See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/274073555

Anticipatory and Post-Event Rumination in

Social Anxiety Disorder: A Review of the
Theoretical and Empirical Literature

Article in Behaviour Change · June 2014

DOI: 10.1017/bec.2014.3

CITATIONS READS

8 1,086

2 authors:

Erika Penney Maree J. Abbott

Foresight Psychology University of Sydney
6 PUBLICATIONS 19 CITATIONS 67 PUBLICATIONS 2,026 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

PhD in Clinical Psychology - Thesis: Investigating the Clinical Utility of the Eating Beliefs
Questionnaire View project

Therapist Stigma Towards Narcissistic Personality Disorder: Lessons Learned From Borderline
Personality Disorder View project

All content following this page was uploaded by Erika Penney on 26 March 2015.

The user has requested enhancement of the downloaded file.

 1 Anticipatory and Post-Event
2 Rumination in Social Anxiety Disorder:
3 A Review of the Theoretical
4 and Empirical Literature
Q1 5
6 Erika S. Penney, and Maree J. Abbott
7 School of Psychology, The University of Sydney, New South Wales, Australia

8 Social anxiety disorder (SAD) is a psychological disorder characterised by an ex-

9 cessive and persistent fear of social or performance situations, which interferes with
10 daily functioning. Cognitive models of SAD (Clark & Wells, 1995; Hofmann,
11 2007; Rapee & Heimberg, 1997) emphasise the importance of negative pre- and
12 post-event rumination as a maintaining factor in the cycle of SAD. While the link
13 between negative rumination and SAD is well supported by empirical research, little
14 is understood about this cognitively important process; thus, research investigating
15 the predictors of negative rumination in SAD is important to consider. Within the
16 current literature, performance appraisal appears to be the most likely unique cog-
17 nitive predictor of post-event rumination. There is limited research into cognitive
18 predictors of pre-event rumination. Treatments targeting this maintaining factor
19 are important to consider. Suggestions for future research examining the cognitive
20 models of SAD by experimentally manipulating perceived social standards in or-
21 der to examine the impact of high and low perceived social standard on appraisal
22 processes (i.e., threat appraisal and performance appraisal), state social anxiety,
23 and negative pre-event and post-event rumination, are proposed. Implications for
24 theoretical models and efficacious treatments for SAD are discussed.

25 ! Keywords: social anxiety disorder, rumination, appraisal

26 Social Anxiety Disorder

27 Social anxiety disorder (SAD) is a psychological disorder characterised by an ex-
28 cessive and persistent fear of social or performance situations, in which an individ-
29 ual fears being negatively evaluated by others (American Psychiatric Association
30 [APA], 2013). Feared social situations include social interactions (e.g., a conver-
31 sation with unfamiliar others), social performances (e.g., a speech), and being ob-
32 served (e.g., eating in public).When an individual with SAD encounters such a situ-
33 ation, they experience cognitive, affective, behavioural, and physiological symptoms
34 of anxiety, and frequently use avoidance or related safety strategies to manage these
35 symptoms. 79
36 A diagnosis of SAD is warranted when anxiety symptoms and avoidance interfere
37 significantly with one’s social or occupational functioning and cause significant distress
38 in daily life (APA, 2013). Social anxiety exists on a spectrum and can be experienced

Address for correspondence: Dr Maree Abbott, School of Psychology (K01), The University of Sydney NSW
2006, Australia. Email: [email protected]

Behaviour Change Volume 31 Number 2 2014 pp. 79–101 ⃝ c The Author(s), published by Cambridge
University Press on behalf of Australian Academic Press Pty Ltd 2014 doi 10.1017/bec.2014.3
 Erika S. Penney and Maree J. Abbott

39 by individuals with and without diagnosed SAD (Turner, Beidel, Dancu, & Stanley,
40 1989); however, both SAD and subclinical social anxiety are associated with impaired
41 daily functioning (Crum & Pratt, 2001).
42 In recent empirical studies, researchers have commonly used both clinical SAD
43 samples and social anxiety analogue samples (Stopa & Clark, 2000, 2001). Thus,
44 for the purpose of this review, SAD will refer only to individuals who meet DSM-
45 IV-TR (APA, 2000) or DSM-V (APA, 2013) criteria for Social Anxiety Disorder,
46 and social anxiety will otherwise refer to non-clinical levels of distress and anxiety
47 associated with social evaluative situations. The main difference between subclinical
48 and clinical social anxiety is the intensity and severity of symptoms, and the distress
49 and interference of those symptoms on everyday functioning (Crum & Pratt, 2001).
50 According to epidemiological research, SAD is one of the most prevalent psycho-
51 logical disorders in Western countries (Kessler, Berglund, Demler, Jin, & Walters,
52 2005). SAD is the fourth most common psychiatric disorder in Australia, with a
53 lifetime prevalence of 10.6% (Australian Bureau of Statistics [ABS], 2007). Within
54 Australia, 78.1% of individuals with SAD have at least one other comorbid disorder;
55 53.4% of which include comorbid Axis I anxiety disorders (Lampe, Slade, Issakidis,
56 & Andrews, 2003). Additionally, 35.8% of individuals with SAD also met criteria
57 for avoidant personality disorder (Lampe et al., 2003). A review of epidemiologi-
58 cal research found women were more likely to develop SAD than men, with the
59 female-to-male ratio reported to be approximately 3:2 (Furmark, 2002).
60 In community settings, SAD is associated with less education, lower salaries, less
61 social support, and being unmarried (Furmark, 2002; Heimberg, Hope, Dodge, &
62 Becker, 1990). SAD has an earlier onset than most anxiety disorders, typically begin-
63 ning in childhood or early adolescence, and often occurring before the development
64 of other psychiatric disorders (Weissman et al., 1996). The experience of SAD, in
65 association with at least one other comorbid disorder, has been found to significantly
66 increase the risk of attempting suicide (Weissman et al., 1996).

67 Efficacious Treatments
68 The most widely used treatment for SAD is group Cognitive Behavioural Therapy
69 (gCBT; Heimberg & Becker, 2002); an adaptation of CBT for anxiety disorders,
70 which includes cognitive restructuring, behavioural experiments, and graded exposure
71 techniques (Beck, Emery, & Greenberg, 1985). Despite being the gold standard, these
72 treatments have received modest results, with one study showing 58% of participants
73 significantly improved after the 12-week gCBT program (Heimberg et al., 1998), and
74 another finding that 51.7% of participants no longer met criteria for SAD after gCBT
75 (Davidson et al., 2004). Additionally, gCBT plus added social skills training has been
76 found to have similar effect sizes from pre- to 12-month follow-up as gCBT treatments
77 alone (Herbert, Rheingold, & Goldstein, 2002).
78
80 Treatment for SAD typically includes an exposure component and a cognitive
79 therapy (CT) component; however, the additive therapeutic effect of CT over ex-
80 posure alone remains controversial. In a meta-analysis of psychological treatments
81 for SAD, Taylor (1996) found that the effect sizes for CT and exposure were not
82 significantly different than exposure alone, although there was a trend for CT and
83 exposure to show the largest effects. Meta-analyses by Feske and Chambless (1995)
84 and Gould, Buckminster, Pollack, Otto, and Yapp (1997) found no evidence for the
85 additive effects of CT over exposure alone.

