اضطراب اجتماعی و ادراک بین فردی: روابط تجزیه و تحلیل مدل های اجتماعی

Cognitive models of social phobia posit that an individual’s negative beliefs about the way he or she is perceived by others (metaperceptions) are a core feature of the disorder. The social relations model (Kenny, 1994) Interpersonal perception: A social relations analysis. New York: Guildford) was used to analyze interpersonal perception data collected following unstructured social interactions in 62 socially anxious (SA) and 62 not socially anxious (NSA) individuals. Using this model, the interpersonal perceptions were analyzed to evaluate whether pathological levels of social anxiety are associated with self-perceptions, metaperceptions, and perceptions from others. SA participants saw themselves negatively and believed others saw them negatively. Although seen as more nervous by others, SA participants were not seen as less likeable. A mediational model demonstrated that the negative metaperceptions of SA individuals were more a function of their own self-perceptions than the negative perceptions of others. These findings were not attributable to depressive symptoms. Implications for theory and treatment of social phobia are discussed.

Social phobia (also known as social anxiety disorder) is characterized by an intense fear of being negatively evaluated by others, in conjunction with avoidance of (or distress in) situations in which this fear is activated (American Psychiatric Association, 1994). Cognitive biases have been implicated in the etiology and maintenance of the disorder (Clark and Wells, 1995, Foa, Franklin, Perry and Herbert, 1996 and Rapee and Heimberg, 1997), including attention and memory biases (Amir, Foa and Coles, 1998, Amir, Foa and Coles, 2001, Asmundson and Stein, 1994, Foa, Gilboa-Schechtman, Amir and Freshman, 2000, Gilboa-Schechtman, Foa and Amir, 1999, Hackmann, Clark and McManus, 2000 and Mattia, Heimberg and Hope, 1993) and interpretative biases for socially relevant information (Alden and Wallace, 1995, Mellings and Alden, 2000 and Stopa and Clark, 2000). Evaluative and interpretative biases may be particularly revealing of how persons with social phobia process information in social situations and then use this information to guide their expectations and future actions. Alden and Wallace (1995) had 32 patients with generalized social phobia (GSP) and 32 non-clinical controls participate in a dyadic “get acquainted” interaction with an experimental assistant who engaged in either positive or negative social behavior. The patients with GSP displayed a negative bias in their appraisals of some, but not all, aspects of their social performance; they displayed, however, a positive bias when appraising their partner’s performance (Alden & Wallace, 1995). Subsequent work by this research group has confirmed that these biases in social judgments are driven by selective attention to negative self-related information (Mellings & Alden, 2000). Stopa and Clark (2000) found that persons with GSP, compared to other anxiety disordered or non-anxious subjects, interpreted ambiguous social events (e.g. someone looks out of the window while in conversation with them) in a negative fashion. Furthermore, persons with GSP interpreted mildly negative social events (e.g. someone disagrees with their opinion on a subject) in a negative self-referential (“I’m boring”) and catastrophic (“I’ll lose my job”) fashion. These interpretations are hypothesized to lead to anxious apprehension and avoidance of future social opportunities. Although data provided by studies of hypothetical social interactions are useful, studies that directly assess interpersonal perceptions during actual social interactions provide much-needed in vivo confirmatory data. These studies can also assess how others perceive persons with high social anxiety. This information might be especially valuable—for example, in cognitive behavioral therapy—for helping patients with social phobia understand how others actually see them. These kinds of studies have been rare, presumably because of the difficulty in designing and interpreting them. Fortunately, recent innovations in the analysis of social interaction data can be applied to this problem, providing a structure for testing and comprehending not only the interpersonal perceptions of the individual with high social anxiety, but also those of the person with whom he or she is interacting. 