- Email: [email protected]
What determines observer-rated social performance in individuals with social anxiety disorder?
Journal of Anxiety Disorders 24 (2010) 830–836
Contents lists available at ScienceDirect
Journal of Anxiety Disorders
What determines observer-rated social performance in individuals with social anxiety disorder? Stephan Stevens a , Mareike Hofmann a , Sonja Kiko b , Anna Katharina Mall b , Regina Steil c , Martin Bohus b,∗,1 , Christiane Hermann a,∗∗,1 a b c
University of Giessen, Germany Central Institute of Mental Health, Mannheim, Germany University of Frankfurt, Germany
a r t i c l e
i n f o
Article history: Received 3 March 2010 Received in revised form 31 May 2010 Accepted 4 June 2010 Keywords: Social anxiety disorder Social performance Determinants
a b s t r a c t Clark and Wells (1995) proposed that cognitive variables and safety behaviors are related to social performance in social anxiety disorder (SAD). Here, we tested this relationship by concurrent assessment of cognitive, behavioral, and physiological variables and social performance in a prototypical social interaction situation. 103 participants with SAD and 23 healthy controls interacted with a confederate. Anxiety, self-focused attention, cognitions, and safety behaviors were assessed by self-report and by confederate ratings. Social performance was evaluated by independent observers using a behavioral coding system. Social performance was predicted using two regression models for self-report and confederate ratings. Between-group differences in social performance disappeared when talking time was taken into account. Talking time emerged as the most powerful predictor of social performance (54% and 58% accounted variance). Positive cognitions, self-focused attention and safety behaviors accounted for an additional, but marginal amount of variance. Reduced talking time might represent a safety behavior and may be considered an easy to measure final common behavioral outcome of cognitive processes underlying social anxiety. © 2010 Elsevier Ltd. All rights reserved.
1. Introduction Social performance refers to overt behavior in a social situation that is observable to others and that likely is a primary source of information for others’ judgment. By contrast, social skills are defined as the knowledge and availability of behaviors that the individual can flexibly and appropriately adjust depending on the social situation (Fydrich & Bürgener, 1999). Hopko, McNeil, Zvolensky, and Eifert (2001) have suggested that the term “social performance” should be used when describing an individual’s behavior in observational studies, because the term “social skill” not only refers to the actual behavior, but implies that the individual may not be able to show adequate behavior despite having the behavior repertoire and knowledge (see also Bögels & Voncken, 2008). Observational
∗ Corresponding author at: Department of Psychosomatic Medicine, Central Institute of Mental Health Mannheim, Square J5, D-68159 Mannheim, Germany. Tel.: +49 621 1703 4001; fax: +49 621 1703 4005. ∗∗ Corresponding author at: Department of Clinical Psychology and Psychotherapy, Justus-Liebig-University Giessen, Otto-Behaghel-Str. 10F, D-35397 Giessen, Germany. Tel.: +49 641 9926081; fax: +49 641 9926099. E-mail addresses: [email protected] (M. Bohus), [email protected] (C. Hermann). 1 Both of these authors contributed equally to this manuscript. 0887-6185/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.janxdis.2010.06.005
studies do not usually allow distinguishing between lack of ability and situational inhibiting factors (e.g., anxiety). Indeed, social performance may be more relevant for understanding and treating social anxiety because it determines how an individual is perceived by others (Alden & Taylor, 2004). A standard approach in examining putative differences in social performance between individuals with and without social anxiety is exposing participants to common social situations such as a speech or an interaction. Discrepancies between self versus other ratings of social performance have supported the notion of a biased perception of the self as suggested by modern cognitive theories of social anxiety disorder (SAD; Clark & Wells, 1995; Rapee & Heimberg, 1997). Most consistently, individuals with SAD underestimate their actual performance when compared to observer performance ratings in speech (Rapee & Lim, 1992) as well as in interaction situations (Stopa & Clark, 1993). Moreover, socially anxious and nonanxious participants have been compared with regard to specific overt behaviors (e.g., gaze contact, pauses during speech: Hofmann, Gerlach, Wender, & Roth, 1997) or overall impression of their performance (Norton & Hope, 2001) as perceived by independent observers or the confederate. Individuals with SAD seem to perform worse in social interaction situations compared to healthy participants (Baker & Edelmann, 2002; Norton & Hope, 2001; Voncken & Bogels, 2008), participants with other
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
anxiety disorders (Fydrich, Chambless, Perry, Buergener, & Beazley, 1998) and individuals with dysthymia (Norton & Hope, 2001). For speech tasks, the findings are more complex. Some studies suggest lower performance in individuals with SAD compared to controls (Moscovitch & Hofmann, 2007), others failed to detect differences (Voncken & Bogels, 2008). In their review on the effects of social anxiety on social performance, Strahan and Conger (1999) underline that the performance of socially anxious individuals is often comparable to the one of healthy controls. They propose that this may reflect the highly idiosyncratic nature of situationally elicited fear. Between group differences may not consistently be obtained due to high within-group variability with regard to elicited fear. They also suggest that social performance is disrupted depending on cognitive and physiological arousal. Consistent with this view, current cognitive models of social anxiety assume that actual social performance should be reduced due to anxiety-related inhibition of situationally adequate behaviors (Clark & Wells, 1995; Rapee & Heimberg, 1997) and therefore is likely to depend on the specific situation (e.g., Beidel, Turner, & Dancu, 1985; Rapee & Lim, 1992). Clark and Wells (1995) proposed that individuals rely on safety behaviors to cope with anxiety related symptoms, which are subsequently interpreted by others as unfriendly or arrogant (Alden & Wallace, 1995). Hence, safety behaviors and cognitive factors, as suggested in the Clark and Wells model, may account for the variance of social performance across situations. If anxiety inhibits social performance, observed social performance should depend on the severity of the cognitive, behavioral and physiological anxiety response. Interestingly, the relationship between model-derived cognitive, behavioral, and physiological variables and social performance has not been studied extensively. It is unclear which factors contribute to social performance as perceived by others. In socially anxious individuals participating in an opposite-sex interaction, higher self and confederate ratings of anxiety were related to lower perceived social performance (Beidel et al., 1985). In a conversation task, self-reported negative cognitions were the only significant predictor of a social performance score as rated by independent video raters (Norton & Hope, 2001). Furthermore, greater selffocused attention has been related to low social performance in a speech as rated by the participant and the audience, but only when the participants lacked confidence in their social skills prior to the task (Burgio, Merluzzi, & Pryor, 1986). Finally, self-reported anxiety and physiological arousal have been associated with lower observer rated social performance, but only when impression management demands were low (Sheffer, Penn, & Cassisi, 2001). To summarize, there is some evidence for a potential influence of situational anxiety, cognitive processes, safety behaviors and selffocused attention on how an individual is perceived by others in a social situation. However, more generalized conclusions are difficult because, across studies, the reporting source of the predictor and the criterion variables (participants themselves, confederates, independent observers) greatly varied as did the measure used for determining social performance (single-item, standard behavioral ratings systems). In the present study, participants with SAD and healthy controls (HC) participated in an interaction task. Based on the Clark and Wells (1995) model, core cognitive (anxiety, negative/positive cognitions, and self-focused attention), behavioral (safety behaviors) and physiological (heart rate, perceived physical symptoms) variables were measured by self-report of the participant and by confederate report. Since the ratings of the various variables may affect the rating of perceived social performance and vice versa, independent observers assessed social performance using a standard behavioral coding system. The primary goal of this study was to evaluate variables derived from the cognitive model as predictors of observer rated social performance, taking into account
831
Table 1 Sociodemographic data and clinical characteristics of individuals with SAD and healthy controls (HC).
Gender (N, % male) Age (years) Years of education CES-D SPS SIAS FNE
SAD (n = 103) M (SD)
HC (n = 23) M (SD)
42 (40.8%) 37.8 (10.2) 11.73 (2.06) 21.7 (11.1) 35.4 (14.1) 46.3 (13.5) 63.7 (9.9)
9 (45%) 38.2 (9.2) 12.24 (1.3) 6.7 (3.8) 9.1 (8.4) 19.5 (10.6) 36.8 (9.2)
2 (1,123) = .12, n.s. t(1,121) = .1, n.s. t(1,121) = .2, n.s. F(1,123) = 27.09** F(1,123) = 71.36** F(1,123) = 81.42** F(1,123) = 117.69**
Note: CES-D: Center for Epidemiological Studies-Depression Scale; SIAS: Social Interaction Anxiety Scale; SPS: Social Phobia Scale; FNE: Fear of Negative Evaluation Scale. **p < .001.
the reporting source, i.e. the participant him/herself or the confederate as a proxy of the observer’s perspective. Specifically, we predicted individuals with SAD to perform worse during social interaction compared to HC using a standardized behavioral coding system. Furthermore, as suggested by Clark and Wells (1995), we expected individuals with SAD to report more dysfunctional cognitions, greater self-focused attention, as well as more safety behaviors, and to be more physiologically aroused which in turn would impair social performance. We further expected that these cognitive, behavioral and physiological variables would emerge as significant predictors for social performance both when assessed by self-report and when judged by the confederate.