Behaviour Change
 Rumination in Social Anxiety Disorder

86 Feske and Chambless (1995) propose that their failure to detect differences be-
87 tween CBT and exposure does not necessarily mean that the CBT approach lacks the
88 potential to improve over exposure, as more powerful cognitive-based interventions
89 are developed.
90 More recent research suggests that CT can enhance efficacy of SAD treatments
91 (Clark et al., 2006; Rapee, Gaston, & Abbott, 2009). Clark et al. (2006) found that
92 CT and exposure were consistently superior to exposure alone on a varied range of
93 SAD measures. Additionally, the CT program used in this study involved fewer total
94 exposure exercises than the exposure program, suggesting that less overall exposure
95 may be needed when embedded within a CT framework. This is hypothesised to
96 be due to some of the therapeutic differences in CT over exposure alone, including
97 framing exposure as a behavioural experiment, elimination of safety behaviours, and
98 the experiential demonstration that self-focused attention and safety and avoidance
99 behaviours can have unhelpful effects (Clark et al., 2006).
100 Extending these findings, Rapee et al. (2009) compared a traditional CBT package
101 (similar to that used by Clark et al., 2006) to an enhanced CBT package, which in-
102 cluded an overlap between the cognitive (i.e., cognitive restructuring) and behavioural
103 (i.e., in vivo exposure) elements through the use of hypothesis testing, evidence gath-
104 ering and restructuring techniques targeting broader core beliefs and life scripts, and
105 the systematic elimination of safety and avoidance behaviours. The enhanced pro-
106 gram showed significantly greater effects over the traditional CBT program in terms
107 of both statistical and clinical significance.
108 These results support hypotheses of cognitive models of SAD, which suggest that a
109 greater understanding of SAD and key mediating factors (e.g., over-evaluation of the
110 likelihood and cost of negative evaluation and use of safety behaviours) will increase
111 treatment efficacy. Improved understanding of key maintaining factors underlying
112 mechanisms and processes is invaluable in the continuing development of comprehen-
113 sive theoretical models and efficacious treatments. For instance, there has been limited
114 research examining some of the key cognitive factors in SAD (e.g., pre- and post-event
115 rumination) and their relationship with cognitive biases and perceived expected stan-
116 dards, though the impact on heightening state social anxiety is well founded.

117 Cognitive Models of Social Anxiety Disorder

118 Contemporary cognitive models of SAD (Clark & Wells, 1995; Rapee & Heimberg,
119 1997; Hofmann, 2007) have been used to provide theoretical frameworks to under-
120 stand the key maintaining processes involved in the disorder. These models posit that
121 dysfunctional beliefs, biased processing of social information, safety behaviours, and
122 avoidance of anxiety-provoking situations contribute to the development and mainte-
123 nance of SAD. The Clark and Wells (1995) and Rapee and Heimberg (1997) models
124 are theoretically similar, with Clark and Wells focusing on the maintaining factors
125 of SAD, and Rapee and Heimberg focusing on the cognitive processes maintaining
126 state anxiety experienced by individuals with SAD. Hofmann (2007) also highlights 81
127 a primary role for high perceived social standards and poorly defined social goals in
128 triggering anticipatory anxiety before a social event.

129 Clark and Wells (1995)

130 Clark and Wells’ (1995) cognitive model suggests that individuals with SAD are
131 more likely to perceive social events as threatening due to cognitive biases that they

Behaviour Change
 Erika S. Penney and Maree J. Abbott

FIGURE 1
Clark’s (2001) cognitive model of social anxiety disorder (adapted from Clark & Wells, 1995).
Printed with permission from the publisher: Guildford Press.

132 hold about themselves, others, and the world. For instance, individuals with SAD
133 are hypothesised to hold excessively high standards for social performance, which
134 naturally provoke anxiety, as they are difficult to achieve and thus promote the
135 perception of failure (Clark & Wells, 1995).
136 Individuals with SAD also hold conditional beliefs concerning social evalua-
137 tion, such as, ‘If I make mistakes, then others will reject me’ (Clark & Wells,
138 1995). These beliefs are the result of a biased assumption that ‘What (I think) others
139 think of me, must be the truth’. These stringent rules for interactions provoke anxiety
140 and activate the expectation of negative social evaluation and rejection from others.
141 Individuals with SAD are thought to have ‘unstable self-schemata’ (p. 76), meaning
142 that while they have negative beliefs about themselves (e.g., ‘I’m inadequate’), these
143 beliefs are conditional on specific situations and are more marked in social situations,
144 compared to less threatening situations, such as being with close friends (Clark &
145 Wells, 1995).
146 As shown in Figure 1, Clarke and Wells’ (1995) model posits several key maintain-
147 ing processes involved in SAD. Individuals enter a social situation, which activates
148 unhelpful assumptions that the social situation is threatening, resulting in a shift
149 in attention and a detailed monitoring of the self. This self-focused attention fur-
150 ther promotes physiological and cognitive symptoms of anxiety, such as blushing
151
82 and the activation of further dysfunctional beliefs. Individuals with SAD then use
152 this somatic and cognitive information to create a negative mental representation
153 of themselves (as they believe would be viewed by the audience), resulting in the
154 individual not attending to potential positive social cues from others (Clark & Wells,
155 1995).
156 Additionally, individuals with SAD are known to engage in safety behaviours
157 (Kim, 2005; Wells et al., 1995), which can either involve overt behaviours (e.g.,
158 avoidance) or internal mental processes (e.g., rehearsing what to say next; Clark,

Behaviour Change
 Rumination in Social Anxiety Disorder

159 2001; Clark & Wells, 1995). These safety behaviours maintain SAD in four ways: (1)
160 a social success may be attributed to the safety behaviour rather than the inference
161 that the situation was less threatening than originally perceived; (2) the individual’s
162 fears may be exacerbated (e.g., holding arms together to hide sweating may produce
163 more sweating); (3) safety behaviours direct attention to the self, which exhausts
164 cognitive capacity that may be used to process objective information; and (4) safety
165 behaviours can produce the feared result (e.g., an individual may rehearse what to say
166 next to avoid appearing foolish, but by doing so may appear slower or preoccupied;
167 Clark, 2001; Clark & Wells, 1995).
168 Clark and Wells (1995) posit that individuals with SAD engage in extensive
169 anticipatory rumination prior to a social or performance event. Such rumination
170 often focuses on past failures and negative predictions of performance in the social
171 situation. Anticipatory social threat places individuals in a self-focused processing
172 mode, where one expects to perform poorly, and engages in cognitive biases, which
173 prevent one from perceiving external signs of acceptance from others.
174 After the event, individuals are hypothesised to engage in a ‘post-mortem’ (i.e.,
175 post-event rumination; Clark & Wells, 1995, p. 74) of the social event, where the
176 dominant images and thoughts focus on perceived failure and negative images of them-
177 selves that were experienced during the event itself. The model predicts a relationship
178 between performance self-appraisal and negative post-event rumination, such that
179 the more negative one’s self-appraisal of their performance, the more frequently an
180 individual will engage in post-event rumination (Clark & Wells, 1995).