1.1. The social relations model One of the inherent difficulties in studying interpersonal interactions is that both the perceiver and his or her target may influence perceptions in these contexts. For example, Alice’s ratings of Bill’s sociability is partly determined by Bill’s behavior and partly determined by how Alice views others in general. The social relations model (SRM; Kenny, 1994) provides both a theoretical and data analytic framework to account for the non-independence of the perceiver and the target. In the SRM, three sources of variance (denoted as perceiver, target, and relationship) contribute to any perception of another person’s personality or behavior. The perceiver effect is the portion of a rating due to how the person sees others in general. The target effect represents how the target is viewed by others in general. Finally, the relationship effect is the unique variance due to the pairing of a particular perceiver and target. The relationship effect also typically includes what is considered “error” variance. Using the example of Alice’s rating of Bill’s likeability, the perceiver effect would be the degree to which Alice tends to see others in general as likeable. Alice might be unfriendly and generally sees others as not very likeable. The target effect would be the extent to which Bill is viewed by others as likeable. Bill may be affable in social interactions, which leads others to think he is likeable. Finally, the relationship effect would account for Alice’s rating of Bill’s likeability beyond her tendency to rate others as likeable and beyond his tendency to be seen as likeable by others. Across all their interaction partners, Alice may perceive Bill as more likeable than she finds most people, but less likeable than Bill is usually seen by others. In addition to perceptions of other people’s traits, the SRM can be applied to how people think they are seen by others (metaperceptions), which is a key factor in social anxiety. Consider an example of Bill’s metaperception of Alice for intelligence. That is, how Bill thinks Alice rated him on intelligence. The perceiver effect is the degree to which a person believes others see him or her in general. Bill may believe that others generally see him as unintelligent. The metaperception target effect represents others’ beliefs about one’s tendencies as a rater. For many of her interaction partners, Alice may seem a harsh rater and, therefore, people consistently believe she will rate them low on intelligence. Finally, the relationship effect accounts for the unique variation between the metaperceiver and his or her target. Bill may think that Alice sees him as particularly unintelligent, beyond his tendency to believe others see him as unintelligent and beyond Alice’s tendency to be seen as someone who rates others as unintelligent. Similar to other-perceptions and metaperceptions, the effects can also be estimated for self-perceptions. Although people are generally consistent in the way they see themselves across social interactions (indicated by a large perceiver effect for self-perceptions), some situations may make interaction partners influential for one’s self-perceptions (a target effect for self-perceptions). For example, if one’s interaction partner is responsible for providing evaluative feedback, people may see themselves differently following interactions with this partner. A relationship effect for self-perceptions would indicate that the way people see themselves is largely driven by their unique relationship with their interaction partner. Most relevant to the present investigation, the perceiver, target, and relationship effects can be correlated with individual difference variables (e.g. social anxiety). In this way, researchers can identify whether certain types of interpersonal perceptions are associated with certain personality traits. Consider the example of social anxiety and perceptions of friendliness: the SRM allows one to estimate the degree to which perceiver, target, and relationship effects for friendliness are related to levels of social anxiety. Specifically, by correlating social anxiety with perceiver effects for friendliness, one can assess whether people who are higher in social anxiety tend to see other people as unfriendly. Alternatively, correlating social anxiety with target effects for friendliness would demonstrate whether people with higher anxiety were seen by others as less friendly. Similar correlations can be estimated for metaperceptions and self-perceptions. In sum, the SRM provides a unique perspective for the study of interpersonal perceptions, including how people see themselves, see others, are seen by others, and think they are seen by others. 1.2. The present research The SRM was used to study the interpersonal perceptions of people with and without high social anxiety following unscripted social interactions. Of particular interest to this study were the negative biases in both self-perceptions and metaperceptions, which have been central to cognitive models of social phobia (Clark and Wells, 1995 and Rapee and Heimberg, 1997). Specifically, these models propose that individuals with social phobia over-emphasize negative self-perceptions, while often ignoring feedback from interaction partners, when determining how they are seen by others. Although self-perceptions, other-perceptions, and metaperceptions have been often investigated individually, the present research assesses all three