2. Method 2.1. Participants 103 participants with SAD were recruited by newspaper advertisements asking for people aged from 18 to 60 with fear of social situations and offering cognitive-behavioral treatment for participating in the Social Phobia Intervention Study of Mannheim (SOPHISMA). Given the primary focus on the treatment study, only 23 HC were recruited randomly from a list served by the registration office of Mannheim, Germany, for purposes of comparison. There were no significant group differences with regard to age, gender and years of education (Table 1). Interested persons were contacted for a telephone screening and were then invited for a structured clinical interview (duration: approx. 2 h), conducted by three trained clinical psychologists. Axes I and II disorders according to DSM-IV were assessed using the German version of SCID-I interview (Wittchen, Wunderlich, Gruschwitz, & Zaudig, 1997) and the German version of SCID-II interview for personality disorders (Fydrich et al., 1997). Inter-rater reliability for SAD diagnosis resulted in a kappa coefficient of .7. Thirty-seven participants with SAD met DSM-IV criteria for Avoidant Personality Disorder (APD). Inclusion criterion for the SAD group was a primary diagnosis of current SAD according to DSM-IV (minimum duration one year). No current Axis I or lifetime Axis II disorder was allowed for HC. Further exclusion criteria for the SAD group were a lifetime diagnosis of schizophrenia, bipolar disorder, anorexia nervosa, current substance abuse or dependence, current suicidal crises or psychological intervention. Twenty-eight participants with SAD had a comorbid affective disorder (major depression or dysthymia; 27.2%), and 13.6% at least one additional anxiety disorder. The study was approved by the Research Ethics Board of the University of Heidelberg and each subject provided written informed consent after the procedures had been fully explained.
832
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
2.2. Questionnaires To assess social anxiety severity, participants filled out the Social Interaction Anxiety Scale (SIAS; Stangier, Heidenreich, Berardi, Golbs, & Hoyer, 1999) and the Social Phobia Scale (SPS; Stangier et al., 1999). SIAS (20 items) measures interaction anxiety (e.g., “Talking to strangers makes me nervous”), SPS (20 items) measures anxiety in performance situations (e.g., “It would be difficult for me to drink in a small group”) using a five point Likert scale ranging from 0 (“not at all typical for me”) to 4 (“extremely typical for me”). The Fear of Negative Evaluation Scale (FNE; Leary & Meadows, 1991; German version: Vormbrock & Neuser, 1983) contains 20 items (e.g., “I am dealing to much with thoughts about what others might think about me”) to be rated on a four point Likert scale ranging from 1 (is never the case) to 4 (is always the case). For measuring depressive symptoms, the Center of Epidemiological Studies-Depression Scale (CES-D; Radloff, 1977; German Version: Hautzinger & Bailer, 1993) was used. The CES-D (15 items; e.g., “During the last week, I talked less than before”) assesses depressive symptoms during the last week on a four point Likert scales ranging from 0 (rarely) to 3 (mostly). 2.3. Measures obtained in the behavioral test 2.3.1. Self-report Anxiety level: A Visual Analogue Scale (VAS) was used to assess self-report anxiety. Participants indicated how anxious they felt from 0 (“no anxiety at all”) to 100 (“worst anxiety I can imagine”). Perceived physical symptoms: Participants received a list of 10 physical sensations including heart beating, sweating, trembling, breathlessness, pain or pressure in the chest, nausea/feeling queasy, dizziness/numbness/feeling faint, fear of losing control, blushing and muscle tension. They indicated on a 5-points Likert scale ranging from 0 (not at all present) to 4 (very much present), how much they perceived these symptoms. Safety behaviors: The German version of the Social Behaviour Questionnaire (SBQ; Clark, 1995; German: Stangier et al., 1996) was used to assess safety behaviors. From the original 27 items, 5 items were omitted because they did not fit the tasks (e.g., “I drink alcohol to cope with the situation”). On a 4-point Likert scale ranging from 0 (not at all) to 3 (very much), participants answered how much they had engaged in the listed behaviors (e.g., “I will avoid eye contact”). (Dys-)Functional cognitions: Cognitions were assessed with the Social Interaction and Self-Statement Test (SISST; Glass, Merluzzi, Biever, & Larsen, 1982), a thirty item questionnaire (e.g., “The confederate will not be interested in me”) measuring the frequency of positive and negative cognitions during conversation on a 4-point Likert scale ranging from 0 (occurred almost never) to 4 (occurred very often). Focus of attention (FAQ): The Focus of Attention Questionnaire Self-Scale (FAQ-self; Woody, 1996) was used to assess internally focussed attention (10 items; e.g., “I attended to my bodily reactions (e.g., my heartbeat)”). Participants judged on a 5-point Likert scale ranging from 0 (not at all) to 4 (very much), how much they attended to different aspects of the self. 2.3.2. Confederate report Anxiety level: Confederates rated the perceived anxiety level of the participant on a 6-point Likert scale ranging from 0 “not at all” to 5 “very much.” Self-focused-attention: Confederates were asked to estimate the extent of self-focused attention ranging from 0 (=not at all) to 100 (=fully self-focused). Safety behaviors: Using the SBQ (Stangier et al., 1996), which was modified for the use in observers (SBQ-other), confederates rated the intensity of safety behaviors they noted. Consistent with
the self-report version, the SBQ-other contains a list of 22 safety behaviors. The confederate rated on a 4-points Likert scale ranging from 0 (not at all) to 3 (very much), how much the participant engaged in each of these behaviors (e.g., “I will avoid eye contact”). 2.3.3. Behavioral and physiological variables Total talking time: The total time of talking (in seconds) during the interaction was stop watched via video analysis after the experiment. Observer rated social performance: The “Ratingskala für soziale Kompetenz” (Fydrich & Burgener, 1999; Fydrich et al., 1998) is a standardized rating system for social performance in social interaction situations and represents the German version of the Social Performance Rating Scale (SPRS; Trower, Bryant, & Argyle, 1978). Gaze direction, voice quality, speech pressure (monosyllabic versus responses preventing discourse), discomfort (agitation and nervousness) and conversation flow were rated on 5-point Likert scales (1 = very good; 5 = very poor; consistent with German school grades), with higher scores indicating less social performance. The total score was calculated as the sum score of the five subscales. Raters were trained using 10 SPRS training videos, only accepting raters (N = 7) achieving a satisfactory level of agreement (r > .9) with standard ratings. Observers were blind to the participants’ diagnosis. Heart rate. With a sample rate of 2048-Hz an electrocardiogram was recorded from two electrodes attached to the chest with the Active Two recorder (www.biosemi.com). One electrode was placed on the right collarbone and one on the lowest rib on the left side while participants were seated during interaction. Heart rate was determined using ANSLAB software (www.psycho.unibas.ch/psychophysiologie/anslab.html). For the interaction, the average heart rate was computed. 2.4. Behavioral test and procedure After the clinical interview, participants were given a set of questionnaires (SIAS, SPS, SPAI, and CES-D) to be completed at home and were scheduled for the laboratory session. The laboratory session took about 2–3 h and was run by a same-sex experimenter. Participants were asked to complete two behavioral tests. For the purpose of this study, only the test involving a social interaction is reported since the German version of the SPRS has only been validated for coding behaviors during a social interaction. The interaction task was to initiate and maintain a conversation sitting in a train with an opposite-sex stranger. The interaction task was always the first behavioral test and consisted of four phases: sitting quietly (5 min), anticipating the conversation (1.5 min), introduction of the confederate and role playing the situation (5 min) and the recovery after the conversation (10 min). The role play partners always were opposite-sex confederates, all of them being trained psychology undergraduate students. The confederates entered the room after the anticipation phase. Confederates were trained to behave friendly but passively, not to start the conversation and not to say more than three sentences at one time. The role play partners were allowed to ask questions or to revive the conversation if the participant did not say anything for longer than thirty seconds. After the end of the conversation, the confederate left the room and completed the ratings and questionnaire. All dependent measures except for heart rate were assessed immediately after the conversation. Average heart rate during the conversation was computed offline. 2.5. Data analysis A multivariate Analysis of Variance (MANOVA) followed by univariate ANOVAs was conducted to examine between group dif-
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836 Table 2 Means (standard deviations) of self-reported and confederate rated cognitive, behavioral, and physiological measures for participants with social anxiety disorder (SAD) and healthy controls (HC). SAD M (SD)
HC M (SD)
Self-report Anxiety (VAS) Self-focused attention (FAQ) Safety behaviors (SBQ) Positive cognitions (SISST) Negative cognitions (SISST) Perceived body symptoms (0–40)
37.0 (26.7) 20.00 (8.0) 26.4 (8.8) 19.5 (10.5) 30 (13.5) 10.00 (13.1)
19.4 (26.9) 11.00 (8.0) 17.6 (6.6) 25.5 (9.0) 10.5 (12.0) 3.00 (2.00)
Confederate rating Anxiety (0–5) Self-focused attention (0–100) Safety behaviors (SBQ-other)
3.9 (2.52) 53.03 (21.51) 20.61 (9.46)
0.39 (0.87) 36.67 (19.51) 10.54 (3.96)
Absolute talking time (s) Heart rate (bpm) during interaction
118.6 (55.1)
159.8 (39.3)
83.2 (13.2)
85.5 (10.9)
833
p < .001), being more self-focused (F(1,123) = 7.7, p < .001) and being more anxious as compared to controls (F(1,123) = 31.13, p < .001) (Table 2). Social performance and talking time: Social performance of participants with SAD (M = 13.5, SD = 4.2) was rated as significantly poorer than that of HC (M = 10.1, SD = 3.2; F(1,123) = 7.5, p = .01, p 2 = .07; higher scores reflecting lower performance). As individuals with SAD talked significantly less compared to HC (F(1,123) = 7.64, p = .007), this analysis was repeated with absolute talking time as covariate. The between group effect failed to reach significance when considering absolute talking time (F(1,123) = 1.40, p = .24, p 2 = .01), which was revealed as significant covariate (F(1,123) = 81.92, p < .001, p 2 = .42). Heart rate: No group differences emerged for heart rate (F(1,90) = .31, p = .58). 3.3. Relationship between self- and confederate report
Note: VAS: Visual Analoque Scale; FAQ: Focus of Attention Questionnaire; SBQ: Safety Behavior Questionnaire; SISST: Social Interaction and Self-Statement Test.
Self- and confederate ratings for anxiety, safety behaviors, and self-focused attention were significantly correlated at a moderate to high level (Table 3).
ferences (SAD versus HC) with regard to all self- and confederate rated variables. The relationship between self- and confederate report was evaluated using Pearson’s correlations. For predicting observer rated social performance, two regression analyses were conducted using either the self-report or the confederate ratings as predictor variables. An alpha level of .05 was used for all statistical tests.