181 Rapee and Heimberg (1997)

182 Rapee and Heimberg’s (1997) cognitive-behavioural model (see Figure 2) is similar to
183 the Clark and Wells (1995) model, but begins with the premise that socially anxious
184 individuals attach significant importance to being positively evaluated by others, yet
185 they assume that any evaluation will be inherently negative. When a socially anxious
186 event is perceived or experienced, an individual with SAD is argued to create a mental
187 representation of themselves as would be seen by a social other, created by memories
188 of past experiences, internal cues (e.g., physiological anxiety symptoms), and external
189 cues (e.g., perceived indicators of negative evaluation; Rapee & Heimberg, 1997).
190 Although pre- and post-event rumination is not explicitly named in this model, it
191 can be inferred that the content of pre- and post-event rumination is the expectation
192 of a negative outcome and the generation of the negative internal mental represen-
193 tation of the self as seen by an audience in anticipation, during and after a social
194 event.
195 Unlike the Clark and Wells’ (1995) model, which postulates that attentional re-
196 sources will be allocated to one’s internal representation, Rapee and Heimberg (1997)
197 argue that the individual’s attentional resources are split to monitor both the self-
198 image and potential external signs of threat simultaneously (e.g., signs of disinterest). 83
199 This self-image is then compared to a perceived standard of audience expectation.
200 The higher the discrepancy between the self-image and perceived standard, the greater
201 the likelihood that the individual will anticipate negative social outcomes in terms
202 of probability and cost, which in turn results in behavioural, cognitive, and physi-
203 cal symptoms of anxiety. Anxiety symptoms and any external indictors of threat are
204 perceived and reintegrated into the mental representation of the self, resulting in
205 maintenance of the social anxiety and ruminative cycle (Rapee & Heimberg, 1997).

Behaviour Change
 Erika S. Penney and Maree J. Abbott

FIGURE 2
Rapee & Heimberg’s (1997) cognitive-behavioural model of social anxiety disorder. Printed
with permission from the publisher: Elsevier.

84
206 Hofmann (2007)
207 Hofmann’s (2007) model (Figure 3) extends the theoretical work of the other models.
208 This model posits that individuals with SAD feel social apprehension due to the
209 perception that social standards are high. Individuals then desire to meet these high
210 social standards, but simultaneously doubt their ability to do so, leading to further social
211 apprehension, anticipatory anxiety, and self-focused attention. In turn, a number of

Behaviour Change
 Rumination in Social Anxiety Disorder

FIGURE 3
Hofmann’s (2007) model of social anxiety disorder. Printed with permission from the pub-
lisher: Routledge.

212 cognitive processes are triggered, resulting in an overestimation of the probability and
213 cost of a poor social outcome (Hofmann, 2007).
214 This theory is consistent with Clark and Wells’ (1995) model, which posits that
215 individuals with SAD believe they will act in a humiliating way, resulting in per-
216 ceived negative social costs. Additionally, individuals with SAD hold a poor view
217 of themselves as a social object, and thus believe they have insufficient social skills
218 and low control over their anxiety (Hofmann, 2007). These cognitive factors result
219 in individuals experiencing further social apprehension and engaging in avoidance
220 or safety behaviours to cope with the experience of excessive anxiety. After a social
221 event, individuals are hypothesised to engage in post-event rumination similar to that
222 described in both Clark and Wells (1995) and Rapee and Heimberg (1997). These 85
223 processes are thought to play a role in the cycle and maintenance of social anxiety
224 (Hofmann, 2007).

225 Rumination in SAD

226 The three main cognitive models of SAD are theoretical frameworks that emphasise
227 the importance of negative rumination as a maintaining factor in the cycle of social

Behaviour Change
 Erika S. Penney and Maree J. Abbott

228 anxiety. The process of negative rumination occurs when an individual engages in an
229 excessive mental review of their performance before or after a social event (Brozovich
230 & Heimberg, 2008). Researchers have found that excessive and persistent negative
231 rumination is engaged in by individuals with SAD both before and after a social-
232 evaluative event (Rachman, Gruter-Andrew, & Shafran, 2000; Vassilopoulos, 2004).
233 It is theorised that negative rumination may be one of the key cognitive processes
234 preventing individuals from becoming desensitised to social threat, despite regular
235 exposure to seemingly benign social events (Clark & Wells, 1995; Hofmann, 2007;
236 Rapee & Heimberg, 1997).
237 In a recent review examining rumination and its role in SAD, it was found that
238 negative rumination (related to a social or performance event) leads to excessive
239 worry about one’s performance, a belief that the performance was unsatisfactory, and
240 memory retrieval of previous perceived social failures (Brozovich & Heimberg, 2008).
241 Negative rumination before and after a social event, and interpreting oneself as a social
242 failure, maintains SAD despite exposure to social events with objectively neutral or
243 positive outcomes. Rumination further maintains SAD by perpetuating negative self-
244 impressions and negative predictions for future events (Brozovich & Heimberg, 2008).
245 Thus, it is important to understand both pre- and post-event rumination processes
246 when investigating SAD.
247 There is a wealth of literature investigating rumination related to depressive symp-
248 toms (e.g., Nolen-Hoeksema, 1987, 2004; Nolen-Hoeksema, Parker, & Larson, 1994;
249 Smets, Wessel, Schreurs, & Raes, 2012), and it is important to distinguish this cog-
250 nitive process from rumination in SAD. Rumination related to depressive symptoms
251 has not been found to be predictive of post-event rumination related to social anxiety
252 (Kocovski & Rector, 2008), suggesting that the content of rumination has diag-
253 nostic specificity. It is important to note that while there is significant comorbidity
254 between SAD and depression (Perugi, Frare, Toni, Mata, & Akiskal, 2001), there
255 are significant conceptual distinctions in the rumination process. Clark and Wells
256 (1995) propose that rumination in SAD focuses on perceived social failures, while re-
257 sponse styles theory (Nolen-Hoeksema, 1991) proposes that rumination in depression
258 focuses on depressive symptoms, emotions, and consequences. While social anxi-
259 ety and depression may share the cognitive process of rumination, the content of
260 the rumination is diagnostically distinct, and must be treated that way in empirical
261 investigations.
262 Cognitive-based interventions for SAD have been shown to be beneficial in the
263 reduction of cognitive symptoms, such as persistent negative rumination (Abbott &
264 Rapee, 2004). Participants with SAD reported significantly lower levels of negative
265 rumination, and significantly more positive appraisals of their speech performances
266 after a 12-week course of gCBT (Abbott & Rapee, 2004). Thus, a better understanding
267 of this key cognitive process would aid the development of both theoretical models
268 and applied treatments.
86
269 Post-Event Rumination
270 The association between social anxiety and post-event rumination has been well
271 established in the literature (Fehm, Schneider, & Hoyer, 2007; Kocovski & Rector,
272 2007; McEvoy & Kingstep, 2006; Rachman et al., 2000). A widely used measure of
273 post-event rumination is the Thoughts Questionnaire (TQ) adapted by Abbott and
274 Rapee (2004) from Edwards, Rapee, and Franklin (2003). This is a self-report measure

