, it becomes important to maintain that meaning in order to sustain a coherent, cohesive view of the world. Affect control theory derives from sym- bolic interactionism (Heise 1977, 1979; Smith-Lovin and Heise 1988); it models mathematically the way individuals create events in order to maintain fundamental meanings. Like symbolic interactionism, af- fect control theory focuses on the individual’s definition of the situation, which includes the perception of self, others, behaviors, and events as symbols with meanings. Within this perspective, events can temporarily alter meanings ascribed to self or others to create new, transient impressions of those persons. The extent to which these transient impres- sions differ from the original fundamental meanings is called a deflection. When events deflect these transient impressions away from the fundamental meanings we attach to people’s identities and behaviors, we must generate new events to restore these mean- ings.2 In other words, the theory predicts that people construct events to confirm fundamen- tal meanings regarding self and others; they respond to social life in ways that control their feelings about reality. Parallel developments in psychological theory are consistent with the emphasis on identity maintenance in affect control theory. These psychological approaches postulate a basic motivation for cognitive and perceptual consistency: Classic formulations include Festinger’s (1957) theory of cognitive disso- nance and Heider’s (1958) balance theory. These theories contend that people attempt to maintain consistent beliefs and attitudes; they suggest that inconsistency actually produces a state of discomfort or dissonance. Psycholo- gists who have applied consistency theory directly to the concept of the self include Lecky (1945), Backman and Secord (1962), 2 Heise (1979, 1986) uses the term “sentiments” to refer to fundamental meanings on the evaluation, potency, and activity dimensions. Aronson and Carlsmith (1962), Shrauger (1975) and Swann (1983, 1985, 1987). Swann in particular has articulated the basic principles of affect control for psychological research questions, showing how people seek consistency with respect to self-knowledge and self-beliefs. All of these lines of sociological and psychological thought have in common the tenet that people perceive, seek out, and create events to sustain the meanings which they hold about themselves and their social worlds. According to these perspectives, this self-knowledge provides people with the necessary ability to interpret, predict, and control their life circumstances. Identity Negotiation An underlying premise of both the self- enhancement and the identity maintenance perspective is the assumption that people actively involve themselves in a process of identity negotiation. Numerous strategies for negotiation of self have been suggested by researchers within the two perspectives. Among the strategies proposed are biased perceptions of self and others (e.g. Klein 1989; Weinstein 1980), manipulation of environmental surroundings (e.g., Vinsel et al. 1980), choice of activities (e.g., Baumeis- ter and Tice 1982), differential amount of effort (e.g., Aronson and Carlsmith 1962; Silverman 1964), management of self- presentation (e.g., Goffman 1959), and selec- tive social interaction (e.g., Backman and Secord 1962; Pelham and Swann 1989; Swann 1985, 1987; Swann et al. 1987). If we negotiate our identities by structuring our perceptions and our interactions, do we negotiate for identities that are stable, or do we strive for an ever-enhanced view of self? For people with high self-esteem, the predic- tions of self-consistency theory and self- enhancement theory are identical: these indi- viduals will attempt actively to maintain their already positive self-images. On the other hand, the two perspectives make conflicting predictions for the behavior of individuals with low self-esteem: identity maintenance theories predict attempts at verification of the negative self-image, whereas self-enhance- ment theory predicts pursuit of an enhanced image of self. The Cognitive-Affective Crossfire In an attempt to reconcile conflicting
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 15 findings in the literature regarding identity maintenance and self-enhancement theories, Shrauger (1975) proposed a compromise: “When cognitive reactions are being as- sessed, findings favor the consistency model; whereas when affective reactions are being considered, they follow self-enhancement predictions” (p. 592). On the basis of Shrauger’s suggestion that cognitive and affective reactions to self-evaluative informa- tion are independent of one another, Swann and his associates (Swann 1987; Swann et al. 1987) posited a “cognitive-affective” cross- fire in which persons with negative self- identities must struggle when faced with self-relevant information. According to Shrauger (1975) and Swann et al. (1987), people respond affectively in a manner consistent with self-enhancement theory; they react cognitively, however, in accordance with identity maintenance theory. As noted above, the predictions of self-enhancement and of identity maintenance will be the same for persons with high self-esteem. For persons with low self-esteem, however, any self-relevant information leads to conflicting cognitive and emotional responses. Informa- tion that affirms their feelings of low self-worth is cognitively confirming but emotionally distressing. On the other hand, information that is too positive is affectively pleasing but cognitively distressing. Swann and his colleagues developed an impressive research program to support their theoretical arguments. They found that when individuals with high and low self-esteem were provided with either positive or negative feedback, the affective responses seemed to be in line with self-enhancement predictions, whereas the cognitive responses seemed to be more congruent with identity maintenance predictions (Swann et al. 1987). Even people with overall high self-esteem sought out unfavorable feedback pertaining to areas in which they held negative self-views (Swann et al. 1989). The tendency to prefer negative but consistent information was eliminated when subjects were deprived of the ability to retrieve and compare the feedback with the representation of self in memory; when cognitively loaded, subjects with both high and low self-esteem preferred the positive feedback. Affect Control Theory and the Cognitive-Affective Crossfire Is there really a cognitive-affective cross- fire? Swann et al. (1987) emphasize the independence of the cognitive and the affec- tive processes, using the metaphor of a war between the two systems for control over behavior. Affect control theory predicts the pattern of Swann’s results without positing a battle between emotion and cognition. In affect control theory, affective and cognitive systems are neither parallel nor conflicting; instead, they are intricately codependent. Cognitive categorizations of people and events create affective reactions which then form the basis for emotional, cognitive, and behavioral responses. Affect control theory produces these pre- dictions about impression change and behav- ioral reactions with a computer simulation program called INTERACT (Heise and Lewis 1988; Smith-Lovin and Heise 1988). The program contains the formal structure of the theory and empirical estimates of its parame- ters. Specifically, it consists of equations that show how impressions change on three dimensions of meaning (evaluation, potency, and activity, or EPA) for actor, object-person, and behavior (Smith-Lovin 1988a). The three dimensions are adopted from Osgood’s clas- sic factor-analytic work on affective meaning in developing the semantic differential scale (Osgood, May, and Miron 1975; Osgood, Suci, and Tannenbaum 1957). These dimen- sions of meaning are useful because they are measurable; they represent the structure of affective meaning across cultural and social boundaries; they apply to a wide variety of social objects; and they correspond roughly to the status, power, and expressivity dimen- sions that other social researchers have recognized as fundamental to social interac- tion (Kemper 1978; Kemper and Collins 1990). Affect control theory then uses a mathe- matical transformation to represent the propo- sition that people maintain fundamental meanings: the program solves for the three- number EPA profile of possible behaviors that would return the identities of the actor and the object-person as close as possible to their positions before the event was processed (Heise 1987). The program contains the current form of the theory-the equations and the mathematical transformation- and allows