types of perceptions in one study. To calculate the variance components in an SRM analysis, interpersonal perceptions must be made for multiple targets. In the present study, groups consisted of four unacquainted participants (two with high and two with normative levels of social anxiety) who completed round–robin interactions in each experimental session. Each participant interacted with each other participant for 5 min and then immediately recorded self-perceptions, perceptions of their partner, and metaperceptions of their partner’s ratings. These perceptions were made for a broad array of personality characteristics, thereby enabling a suitable “multi-trait” approach to the study. Included were descriptors that should (e.g. sociable, nervous) and should not (e.g. intelligent, reliable) be related to social anxiety. Although no laboratory task is perfect, brief unscripted social interaction has some strengths in that it involves (1) “live” interactions as opposed to imagining scenarios, (2) real participants (i.e. no confederates are used), and (3) a common experience with face validity and direct relevance for persons with social phobia, i.e. meeting new people. The clearest predictions in this study were in reference to how people with high levels of social anxiety would make interpersonal perceptions. We hypothesized that higher levels of social anxiety should be negatively correlated with perceiver effects in self-perceptions and metaperceptions for desirable characteristics (e.g. likeability). Although less certain, it seemed likely that this general negativity would also be reflected in ratings of their partners, which should yield negative correlations between social anxiety and perceiver effects in other-perceptions. Throughout, we hypothesized that these effects would be detected even when adjusting for the most frequent and obvious potential confound, depressive symptoms (Cox, Rector, Bagby, et al., 2000 and Stein, Tancer, Gelernter, Vittone and Uhde, 1990). It was less obvious how people with high levels of social anxiety would be seen by their interaction partners. Indeed, research reviewed by Rapee and Heimberg (1997) suggests mixed findings when examining other people’s perceptions of those with high social anxiety. We expected to find that people with greater social anxiety would be seen differently by their interaction partners. This prediction would be supported by correlations between target effects in other-perceptions and levels of social anxiety. If people with high levels of social anxiety are seen differently by others, how consistent will this discrimination be across the perceptions of various personality characteristics? It seemed likely that differential ratings of characteristics related to anxiety (e.g. nervous, relaxed) and extroversion (e.g. sociability, quietness) should be a sine qua non in perceptions of socially anxious (SA) individuals. However, it seems less probable that they would be seen differently on traits unrelated to their anxiety (e.g. conscientiousness or intelligence). The extent to which subjects would be viewed as likeable was unknown, and open to empirical testing. One benefit of assessing self-perceptions, other-perceptions, and metaperceptions with the SRM is that it allows one to test mediational models among the variables. A hallmark of social phobia is the belief that one is seen negatively by their interaction partners (i.e. the predicted negative correlation between social anxiety and perceiver effects in metaperceptions). What leads people with social anxiety to believe they are seen this way? Several models have been proposed to explain how people construct metaperceptions (Kenny, 1994; Kenny & DePaulo, 1993). One possibility is that people with high social anxiety receive negative feedback (either verbal or non-verbal) from their interaction partners. This might be a “kernel-of-truth” explanation that people with social anxiety think others dislike them because, in fact, they are disliked by others. If this were the case, one would expect target effects in other-perceptions to mediate the relationship between social anxiety and perceiver effects in metaperceptions. Although most models of social phobia discuss this process as a possible explanation for negative beliefs of the SA, results have been inconclusive as to whether people with social phobia actually perform more poorly in social situations (Rapee & Heimberg, 1997) In contrast to the kernel-of-truth hypothesis, self-focused cognitive models of social phobia predict that the person’s own self-perceptions lead to beliefs about the negative intentions of others (e.g. Clark and Wells, 1995 and Rapee and Heimberg, 1997). Furthermore, reviews of research on metaperceptions support this perspective over other models (Kenny, 1994; Kenny & DePaulo, 1993). Therefore, we expect that self-perceptions will mediate the relationship between social anxiety and perceiver effects in metaperception.