Behaviour Change
 Rumination in Social Anxiety Disorder

275 that assesses the frequency of post-event rumination during the week following a social
276 performance or interaction, rather than general levels of rumination. The items were
277 designed to reflect the type of content hypothesised by Clark and Wells (1995) and
278 found by Rachman et al. (2000) to be dominant in the ruminative process, including
279 self-perceptions of anxiety, poor performance, and perceived negative evaluations
280 from others. The TQ has been found to have strong psychometric properties, and
281 there is strong empirical evidence for the association between negative rumination
282 (as measured by the TQ) and social anxiety (see Abbott & Rapee, 2004; Edwards
283 et al., 2003; Rapee & Abbott, 2007; Zou & Abbott, 2012).
284 Individuals higher in social anxiety engage in greater levels of negative rumination
285 after a social event, such as an impromptu speech performance (Edwards et al., 2003),
286 or social interaction (Mellings & Alden, 2000). Mellings and Alden (2000) asked
287 individuals with high and low social anxiety to participate in a 10-minute conversation
288 with an opposite sex confederate in a laboratory setting. Individuals with higher
289 social anxiety reported significantly more post-event rumination the day after the
290 conversation task. Additionally, increased post-event rumination was associated with
291 increased negative appraisal of one’s social performance (Mellings & Alden, 2000).
292 This is consistent with cognitive models (e.g., Hofmann, 2007), which state that
293 during the ruminative process an individual generates negative self-images and poor
294 performance appraisals, which increases future social apprehension and maintains
295 the cycle of SAD. One limitation of Mellings and Alden’s (2000) study was that it
296 utilised a non-clinical social anxiety analogue sample and thus caution is warranted
297 in generalising these findings to individuals with SAD.
298 More recently, empirical research in this field has emerged, investigating the role
299 of post-event rumination within samples of individuals diagnosed with SAD (Abbott
300 & Rapee, 2004; Kocovski & Rector, 2008; Perini, Abbott, & Rapee, 2006; Rapee
301 & Abbott, 2007). In a study by Abbott and Rapee (2004), participants who met
302 DSM-IV-TR criteria for a principle diagnosis of SAD were more likely than non-
303 anxious controls to engage in negative rumination immediately after an impromptu
304 speech task, as well as in the week following the speech. Moreover, the clinical
305 group continued to perceive the speech task in negative terms over the subsequent
306 week, while the control group viewed their performance on the task more positively,
307 suggesting that negative rumination maintained negative self-appraisals (Abbott &
308 Rapee, 2004). Similarly, Kocovski and Rector (2008) found that SAD participants
309 with higher levels of social anxiety reported higher levels of post-event rumination
310 after the first gCBT exposure task, regardless of the social anxiety measure used (i.e.,
311 the Social Phobia Scale, Social Interaction Anxiety Scale, Liebowitz Social Anxiety
312 Scale).
313 Evidence indicates that individuals with higher social anxiety (within both clinical
314 and non-clinical groups) engage in more negative rumination after a socially anxious
315 event. However, this rumination may not be solely accounted for by poorer social
316 performances, as research has found that socially anxious individuals consistently rate 87
317 their own performance more negatively than both objective raters and comparison
318 control groups, and effects are maintained when variance due to objective performance
319 ratings is partialled out (Abbott & Rapee, 2004; Dannahy & Stopa, 2007; Mellings
320 & Alden, 2000; Rapee & Lim, 1992).
321 While there is emerging research linking post-event rumination and SAD, there
322 is limited research in the area of pre-event rumination and SAD. This is a signifi-
323 cant limitation of the empirical literature given the importance of this hypothesised

Behaviour Change
 Erika S. Penney and Maree J. Abbott

324 cognitive process in maintaining the cycle of SAD (Clark & Wells, 1995; Hofmann,
325 2007; Rapee & Heimberg, 1997).

326 Pre-Event Rumination

327 In their cognitive model of SAD, Clark and Wells (1995) and Wells and Clark (1997)
328 hypothesised that individuals with SAD would engage in a form of cognitive processing
329 prior to entering a social event that involves biased thinking and subsequent anxious
330 affect. Similar to how individuals with SAD are thought to think through a ‘post-
331 mortem’ after an event, so too do individuals go through a ‘pre-mortem’ before the
332 event, ruminating on past failures, negative self-images, and predictions of rejection
333 and negative evaluation from others. Thus, individuals are already in a negative self-
334 focused processing mode and expecting failure before they enter the social situation,
335 and may be less receptive to objective indicators of acceptance (Clark & Wells, 1995;
336 Wells & Clark, 1997). Although not explicitly named in their earlier account, Clark
337 and Wells (1995) appear to be referring to pre-event rumination.
338 Vassilopoulos (2004) was one of the first researchers to empirically demonstrate the
339 positive relationship between greater social anxiety and greater levels of negative pre-
340 event rumination, even after controlling for the effects of trait anxiety and depression.
341 Highly socially anxious individuals reported that they experienced frequent negative
342 thoughts about a future social event that were intrusive, difficult to control, and
343 interfered with their ability to concentrate. Additionally, it was found that while both
344 high and low socially anxious groups made predictions regarding cost and consequences
345 about future social events, these were significantly more negative in socially anxious
346 participants (Vassilopoulos, 2004).
347 In a study by Brown and Stopa (2006), undergraduate students who scored low
348 and high in social anxiety were asked to engage in either a distraction task or a pre-
349 event rumination period before giving a brief speech. Socially anxious participants
350 reported more anxiety and greater frequency of negative self-images than non-anxious
351 participants after engaging in pre-event rumination. Despite this, both low and high
352 social anxiety groups rated their speech performance as better after an anticipated
353 (compared to an unanticipated) speech, suggesting that individuals may hold positive
354 meta-beliefs about the function of anticipation and rumination before a social perfor-
355 mance (Brown & Stopa, 2006). This may also be indicative of the role that pre-event
356 rumination (and positive meta-beliefs about rumination) has in maintaining social
357 anxiety.
358 Limitations of this and similar research (e.g., Vassilopolous, 2004, 2005; Wong
359 & Moulds, 2009, 2011, 2012a, 2012b) are inherent in the use of undergraduate
360 students, and high and low socially anxious groups created using the Fear of Negative
361 Evaluation (FNE) scale (Watson & Friend, 1969), a 30-item self-report measure that
362 assesses apprehension, distress, and expectation of negative evaluation. The FNE
363
88 assesses cognitions associated with social anxiety rather than specific symptoms of SAD,
364 such as fear and avoidance of social situations, and therefore cannot be considered
365 a diagnostic measure. Additionally, other clinical groups have also demonstrated
366 high fear of negative evaluation, and thus this measure may not be specific to SAD
367 (Rodebaugh et al., 2004).
368 Recent studies in the area of negative rumination have included an experimental
369 state anxiety task; for example, a speech or conversation task (e.g., Abbott & Rapee,
370 2004; Kashdan & Roberts, 2007; Perini et al., 2006; Rapee & Abbott, 2007). These

Behaviour Change
 Rumination in Social Anxiety Disorder

371 studies address limitations of previous studies that have relied heavily on self-report
372 measures, by controlling for frequency, content, and intensity of post-event rumina-
373 tion (e.g., Kocovski & Rector, 2007; Fehm et al., 2007; Wong & Moulds, 2010).
374 Despite this improvement in study design within this field, relatively few studies have
375 completed experimentally designed research in the area of pre-event rumination, and
376 those few that have (e.g., Brown & Stopa, 2006), did not use a clinical participant
377 group.
378 Research into pre-event rumination has been inconsistent and has focused on
379 social anxiety analogue samples, rather than participants with SAD. Further research
380 is needed to specifically investigate the role of pre-event rumination in SAD samples,
381 utilising state anxiety tasks (e.g., a speech) within experimentally-designed research.
382 Future research is needed to extend the literature in this area by not only examining
383 the link between SAD and pre-event rumination, but the cognitive and appraisal
384 processes that might predict this pervasive symptom of SAD.