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
16 SOCIAL PSYCHOLOGY QUARTERLY us to link verbal descriptions (identity and behavior labels defining the situation) to their numerical EPA values.3 To illustrate, let us consider an INTER- ACT simulation of two people receiving socially relevant feedback. Person 1 thinks of herself as highly skilled socially-a real Sweetheart.4 Person 2 considers herself so- cially incompetent-more of a Clod. When female undergraduates5 rate the two, a Sweetheart receives an evaluation of 2.85 on a scale that ranges from -4 (anchored by the terms “infinitely bad, awful”) to + 4 (“infi- nitely good, nice”), whereas Clods are rated – 1.30. Sweetheart and Clod have potency ratings (on a scale from -4, “infinitely powerless, shallow” to +4, “infinitely pow- erful, deep”) of 1.47 and -0.52 respectively, and activity ratings (on a scale from -4, “infinitely slow, old, quiet” to +4, “infi- nitely fast, young, noisy”) of 1.28 and 0.13. Affect control theory refers to the shared affective meanings associated with these identity structures as findamental sentiments. Events (such as social interactions) create 3 We use undergraduates’ average ratings of identity and behavior concepts on semantic differential scales to locate them in the three-dimensional EPA space. For a description of the study that collected these EPA “dictionaries” of social concepts, see Heise and Lewis (1988) and Smith-Lovin and Heise (1988). 4 We use capitalized words to indicate social identities and behaviors that have been rated on the EPA dimensions and are available for use in the INTERACT program. 5 These ratings come from a study reported in Smith-Lovin and Heise (1988). The measurement of fundamental sentiments by averaging across the ratings of 30 to 40 individuals is based on the assumption that these fundamental affective meanings are largely shared. This assumption is supported by a large literature that indicates substantial agreement across widely varying social groups on the prestige of occupations, the seriousness of crimes, and the semantic differential ratings of a wide variety of concepts (see review in Smith-Lovin 1988a). If we assume that most people in our culture agree that Mothers (in general) are nicer than Mafiosi, Bankers are more powerful than Babies, and Teenagers are more lively than Teetotalers, then we can average across the transient impressions of many individuals to assess the fundamental sentiments associ- ated with the identity (or behavior or emotion). The deviations from the fundamental sentiment will cancel out across individuals unless they have all experienced a similar event that changes their transient impressions in the same direction (e.g., one would not want to assess the fundamental sentiments toward Politicians immediately after a scandal has been publicized). This assumption about shared cultural fundamental sentiments allows us to use measurements on an undergraduate population to discuss social dynamics more generally. transient impressions as a result of the self-identities and the events experienced. For example, we would expect a Sweetheart who has just been Praised to view her identity somewhat differently from a Sweetheart who had just been Criticized. A Clod who has experienced the same types of feedback would feel different from the Sweetheart because of her different fundamental self- identity. How events transform our funda- mental views of ourselves into new transient meanings is described by structural equations that were developed by using the event as the unit of analysis (Heise and Smith-Lovin 1981; Smith-Lovin 1988a, 1988b). When the young women (the Sweetheart and the Clod) receive either clearly positive or clearly negative feedback about their social selves, how do they respond? Both funda- mental sentiments and transient impressions are represented in affect control theory by their loadings on Osgood’s evaluation, po- tency, and activity (EPA) dimensions. When an actor’s transient impression (produced by the current interaction) differs from his or her fundamental sentiment, a deflection is pro- duced. Deflection is defined as the sum of the squared differences between the fundamental sentiment and the transient impression on each of the three EPA dimensions. Thus, in the three-dimensional semantic space, a deflection corresponds roughly to the distance between the fundamental sentiment and the transient impression. It represents the extent to which the event has dislocated the social actor’s view of himself and others in the interaction. According to affect control the- ory, deflection is related directly to one’s fundamental identity: the more discrepant the feedback, the greater the deflection. Thus INTERACT predicts that both the Clod and the Sweetheart experience greater deflection when exposed to feedback that violates their previously held self-images. These predic- tions are depicted in Figure la. Emotions accompany deflections and sig- nal to the interactant how this experience did or did not confirm his or her expectations as based on the current definition of the situation (Smith-Lovin 1990; Smith-Lovin and Heise 1988). The character of an emotional re- sponse is determined by both the size and the direction of the deflection as well as by the transient impression produced by the situa- tion. In other words, the two components of emotions in the model are 1) the affective
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 17 6 5 – D 4-I e f e 3 t 0 ?2 Positive Negative Feedback Feedback -Sweetheart Clod Figure la. Predicted Deflection for Sweetheart and Clod after Positive and Negative Feedback meanings created by the interaction and 2) the distance between those meanings and the affective meanings initially associated with the identities involved in the interaction. Thus, when deflections are small (i.e., when events are confirming), an actor’s emotional response is determined largely by his or her identity (i.e., the fundamental sentiment). When events are disconfirming, however, the nature of this deflection heavily determines the character of the emotional response. Considering only the evaluation component of emotion, the equation predicting the emo- tional response in affect control theory can be reduced to E = 2T – I or T + (T – I) where E is the evaluation of the emotion, T is the transient impression produced by the social interaction, and I is the fundamental self-identity (Averett and Heise 1987). Thus we see that emotional response will be influenced both by the impressions created by the current situation (T) and by the discrep- ancy between one’s fundamental self-identity and current events (T – I) (Averett and Heise 1987; Smith-Lovin 1990). When social events fail to confirm fundamental identity, the nature of one’s transient feelings in the current setting (T) is very important in predicting the evaluative content of one’s emotional response. If we take a negative transient impression and subtract from it a positive self-identity (I), a quite negative emotion will be the result. If a person with a negative self-identity is deflected to a tempo- rarily positive self-image, the emotion will be positive (T will be high, and the subtraction of a negative I will add to this positivity). The emotion is positive both because of the pleasant transient impression and because it is so far above the normally negative fundamen- tal self-image. On the other hand, when social events confirm the sentiments of one’s fundamental identity, or T = I, then E = 2T – I = 21 – I = I. In other words, when one is not deflected, emotional experiences mirror one’s funda- mental identity. INTERACT predictions for the emotional reactions of our Sweetheart and our Clod illustrate this logic. Figure lb shows how a Sweetheart who is Praised feels an emotion that is consistent with her typical identity: she is happy. On the other hand, a criticized Sweetheart feels an emotion that is more reflective of her transient impression-a more negative emotion. Similarly, when the Clod is Praised, she also feels good because the discrepancy places greater weight on current events in determining her feelings. When the 0 -0.2 N e 9 a -0.4 V e -0.6 E – m 0 0.8 0 n -1.’ -1.2 – – -i.A.. __ Positive Negative Feedback Feedback -Sweetheart –Clod Figure lb. Predicted Negative Emotion for Sweetheart and Clod after Positive and Negative Feedback
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