3.4. Prediction of observer-rated social performance
3. Results 3.1. Clinical measures The MANOVA revealed a main effect of group (F(4,123) = 38.87, p < .001; p 2 = .6), with univariate follow-up ANOVAs indicating that individuals with SAD reported significantly more social fears (SIAS, SPS), fear of negative evaluation (FNE) and depressive symptoms (CES-D) compared to HC (p’s < .05; Table 1). 3.2. Behavioral test Self-report ratings: Using all self-report cognitive, behavioral and physiological variables measured during the behavioral test as dependent measures, the MANOVA revealed a significant between group effect (F(8,123) = 5.2, p < .001; p 2 = .28). Participants with SAD endorsed higher self-reported anxiety (F(1,123) = 18.91, p < .001), fewer positive cognitions (F(1,123) = 7.48, p < .01), more negative cognitions (F(1,121) = 43.38, p < .001), a greater internal focus of attention (F(1,123) = 22.09, p < .001), more safety behaviors (F(1,118) = 29.59, p < .001) and perceived more physical symptoms (Table 2). Confederate ratings: The overall MANOVA revealed a significant group effect (F(3,123) = 10.09, p < .001; p 2 = .22). Individuals with SAD were rated as showing more safety behaviors (F(1,123) = 16.52,
As absolute talking time differed between groups, hierarchical regressions were computed with absolute talking time entered in step one and the model variables entered in step two. Model 1—self-ratings as predictors: Absolute talking time emerged as the best predictor for observer rated social performance (F(1,123) = 80.32, p < .001; R2 = .49, Table 4). Above and beyond talking time, only self-reported positive cognitions further increased the amount of accounted variance (F(8,123) = 12.5, p < .001). However, although significant, the change in R2 was only marginal (R2 = .07). Model 2—confederate ratings as predictors: Consistent with the regression model using self-ratings, absolute talking time contributed most to observer rated social performance (F(1,123) = 90.43, p < .001; R2 = .48; Table 4). Confederate rated self-focused attention further increased the amount of explained variance (F(4,123) = 30.96, p < .001). Yet, R2 increased only slightly (R2 = .09). 3.5. Relationship between talking time and Clark and Wells (1995) variables As absolute talking time turned out to be the strongest predictor of observer rated social performance regardless of whether participants’ self-report or the confederate’s ratings were included as predictors, zero-order correlations between talking time and the participant’s self-report and the confederate’s ratings were computed. Longer talking time was significantly associated with a lower confederate’s rating of the participant’s safety behaviors (Table 5). A similar pattern was obtained for the participants’ self-ratings of anxiety, safety behaviors, and negative cognitions with the obtained correlation coefficients being small in size. Self-
Table 3 Bivariate correlations between self and confederate rated variables.
Self-report anxiety Self-report self-focused attention Self-report safety behaviors * **
p < .05. p < .001.
Confederate rating anxiety
Confederate rating self-focused attention
Confederate rating safety behaviors
.54** .36** .40**
.22* .35** .23*
.35** .26** .28**
834
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
Table 4 Regression analysis predicting observer rated social performance during conversation using participants self-report and confederate ratings. Self-report variable
B
Step 1 Absolute talking time (s)
SE B

R2
−.05
.01
−.68***
.49***
−.05 −.03 1.38 −1.14 .71 −.54 −.01 −.03
.01 .02 .87 .56 .57 .53 1.28 .03
−.64*** −.20 .22 −.17** .15 −.11 .00 −.1
.54***
Confederate ratings Step 1 Absolute talking time (s)
−.05
.01
−.67***
.48***
Step 2 Absolute talking time (s) Anxiety (0–5) Self-focused attention (FAQ) Safety behaviors (SBQ-other)
−.04 −.70 .05 1.90
.01 .45 .02 1.14
−.52*** −.14 .25** .19*
Step 2 Absolute talking time (s) Anxiety (VAS) Perceived physical symptoms Positive cognitions (SISST) Negative cognitions (SISST) Self-focused attention (FAQ) Safety behaviors (SBQ) Heart rate (bpm)
.58***
Note: VAS: Visual Analogue Scale; FAQ: Focus of Attention Questionnaire; SBQ: Safety Behavior Questionnaire; SISST: Social Interaction and Self-statement Test. ***p < .001, **p < .05, *p < .1.
Table 5 Zero-order correlations among self-report and confederate rated variables, absolute talking time and social performance. Social performance
Talking time
Self-report Anxiety (VAS) Self-focused attention (FAQ) Safety behaviors (SBQ) Positive cognitions (SISST) Negative cognitions (SISST) Perceived body symptoms (0–40)
.23* .20* .24* −.20* .30* .13
−.24* −.19* −.20* .18 −.23* −.11
Confederate rating Anxiety (0–5) Self-focused attention (0–100) Safety behaviors (SBQ-other) Heart rate (bpm) during interaction Social performance
.47** .56** .57** −.1 –
−.46** −.48** −.48** .14 .67**
Note: VAS: Visual Analogue Scale; FAQ: Focus of Attention Questionnaire; SBQ: Safety Behavior Questionnaire; SISST: Social Interaction and Self-statement Test. ***p < .001,**p < .05, *p < .1.
reported positive cognitions were significantly related to greater observer-rated social performance, though at a low level. 4. Discussion Individuals with SAD and HC participated in a social getting acquainted task. As expected, participants with SAD reported more anxiety, more physical symptoms, negative cognitions, a greater self-focus, more safety behaviors, and less positive cognitions than HC. Yet, heart rate did not differ between groups. Participants’ selfreport was echoed by the confederate’s ratings which revealed a significant difference between SAD participants and HC regarding anxiety, focus of attention and safety behaviors. The agreement between self-report and confederate anxiety rating was high, the correlations between self- and confederate rated self-focused attention and safety behaviors were moderate in size. As expected, social performance of the individuals with SAD was significantly lower as rated by independent observers using a well-established behavioral coding system which takes into account several specific behaviors (e.g., gaze) that an observer has to score to yield a
measure of social performance. Intriguingly, when controlled for absolute talking time, the behavioral observation ratings did no longer significantly differ between SAD and HC. The importance of absolute talking time is further underlined by the finding that it emerged as the best predictor of observer rated social performance, regardless of whether self-report or confederate rated cognitive and behavioral variables as derived from the cognitive model of SAD by Clark and Wells (1995) were included in the regression model. Furthermore, there was only a marginal increase in explained variance when self-rated negative cognitions were included. Similarly, safety behaviors and self-focused attention as observed by the confederate added only marginally to the amount of accounted variance in social performance. The finding that absolute talking time was the most powerful predictor of social performance as perceived by an observer is rather surprising. Clearly, this suggests that the time a person spends talking during a social interaction and the way he/she will be perceived as closely related. Although this finding is contrary to the suggestion that social performance is not equivalent to mere talking time (Fydrich et al., 1998), observer rated social performance seems to be highly related to the time the individual talks. Previous studies did not explicitly control for talking time when examining social performance in SAD, but some studies assessed the amount of time the participants talked. Baker and Edelmann (2002) found no difference between participants with SAD and HC in absolute talking time in a speech. However, Hofmann et al. (1997) reported individuals with SAD to pause more frequently than HC during a speech, which was interpreted as worse social performance. For interaction situations, only studies using analogue samples are available. Highly socially anxious individuals talked less than nonanxious individuals (Burgio et al., 1986) and time to their first utterance was significantly longer (Thompson & Rapee, 2002). In the Thompson and Rapee (2002) study, time to first utterance was used as an objective measure in addition to the behavioral ratings of social performance by two independent observers. Interestingly, high socially anxious people were rated as performing worse in an unstructured social interaction and needed more time until their first utterance. Yet, the relationship between the delayed onset of a conversation (and presumably a concomitant shorter talking time) and perceived social performance was not further investigated in this study. Our results suggest that, at least in interaction situations, differences in
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
observer perceived social performance might be accounted for by differences in absolute talking time. In line with such a conclusion, an earlier study focusing on predictors of successful dating using behavioral observation of high and low frequency daters during a conversation showed that for males absolute talking time, and for females eye contact and talking time were the best predictors of global social performance as judged by independent observers using a one item rating (Glasgow & Arkowitz, 1975). The putative pivotal role of absolute talking time in accounting for perceived social performance raises several issues. First, talking time may represent a meaningful behavior in and by itself. Individuals with SAD seem to talk less than HC. Talking less may function as a safety or avoidance behavior in an attempt to reduce the risk of making foolish or inappropriate statements (Thompson & Rapee, 2002). This interpretation is in line with the cognitive model of SAD by Clark and Wells (1995), assuming that safety behaviors directly influence perceived social performance. Due to being easy to observe, talking time may be particularly detrimental to the impression observers may get. Second, absolute talking time may be considered an easy to measure proxy of social performance. Possibly, talking time reflects the combined inhibitory effect of activated dysfunctional cognitions, self-focused attention, anxiety and safety behaviors. This hypothesis is indirectly supported by the obtained moderate to large correlations between talking time and the confederate ratings of these variables. Notably, a similar pattern of correlations was obtained for the participants’ self-ratings, with these correlations being smaller in magnitude as is to be expected due to the different reporting source. Clearly, the relationship between talking time and the cognitive-behavioral anxiety response does not allow conclusions about the underlying direction of influence. Whether talking time mediates the disruptive influence of the cognitive, behavioral, and physiological anxiety response on social performance needs to be studied in more