Modelling relationships between cognitive variables during and following public speaking in participants with social phobia

Abstract

Cognitive models of social phobia predict that several cognitive processes will mediate the relationship between trait levels of social anxiety and the extent of anxiety experienced in a specific social-evaluative situation. The current study aimed to provide a test of these relationships. Over 200 clinical participants with social phobia completed measures of their general social anxiety and a week later performed a brief impromptu speech. They completed a measure of state anxiety in response to the speech as well as questionnaires assessing several cognitive constructs including focus of perceived attention, perceived performance, and probability and cost of negative evaluation. A week later, they completed measures of negative rumination experienced over the week, as well as a measure of the recollection of their perceived performance. Path analysis provided support for a model in which the cognitive factors mediated between general social anxiety and the degree of anxiety experienced in response to the speech. A second model supported the theory that negative rumination mediated between characteristic social anxiety and negative bias in the recollection of performance.

Introduction

Some of the earliest behavioural formulations of social phobia highlighted individuals’ realisation of their social limitations due to a lack of social skills (Curran, 1977; Twentyman & McFall, 1975). Later cognitive formulations pointed to the central role of specific thinking distortions made by people with social phobia, such as perceived inadequacies and fears of negative evaluation (Beck, Emery, & Greenberg, 1985; Hartman, 1983). One of the more influential theories was proposed by Schlenker and Leary (1982), who suggested that the extent of anxiety in a social situation was determined by the discrepancy between an individual's perception of the audience's expectations and his or her own ability to meet those demands.

More recently, models of social phobia have begun to integrate some of these features together with ideas from research into information processing to provide more detailed predictions (Clark & Wells, 1995; Rapee & Heimberg, 1997). According to Clark and Wells (1995), when socially phobic individuals enter a social situation, they shift their attention inward and focus on their anxious responding to produce a distorted image of the way they are presenting in the situation. In the longer term, they engage in avoidance and safety behaviours to protect themselves from threat, and in this way maintain their negative self-perceptions over time. Following a social event, according to this model, individuals with social phobia continue to ruminate about their perceived shortcomings, which further helps to maintain their distorted image of themselves. A very similar model was proposed by Rapee and Heimberg (1997), who argued that people with social phobia are characterised by a negatively biased internal representation of themselves, together with a tendency to preferentially focus attention towards that mental representation or supporting external evidence. In turn, the negative mental representation leads to overestimates of the likelihood and consequences of negative evaluation from others, thereby leading to anxious symptomatology and behaviours (such as subtle avoidance) that ultimately provide support for the negative mental representation.

A growing literature has provided correlational support for specific aspects of these models. Most commonly, research has shown that people with social phobia report higher or more extreme levels of the various cognitive processes than do people low in social anxiety. For example, a wealth of research has shown that people with social phobia report a greater expectation for negative evaluation than others (Foa, Franklin, Perry, & Herbert, 1996; Halford & Foddy, 1982; Hirsch & Clark, 2004; Wilson & Rapee, 2005a). People with social phobia have also been shown to hold a more negative view of themselves than people without social phobia. Several studies have shown that highly socially anxious individuals underestimate their public speaking and interacting abilities to a greater extent than low anxious individuals or even those with other anxiety disorders (Alden & Wallace, 1995; Rapee & Lim, 1992; Stopa & Clark, 1993). These biases have been shown to extend to perceptions of one's own appearance as well as other social performance abilities (Alden & Wallace, 1995; Montgomery, Haemmerlie, & Edwards, 1991; Rapee & Abbott, 2006). Evidence for biased attentional focus in social anxiety has provided slightly less consistent results, possibly due to suggestions that social anxiety is characterised by attention that is divided between an excessive internal and selective external focus (Rapee & Heimberg, 1997). Some research has shown that socially anxious individuals report higher levels of self-consciousness (Hope, Gansler, & Heimberg, 1989) or take an observer perspective with respect to their own performance (Hackmann, Surawy, & Clark, 1998; Spurr & Stopa, 2003). However, using information-processing methodology, research has variously shown that people with social phobia demonstrate preferential attentional focus either towards negative facial expressions (Gilboa-Schechtman, Foa, & Amir, 1999; Mogg & Bradley, 2002; Pishyar, Harris, & Menzies, 2004; Sposari & Rapee, 2007) or away from emotional facial expressions (Chen, Ehlers, Clark, & Mansell, 2002; Mansell, Clark, Ehlers, & Chen, 1999). Finally, following a social interaction, research has shown that socially anxious people engage in more extensive negative rumination than do controls (Abbott & Rapee, 2004; Dannahy & Stopa, 2007; Edwards, Rapee, & Franklin, 2003; Mellings & Alden, 2000).

Hence considerable evidence supports the overall suggestion that individuals high in social anxiety will engage in these cognitive processes to a greater extent than those lower in social anxiety. However, the real strength of cognitive models of social anxiety is the prediction that these cognitive processes will mediate between an individual's characteristic (trait) level of social anxiousness and his or her experienced anxiety in a social performance situation (Rapee & Heimberg, 1997). In other words, the models predict that a socially anxious person will experience heightened anxiety in a situation of social evaluation “because” she or he processes information in these particular ways. Empirical evidence testing these mediational relationships has been sparse.

Some research that has manipulated the negativity of self-images has suggested that these negative perspectives may play a causal role in maintaining social anxiety (Hirsch, Clark, Mathews, & Williams, 2003; Hirsch, Mathews, Clark, Williams, & Morrison, 2006). While this research has gone beyond demonstration of simple mediation, the implication is that socially anxious people hold more negative mental representations of themselves in social situations and, in turn, these negative representations lead to heightened anxiety. In contrast, demonstrating the casual status of self-consciousness in the production of social anxiety has been difficult (Bögels, Rijsemus, & De Jong, 2002; Woody & Rodriguez, 2000) and has questioned the extent to which biases in self-focussed attention may mediate between trait social anxiety and situational anxiety. Finally, some research has indicated that the association between social anxiety and later negative rumination may be mediated by an individual's perception of their performance immediately following delivery of a speech (Abbott & Rapee, 2004; Perini, Abbott, & Rapee, 2006).

Given the relative lack of data examining the relationships between cognitive features of social anxiety and the lack of data demonstrating that these cognitive variables mediate between characteristic social anxiety and anxious responding in social situations, the current study aimed to test some of these relationships. Specifically, participants with social phobia were required to present a brief impromptu speech and reported on several theoretically relevant cognitive constructs as well as on the degree of state anxiety experienced during the speech. One week later participants completed two further measures, a measure of the extent to which they ruminated in a negative fashion over the speech throughout the week and a measure of their recall of the quality of their speech performance. Two path analyses were conducted: first to test prediction of variance in state anxiety during the speech and second to test prediction of variance in speech performance recall one week later.