18 SOCIAL PSYCHOLOGY QUARTERLY Clod is Criticized, she experiences little deflection and feels an emotion more consis- tent with her identity-unhappiness. Rather than assuming that all events have the same emotional impact on everyone, affect control theory stresses the importance of identity in determining emotion. In addi- tion, rather than viewing cognitive and affective processes as separate and conflicting systems, this theory hypothesizes that they are components of a complex integrated process. From the theory we can derive the following hypotheses about the relationship between fundamental self-evaluation and re- sponses to feedback about performance: Hypothesis 1. There will be a positive relation- ship between the evaluative valence of self- relevant feedback and the evaluative valence of emotional response. Hypothesis 2. There will be an interaction between the evaluation of a subject’s self- identity and the evaluative content of self- relevant feedback in its effect on deflection. When the feedback and the subject’s self-identity match, deflection will be low, and the feedback will be viewed as accurate and competent. Conversely, when the feed- back is very inconsistent with the subject’s self-image, deflection will be high. Affect control theory also can generate predictions about the degree to which the person who receives self-relevant feedback will like the evaluator. The equation for predicting the evaluation of an actor (in this case, the evaluator) shows a strong effect of behavior evaluation (Smith-Lovin 1988a, p. 47). In other words, how much we like someone is influenced strongly by how nicely we think they behaved. This point would lead us to predict that, everything else equal, an evaluator who gives positive feedback will be liked more than one who delivers criticism. In addition, however, the equation contains a multiplicative interaction effect involving behavior evaluation and object evaluation (Smith-Lovin 1988a, pp. 47-9). In this case, the object of the feedback is the experimental subject (the Clod or the Sweetheart). If that subject has a positive self-image (the Sweet- heart), giving her positive feedback leads to even higher evaluation for the rater (because a positive times a positive leads to a positive effect). Criticizing a Sweetheart would lower the rater’s evaluation, however, because the negative behavior would be multiplied by a positive self-image of the object of the criticism. For an object-person with a funda- mentally negative self-image, the pattern would be reversed: praise would raise the evaluation of the rater less, because the positive behavior would be multiplied by a negative object-person evaluation. Criticism would be a negative behavior directed at a negative object-person. The two negatives would cancel each other in the multiplicative interaction, resulting in a net increment to the evaluation of the Critic. This logic leads us to two additional hypotheses about the evalua- tion of the feedback author: Hypothesis 3: There will be a positive main effect of feedback evaluative valence on the evaluation of and liking for the author of the feedback. Hypothesis 4: There will be an interaction between the evaluative valence of a subject’s self-concept and the valence of the feedback received in its effect on liking for the evaluator. The hypotheses stated above should hold whenever the following scope conditions are met: Subjects will experience the predicted affective and cognitive reactions only if the definition of the situation used to generate these reactions is salient to the research participants. Social self-esteem (with funda- mental sentiments similar to those of Sweet- heart and Clod) must be evoked by the experimental situation. Furthermore, the sub- jects must perceive the identity-relevant feedback as Praise or Criticism (or as other acts associated with similar fundamental sentiments. We attempt to structure the experimental situation so that these conditions will be met. STUDY 1: PREDICTING EMOTION AND DEFLECTION The purpose of this first study was to test Hypotheses 1 through 4, derived from affect control theory. We use a modification of the Swann et al. (1987) paradigm, in which individuals with high and low self-esteem were provided with either positive or negative feedback about themselves. Cognitive reac- tions (e.g., the subjects’ perceptions of the accuracy of the feedback) and emotional reactions (e.g., the subjects’ levels of nega- tive affect) were measured in counterbalanced order.
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 19 Subjects and Setting Experimental subjects were 75 undergradu- ate students who volunteered to participate in the study in exchange for extra course credit in one of several undergraduate courses at a northeastern college. The experimental set- ting was a large, comfortable room with two one-way mirrors on one wall, approximately two feet apart. Next to the experimental room, on the other side of the one-way mirrors, were two connecting rooms, each with a view of the experimental room. Procedures A female experimenter introduced the purpose of the study as an experiment in interpersonal communications. Subjects were told that the study would compare the accuracy of first impressions based on nonverbal cues with the accuracy of first impressions based on verbal cues. Subjects first were asked to fill out “background information sheets,” which included the short form of the Texas Social Behavior Inventory (Helmreich, Spence, and Stapp 1974). To maximize privacy and to encourage greater candor, the experimenter left the room while the subject filled out this sheet. When the subject had finished the questionnaire, the experimenter reentered the room. The subject then was asked to deliver a prepared speech. The experimenter informed the subject that he or she would be watched by two observers, one behind each of the two one-way mirrors. The experimenter added that the anonymous observers could not hear through the one-way mirrors and therefore would not be biased by any verbal cues. The subject was told that after the speech, he or she would be able to examine and evaluate the ratings from both observers. The speech was a passage from Jonathan Livingston Seagull that took approximately three minutes to read. After the speech, the experimenter told the subject that the observ- ers would need two or three minutes to transfer their ratings onto summary sheets. During this time, she told the subject about the second phrase of the experiment. This phase was presented as a study of the accuracy of first impressions that are based on verbal behavior. The experimenter explained that the subject would be asked to hold a conversation with another student, which would be tape-recorded for analysis at another time .6 After the three-minute interval, the experi- menter left the experimental room and returned with two rater feedback summary sheets. In actuality, the feedback was pre- pared ahead of time and was identical for all subjects (except that the gender of the pronouns was matched to the subject’s gender). The feedback summary sheets con- sisted of a series of seven-point Likert scale questions at the top, with a space for general comments at the bottom. On the positive feedback sheet, Rater A characterized the subject as a 7 (“considerably higher than the average student”-the highest category) on social expertise, communication skills, and confidence scales. The rater gave the subject a 5 (“slightly higher”) on interpersonal sensitivity and a 6 (“somewhat higher”) on leadership ability. The subject was not rated at or below the midpoint (4) of any scale. In addition, the positive feedback sheet con- tained these remarks in longhand under a section labeled “General Comments”: The presenter is clearly a poised and graceful speaker. Her posture during the reading gave me the impression that she is at ease in social situations. The way she communicated through movements of her head made me really want to hear what she was saying! I imagine the presenter is very outgoing, with many close friends. That’s about all I could say about her. On the negative feedback sheet, Rater B gave the subject no rating higher than the midpoint of the scale. On leadership ability, communi- cation skills, and confidence, the rating was the next-to-lowest category (2). In the “Gen- eral Comments” section, rater B wrote: From her gestures, it appeared that the presenter was nervous and uncomfortable. This was especially evident by the way she looked down when she was speaking. It is hard to tell after such a brief observation, but I think that this presenter is probably awkward around groups of people and does not do well speaking to strangers. Observers were identified as “Rater A” and “Rater B” and were always presented as 6 This portion of the cover was necessary for maintaining consistency of procedures between Studies 1 and 2. The justification of the cover will become apparent during the reading of Study 2.