detail. Possibly, talking less may also further activate dysfunctional cognitions, safety behaviors, and physiological arousal. Third, our findings raises questions about how social performance is best operationalized when assessed. Here, observers rated five different behaviors (e.g., voice quality, gaze direction, and conversation flow). Yet, talking time alone best accounted for overall perceived social performance. Possibly, talking time overshadows the perception of behavioral nuances even though observers may not be aware of this effect, underlining the need to assess absolute talking time in future studies. Few cognitive variables as outlined by cognitive models of SAD were found to contribute to perceived social performance above and beyond absolute talking time. Positive cognitions were an additional significant predictor in the self-rating model, confederate rated self-focused attention, and safety behaviors emerged as predictors in the confederate rating model. Clearly, in either model, the increase in accounted variance of perceived social performance was marginal. Consistent with our findings, Beidel et al. (1985) also observed a negative association between positive cognitions and social performance suggesting that adaptive behavior in a social situation is promoted by positive cognitions. Our findings further suggest that self-focused attention and safety behaviors (both triggered by dysfunctional cognitions that are not readily observable) as perceived by the confederate are associated with lower social performance as perceived by an observer (for a review Alden & Taylor, 2004; Papsdorf & Alden, 1998; Voncken, Alden, Bogels, & Roelofs, 2008). Hence, as suggested by Clark and Wells (1995), social performance is compromised if self-focused attention is high and safety behaviors are shown. Somewhat surprisingly, situational anxiety did not predict observer-rated social performance, at least when talking time was included as predictor. This is consistent with the notion that individuals with SAD are prone to overestimate the visibility of their anxiety symptoms. Nonetheless, situational anxi-
835
ety is important since it is associated with less talking time (Table 5), and possibly with other safety behaviors which, in turn, lets people appear less socially apt. Our study has some limitations which have to be addressed. First, for practical reasons, not all videos were rated by all independent observers. Hence inter-rater reliability for the social performance measure could not be determined. However, all observers were trained such that they achieved an agreement with a standardized training video of greater than .9. Second, the Selfreport Anxiety Scale and the confederate anxiety rating did not have the same format. This was primarily due to the fact that the participants’ ratings had to be consistent with other scales used in the larger treatment study. Yet, the result pattern for self-report and confederate rated cognitive model variables is quite consistent, with the lower correlations observed for the self-report data most likely reflecting that confederate and observers both rated the participants from the “other” perspective. Using the confederate ratings as a proxy for the “others” perspective was necessary in order to prevent mutual confounding influences of rating social performance using a behavioral coding system and rating the core variables of the Clark and Wells’ (1995) model of SAD. In summary, this is one of the first studies assessing cognitive, behavioral, and physiological variables derived from the cognitive model of SAD by Clark and Wells (1995) in a social interaction situation to predict observer ratings of social performance. Positive cognitions, self-focused attention and safety behaviors were associated with lower perceived social performance as proposed by Clark and Wells (1995). As lower social performance might contribute to interpersonal problems in individuals with SAD, our results underline the importance of interventions targeting at reducing self-focused attention, safety behaviors, and dysfunctional cognitions. As suggested by the obtained zero-order correlations, reduced self-focused attention, fewer safety behaviors, and less negative cognitions may be beneficial by increasing talking time which, in our study, emerged as the strongest predictor of observer-perceived social performance. In fact, talking time might be a useful and easy to measure index of social performance as perceived by others and may represent the final common behavioral outcome of the cognitive mechanisms underlying social anxiety. Acknowledgement This study was funded by a grant of the Deutsche Forschungsgemeinschaft (BO 1487/7) to MB and CH. References Alden, L. E., & Taylor, C. T. (2004). Interpersonal processes in social phobia. Clinical Psychology Review, 24, 857–882. Alden, L. E., & Wallace, S. T. (1995). Social phobia and social appraisal in successful and unsuccessful social interactions. Behaviour Research and Therapy, 33, 497–505. Baker, S. R., & Edelmann, R. J. (2002). Is social phobia related to lack of social skills? Duration of skill-related behaviours and ratings of behavioural adequacy. British Journal of Clinical Psychology, 41, 243–257. Beidel, D. C., Turner, S. M., & Dancu, C. V. (1985). Physiological, cognitive and behavioral aspects of social anxiety. Behaviour Research and Therapy, 23, 109–117. Bögels, S. M., & Voncken, M. (2008). Social skills training versus cognitive therapy for social anxiety disorder characterized by fear of blushing, trembling, or sweating. International Journal of Cognitive Therapy, 1, 138–150. Burgio, K. L., Merluzzi, T. V., & Pryor, J. B. (1986). Effects of performance expectancy and self-focused attention on social-interaction. Journal of Personality and Social Psychology, 50, 1216–1221. Clark, D. M. (1995). Social Behaviour Quesstionnaire (SBQ). Unpublished Work. Department of Psychiatry, Oxford University, Oxford, UK. Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In: R. G Heimberg, M. R. Liebowitz, D. A. Hope, & F. R. Schneier (Eds.), Social phobia: diagnosis, assessment and treatment. New York: Guilford. Fydrich, T., & Bürgener, F. (1999). Ratingskalen für soziale Kompetenz. In: J. Margraf, & K. Rudolf (Eds.), Soziale Kompetenz-Soziale Phobie: Anwendungsfelder,
836
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
Entwicklungslinien, Erfolgsaussichten. Hohengehren: Baltmansweiler: Schneider, pp. 81–96 Fydrich, T., Chambless, D. L., Perry, K. J., Buergener, F., & Beazley, M. B. (1998). Behavioral assessment of social performance: a rating system for social phobia. Behaviour Research and Therapy, 36, 995–1010. Fydrich, T., Renneberg, B., Schmitz, B., & Wittchen, H.-U. (1997). Strukturiertes Klinisches Interview für DSM-IV Achse II. Persönlichkeitsstörungen (SKID II). Göttingen: Hogrefe. Glasgow, R. E., & Arkowitz, H. (1975). The behavioral assessment of male and female social competence in dyadic heterosexual interaction. Behavior Therapy, 6, 488–498. Glass, C. R., Merluzzi, T. V., Biever, J. L., & Larsen, K. H. (1982). Cognitive assessment of social anxiety—development and validation of a self-statement questionnaire. Cognitive Therapy and Research, 6, 37–55. Hautzinger, M., & Bailer, M. (1993). Allgemeine Depressions-Skala (ADS). Weinheim: Beltz. Hofmann, S. G., Gerlach, A. L., Wender, A., & Roth, W. T. (1997). Speech disturbances and gaze behavior during public speaking in subtypes of social phobia. Journal of Anxiety Disorders, 11, 573–585. Hopko, D. R., McNeil, D. W., Zvolensky, M. J., & Eifert, G. H. (2001). The relation between anxiety and skill in performance-based anxiety disorders: a behavioral formulation of social phobia. Behavior Therapy, 32, 185–207. Leary, M. R., & Meadows, S. (1991). Predictors, elicitors, and concomitants of social blushing. Journal of Personality and Social Psychology, 60, 254–262. Moscovitch, D. A., & Hofmann, S. G. (2007). When ambiguity hurts: social standards moderate self-appraisals in generalized social phobia. Behaviour Research and Therapy, 45, 1039–1052. Norton, P. J., & Hope, D. A. (2001). Kernels of truth or disorted perceptions: self and observer ratings of social anxiety and performance. Behavior Therapy, 32, 765–786. Papsdorf, M., & Alden, L. (1998). Mediators of social rejection in social anxiety: similarity, self-disclosure, and overt signs of anxiety. Journal of Research in Personality, 32, 351–369. Radloff, L. S. (1977). The CES-D Scale: a self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.
Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral model of anxiety in social phobia. Behaviour Research and Therapy, 35, 741–756. Rapee, R. M., & Lim, L. (1992). Discrepancy between self and observer ratings of performance in social phobics. Journal of Abnormal Psychology, 101, 728–731. Sheffer, C. E., Penn, D. L., & Cassisi, J. E. (2001). The effects of impression management demands on heart rate, self-reported social anxiety, and social competence in undergraduate males. Journal of Anxiety Disorders, 15, 171–182. Stangier, U., Ehlers, A., Heidenreich, T., & Clark, D. M. (1996). Fragebogen zu Sicherheitsverhalten in sozialen Situationen. Frankfurt am Main: Johann Wolfgang von Goethe Universität. Stangier, U., Heidenreich, T., Berardi, A., Golbs, U., & Hoyer, J. (1999). Die Erfassung sozialer Phobie durch die Social Interaction Anxiety Scale (SIAS) und die Social Phobia Scale (SPS). Zeitschrift für Klinische Psychologie, 28, 28–36. Stopa, L., & Clark, D. M. (1993). Cognitive-processes in social phobia. Behaviour Research and Therapy, 31, 255–267. Strahan, E. Y., & Conger, A. J. (1999). Social anxiety and social performance: why don’t we see more catastrophes? Journal of Anxiety Disorders, 13, 399–416. Thompson, S., & Rapee, R. M. (2002). The effect of situational structure on the social performance of socially anxious and non-anxious participants. Journal of Behavior Therapy and Experimental Psychiatry, 33, 91–102. Trower, P., Bryant, B., & Argyle, M. (1978). Social skills and mental health. Pittsburgh: University of Pittsburgh Press. Voncken, M. J., Alden, L. E., Bogels, S. M., & Roelofs, J. (2008). Social rejection in social anxiety disorder: the role of performance deficits, evoked negative emotions and dissimilarity. British Journal of Clinical Psychology, 47, 439–450. Voncken, M. J., & Bogels, S. M. (2008). Social performance deficits in social anxiety disorder: reality during conversation and biased perception during speech. Journal of Anxiety Disorders, 22, 1384–1392. Vormbrock, F., & Neuser, J. (1983). Konstruktion zweier spezifischer TraitFragebögen zur Erfassung von Angst in sozialen Situationen (SANB und SVSS). Diagnostica, 29, 98–106. Wittchen, H. U., Wunderlich, U., Gruschwitz, S., & Zaudig, M. (1997). SKID-I. Strukturiertes klinisches Interview für psychische Störungen für DSM-IV. Göttingen: Hogrefe. Woody, S. R. (1996). Effects of focus of attention on anxiety levels and social performance of individuals with social phobia. Journal of. Abnormal Psychology, 105, 61–69.