385 Predictors of Rumination

386 While the link between negative rumination and SAD is well supported by empirical
387 research and clinical observation, little is understood about this cognitively important
388 process, which has received relatively little empirical attention. Within the current
389 literature, performance appraisal and fear of negative evaluation have been the most
390 commonly studied predictors of negative rumination.
391 It has been argued that there are striking similarities between pre-event and
392 post-event processing (Vassilopolous, 2004). These two processes appear to have
393 similar cognitive processes in terms of frequency, pervasiveness, intrusiveness, and
394 observer-perspective imagery (Vassilopolous, 2004). Future research should aim to
395 clarify whether pre-event rumination can be predicted by the same symptom, cogni-
396 tive and appraisal variables found to significantly predict post-event rumination (e.g.,
397 performance and threat appraisal, and state and trait anxiety), thereby enhancing the
398 understanding of SAD and contributing to the assessment and extension of cognitive
399 models.

400 Threat appraisal.Cognitive models of anxiety hypothesise that biases in information

401 processing (e.g., a hypersensitivity to threat or perceived danger) play a significant role
402 in the aetiology and maintenance of anxiety disorders (Beck, 1976; Beck et al., 1985).
403 This perceived threat or danger is typically related to subjective cost and probability
404 appraisals (Butler & Mathews, 1983). Clark and Wells (1995) posit that individuals
405 with SAD believe that (1) they have a greater likelihood of behaving in a socially
406 unacceptable way, and (2) that this will lead to disastrous consequences or social
407 costs, in terms of status, worth, and rejection. Similarly, if a situation was ambiguous,
408 anxious individuals are more likely to interpret this situation as more threatening, in
409 terms of cost and probability (Butler & Mathews, 1983), and experience greater state 89
410 anxiety (Moscovitch & Hofmann, 2007), than non-anxious controls.
411 Rapee and Heimberg (1997) posit that individuals determine the threat level of
412 a social situation according to the perceived probability of negative evaluation from
413 others, and perceived consequences or cost of that negative evaluation. This perceived
414 likelihood and cost has been found to be predictive of self-reported anxiety and levels
415 of avoidance (Zane & Williams, 1993). McManus, Clark, and Hackmann (2000)
416 have found that excessively high estimates of perceived likelihood and cost of poor

Behaviour Change
 Erika S. Penney and Maree J. Abbott

417 social outcome and evaluation during social events is specific to SAD, and not due to
418 general high anxiety or depression.
419 Rapee and Abbott (2007) developed a measure to assess the perceived probability
420 and consequences of negative evaluation in the context of a speech performance.
421 Participants were asked to indicate what they believed an objective rater would think
422 of their performance in terms of (1) the likelihood that the rater would have a
423 certain thought (e.g., ‘the rater will think you look silly’), and (2) the cost of negative
424 evaluation from the rater in terms of how distressing that would be. This measure
425 was found to have good psychometric properties and excellent internal consistency.
426 Greater perceived negative probability and cost estimates were found to be predictive
427 of increased negative rumination post speech (Rapee & Abbott, 2007).
428 As cognitive models posit that expectation of negative social outcomes occur
429 before the social event (Clark & Wells, 1995; Hofmann, 2007; Rapee & Heimberg,
430 1997), it is likely that negative probability and cost is likely to be predictive of
431 both pre- and post-event rumination. Empirical evidence supports this predictive
432 relationship between more negative probability and cost estimates and greater post-
433 event rumination (Rapee & Abbott, 2007).While there is evidence of an association
434 between anticipatory anxiety and increased negative expected likelihood and cost
435 (Butler & Mathews, 1987), there is limited research assessing whether pre-event
436 rumination in SAD is predicted by higher expectation of negative probability and
437 cost. It is reasonable to assume pre- and post-event rumination processes are similar,
438 as hypothesised in cognitive models (Clark & Wells, 1995; Rapee & Heimberg, 1997;
439 Hofmann, 2007); nonetheless, empirical research has yet to determine the validity of
440 this prediction.

441 Performance appraisal. Clark and Wells (1995; Clark, 2001) hypothesise that a rela-
442 tionship exists between negative rumination and poor self-appraisal of performance.
443 Individuals with SAD are often anxious in social situations as they expect to perform
444 poorly, and many studies support the link between social anxiety and poor self-ratings
445 of performance (Beidel, Turner, & Dancu, 1985; Dannahy & Stopa, 2007; Mellings
446 & Alden, 2000; Stopa & Clark, 1993). While individuals with SAD typically rate
447 their performance as poorer than controls, there is evidence that individuals do not
448 objectively perform worse than non-anxious individuals (Beidel et al., 1985; Rapee
449 & Lim, 1992) or, where any differences exist between SAD and non-anxious samples
450 in terms of objective performance, they are relatively small. Instead, individuals with
451 SAD typically rate their performance as of a poorer standard than an objective blind
452 rater of their performance (Chen, Rapee, & Abbott, 2013; Heimberg, Hope, Dodge,
453 & Becker, 1990; Rapee & Lim, 1992).
454 One frequently used measure of performance appraisal is the Performance Ques-
455 tionnaire (PQ), developed by Rapee and Lim (1992). This measure has been used to
456 measure a participant’s self-ratings, as well as an observer’s rating of the participant’s
457
90 performance, allowing for a direct comparison between self-appraisal of performance
458 and objective appraisal of performance. This measure was found to have good internal
459 consistency and reliability (Rapee & Lim, 1992).
460 Rapee and Lim (1992) found a significant discrepancy between self-ratings and
461 observer-ratings of performance on global features (e.g., overall impression on audi-
462 ence), such that individuals with SAD rated their own performance much more poorly
463 than did an observer. A strength of this study is that it used a clinical population of
464 individuals with diagnosed SAD as well as a non-anxious comparison group (Rapee &