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
20 SOCIAL PSYCHOLOGY QUARTERLY the same gender as the subject. Subjects received the two sets of feedback one at a time. After the subject had read the first set, the experimenter asked him or her to fill out a questionnaire assessing first affective and then cognitive reactions to the feedback.7 The order of the positive and the negative feedback was counterbalanced such that one half of the subjects read and responded to the negative feedback first and the other half read and responded to the positive feedback first. The feedback manipulation is the order in which the two types of feedback were received. When they had completed the requested measures, the subjects were thanked and thoroughly debriefed. Reactions during post- experimental interviews showed that most subjects found the cover story fully convinc- ing. We excluded from analysis the data provided by two subjects when we discovered during the debriefing that they were suspi- cious of some element of the cover story. Self-Concept Measures We used the short form of the Texas Social Behavior Inventory (TSBI; Helmreich et al. 1974) as the measure of self-concept. The scale measures social self-esteem with items like “I would describe myself as socially unskilled,” “I usually expect to succeed in the things I do,” and “I feel that I can confidently approach and deal with anyone I meet.” The TSBI is a 16-item test scored on a six-point Likert scale with a total range of 16 to 80. We chose the TSBI because of its focus on “social self-esteem.” As stated above, we administered it as part of a “background information” questionnaire at the beginning of the study, but did not score it until after the experiment. Therefore the experimenter was blind to the subject’s score during the experimental procedure. The designation of an individual as a person with high or low self-concept is not 7 Swann et al. (1987) found that the affective responses vanished when measured after assessment of the cognitive responses. Either the passage of time alone or the actual cognitive processing of reactions to the self-relevant feedback was responsible for diminishing the affective reactions to the positive and negative evaluations. The cognitive responses, however, remained consistent irrespective of the order of assessment. The “cognitive-affective crossfire” was found only when affective responses were assessed before cognitive responses. trivial in this research. For designation of low self-concept, Swann (1987) recommended the use of cutoffs at the lower 10 to 30 percent of undergraduate samples. Swann et al. (1987) used scores of 51 and 66 respectively as the cutoffs for low and high self-concept. For purposes of replication, we used the same cutoff points in both of the studies described here.8 Subjects scoring between 51 and 66 were regarded as individuals with medium self-concepts. Cognitive and Affective Reactions to Feedback We used Zuckerman and Lubin’s (1965) Multiple Affect Adjective Check List (MAACL) to assess affective responses to the feedback. The checklist is scored to provide subscales for anxiety, depression, and hostil- ity, which combine to form a subscale for overall negative affect. To measure subjects’ cognitive reactions to the rater feedback, we used the series of items developed by Swann et al. (1987) to assess the perceived accuracy of the feedback, the perceived competence of the rater, and the subject’s attraction to the evaluator. We employed only the items retained by Swann et al. (1987) in their final analysis.9 RESULTS OF STUDY 1 Self-Concept Out of a possible 16-80 range of scores on the TSBI, our sample range was 39-79, with a mean of 61.2. These scores are slightly higher than those reported by Swann et al. (1987) and higher than the normative scores reported by Helmreich et al. (1974). Of our 75 subjects, 20 are classified as having high self-esteem, 50 with medium self-esteem and five with low self-esteem. 8 We repeated the analyses using cutoffs correspond- ing to the top and the bottom quartile of the sample and to the top and the bottom quartile of normative data reported by Helmreich et al. 1974. We also used self-concept as a continuous variable and conducted a multiple regression analysis. These changes in categorization of the indepen- dent variable did not alter the substantive results. 9 Because of possible interference from our cover story (i.e., our alleged interest in comparing first impression techniques and raters), we used only the cognitive responses to the Rater, and not those pertaining to the method. Thus Swann’s measures of “technique diagnos- ticity” and self/other attribution were omitted from the analyses.
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 21 Cognitive Reactions to Positive and Negative Feedback The subjects’ mean cognitive responses to the feedback are reported in Table 1.10 Each of the subscales titled “Rater Competence” and “Rater Accuracy” was entered into a 2 (feedback: positive, negative) x 3 (self- concept: low, medium, high) least-squares analysis of variance (ANOVA). We predicted a significant self-concept x feedback interac- tion (Hypothesis 2). Specifically, we pre- dicted that individuals with high self-concept would perceive the positive raters as rela- tively accurate and competent, whereas those with low self-concepts would perceive the negative rater as relatively accurate and competent. We made no theoretical predic- tions for individuals in the medium self- concept category. We used Bonferroni’s t procedure (Kirk 1968) on all comparisons to control for experimentwise error. As predicted, we found the self-concept x feedback interaction significant for both rater accuracy (F(2,72) = 16.88, p<.001) and rater competence (F(2,73) = 3.46, p<.05). Planned compari- sons among means revealed that individuals with low self-concept who received negative feedback found the rater more accurate (p<.0001) than did persons with high self- concept, whereas the reverse was true in the positive feedback condition for both accuracy (p<.05) and rater competence (p<.05). In keeping with the results reported by Swann et al. (1987), we also found a significant main effect of feedback on rater competence (F(1,73) = 4.46, p<.05) and rater accuracy (F(1,72) = 9.95, p<.01). Across all levels of self-concept, subjects rated the positive rater as both more compe- tent and more accurate than the negative rater. If one considers “competent” and “accurate” to be affectively positive traits (as well as cognitive judgments), this main effect of feedback condition on the cognitive measures supports our Hypothesis 4. Attraction to the Rater In Hypothesis 3, we also predicted both a significant main effect of feedback on attrac- 10 We used the least-square estimates of the marginal means to correct for the unbalanced design. It is these least square means that are reported in Tables 1 and 2. Table 1. Mean Cognitive Responses to Feedback by Self-Concept Positive Feedback Low Medium High Self- Self- Self- Concept Concept Concept Rater Competence 17.0 25.0 28.6 Rater Accuracy 26.0 28.7 37.0 Attractiveness 12.5 14.4 16.5 Negative Feedback Low Medium High Self- Self- Self- Concept Concept Concept Rater Competence 22.5 16.9 16.9 Rater Accuracy 34.5 19.6 11.2 Attractiveness 12.0 9.3 7.9 tion and a significant self-concept x feed- back interaction. Table 1 reports the mean responses to a subscale assessing the partici- pants’ attraction to the rater. We also entered the attraction subscale into a 2 (feedback: positive, negative) x 3 (self-concept: low, medium, high) ANOVA. The pattern of means displayed in Table 1 confirms these predictions. Analysis of vari- ance revealed a significant main effect of feedback on attraction (F(1,73) = 25.32, p<.0001). Subjects at all levels of self- concept displayed a strong tendency to perceive the positive rater as attractive. Nonetheless, we also found a significant self-concept x feedback interaction (F(2,73) = 5.73, p<.01). Planned comparisons show that individuals with high self-concept per- ceived the positive rater as significantly more attractive than did the persons with low self-concept (p<.05), whereas those with low self-concepts considered the negative rater as significantly more attractive than did those with high self-concepts (p<.05). Affective Reactions to Positive and Negative Feedback The measures of affective reactions were entered into 2 (feedback: positive, negative) x 3 (self-concept: low, medium, high) least-squares ANOVAs. The means for these affective responses are reported in Table 2. As predicted by Hypothesis 1, subjects reported experiencing greater levels of nega- tive affect after receiving negative feedback than after receiving positive feedback. This outcome resulted in a significant effect of the