Contents lists available at ScienceDirect
Journal of Anxiety Disorders
What determines observer-rated social performance in individuals with social anxiety disorder? Stephan Stevens a , Mareike Hofmann a , Sonja Kiko b , Anna Katharina Mall b , Regina Steil c , Martin Bohus b,∗,1 , Christiane Hermann a,∗∗,1 a b c
University of Giessen, Germany Central Institute of Mental Health, Mannheim, Germany University of Frankfurt, Germany
a r t i c l e
i n f o
Article history: Received 3 March 2010 Received in revised form 31 May 2010 Accepted 4 June 2010 Keywords: Social anxiety disorder Social performance Determinants
a b s t r a c t Clark and Wells (1995) proposed that cognitive variables and safety behaviors are related to social performance in social anxiety disorder (SAD). Here, we tested this relationship by concurrent assessment of cognitive, behavioral, and physiological variables and social performance in a prototypical social interaction situation. 103 participants with SAD and 23 healthy controls interacted with a confederate. Anxiety, self-focused attention, cognitions, and safety behaviors were assessed by self-report and by confederate ratings. Social performance was evaluated by independent observers using a behavioral coding system. Social performance was predicted using two regression models for self-report and confederate ratings. Between-group differences in social performance disappeared when talking time was taken into account. Talking time emerged as the most powerful predictor of social performance (54% and 58% accounted variance). Positive cognitions, self-focused attention and safety behaviors accounted for an additional, but marginal amount of variance. Reduced talking time might represent a safety behavior and may be considered an easy to measure final common behavioral outcome of cognitive processes underlying social anxiety. © 2010 Elsevier Ltd. All rights reserved.
1. Introduction Social performance refers to overt behavior in a social situation that is observable to others and that likely is a primary source of information for others’ judgment. By contrast, social skills are defined as the knowledge and availability of behaviors that the individual can flexibly and appropriately adjust depending on the social situation (Fydrich & Bürgener, 1999). Hopko, McNeil, Zvolensky, and Eifert (2001) have suggested that the term “social performance” should be used when describing an individual’s behavior in observational studies, because the term “social skill” not only refers to the actual behavior, but implies that the individual may not be able to show adequate behavior despite having the behavior repertoire and knowledge (see also Bögels & Voncken, 2008). Observational
∗ Corresponding author at: Department of Psychosomatic Medicine, Central Institute of Mental Health Mannheim, Square J5, D-68159 Mannheim, Germany. Tel.: +49 621 1703 4001; fax: +49 621 1703 4005. ∗∗ Corresponding author at: Department of Clinical Psychology and Psychotherapy, Justus-Liebig-University Giessen, Otto-Behaghel-Str. 10F, D-35397 Giessen, Germany. Tel.: +49 641 9926081; fax: +49 641 9926099. E-mail addresses: [email protected] (M. Bohus), [email protected] (C. Hermann). 1 Both of these authors contributed equally to this manuscript. 0887-6185/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.janxdis.2010.06.005
studies do not usually allow distinguishing between lack of ability and situational inhibiting factors (e.g., anxiety). Indeed, social performance may be more relevant for understanding and treating social anxiety because it determines how an individual is perceived by others (Alden & Taylor, 2004). A standard approach in examining putative differences in social performance between individuals with and without social anxiety is exposing participants to common social situations such as a speech or an interaction. Discrepancies between self versus other ratings of social performance have supported the notion of a biased perception of the self as suggested by modern cognitive theories of social anxiety disorder (SAD; Clark & Wells, 1995; Rapee & Heimberg, 1997). Most consistently, individuals with SAD underestimate their actual performance when compared to observer performance ratings in speech (Rapee & Lim, 1992) as well as in interaction situations (Stopa & Clark, 1993). Moreover, socially anxious and nonanxious participants have been compared with regard to specific overt behaviors (e.g., gaze contact, pauses during speech: Hofmann, Gerlach, Wender, & Roth, 1997) or overall impression of their performance (Norton & Hope, 2001) as perceived by independent observers or the confederate. Individuals with SAD seem to perform worse in social interaction situations compared to healthy participants (Baker & Edelmann, 2002; Norton & Hope, 2001; Voncken & Bogels, 2008), participants with other
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
anxiety disorders (Fydrich, Chambless, Perry, Buergener, & Beazley, 1998) and individuals with dysthymia (Norton & Hope, 2001). For speech tasks, the findings are more complex. Some studies suggest lower performance in individuals with SAD compared to controls (Moscovitch & Hofmann, 2007), others failed to detect differences (Voncken & Bogels, 2008). In their review on the effects of social anxiety on social performance, Strahan and Conger (1999) underline that the performance of socially anxious individuals is often comparable to the one of healthy controls. They propose that this may reflect the highly idiosyncratic nature of situationally elicited fear. Between group differences may not consistently be obtained due to high within-group variability with regard to elicited fear. They also suggest that social performance is disrupted depending on cognitive and physiological arousal. Consistent with this view, current cognitive models of social anxiety assume that actual social performance should be reduced due to anxiety-related inhibition of situationally adequate behaviors (Clark & Wells, 1995; Rapee & Heimberg, 1997) and therefore is likely to depend on the specific situation (e.g., Beidel, Turner, & Dancu, 1985; Rapee & Lim, 1992). Clark and Wells (1995) proposed that individuals rely on safety behaviors to cope with anxiety related symptoms, which are subsequently interpreted by others as unfriendly or arrogant (Alden & Wallace, 1995). Hence, safety behaviors and cognitive factors, as suggested in the Clark and Wells model, may account for the variance of social performance across situations. If anxiety inhibits social performance, observed social performance should depend on the severity of the cognitive, behavioral and physiological anxiety response. Interestingly, the relationship between model-derived cognitive, behavioral, and physiological variables and social performance has not been studied extensively. It is unclear which factors contribute to social performance as perceived by others. In socially anxious individuals participating in an opposite-sex interaction, higher self and confederate ratings of anxiety were related to lower perceived social performance (Beidel et al., 1985). In a conversation task, self-reported negative cognitions were the only significant predictor of a social performance score as rated by independent video raters (Norton & Hope, 2001). Furthermore, greater selffocused attention has been related to low social performance in a speech as rated by the participant and the audience, but only when the participants lacked confidence in their social skills prior to the task (Burgio, Merluzzi, & Pryor, 1986). Finally, self-reported anxiety and physiological arousal have been associated with lower observer rated social performance, but only when impression management demands were low (Sheffer, Penn, & Cassisi, 2001). To summarize, there is some evidence for a potential influence of situational anxiety, cognitive processes, safety behaviors and selffocused attention on how an individual is perceived by others in a social situation. However, more generalized conclusions are difficult because, across studies, the reporting source of the predictor and the criterion variables (participants themselves, confederates, independent observers) greatly varied as did the measure used for determining social performance (single-item, standard behavioral ratings systems). In the present study, participants with SAD and healthy controls (HC) participated in an interaction task. Based on the Clark and Wells (1995) model, core cognitive (anxiety, negative/positive cognitions, and self-focused attention), behavioral (safety behaviors) and physiological (heart rate, perceived physical symptoms) variables were measured by self-report of the participant and by confederate report. Since the ratings of the various variables may affect the rating of perceived social performance and vice versa, independent observers assessed social performance using a standard behavioral coding system. The primary goal of this study was to evaluate variables derived from the cognitive model as predictors of observer rated social performance, taking into account
831
Table 1 Sociodemographic data and clinical characteristics of individuals with SAD and healthy controls (HC).
Gender (N, % male) Age (years) Years of education CES-D SPS SIAS FNE
SAD (n = 103) M (SD)
HC (n = 23) M (SD)
42 (40.8%) 37.8 (10.2) 11.73 (2.06) 21.7 (11.1) 35.4 (14.1) 46.3 (13.5) 63.7 (9.9)
9 (45%) 38.2 (9.2) 12.24 (1.3) 6.7 (3.8) 9.1 (8.4) 19.5 (10.6) 36.8 (9.2)
2 (1,123) = .12, n.s. t(1,121) = .1, n.s. t(1,121) = .2, n.s. F(1,123) = 27.09** F(1,123) = 71.36** F(1,123) = 81.42** F(1,123) = 117.69**
Note: CES-D: Center for Epidemiological Studies-Depression Scale; SIAS: Social Interaction Anxiety Scale; SPS: Social Phobia Scale; FNE: Fear of Negative Evaluation Scale. **p < .001.
the reporting source, i.e. the participant him/herself or the confederate as a proxy of the observer’s perspective. Specifically, we predicted individuals with SAD to perform worse during social interaction compared to HC using a standardized behavioral coding system. Furthermore, as suggested by Clark and Wells (1995), we expected individuals with SAD to report more dysfunctional cognitions, greater self-focused attention, as well as more safety behaviors, and to be more physiologically aroused which in turn would impair social performance. We further expected that these cognitive, behavioral and physiological variables would emerge as significant predictors for social performance both when assessed by self-report and when judged by the confederate.
2. Method 2.1. Participants 103 participants with SAD were recruited by newspaper advertisements asking for people aged from 18 to 60 with fear of social situations and offering cognitive-behavioral treatment for participating in the Social Phobia Intervention Study of Mannheim (SOPHISMA). Given the primary focus on the treatment study, only 23 HC were recruited randomly from a list served by the registration office of Mannheim, Germany, for purposes of comparison. There were no significant group differences with regard to age, gender and years of education (Table 1). Interested persons were contacted for a telephone screening and were then invited for a structured clinical interview (duration: approx. 2 h), conducted by three trained clinical psychologists. Axes I and II disorders according to DSM-IV were assessed using the German version of SCID-I interview (Wittchen, Wunderlich, Gruschwitz, & Zaudig, 1997) and the German version of SCID-II interview for personality disorders (Fydrich et al., 1997). Inter-rater reliability for SAD diagnosis resulted in a kappa coefficient of .7. Thirty-seven participants with SAD met DSM-IV criteria for Avoidant Personality Disorder (APD). Inclusion criterion for the SAD group was a primary diagnosis of current SAD according to DSM-IV (minimum duration one year). No current Axis I or lifetime Axis II disorder was allowed for HC. Further exclusion criteria for the SAD group were a lifetime diagnosis of schizophrenia, bipolar disorder, anorexia nervosa, current substance abuse or dependence, current suicidal crises or psychological intervention. Twenty-eight participants with SAD had a comorbid affective disorder (major depression or dysthymia; 27.2%), and 13.6% at least one additional anxiety disorder. The study was approved by the Research Ethics Board of the University of Heidelberg and each subject provided written informed consent after the procedures had been fully explained.