Behaviour Change
 Rumination in Social Anxiety Disorder

465 Lim, 1992). The finding that individuals with SAD rated their performance as poorer
466 than an objective observer is consistent with cognitive models, which posit that indi-
467 viduals with SAD believe their own abilities are less than the expected standard and
468 that, as such, their performance will ‘fall short’ of expected audience standards (Rapee
469 & Heimberg, 1997). This discrepancy helps maintain anxiety and prevents individu-
470 als with SAD learning from exposure that social situations are less threatening than
471 originally thought.
472 More recently there is evidence that performance appraisal is not only linked
473 to social anxiety, but is predictive of negative rumination. Research indicates that
474 performance appraisal is the most likely unique predictor of post-event rumination
475 within a clinical sample of SAD participants (Abbott & Rapee, 2004; Chen et al.,
476 2013; Perini et al., 2006; Zou & Abbott, 2012). Abbott and Rapee (2004) investigated
477 performance appraisal and negative rumination in a treatment-seeking sample of
478 individuals with SAD. Participants completed a brief 3-minute speech on a topic
479 of their choice and then completed the PQ (Rapee & Lim, 1992). One week later,
480 participants were asked to complete a measure of rumination (the TQ) and the
481 PQ again. This study found that within individuals with SAD, poorer performance
482 appraisal uniquely predicted greater levels of post-event rumination one week after
483 the speech.
484 Using a similar methodology to Abbott and Rapee (2004), Perini et al. (2006)
485 found performance appraisal fully mediated the relationship between social anxiety
486 and post-event rumination, such that when performance appraisal was statistically
487 controlled, the relationship between social anxiety and post-event rumination was no
488 longer significant, suggesting that one’s perception of performance may be a key factor
489 in understanding the rumination process.
490 Similarly, Chen et al. (2013) conducted structural equation modelling of two
491 primary pathways: (1) showing that trait social anxiety is directly related to negative
492 post-event rumination, and (2) showing a second pathway whereby negative self-
493 appraisals and self-focused attention mediated the relationship between trait social
494 anxiety and negative rumination, again indicating the importance of performance
495 self-appraisal in understanding the rumination process (Chen et al., 2013).
496 Research by Rapee and Abbott (2007) suggests that this relationship is bidi-
497 rectional and may go some way to explaining the vicious cycle in which SAD is
498 maintained. Rapee and Abbott (2007) found that negative self-appraisal mediated
499 the relationship between trait social anxiety and negative post-event rumination;
500 however, they also found that negative rumination explained additional variance in
501 negative self-appraisals one week after a speech, above what was explained by the
502 self-appraisals at the time of the speech. This may help to explain the vicious cycle
503 of SAD, whereby cognitive biases in perception of one’s performance increase the
504 tendency to ruminate negatively on the performance, which further maintains one’s
505 negative self-representations of oneself as a social object (Rapee & Abbott, 2007).
506 There is also emerging evidence that performance appraisal may be associated with 91
507 pre-event rumination (Brown & Stopa, 2006; Vassilopolous, 2005). However, cau-
508 tion is warranted in this area as this research has typically used non-clinical samples,
509 and moreover, Brown and Stopa (2006) and Vassilopolous (2005) experimentally in-
510 duced participants to engage in pre-event rumination or distraction, and thus the exact
511 nature of the relationship between SAD, performance appraisal, and pre-event rumi-
512 nation is unclear. It is still to be determined if the relationship between performance
513 appraisal and post-event rumination will be mirrored in pre-event rumination.

Behaviour Change
 Erika S. Penney and Maree J. Abbott

514 Fear of negative evaluation. The link between social anxiety and fear of negative
515 evaluation is well established in both the literature and cognitive models. Rapee and
516 Heimberg (1997) postulate that it is the likelihood and cost associated with fear of
517 negative evaluation that plays a causal role in cognitive symptoms of SAD, such as
518 negative rumination.
519 Fear of negative evaluation has been found to be associated with increased levels
520 of negative rumination (Brozovich & Heimberg, 2011; Dannahy & Stopa, 2007;
521 Edwards et al., 2003; Fehm et al., 2007; Zou & Abbott, 2012). Dannahy and Stopa
522 (2007) investigated the link between fear of negative evaluation and post-event
523 rumination. Participants were categorised based on their Fear of Negative Evaluation
524 (FNE) scale scores. In a sample of 50 undergraduate students, the high FNE group
525 engaged in more post-event rumination overall than the low FNE group. High FNE
526 participants were also found to be more likely to rate their performance as poor.
527 Unfortunately, fear of negative evaluation was not used in hierarchical regression
528 analysis controlling for other variables, and social anxiety analogue samples were
529 used, limiting the generalisability of findings.
530 Additionally, rather than use the FNE to measure the cognitive symptom of fear of
531 negative evaluation, many studies (e.g., Dannahy & Stopa, 2007) have used the FNE
532 (Watson & Friend, 1969), and the Brief Fear of Negative Evaluation (BFNE) scale
533 (Leary, 1983; e.g., Brown & Stopa, 2006; Makkar & Grisham, 2011) ‘diagnostically’
534 to create their high and low socially anxious groups. As the FNE is not designed
535 to assess broader symptoms of SAD, this is an inappropriate use of these measures.
536 The FNE is less widely used than the BFNE (Leary, 1983), which is considered a
537 more valid measure of fear of negative evaluation (Leary, 1983; Rodebaugh et al.,
538 2004).
539 At present, Zou and Abbott’s (2012) is one of the only published studies using
540 a clinical population to investigate whether fear of negative evaluation is predictive
541 of post-event rumination. In a study comprising 80 undergraduate students (40 with
542 diagnosed SAD), fear of negative evaluation was not found to be a unique predictor
543 of post-event rumination over and above other variables, including performance ap-
544 praisal, trait social anxiety, state anxiety, state mood, or depression (Zou & Abbott,
545 2012). Thus, it is unclear at this stage whether this cognitive variable exerts direct ef-
546 fects on negative rumination or whether it may indirectly impact negative rumination
547 through the process of performance appraisal or other variables.

548 Self-Concept
549 Self-concept refers to one’s perception of oneself, including a cognitive schemata that
550 processes and stores self-relevant information about the ‘self’, such as concrete and
551 abstract memories, beliefs, feelings, and self-evaluations (Campbell, 1990; Campbell
552 & Lavallee, 1993). The self can have both an evaluative component (linked to trait
553
92 self-esteem and global self-reflection) and a knowledge component (i.e., beliefs about
554 one’s attributes; Campbell, 1990).
555 Self-concept is dynamic in that it is the collection of self-representations that
556 are active at any one time, and plays a role in regulating behaviour (Stopa, 2009).
557 Self-concept can be subject to change in response to life-events or therapy (Showers,
558 Limke, & Zeigler-Hilln, 2004), and is associated with self-esteem, such that individuals
559 with low self-esteem often have a less clear concept of who they are (Campbell, 1990;
560 Campbell & Lavallee, 1993).