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
22 SOCIAL PSYCHOLOGY QUARTERLY Table 2. Mean Affective Responses to Feedback by Self-Concept Positive Feedback Low Medium High Self- Self- Self- Concept Concept Concept Depression 5.0 11.7 9.6 Hostility 3.0 7.3 5.5 Anxiety 4.5 5.7 5.1 Negative Affect 12.5 24.6 20.3 Negative Feedback Low Medium High Self- Self- Self- Concept Concept Concept Depression 12.3 14.3 11.6 Hostility 8.7 7.8 6.8 Anxiety 10.3 9.1 5.6 Negative Affect 31.3 31.2 23.9 nature of feedback on anxiety (F(1,71) = 9.02, p<.05) and overall negative affect (F(1,71) = 6.44, p<.01).’1 As predicted, we found no main effects of self-concept on the affective measures, and no interactions between feedback and self- concept predicting negative affect. Subjects did not experience negative affect as a consequence of their self-concept or of the consistency of the feedback they received. DISCUSSION OF STUDY 1 Our findings confirm the predictions of af- fect control theory about the pattern of cogni- tive and affective responses to self-evaluative information. Figures 2a and 2b highlight the consistency between our data and affect control predictions by displaying information from Ta- bles 1 and 2 in a comparative format. All of the subjects reported feeling higher levels of neg- ative affect after having been evaluated unfa- vorably and feeling lower levels of negative affect after being evaluated positively. The same pattern was revealed in all three subscales of negative affect: anxiety, depression, and hostil- ity.t2 1 Inspection of the mean measures of hostility and depression suggests a tendency toward this pattern; ANOVA results indicate a nearly significant main effect of feedback on depression (F(l,71) = 3.58, p= .06) and a significant positive main effect of feedback on hostility (F(1,71)=3.93, p<.05). When the Bonferroni t proce- dure was implemented, however, the main effect on hostility did not remain significant. 12 The effect was significant only for anxiety; the other two scales show strong trends in the same direction. 0 -5 D -1QL e f e e -15- c 0 -20 n/ -25 – Positive Negative Feedback Feedback – Sweetheart Clod Figure 2a. Observed Deflection for High and Low Self-Esteem Participants after Feedback Despite the unpleasant emotional response associated with receiving negative evaluation, subjects with low self-concept still perceived the negative rater as more competent and more accurate than the rater who evaluated them positively. Moreover, they reported higher levels of attraction to the negative rater than did subjects with high self-esteem. In fact, there is no significant difference be- tween the low-self-esteem subjects’ liking of the positive and the negative raters; the people who saw themselves in a negative light were just as likely to like the sharp critic as to like the person who praised them. Now the question arises: If individuals with negative self-concept were given the choice, with whom would they prefer to spend. time-the rater who views them too posi- tively but makes them “feel good,” or the rater who sees them the way they view themselves, and is “competent” and “accu- rate”? Our second study includes a behavioral measure designed to answer this question. The sociological importance of this process lies in its involvement in determining social behavior. The central tenet of affect control theory is that people initiate action in order to minimize deflection. Thus, when events move meanings away from fundamental sentiments, individuals attempt to create events that restore their original meanings. Suppose our Sweetheart and our Clod were both exposed to both types of feedback- Criticism and Praise. Suppose further that each then was allowed to Talk to one of the
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 23 30 7, N7 e g a 25 7 t I V e E 20 m 07 t / o 15 / Positive Negative Feedback Feedback Sweetheart — Clod Figure 2b. Observed Negative Emotion for High and Low Self-Esteem Participants after Feed- back feedback bearers. Whom would each choose? Figure 3 shows INTERACT predictions for the deflections experienced after such choices are made. According to the figure, the Sweetheart will be less deflected if she Talks to the positive rater. The Clod, however, will experience less deflection if she Talks to the negative rater. In each case, INTERACT predicts that each of these two research subjects would be more inclined to converse with the rater who shares her self-image. In more formal terms, we suggest the following: Hypothesis 5. There will be an interaction between subject’s self-concept and the evalua- tive valence of feedback in affecting the choice of a future partner for interaction. Specifically, people will choose to interact with a rater who confirms their identity rather than one who enhances their self-image. In Study 2 we test this self-verification hypothesis in an experimental study of selective interaction. We predict that individ- uals will select interaction partners in such a way as to confirm existing self-views and to minimize confrontation with self-discrepant information. In order to test the selective interaction hypothesis, research subjects must be provided with an opportunity to self- enhance or self-verify. Further, subjects must believe that a resulting interaction actually will occur, and that it will provide future self-relevant information which is anticipated to be either confirming or disconfirming. STUDY 2: SELECTIVE INTERACTION FOR IDENTITY MAINTENANCE The strategy of selective interaction has been suggested as one possible tactic in identity negotiation (Backman and Secord 1962; Swann 1987). As a strategy, selective interaction could be used either to seek identity-confirming information about one’s self or to pursue favorable self-information. Because we believe in the power of social interaction to shape one’s identity, the strategy of selecting an interaction partner that would shape one’s identity positively seems plausible. Yet, the interactionist hy- pothesis generated by affect control simula- tions and tested in this study is that people select interaction partners in order to seek confirmation and to avoid disconfirmation of self-views. To test the selective interaction hypothesis, we employed the same experi- mental design as we used in the previous study. In Study 2, rather than measuring subjects’ cognitive and emotional reactions to the feedback, we asked them to choose from the two raters a conversation partner for the second portion of the study. Subjects, Setting, and Procedures Subjects were 76 undergraduate students who volunteered to participate in the study in exchange for extra credit in one of several lower-division undergraduate courses at a northeastern college. The experimental set- ting was the same as in Study 1. The procedures were identical, with one excep- tion: the positive and the negative feedback were given to the subject with the order of the feedback (which sheet was on top) counter- balanced. After looking at both sets of feedback and before receiving any question- naires, the subjects were asked to choose between the two raters. The experimenter offered the choice in the following way: While you are filling out these questionnaires, I will be in the next room getting things prepared for the second half of the study. In this part of the study we will be audiotaping a conversation between you and another participant to be rated at a future time by different raters who have not been exposed to your nonverbal behavior. I am going to ask one of the raters to come in and serve as a participant in this second part of the study. Because we will be using the tape to generate a second set of first impressions of you, I would like you to choose the rater that