832
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
2.2. Questionnaires To assess social anxiety severity, participants filled out the Social Interaction Anxiety Scale (SIAS; Stangier, Heidenreich, Berardi, Golbs, & Hoyer, 1999) and the Social Phobia Scale (SPS; Stangier et al., 1999). SIAS (20 items) measures interaction anxiety (e.g., “Talking to strangers makes me nervous”), SPS (20 items) measures anxiety in performance situations (e.g., “It would be difficult for me to drink in a small group”) using a five point Likert scale ranging from 0 (“not at all typical for me”) to 4 (“extremely typical for me”). The Fear of Negative Evaluation Scale (FNE; Leary & Meadows, 1991; German version: Vormbrock & Neuser, 1983) contains 20 items (e.g., “I am dealing to much with thoughts about what others might think about me”) to be rated on a four point Likert scale ranging from 1 (is never the case) to 4 (is always the case). For measuring depressive symptoms, the Center of Epidemiological Studies-Depression Scale (CES-D; Radloff, 1977; German Version: Hautzinger & Bailer, 1993) was used. The CES-D (15 items; e.g., “During the last week, I talked less than before”) assesses depressive symptoms during the last week on a four point Likert scales ranging from 0 (rarely) to 3 (mostly). 2.3. Measures obtained in the behavioral test 2.3.1. Self-report Anxiety level: A Visual Analogue Scale (VAS) was used to assess self-report anxiety. Participants indicated how anxious they felt from 0 (“no anxiety at all”) to 100 (“worst anxiety I can imagine”). Perceived physical symptoms: Participants received a list of 10 physical sensations including heart beating, sweating, trembling, breathlessness, pain or pressure in the chest, nausea/feeling queasy, dizziness/numbness/feeling faint, fear of losing control, blushing and muscle tension. They indicated on a 5-points Likert scale ranging from 0 (not at all present) to 4 (very much present), how much they perceived these symptoms. Safety behaviors: The German version of the Social Behaviour Questionnaire (SBQ; Clark, 1995; German: Stangier et al., 1996) was used to assess safety behaviors. From the original 27 items, 5 items were omitted because they did not fit the tasks (e.g., “I drink alcohol to cope with the situation”). On a 4-point Likert scale ranging from 0 (not at all) to 3 (very much), participants answered how much they had engaged in the listed behaviors (e.g., “I will avoid eye contact”). (Dys-)Functional cognitions: Cognitions were assessed with the Social Interaction and Self-Statement Test (SISST; Glass, Merluzzi, Biever, & Larsen, 1982), a thirty item questionnaire (e.g., “The confederate will not be interested in me”) measuring the frequency of positive and negative cognitions during conversation on a 4-point Likert scale ranging from 0 (occurred almost never) to 4 (occurred very often). Focus of attention (FAQ): The Focus of Attention Questionnaire Self-Scale (FAQ-self; Woody, 1996) was used to assess internally focussed attention (10 items; e.g., “I attended to my bodily reactions (e.g., my heartbeat)”). Participants judged on a 5-point Likert scale ranging from 0 (not at all) to 4 (very much), how much they attended to different aspects of the self. 2.3.2. Confederate report Anxiety level: Confederates rated the perceived anxiety level of the participant on a 6-point Likert scale ranging from 0 “not at all” to 5 “very much.” Self-focused-attention: Confederates were asked to estimate the extent of self-focused attention ranging from 0 (=not at all) to 100 (=fully self-focused). Safety behaviors: Using the SBQ (Stangier et al., 1996), which was modified for the use in observers (SBQ-other), confederates rated the intensity of safety behaviors they noted. Consistent with
the self-report version, the SBQ-other contains a list of 22 safety behaviors. The confederate rated on a 4-points Likert scale ranging from 0 (not at all) to 3 (very much), how much the participant engaged in each of these behaviors (e.g., “I will avoid eye contact”). 2.3.3. Behavioral and physiological variables Total talking time: The total time of talking (in seconds) during the interaction was stop watched via video analysis after the experiment. Observer rated social performance: The “Ratingskala für soziale Kompetenz” (Fydrich & Burgener, 1999; Fydrich et al., 1998) is a standardized rating system for social performance in social interaction situations and represents the German version of the Social Performance Rating Scale (SPRS; Trower, Bryant, & Argyle, 1978). Gaze direction, voice quality, speech pressure (monosyllabic versus responses preventing discourse), discomfort (agitation and nervousness) and conversation flow were rated on 5-point Likert scales (1 = very good; 5 = very poor; consistent with German school grades), with higher scores indicating less social performance. The total score was calculated as the sum score of the five subscales. Raters were trained using 10 SPRS training videos, only accepting raters (N = 7) achieving a satisfactory level of agreement (r > .9) with standard ratings. Observers were blind to the participants’ diagnosis. Heart rate. With a sample rate of 2048-Hz an electrocardiogram was recorded from two electrodes attached to the chest with the Active Two recorder (www.biosemi.com). One electrode was placed on the right collarbone and one on the lowest rib on the left side while participants were seated during interaction. Heart rate was determined using ANSLAB software (www.psycho.unibas.ch/psychophysiologie/anslab.html). For the interaction, the average heart rate was computed. 2.4. Behavioral test and procedure After the clinical interview, participants were given a set of questionnaires (SIAS, SPS, SPAI, and CES-D) to be completed at home and were scheduled for the laboratory session. The laboratory session took about 2–3 h and was run by a same-sex experimenter. Participants were asked to complete two behavioral tests. For the purpose of this study, only the test involving a social interaction is reported since the German version of the SPRS has only been validated for coding behaviors during a social interaction. The interaction task was to initiate and maintain a conversation sitting in a train with an opposite-sex stranger. The interaction task was always the first behavioral test and consisted of four phases: sitting quietly (5 min), anticipating the conversation (1.5 min), introduction of the confederate and role playing the situation (5 min) and the recovery after the conversation (10 min). The role play partners always were opposite-sex confederates, all of them being trained psychology undergraduate students. The confederates entered the room after the anticipation phase. Confederates were trained to behave friendly but passively, not to start the conversation and not to say more than three sentences at one time. The role play partners were allowed to ask questions or to revive the conversation if the participant did not say anything for longer than thirty seconds. After the end of the conversation, the confederate left the room and completed the ratings and questionnaire. All dependent measures except for heart rate were assessed immediately after the conversation. Average heart rate during the conversation was computed offline. 2.5. Data analysis A multivariate Analysis of Variance (MANOVA) followed by univariate ANOVAs was conducted to examine between group dif-
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836 Table 2 Means (standard deviations) of self-reported and confederate rated cognitive, behavioral, and physiological measures for participants with social anxiety disorder (SAD) and healthy controls (HC). SAD M (SD)
HC M (SD)
Self-report Anxiety (VAS) Self-focused attention (FAQ) Safety behaviors (SBQ) Positive cognitions (SISST) Negative cognitions (SISST) Perceived body symptoms (0–40)
37.0 (26.7) 20.00 (8.0) 26.4 (8.8) 19.5 (10.5) 30 (13.5) 10.00 (13.1)
19.4 (26.9) 11.00 (8.0) 17.6 (6.6) 25.5 (9.0) 10.5 (12.0) 3.00 (2.00)
Confederate rating Anxiety (0–5) Self-focused attention (0–100) Safety behaviors (SBQ-other)
3.9 (2.52) 53.03 (21.51) 20.61 (9.46)
0.39 (0.87) 36.67 (19.51) 10.54 (3.96)
Absolute talking time (s) Heart rate (bpm) during interaction
118.6 (55.1)
159.8 (39.3)
83.2 (13.2)
85.5 (10.9)
833
p < .001), being more self-focused (F(1,123) = 7.7, p < .001) and being more anxious as compared to controls (F(1,123) = 31.13, p < .001) (Table 2). Social performance and talking time: Social performance of participants with SAD (M = 13.5, SD = 4.2) was rated as significantly poorer than that of HC (M = 10.1, SD = 3.2; F(1,123) = 7.5, p = .01, p 2 = .07; higher scores reflecting lower performance). As individuals with SAD talked significantly less compared to HC (F(1,123) = 7.64, p = .007), this analysis was repeated with absolute talking time as covariate. The between group effect failed to reach significance when considering absolute talking time (F(1,123) = 1.40, p = .24, p 2 = .01), which was revealed as significant covariate (F(1,123) = 81.92, p < .001, p 2 = .42). Heart rate: No group differences emerged for heart rate (F(1,90) = .31, p = .58). 3.3. Relationship between self- and confederate report
Note: VAS: Visual Analoque Scale; FAQ: Focus of Attention Questionnaire; SBQ: Safety Behavior Questionnaire; SISST: Social Interaction and Self-Statement Test.
Self- and confederate ratings for anxiety, safety behaviors, and self-focused attention were significantly correlated at a moderate to high level (Table 3).
ferences (SAD versus HC) with regard to all self- and confederate rated variables. The relationship between self- and confederate report was evaluated using Pearson’s correlations. For predicting observer rated social performance, two regression analyses were conducted using either the self-report or the confederate ratings as predictor variables. An alpha level of .05 was used for all statistical tests.