Behaviour Change
 Rumination in Social Anxiety Disorder

561 Within SAD, it may be important to differentiate different aspects of the self-
562 concept; for example, content of the ‘self’ versus other process characteristics, such as
563 stability and clarity. Self-concept stability refers to the constancy of one’s self-concept
564 over time (Rosenberg, 1965), whereas self-concept clarity refers to whether one’s
565 self-concept is clearly defined, internally consistent, and temporally stable (Campbell
566 et al., 1996).
567 Campbell et al. (1996) developed the Self-Concept Clarity (SCC) scale us-
568 ing a sample of 471 undergraduate students. Participants were first administered a
569 40-item questionnaire designed to measure clarity of self-concept. The final measure
570 was created using 12 items that had good internal consistency and temporal stability,
571 and the scale has been found to be a reliable measure of the relatively stable individual
572 trait of self-concept clarity (Campbell et al., 1996).
573 Zeigler-Hill and Showers (2007) found that the self-concept of compartmentalised
574 individuals (i.e., individuals with a different view of themselves across different situa-
575 tions) is less stable than the self-concept of integrated individuals. This is particularly
576 relevant for social anxiety due to Clark and Wells’ (1995) assertion that individuals
577 with SAD are characterised by ‘unstable self-schemata’ (p. 76). Thus, those with low
578 self-concept clarity or highly compartmentalised self-concepts are believed to pos-
579 sess unstable self-esteem, which causes them to be hypersensitive to external cues to
580 determine their self-worth (Zeigler-Hill & Showers, 2007).
581 While cognitive models of SAD hypothesise the importance of negative self-beliefs
582 in maintaining SAD, most research has focused on the negative content of self-beliefs
583 (e.g., Schulz, Alpers, & Hofmann, 2008; Stopa & Clark, 1993; Tanner, Stopa, & De
584 Houwer, 2006; Turner, Johnson, Beidel, Heiser, & Lydiard, 2003), rather than the
585 clarity or structure of self-beliefs. Stopa, Brown, Luke, and Hirsch (2010) examined
586 structure of the self (e.g., self-organisation, self-complexity, and self-concept clarity)
587 in social anxiety. Socially anxious participants demonstrated less self-concept clarity
588 than non-anxious participants on both self-report and computerised measures of self-
589 consistency and confidence in self-related judgments. This indicates that socially
590 anxious individuals are less certain about who they are and have more difficulty
591 maintaining a consistent and coherent sense of self over time. Self-concept clarity
592 was found to correlate with social anxiety, even after controlling for the effects of
593 depression and self-esteem. Self-concept clarity uniquely predicted social anxiety and
594 accounted for 7% of the unique variance in social anxiety scores (Stopa et al., 2010).
595 Similarly, Wilson and Rapee (2006) found that individuals high in social anxiety
596 not only demonstrate more negative self-descriptions, but they also demonstrate less
597 certainty in their self-descriptions, than individuals low in social anxiety. This cer-
598 tainty in self-concept was not associated with other forms of psychopathology (e.g.,
599 depression, general anxiety, or stress) after controlling for the effects of social anxiety,
600 suggesting that there may be something about social anxiety that uniquely influences
601 the self-concept clarity of individuals (Wilson & Rapee, 2006).
602 This research is consistent with the Clark and Wells’ (1995) model, which suggests 93
603 SAD is associated with ‘unstable self-schemata’ (p. 76), whereby an individual’s self-
604 concept can change based on their environment. For example, individuals with SAD
605 typically have a more negative self-concept in social and performance situations, and a
606 more positive self-concept when alone, indicating that the clarity of their self-concept
607 is less certain and less stable than for those with social anxiety.
608 Campbell et al. (1996) found that lower levels of self-concept clarity were as-
609 sociated with higher levels of negative rumination, including affectively negative,

Behaviour Change
 Erika S. Penney and Maree J. Abbott

610 intrusive self-related thoughts. However, while there is evidence of the link between
611 social anxiety and self-concept clarity (Stop et al., 2010), and self-concept clarity and
612 negative rumination (Campbell et al., 1996), there is limited research on the role
613 of self-concept clarity in negative rumination within a SAD sample. Additionally,
614 methodological limitations in this area have included the use of social anxiety ana-
615 logue samples and correlational designs, indicating a need for experimentally designed
616 research investigating the role of self-concept clarity in predicting levels of pre- and
617 post-event rumination, in individuals with clinically diagnosed SAD.
618 A better understanding of the structure and clarity of one’s self-concept may aid the
619 understanding of the cycle of social anxiety. It is necessary for theoretical models and
620 empirical research to determine whether self-concept clarity in individuals with SAD
621 provides additional explanatory power in understanding the relationship between
622 perceived standards and self-appraisal, state social anxiety, and pre- and post-event
623 negative rumination.

624 Perceived Standards

625 According to Hofmann’s (2007) cognitive model of SAD, individuals are anxious in
626 social evaluative situations because they perceive that the audience or social other has
627 a high expected standard for their social performance. Rapee and Heimberg’s (1997)
628 cognitive model suggests that individuals with SAD form negative self-representations
629 of themselves based on how they believe the social audience is evaluating them. This
630 suggests that negative self-appraisals that make up the self-representation are malleable
631 responses, which are created based on the perceived external feedback that a social
632 peer or audience may give about the expected standard.
633 Consistent with Hofmann (2007), individuals with SAD have been found to
634 demonstrate a greater discrepancy than non-anxious controls between their perceived
635 expected standards and their perceived abilities and skills, which is particularly pro-
636 nounced when an individual is under social threat (Wallace & Alden, 1991, 1995).
637 In a study by Wallace and Alden (1995), socially anxious participants who had a suc-
638 cessful social interaction and received positive feedback about their performance were
639 more likely than non-anxious controls to experience increased anxiety in anticipation
640 of a second interaction. Although one might expect that positive social experiences
641 would raise an individual’s belief in their own social skills, instead individuals with
642 SAD believe that their initial success has led to a higher expected standard in subse-
643 quent tasks (Wallace & Alden, 1995). Thus, it can be concluded that perceived social
644 standards are associated with self-appraisal of performance, perceived social abilities,
645 and state anxiety.
646 Moscovitch and Hofmann (2007) investigated the link between SAD and per-
647 ceived standards. They aimed to examine whether subjective and objective perfor-
648 mance appraisal was moderated by type of standard. Prior to giving a speech, partici-
649
94 pants were shown a video of an exemplar speech and either informed that this was the
650 expected standard (high standard), or that it was higher than the expected standard
651 (low standard), or given no feedback (ambiguous standard; Moscovitch & Hofmann,
652 2007).
653 Within a clinical sample of SAD participants, performance appraisal was poorer
654 in the high and ambiguous standard conditions, compared to non-anxious controls
655 and individuals with SAD in the low standard condition (Moscovitch & Hofmann,
656 2007). Performance appraisal for individuals with SAD in the low standard condition

Behaviour Change
 Rumination in Social Anxiety Disorder

657 was indistinguishable from non-anxious controls. However, individuals with SAD
658 reported significantly higher levels of state anxiety than controls across all conditions
659 (Moscovitch & Hofmann, 2007). This study supports hypotheses from the cognitive
660 models that when individuals with SAD perceive the expected social standard to be
661 high, they experience more state anxiety and poorer self-appraisals of performance.
662 However, it can be questioned whether the ambiguous standard condition was well
663 designed, as it may have been that in the absence of any instructions, the participants
664 assumed that they were expected to perform to the standard of the exemplar video.
665 Thus, assessing perceived audience standards as well as one’s perceived ability to meet
666 such standards is necessary in future replications.
667 Given the evidence that performance appraisal is predictive of negative rumina-
668 tion, it seems reasonable to assume that perceived standards will exert influence on
669 negative rumination, possibly via self-appraisals. Further research is needed to extend
670 the work of Moscovitch and Hofmann (2007) to investigate not only the role of
671 performance standard on performance appraisal, threat appraisal, and state anxiety,
672 but also its impact on pre- and post-event rumination for groups of SAD participants
673 and non-clinical control participants.