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
24 SOCIAL PSYCHOLOGY QUARTERLY 10 8- D e f 6 – e C 4 – 0 n 2- 0 Clod Sweetheart _~ Critic 2 Flatterer Figure 3. Deflection of Sweetheart and Clod after Choosing Interaction Partner you would be the most comfortable talking with for the next few minutes. If the subject did not express an immediate preference, the experimenter reiterated her wish that the subject be comfortable and asked again for a choice. After the choice was made, subjects were carefully debriefed and interviewed. RESULTS OF STUDY 2 Self-Concept The sample scores on the social self-esteem scale (the TSBI) ranged from 38 to 76, with a mean of 61.2. The distribution of scores corresponded closely to that of the sample in Study 1, but again was centered around a higher mean than in the normative data reported by Helmreich et al. (1974) and in the sample used by Swann et al. (1987). Again, we used the same cutoff points as reported by Swann et al. (1987). In Study 2, there were 24 subjects with high self-esteem, 45 with moderate self-esteem, and nine classified as having low self-esteem. Selective Interaction Table 3 presents the choice of rater for individuals in each of the three self-concept categories and for both orders of feedback. We performed log-linear analyses to learn whether either self-concept or order of feedback influenced choice of rater.13 The iterative proportional fitting algorithm was used to estimate the models reported in Table 3. Model 1 is the model of independence, which we estimated to determine whether any systematic relationship existed among the variables in the model. As shown in Table 3, Model 1 does not fit the data, with a likelihood ratio of 16.60 on seven degrees of freedom. Thus we can presume that there is some systematic variation to be explained by the variables in the model. Of the remaining models, the most parsimonious model that fits the data is Model 5, which specifies a relationship between self-concept category and choice of rater and no relationship between order and either self-concept or choice of rater. 14 The data show that choice of positive or negative rater depended on the level of the subject’s social self-esteem, but choice of rater was not influenced by the order of the feedback. In what way was self-concept related to choice of rater? Figure 4 shows the choice of 13 We also performed these analyses using logistic regression and treating self-concept as a continuous variable. The results were not substantively different from those achieved with the log-linear models. 14 The order x choice effect is null by design because of random assignment to conditions, and is included here only for completeness.
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 25 Table 3. Choice of Rater by Self-Concept and Order of Feedback Order of Feedback Positive First Negative First Self-Concept L M H L M H Choice Negative 3 6 5 5 5 3 Positive 0 18 7 1 16 9 Model Fitted Marginals df G2 1 (Order) (Self) (Choice) 7 16.60 2 (Self x Order) (Self x Choice) (Order x Choice) 2 1.04 3 (Self x Choice) (Order x Choice) 4 2.80 4 (Self) (Order x Choice) 6 42.97 5 (Order) (Self x Choice) 5 2.85 interaction partner broken down by self- concept category. Collapsing across order of feedback, we find that 89 percent of persons with low self-concept, 24 percent with medium self-concept, and 33 percent with high self-concept chose the rater who pro- vided negative feedback. In fact, only one of the suibjects with low self-concept chose to talk with the complimentary rater, whereas the great majority of subjects with medium and high self-concept preferred the positive rater. DISCUSSION OF STUDY 2 Our findings suggested that the subjects selected as a conversation partner the rater who viewed them most like the way they perceived themselves. With only one excep- tion, individuals with low self-concept chose to interact with the rater who rated them unfavorably, whereas the great majority of persons with both medium and high self- concept preferred the rater who evaluated them positively. Further, persons with me- dium and high self-concept behaved similarly with respect to choice of interaction partner. These results, however, do not necessarily establish that subjects chose raters whose views were consistent with their own in order to confirm their own self-view. During the postexperimental interviews we became aware of at least one alternative motive for choosing a conversation partner: some of the subjects reported choosing the negative rater in order to change that rater’s opinion of them. To explore further this strategy of identity negotiation, we analyzed data from a post- choice questionnaire given to 44 of the subjects in Study 2.15 The questionnaire asked 15 Because this questionnaire was considered explor- the subjects why they made the choice they made. Subjects were categorized by their responses into two groups: 1) those who anticipated liking the rater they chose and 2) those who wanted to change the rater’s impression. All 32 of the 44 subjects who chose the positive rater (one low self-concept, 22 medium self-concept, and nine high self- concept) stated that they did so because they expected to like the positive rater. Twelve of the 44 subjects chose the negative rater, four in each of the three self-concept categories. Three of the four respondents with low self-concept reported choosing the negative rater because they expected to like that rater. The other reported choosing the negative rater to change his or her mind. The pattern was exactly reversed for the respondents with medium and high self-concept. In each of these categories, three respondents reported choosing the negative rater in order to change that rater’s opinion, and only one respondent reported making the choice because he or she expected to like the negative rater. In other words, these subjects used selec- tive interaction in more way than one as a strategy for identity negotiation. In the predominant strategy, subjects chose interac- tion partners who verified their self image. A few subjects with medium and high social self-concept, however, chose the negative rater in an attempt to change his or her mind. Thus an affect control perspective on choice of interaction partner is strengthened by these exploratory findings concerning the motiva- tion for the interaction choices. Experimental subjects used both passive (selection of alters atory, we administered the postchoice questionnaires only when time allowed. Thus the 44 respondents are a nonrandom subsample of the entire sample of 76.