3.4. Prediction of observer-rated social performance
3. Results 3.1. Clinical measures The MANOVA revealed a main effect of group (F(4,123) = 38.87, p < .001; p 2 = .6), with univariate follow-up ANOVAs indicating that individuals with SAD reported significantly more social fears (SIAS, SPS), fear of negative evaluation (FNE) and depressive symptoms (CES-D) compared to HC (p’s < .05; Table 1). 3.2. Behavioral test Self-report ratings: Using all self-report cognitive, behavioral and physiological variables measured during the behavioral test as dependent measures, the MANOVA revealed a significant between group effect (F(8,123) = 5.2, p < .001; p 2 = .28). Participants with SAD endorsed higher self-reported anxiety (F(1,123) = 18.91, p < .001), fewer positive cognitions (F(1,123) = 7.48, p < .01), more negative cognitions (F(1,121) = 43.38, p < .001), a greater internal focus of attention (F(1,123) = 22.09, p < .001), more safety behaviors (F(1,118) = 29.59, p < .001) and perceived more physical symptoms (Table 2). Confederate ratings: The overall MANOVA revealed a significant group effect (F(3,123) = 10.09, p < .001; p 2 = .22). Individuals with SAD were rated as showing more safety behaviors (F(1,123) = 16.52,
As absolute talking time differed between groups, hierarchical regressions were computed with absolute talking time entered in step one and the model variables entered in step two. Model 1—self-ratings as predictors: Absolute talking time emerged as the best predictor for observer rated social performance (F(1,123) = 80.32, p < .001; R2 = .49, Table 4). Above and beyond talking time, only self-reported positive cognitions further increased the amount of accounted variance (F(8,123) = 12.5, p < .001). However, although significant, the change in R2 was only marginal (R2 = .07). Model 2—confederate ratings as predictors: Consistent with the regression model using self-ratings, absolute talking time contributed most to observer rated social performance (F(1,123) = 90.43, p < .001; R2 = .48; Table 4). Confederate rated self-focused attention further increased the amount of explained variance (F(4,123) = 30.96, p < .001). Yet, R2 increased only slightly (R2 = .09). 3.5. Relationship between talking time and Clark and Wells (1995) variables As absolute talking time turned out to be the strongest predictor of observer rated social performance regardless of whether participants’ self-report or the confederate’s ratings were included as predictors, zero-order correlations between talking time and the participant’s self-report and the confederate’s ratings were computed. Longer talking time was significantly associated with a lower confederate’s rating of the participant’s safety behaviors (Table 5). A similar pattern was obtained for the participants’ self-ratings of anxiety, safety behaviors, and negative cognitions with the obtained correlation coefficients being small in size. Self-
Table 3 Bivariate correlations between self and confederate rated variables.
Self-report anxiety Self-report self-focused attention Self-report safety behaviors * **
p < .05. p < .001.
Confederate rating anxiety
Confederate rating self-focused attention
Confederate rating safety behaviors
.54** .36** .40**
.22* .35** .23*
.35** .26** .28**
834
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
Table 4 Regression analysis predicting observer rated social performance during conversation using participants self-report and confederate ratings. Self-report variable
B
Step 1 Absolute talking time (s)
SE B

R2
−.05
.01
−.68***
.49***
−.05 −.03 1.38 −1.14 .71 −.54 −.01 −.03
.01 .02 .87 .56 .57 .53 1.28 .03
−.64*** −.20 .22 −.17** .15 −.11 .00 −.1
.54***
Confederate ratings Step 1 Absolute talking time (s)
−.05
.01
−.67***
.48***
Step 2 Absolute talking time (s) Anxiety (0–5) Self-focused attention (FAQ) Safety behaviors (SBQ-other)
−.04 −.70 .05 1.90
.01 .45 .02 1.14
−.52*** −.14 .25** .19*
Step 2 Absolute talking time (s) Anxiety (VAS) Perceived physical symptoms Positive cognitions (SISST) Negative cognitions (SISST) Self-focused attention (FAQ) Safety behaviors (SBQ) Heart rate (bpm)
.58***
Note: VAS: Visual Analogue Scale; FAQ: Focus of Attention Questionnaire; SBQ: Safety Behavior Questionnaire; SISST: Social Interaction and Self-statement Test. ***p < .001, **p < .05, *p < .1.
Table 5 Zero-order correlations among self-report and confederate rated variables, absolute talking time and social performance. Social performance
Talking time
Self-report Anxiety (VAS) Self-focused attention (FAQ) Safety behaviors (SBQ) Positive cognitions (SISST) Negative cognitions (SISST) Perceived body symptoms (0–40)
.23* .20* .24* −.20* .30* .13
−.24* −.19* −.20* .18 −.23* −.11
Confederate rating Anxiety (0–5) Self-focused attention (0–100) Safety behaviors (SBQ-other) Heart rate (bpm) during interaction Social performance
.47** .56** .57** −.1 –
−.46** −.48** −.48** .14 .67**
Note: VAS: Visual Analogue Scale; FAQ: Focus of Attention Questionnaire; SBQ: Safety Behavior Questionnaire; SISST: Social Interaction and Self-statement Test. ***p < .001,**p < .05, *p < .1.
reported positive cognitions were significantly related to greater observer-rated social performance, though at a low level. 4. Discussion Individuals with SAD and HC participated in a social getting acquainted task. As expected, participants with SAD reported more anxiety, more physical symptoms, negative cognitions, a greater self-focus, more safety behaviors, and less positive cognitions than HC. Yet, heart rate did not differ between groups. Participants’ selfreport was echoed by the confederate’s ratings which revealed a significant difference between SAD participants and HC regarding anxiety, focus of attention and safety behaviors. The agreement between self-report and confederate anxiety rating was high, the correlations between self- and confederate rated self-focused attention and safety behaviors were moderate in size. As expected, social performance of the individuals with SAD was significantly lower as rated by independent observers using a well-established behavioral coding system which takes into account several specific behaviors (e.g., gaze) that an observer has to score to yield a
measure of social performance. Intriguingly, when controlled for absolute talking time, the behavioral observation ratings did no longer significantly differ between SAD and HC. The importance of absolute talking time is further underlined by the finding that it emerged as the best predictor of observer rated social performance, regardless of whether self-report or confederate rated cognitive and behavioral variables as derived from the cognitive model of SAD by Clark and Wells (1995) were included in the regression model. Furthermore, there was only a marginal increase in explained variance when self-rated negative cognitions were included. Similarly, safety behaviors and self-focused attention as observed by the confederate added only marginally to the amount of accounted variance in social performance. The finding that absolute talking time was the most powerful predictor of social performance as perceived by an observer is rather surprising. Clearly, this suggests that the time a person spends talking during a social interaction and the way he/she will be perceived as closely related. Although this finding is contrary to the suggestion that social performance is not equivalent to mere talking time (Fydrich et al., 1998), observer rated social performance seems to be highly related to the time the individual talks. Previous studies did not explicitly control for talking time when examining social performance in SAD, but some studies assessed the amount of time the participants talked. Baker and Edelmann (2002) found no difference between participants with SAD and HC in absolute talking time in a speech. However, Hofmann et al. (1997) reported individuals with SAD to pause more frequently than HC during a speech, which was interpreted as worse social performance. For interaction situations, only studies using analogue samples are available. Highly socially anxious individuals talked less than nonanxious individuals (Burgio et al., 1986) and time to their first utterance was significantly longer (Thompson & Rapee, 2002). In the Thompson and Rapee (2002) study, time to first utterance was used as an objective measure in addition to the behavioral ratings of social performance by two independent observers. Interestingly, high socially anxious people were rated as performing worse in an unstructured social interaction and needed more time until their first utterance. Yet, the relationship between the delayed onset of a conversation (and presumably a concomitant shorter talking time) and perceived social performance was not further investigated in this study. Our results suggest that, at least in interaction situations, differences in
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
observer perceived social performance might be accounted for by differences in absolute talking time. In line with such a conclusion, an earlier study focusing on predictors of successful dating using behavioral observation of high and low frequency daters during a conversation showed that for males absolute talking time, and for females eye contact and talking time were the best predictors of global social performance as judged by independent observers using a one item rating (Glasgow & Arkowitz, 1975). The putative pivotal role of absolute talking time in accounting for perceived social performance raises several issues. First, talking time may represent a meaningful behavior in and by itself. Individuals with SAD seem to talk less than HC. Talking less may function as a safety or avoidance behavior in an attempt to reduce the risk of making foolish or inappropriate statements (Thompson & Rapee, 2002). This interpretation is in line with the cognitive model of SAD by Clark and Wells (1995), assuming that safety behaviors directly influence perceived social performance. Due to being easy to observe, talking time may be particularly detrimental to the impression observers may get. Second, absolute talking time may be considered an easy to measure proxy of social performance. Possibly, talking time reflects the combined inhibitory effect of activated dysfunctional cognitions, self-focused attention, anxiety and safety behaviors. This hypothesis is indirectly supported by the obtained moderate to large correlations between talking time and the confederate ratings of these variables. Notably, a similar pattern of correlations was obtained for the participants’ self-ratings, with these correlations being smaller in magnitude as is to be expected due to the different reporting source. Clearly, the relationship between talking time and the cognitive-behavioral anxiety response does not allow conclusions about the underlying direction of influence. Whether talking time mediates the disruptive influence of the cognitive, behavioral, and physiological anxiety response on social performance needs to be studied in more