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
26 SOCIAL PSYCHOLOGY QUARTERLY p 120 e r c 1 00 e n t 80- R a 60 t e r 40- 0 h o 20- C e 0 Low Medium High Negative 1 Positive Figure 4. Interaction Partner Choice by Social Self-Concept who already agreed with their self-image) and active (selection of alters for corrective behavioral action) behavioral strategies to accomplish identity maintenance. GENERAL DISCUSSION The first purpose of this research was to test affect control theory hypotheses about emo- tional and cognitive reactions of individuals when confronted with new identity-relevant in- formation. Our hypothesis was that emotional and cognitive responses to personal evaluations are not mirror reflections of one another, nor are they independent, conflicting systems. In- stead, interactionist-based affect control theory predicts that these responses are related in sys- tematic ways. Cognitive definitions and affec- tive responses create a control system whereby emotions signal events that confirm or discon- firm self-identities, and cognitive processing re- veals the degree of inconsistency between the existing social situation and expectations from preexisting schemas. The second goal of the present research was to explore the behavioral implications of identity maintenance. Affect control theory predicts that interaction partners will be selected to provide consistency in self- schemas. Social interaction will serve as an active force to maintain self-identities. Our experimental situation reveals striking evi- dence for these phenomena. The desire for a stable identity seems to be an important motivation for determining our cog- nitive reactions to events as well as our active creation of new events. The notion that indi- viduals take an active role in maintaining a con- sistent, coherent view of self should not be sur- prising to sociologists or other observers of social behavior. The existing of both self-enhancing and self-handicapping behavior is confirmed by empirical research as well as by casual obser- vation. Our findings of identity-maintaining be- havior lead to the prediction that identities are rather invariant. In fact, people maintain sur- prisingly stable concepts of self (Mortimer and Lorence 1981; O’Malley and Bachman 1983). This stability carries implications for important sociological concerns. A well-defined identity makes it possible to experience the continuity of self as well as the differentiation between self and other. A sense of self-continuity permits the anticipa- tion of future consequences for present actions. The feeling of differentiation be- tween self and others allows for the percep- tion of the evaluations of others as well as for the development of a “generalized other.” Each of these processes is crucial to the motivation to observe social rules. Similarly, the continuity and the differentiation provided by a stable identity allow us to self-classify and operate within social roles. Whereas self-stability may be crucial to
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
SELECTIVE INTERACTION AND IDENTITY 27 societal functioning, the active maintenance of negative identities may be problematic for individual social functioning. Findings like those reported in Study 1 imply that the maintenance of negative identities will be emotionally distressing for the individuals involved; Study 2 suggests that treatment of people with chronically low self-esteem may be difficult. When an individual is in structural surroundings that foster feelings of low self-worth (e.g., a child in an abusive home), those feelings may be highly resistant to change. Such abused people will reject as inaccurate any information designed to foster a sense of self-worth. People with low self-esteem will perpetuate their distress by choosing abusive, critical interaction part- ners. Because affect control theory suggests that only alters who view themselves in a fundamentally negative way will be likely to engage consistently in such negative behav- iors, we predict that people with low self esteem will choose one another as interaction partners. This homophily principle leads to a perversely stable social structure: people who think negatively of themselves are locked into dissatisfying, unpleasant relations with simi- lar others, whereas alters with high self esteem have little incentive to intrude into these persistently negative interactions. This study demonstrates the importance of merging social cognitive ideas with a more social interactional focus to examine identity maintenance alternatives other than cognitive processing. Individuals can participate in the behavioral selection of information, not merely in cognitive selection. Outside the laboratory people have opportunities to play much more active roles in determining exposure to information and in eliciting certain types of information. We think that an interesting opportunity for.sociologists would be to extend this research beyond the study of assimilation/accommodation outcomes to analysis of the social mechanisms we use to attend to and discount information (e.g., reinterpretation, second opinions, avoidance, “fishing,” choice of friends, invalidation) and of the social consequences of these mecha- nisms. Such research would require more flexible laboratory procedures or field studies to examine the strategies that individuals adopt to shape incoming information and define their social environments. REFERENCES Aronson, Elliot and John M. Carlsmith. 1962. “Perfor- mance Expectancy as a Determinant of Actual Performance.” Journal of Abnormal and Social Psychology 65:178-82. Averett, Christine and David R. Heise. 1987. “Modified Social Identities: Amalgamations, Attributions and Emotions.” Journal of Mathematical Sociology 13 (1-2): 103-32. Backman, C.W. and Secord, P.F. 1962. “Liking, Selective Interaction, and Misperception in Congruent Interpersonal relations.” Sociometry 25:321-335. Baumeister, R.F. 1986. Identity: Cultural Change and the Struggle for Self. New York: Oxford University Press. Baumeister, R.F. and D. Tice. 1982. “Self-Esteem and Responses to Success and Failure: Subsequent Perfor- mance and Intrinsic Motivation. Journal of Personality 53:450-67. Becker, E. 1968. The Structure of Evil. New York: Bazilier. Bem, Daryl. 1967. “Self Perception: An Alternative Interpretation of Cognitive Dissonance phenomena.” Psychological Review 74:183-200. Cooley, Charles H. 1964. Human Nature and the Social Order. New York: Schocken. Festinger, L. 1957. A Theory of Cognitive Dissonance. Stanford: Stanford University Press. Goffman, Erving. 1959. The Presentation of Self in Everyday Life. Garden City, NY: Doubleday. Heider, F. 1958. The Psychology of Interpersonal Relations. New York: Wiley. Heise, David R. 1977. “Social Action as the Control of Affect.” Behavioral Science 22:163-77. . 1979. Understanding Events: Affect and the Construction of Social Action, New York: Cambridge University Press. . 1986. “Modeling Symbolic Interaction.” Pp. 291-309 in Approaches to Social Theory, edited by S. Lindenberg, J.S. Coleman and S. Novak. New York: Russell Sage Foundation. _____. 1988. “Affect Control Theory: Concepts and Model.” Pp. 1-34 in Analyzing Social Interaction: Advances in Affect Control theory, edited by Lynn Smith-Lovin & David R. Heise. New York: Gordon and Breach Science Publishers. Heise, David R. and Elsa Lewis. 1988. Introduction to Interact. Raleigh: National Collegiate Software Clear- inghouse. Heise, David R. and Lynn Smith-Lovin. 1981. “Impres- sions of Goodness, Powerfulness and Liveliness from Discerned Social Events.” Social Psychology Quar- terly 44 (2):93-116. Helmreich, R., J.T. Spence and J. Stapp. 1974. “Short Form of the Texas Social Behavior Inventory, an Objective Measure of Self-Esteem.” Bulletin of the Psychonomic Society 4: 473-75. Hewitt, John P. 1979. Self and Society: A Symbolic Interactionist Social Psychology. Boston: Allyn and Bacon, Inc. Jones, S.C. 1973. “Self- and Interpersonal Evaluations: Esteem Theories versus Consistency Theories.” Psy- chological Bulletin 17:185-99. Kemper, Theodore D. 1978. A Social Interactional Theory of Emotions. New York: Wiley. Kemper, Theodore D. and Randall Collins. 1990. “Dimensions of Microinteraction.” American Journal of Sociology 96 (1):32-68. Kirk, 1968. Experimental Design. Belmont: Brooks/Cole Publishing Co. Klein, W.M. 1989. “The Justification of Motivated
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms
28 SOCIAL PSYCHOLOGY QUARTERLY Beliefs about Others.” Paper presented at the annual meeting of the Eastern Psychological Association, Boston. Langer, E.J. and J. Roth. 1975. “Heads I Win, Tails It’s Chance: The Illusion of Control as a Function of the Sequence of Outcomes in a Purely Chance Task.” Journal of Personality and Social Psychology 32: 951-55. Lecky, P. 1945. Self-Consistency. New York: Island Press. McDougall, W. 1932. The Energies of Men. London: Methuen. Mead, George H. 1934. Mind, Self, and Society, Chicago: University of Chicago Press. Mortimer, J.T. and J. Lorence. 1981. “Self-Concept Stability and Change from Late Adolescence to Early Adulthood.” Research in Community and Mental Health 2:5-42. O’Malley, R.M. and J.G. Bachman, 1983. “Self- Esteem: Change and Stability between the Ages of 13 and 23.” Developmental Psychology 19:257-286. Osgood, Charles E., W.H. May, and M.S. Miron. 1975. Cross-Cultural Universals of Affective Meaning. Ur- bana: University of Illinois Press. Osgood, Charles E., George C. Suci, and Perry H. Tannenbaum. 1957. The Measurement of Meaning. Urbana: University of Illinois Press. Pelham, B.W. and W.B. Swann Jr. 1989. “From Self-Consciousness to Self-Worth: The Sources and Structure of Self-Esteem.” Journal of Personality and Social Psychology 57:672-80. Rosenberg, Morris. 1988. “Self-Objectification: Rele- vance for the Species and Society.” Sociological Forum 3:548-65. Shrauger, J.S. 1975. “Responses to Evaluation as a Function of Initial Self-Perceptions.” Psychological Bulletin 82:581-96. Silverman, I. 1964. “Self-Esteem and Differential Responsiveness to Success and Failure.” Journal of Social Psychology 69:115-19. Smith, M.B. 1968. “The Self and Cognitive Consis- tency.” Pp. 366-372 in Theories of Cognitive Consistency: A Sourcebook, edited by E. Abelson, E. Aronson, W.J. McGuire, T.M. Newcomb, M.J. Rosenberg, and P.H. Tannenbaum. Chicago: Rand McNally. Smith-Lovin, Lynn 1988a. “Affect Control Theory: An Assessment. ” Pp. 171-92 in Analyzing Social Interac- tion: Advances in Affect Control Theory, edited by Lynn Smith-Lovin and David R. Heise (eds.). New York: Gordon and Breach Science Publishers. . 1988b. “Impressions From Events.” Pp. 35-70 in Analyzing Social Interaction: Advances in Affect Control Theory, edited by Lynn Smith-Lovin and David R. Heise. New York: Gordon and Breach Science Publishers. . 1990. Emotion as Confirmation and Disconfir- mation of Identity: An Affect Control Model. In Research Agendas in the Sociology of Emotions, edited by Theodore D. Kemper. New York: SUNY Press. Smith-Lovin, Lynn, and David R. Heise (Eds.). 1988. Analyzing Social Interaction: Advances in Affect Control Theory. New York: Gordon and Breach Science Publishers. Stryker, Sheldon. 1980. Symbolic Interaction: A Social Structural Version. Menlo Park: Benjamin Cummings Publishing Co. Swann, William B. Jr. 1983. “Self-Verification: Bring- ing Social Reality into Harmony with Self.” Pp. 33-66 in Psychological Perspectives on the Self, Vol. 2. edited by Jerry Suls and Anthony Greenwald. Hillsdale: Lawrence Erlbaum. . 1985. “The Self as Architect of Reality,” Pp. 100-25 in The Self and Social Life, edited by Barry R. Schlenker. New York: McGraw Hill. . 1987. “Identity Negotiation: Where Two Roads Meet.” Journal of Personality and Social Psychology 53:1038-51. Swann, William B., Jr., John J. Griffin Jr., Steven C. Predmore, and Bebe Gaines. 1987. “The Cognitive- Affective Crossfire: When Self-Consistency Confronts Self-Enhancement.” Journal of Personality and Social Psychology 52 (5): 881-89. Swann, William B., Jr., Brett W. Pelham, and Douglas S. Krull. 1989. “Agreeable Fancy or Disagreeable Truth: Reconciling Self-Enhancement and Self- Verification.” Journal of Personality and Social Psychology 57(5): 782-91. Tajfel, H. 1981. Human Groups and Social Categories: Studies in Social Psychology. London: Cambridge University Press. Taylor, S.E. and J.D. Brown. 1988. “Illusion and Well-Being: A Social Psychological Perspective on Mental Health. ” Psychological Bulletin 103:193-210. Tesser, Abraham. 1988. “Toward a Self-Evaluation Maintenance Model of Social Behavior.” Pp. 181-227 in Advances in Experimental Social Psychology. Vol. 20, edited by L. Berkowitz. New York: Academic Press. Tesser, A. and Jennifer Campbell. 1983. “Self- Definition and Self-Evaluation Maintenance.” Pp. 1-32 in Psychological Perspectives on the Self. Vol 2. ed. Edited by Jerry Suls & Anthony Greenwald. Hillsdale: Lawrence Erlbaum. Vinsel, A., Barbara B. Brown, Irwin Altman and C. Foss. 1980. “Privacy Regulation, Territorial Displays and Effectiveness of Individual Function.” Journal of Personality and Social Psychology 39:1104-15. Weinstein, N.D. 1980. “Unrealistic Optimism about Future Life Events.” Journal of Personality and Social Psychology 39:806-20. Zuckerman, M. and B. Rubin. 1965. Manual for the Multiple Affect Adjective Check List. San Diego: Educational and Industrial Testing Services. DAWN T. ROBINSON is a Postdoctoral Fellow at Stanford University. She is applying affect control ideas to the study of group interaction and network relations. She is also conducting experiments that explore how identities become salient in interaction situations. LYNN SMITH-LOVIN is Professor of Sociology at the University of Arizona. She currently is conducting a series of studies to test affect control theory and its predictions about the relationships among identities, behaviors and emotions. She also is working on new theoretical ideas that link multiple identities and the production of mixed emotions.
This content downloaded from
132.174.254.159 on Tue, 16 Mar 2021 22:40:35 UTC
All use subject to https://about.jstor.org/terms