detail. Possibly, talking less may also further activate dysfunctional cognitions, safety behaviors, and physiological arousal. Third, our findings raises questions about how social performance is best operationalized when assessed. Here, observers rated five different behaviors (e.g., voice quality, gaze direction, and conversation flow). Yet, talking time alone best accounted for overall perceived social performance. Possibly, talking time overshadows the perception of behavioral nuances even though observers may not be aware of this effect, underlining the need to assess absolute talking time in future studies. Few cognitive variables as outlined by cognitive models of SAD were found to contribute to perceived social performance above and beyond absolute talking time. Positive cognitions were an additional significant predictor in the self-rating model, confederate rated self-focused attention, and safety behaviors emerged as predictors in the confederate rating model. Clearly, in either model, the increase in accounted variance of perceived social performance was marginal. Consistent with our findings, Beidel et al. (1985) also observed a negative association between positive cognitions and social performance suggesting that adaptive behavior in a social situation is promoted by positive cognitions. Our findings further suggest that self-focused attention and safety behaviors (both triggered by dysfunctional cognitions that are not readily observable) as perceived by the confederate are associated with lower social performance as perceived by an observer (for a review Alden & Taylor, 2004; Papsdorf & Alden, 1998; Voncken, Alden, Bogels, & Roelofs, 2008). Hence, as suggested by Clark and Wells (1995), social performance is compromised if self-focused attention is high and safety behaviors are shown. Somewhat surprisingly, situational anxiety did not predict observer-rated social performance, at least when talking time was included as predictor. This is consistent with the notion that individuals with SAD are prone to overestimate the visibility of their anxiety symptoms. Nonetheless, situational anxi-
835
ety is important since it is associated with less talking time (Table 5), and possibly with other safety behaviors which, in turn, lets people appear less socially apt. Our study has some limitations which have to be addressed. First, for practical reasons, not all videos were rated by all independent observers. Hence inter-rater reliability for the social performance measure could not be determined. However, all observers were trained such that they achieved an agreement with a standardized training video of greater than .9. Second, the Selfreport Anxiety Scale and the confederate anxiety rating did not have the same format. This was primarily due to the fact that the participants’ ratings had to be consistent with other scales used in the larger treatment study. Yet, the result pattern for self-report and confederate rated cognitive model variables is quite consistent, with the lower correlations observed for the self-report data most likely reflecting that confederate and observers both rated the participants from the “other” perspective. Using the confederate ratings as a proxy for the “others” perspective was necessary in order to prevent mutual confounding influences of rating social performance using a behavioral coding system and rating the core variables of the Clark and Wells’ (1995) model of SAD. In summary, this is one of the first studies assessing cognitive, behavioral, and physiological variables derived from the cognitive model of SAD by Clark and Wells (1995) in a social interaction situation to predict observer ratings of social performance. Positive cognitions, self-focused attention and safety behaviors were associated with lower perceived social performance as proposed by Clark and Wells (1995). As lower social performance might contribute to interpersonal problems in individuals with SAD, our results underline the importance of interventions targeting at reducing self-focused attention, safety behaviors, and dysfunctional cognitions. As suggested by the obtained zero-order correlations, reduced self-focused attention, fewer safety behaviors, and less negative cognitions may be beneficial by increasing talking time which, in our study, emerged as the strongest predictor of observer-perceived social performance. In fact, talking time might be a useful and easy to measure index of social performance as perceived by others and may represent the final common behavioral outcome of the cognitive mechanisms underlying social anxiety. Acknowledgement This study was funded by a grant of the Deutsche Forschungsgemeinschaft (BO 1487/7) to MB and CH. References Alden, L. E., & Taylor, C. T. (2004). Interpersonal processes in social phobia. Clinical Psychology Review, 24, 857–882. Alden, L. E., & Wallace, S. T. (1995). Social phobia and social appraisal in successful and unsuccessful social interactions. Behaviour Research and Therapy, 33, 497–505. Baker, S. R., & Edelmann, R. J. (2002). Is social phobia related to lack of social skills? Duration of skill-related behaviours and ratings of behavioural adequacy. British Journal of Clinical Psychology, 41, 243–257. Beidel, D. C., Turner, S. M., & Dancu, C. V. (1985). Physiological, cognitive and behavioral aspects of social anxiety. Behaviour Research and Therapy, 23, 109–117. Bögels, S. M., & Voncken, M. (2008). Social skills training versus cognitive therapy for social anxiety disorder characterized by fear of blushing, trembling, or sweating. International Journal of Cognitive Therapy, 1, 138–150. Burgio, K. L., Merluzzi, T. V., & Pryor, J. B. (1986). Effects of performance expectancy and self-focused attention on social-interaction. Journal of Personality and Social Psychology, 50, 1216–1221. Clark, D. M. (1995). Social Behaviour Quesstionnaire (SBQ). Unpublished Work. Department of Psychiatry, Oxford University, Oxford, UK. Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In: R. G Heimberg, M. R. Liebowitz, D. A. Hope, & F. R. Schneier (Eds.), Social phobia: diagnosis, assessment and treatment. New York: Guilford. Fydrich, T., & Bürgener, F. (1999). Ratingskalen für soziale Kompetenz. In: J. Margraf, & K. Rudolf (Eds.), Soziale Kompetenz-Soziale Phobie: Anwendungsfelder,
836
S. Stevens et al. / Journal of Anxiety Disorders 24 (2010) 830–836
Entwicklungslinien, Erfolgsaussichten. Hohengehren: Baltmansweiler: Schneider, pp. 81–96 Fydrich, T., Chambless, D. L., Perry, K. J., Buergener, F., & Beazley, M. B. (1998). Behavioral assessment of social performance: a rating system for social phobia. Behaviour Research and Therapy, 36, 995–1010. Fydrich, T., Renneberg, B., Schmitz, B., & Wittchen, H.-U. (1997). Strukturiertes Klinisches Interview für DSM-IV Achse II. Persönlichkeitsstörungen (SKID II). Göttingen: Hogrefe. Glasgow, R. E., & Arkowitz, H. (1975). The behavioral assessment of male and female social competence in dyadic heterosexual interaction. Behavior Therapy, 6, 488–498. Glass, C. R., Merluzzi, T. V., Biever, J. L., & Larsen, K. H. (1982). Cognitive assessment of social anxiety—development and validation of a self-statement questionnaire. Cognitive Therapy and Research, 6, 37–55. Hautzinger, M., & Bailer, M. (1993). Allgemeine Depressions-Skala (ADS). Weinheim: Beltz. Hofmann, S. G., Gerlach, A. L., Wender, A., & Roth, W. T. (1997). Speech disturbances and gaze behavior during public speaking in subtypes of social phobia. Journal of Anxiety Disorders, 11, 573–585. Hopko, D. R., McNeil, D. W., Zvolensky, M. J., & Eifert, G. H. (2001). The relation between anxiety and skill in performance-based anxiety disorders: a behavioral formulation of social phobia. Behavior Therapy, 32, 185–207. Leary, M. R., & Meadows, S. (1991). Predictors, elicitors, and concomitants of social blushing. Journal of Personality and Social Psychology, 60, 254–262. Moscovitch, D. A., & Hofmann, S. G. (2007). When ambiguity hurts: social standards moderate self-appraisals in generalized social phobia. Behaviour Research and Therapy, 45, 1039–1052. Norton, P. J., & Hope, D. A. (2001). Kernels of truth or disorted perceptions: self and observer ratings of social anxiety and performance. Behavior Therapy, 32, 765–786. Papsdorf, M., & Alden, L. (1998). Mediators of social rejection in social anxiety: similarity, self-disclosure, and overt signs of anxiety. Journal of Research in Personality, 32, 351–369. Radloff, L. S. (1977). The CES-D Scale: a self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.
Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral model of anxiety in social phobia. Behaviour Research and Therapy, 35, 741–756. Rapee, R. M., & Lim, L. (1992). Discrepancy between self and observer ratings of performance in social phobics. Journal of Abnormal Psychology, 101, 728–731. Sheffer, C. E., Penn, D. L., & Cassisi, J. E. (2001). The effects of impression management demands on heart rate, self-reported social anxiety, and social competence in undergraduate males. Journal of Anxiety Disorders, 15, 171–182. Stangier, U., Ehlers, A., Heidenreich, T., & Clark, D. M. (1996). Fragebogen zu Sicherheitsverhalten in sozialen Situationen. Frankfurt am Main: Johann Wolfgang von Goethe Universität. Stangier, U., Heidenreich, T., Berardi, A., Golbs, U., & Hoyer, J. (1999). Die Erfassung sozialer Phobie durch die Social Interaction Anxiety Scale (SIAS) und die Social Phobia Scale (SPS). Zeitschrift für Klinische Psychologie, 28, 28–36. Stopa, L., & Clark, D. M. (1993). Cognitive-processes in social phobia. Behaviour Research and Therapy, 31, 255–267. Strahan, E. Y., & Conger, A. J. (1999). Social anxiety and social performance: why don’t we see more catastrophes? Journal of Anxiety Disorders, 13, 399–416. Thompson, S., & Rapee, R. M. (2002). The effect of situational structure on the social performance of socially anxious and non-anxious participants. Journal of Behavior Therapy and Experimental Psychiatry, 33, 91–102. Trower, P., Bryant, B., & Argyle, M. (1978). Social skills and mental health. Pittsburgh: University of Pittsburgh Press. Voncken, M. J., Alden, L. E., Bogels, S. M., & Roelofs, J. (2008). Social rejection in social anxiety disorder: the role of performance deficits, evoked negative emotions and dissimilarity. British Journal of Clinical Psychology, 47, 439–450. Voncken, M. J., & Bogels, S. M. (2008). Social performance deficits in social anxiety disorder: reality during conversation and biased perception during speech. Journal of Anxiety Disorders, 22, 1384–1392. Vormbrock, F., & Neuser, J. (1983). Konstruktion zweier spezifischer TraitFragebögen zur Erfassung von Angst in sozialen Situationen (SANB und SVSS). Diagnostica, 29, 98–106. Wittchen, H. U., Wunderlich, U., Gruschwitz, S., & Zaudig, M. (1997). SKID-I. Strukturiertes klinisches Interview für psychische Störungen für DSM-IV. Göttingen: Hogrefe. Woody, S. R. (1996). Effects of focus of attention on anxiety levels and social performance of individuals with social phobia. Journal of. Abnormal Psychology, 105, 61–69.