674 Treatments Targeting Negative Rumination in SAD

675 SAD can be a debilitating disorder, and when it is untreated it will typically run a
676 chronic course (Dewit, Ogborne, Offord, & MacDonald, 1999). Cognitive models of
677 SAD label pre- and post-event rumination as key maintaining factors of the disor-
678 der, and thus targeting these cognitive variables may be essential in alleviating the
679 symptoms of SAD. Clark and Wells (1995) suggest that CBT should target cognitive
680 self-appraisals and encourage less negative self-appraisals after performance and eval-
681 uative situations. This in turn may result in reduced negative rumination and help
682 to break the vicious cycle that maintains both negative self-appraisals and negative
683 rumination (Clark & Wells, 1995). Emerging empirical research has also supported
684 the importance of targeting negative rumination during treatment, as increased rumi-
685 nation may be associated with an increased severity of the disorder (Abbott & Rapee,
686 2004; Trew & Alden, 2009).
687 Makkar and Grisham (2013) conducted a study where both high and low socially
688 anxious participants were exposed to false physiological feedback indicating that
689 the participant was experiencing an increase or decrease in their heart rate prior
690 to a speech task. Participants who believed their heart rate was increasing reported
691 higher levels of post-event rumination and poorer appraisals of their performance,
692 relative to participants who believed their heart rate had decreased. In line with
693 cognitive models, these effects were mediated by self-focused attention (Makkar &
694 Grisham, 2013). This suggests that the negative attention and appraisal given to
695 somatic symptoms contributes to the maintenance of SAD and negative rumination,
696 and thus challenging catastrophic interpretations of somatic symptoms may help to 95
697 reduce post-event rumination and the cycle of SAD.
698 Few studies have evaluated whether CBT for SAD reduces the frequency and
699 negativity of pre- and post-event rumination. Evaluating the efficacy of current treat-
700 ments in the reduction of this cognitive symptom is important due to the theoretical
701 importance of this factor in the maintenance of the disorder. Abbott and Rapee
702 (2004) found that individuals with SAD engaged in significantly less negative post-
703 event rumination after a CBT intervention. Additionally, individuals with SAD had

Behaviour Change
 Erika S. Penney and Maree J. Abbott

704 significantly fewer negative self-appraisals of their performance post-treatment com-

705 pared to pre-treatment. That both negative self-appraisals and negative rumination
706 reduced after this CBT intervention is significant, given the empirical and theoretical
707 link between these two variables and their role in SAD.
708 More research is needed to evaluate the specific processes that are linked with
709 a reduction in negative rumination. For instance, it is yet to be determined if per-
710 haps cognitive restructuring reduces perceived standards and negative self-appraisal,
711 which in turn reduces rumination; or if behavioural and exposure exercises demon-
712 strate the lower likelihood and cost of negative consequences in an evaluative sit-
713 uation, thereby reducing the frequency of rumination and the negativity of self-
714 appraisals.

715 Methodological Limitations of Past Research

716 While there is a growing amount of research investigating negative rumination in
717 social anxiety, there are significant methodological limitations within the literature
718 and a paucity of experimental studies. Many of the earlier studies of negative rumina-
719 tion have used self-report methods of inducing social anxiety, by asking participants
720 to recall a time they felt socially anxious (Fehm et al., 2007; Kocovski & Rector,
721 2007; Rachman et al., 2000), and some have used experimental manipulations using
722 state tasks and measures, such as a speech task (Abbott & Rapee, 2004; Cody &
723 Teachman, 2012; Perini et al., 2012) or an interaction task (Brozovich & Heimberg,
724 2011; Dannahy & Stopa, 2007). Despite the growing literature examining post-event
725 rumination in social anxiety, there is a lack of experimentally designed research in
726 the area of pre-event rumination.
727 An additional limitation is that many studies have used social anxiety analogue
728 samples rather than participants with diagnosed SAD. Some studies have used the
Q2 729 Social Interaction Anxiety Scale (SIAS; e.g., Brozovich & Heimberg, 2011; Cody
Q2 730 & Teachman, 2012) or the SADS (e.g., Mellings & Alden, 2000) to create their
731 high and low socially anxious participant groups. While these measures of social
732 anxiety symptoms do correlate highly with diagnoses of SAD (Mattick & Clark,
733 1998; Watson & Friend, 1969), they are not sufficient on their own to diagnose SAD.
734 Of more concern are studies which have used the FNE (e.g., Dannahy & Stopa, 2007;
735 Vassilopoulos, 2005; Wong & Moulds, 2009) or BFNE (e.g., Brown & Stopa, 2006;
736 Makkar & Grisham, 2011) to create their high and low socially anxious groups, as the
737 FNE is not a diagnostic measure of SAD and high scores are also obtained from other
738 Axis I samples, such as depressed participants.
739 While there is experimentally designed research in the area of post-event rumi-
740 nation utilising clinical samples with diagnosed SAD (Abbott & Rapee, 2004; Chen
741 et al., 2012; Kocovski & Rector, 2008; Laposa & Rector, 2011; Perini et al., 2012;
742 Rapee & Abbott 2007; Zou & Abbott, 2012), the current authors are not aware
743
96 of any studies that have employed an experimental design in the area of pre-event
744 rumination utilising a clinical sample.
745 Additionally, the preliminary stage of the research in this field necessarily means
746 there is limited research that has empirically examined processes associated with pre-
747 and post-event rumination in SAD. The exact nature of the relationship between
748 processes, such as inflated perceived standards, performance and threat appraisal,
749 state anxiety, and the ruminative processes, particularly with respect to pre-event
750 rumination is as yet unclear, warranting future research.

Behaviour Change
 Rumination in Social Anxiety Disorder

751 Summary and Future Research Implications

752 This review has provided empirical support for the role of post-event process-
753 ing in maintaining SAD, as hypothesised by contemporary cognitive models.
754 There is evidence to suggest that performance and threat appraisal are predic-
755 tive of post-event rumination in SAD; however, whether this pattern is mirrored
756 for pre-event rumination remains to be empirically evaluated. There is also ev-
757 idence that self-concept clarity and perceived standards may impact state social
758 anxiety, appraisal processes, and negative rumination. Further research is war-
759 ranted to examine the impact of performance standards, threat and performance
760 appraisals, and self-concept variables on models of pre- and post-event rumina-
761 tion. Cognitive models suggest that significant predictors should show compara-
762 ble effects across pre- and post-event processing periods, irrespective of the social
763 task.
764 Future research is needed to address the methodological limitations of past re-
765 search, and examine the key hypotheses posited by the cognitive models of SAD
766 by experimentally manipulating perceived social standards in order to examine the
767 impact of high and low perceived social standards on appraisal processes (i.e., both
768 threat appraisal and performance appraisal), state social anxiety, and pre-event and
769 post-event rumination. Research is needed to clarify whether pre-event rumination
770 can be predicted by the same symptom, cognitive, and appraisal variables that are
771 predictive of post-event rumination. This will allow for a better understanding of
772 SAD and will contribute to cognitive models and treatment protocols for SAD by
773 clarifying the interconnectiveness of pre-event and post-event rumination, and their
774 role in maintaining state social anxiety and SAD.

775 References
776 Abbott, M., & Rapee, R. (2004). Post-event rumination and negative self-appraisal in social phobia
777 before and after treatment. Journal of Abnormal Psychology, 113(1), 136–144.
778 American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th
779 ed., rev.). Washington, DC: American Psychiatric Association.
780 American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th
781 ed.). Washington, DC: Author.
782 Australian Bureau of Statistics. (2007). National survey of mental health and wellbeing: Summary of
783 results. Canberra, Australia: Author. Retrieved from http://www.abs.gov.au/AUSSTATS/abs@.
784 nsf/mf/4326.0
785 Beck, A.T. (1976). Cognitive therapy and emotional disorders. New York: International Universities
786 Press.
787 Beck, A.T., Emery, G., & Greenberg, R. (1985). Anxiety disorders and phobias: A cognitive perspective.
788 New York: Basic Books.
789 Beidel, D.C., Turner, S.M., & Dancu, C.V. (1985). Physiological, cognitive, and behavioral aspects
790 fo social anxiety. Behaviour Research and Therapy, 23, 109–117.
791 Brown, M., & Stopa, L. (2006). Does anticipation help or hinder performance in a subsequent speech? 97
792 Behavioural and Cognitive Psychotherapy, 35, 133–147.
793 Brozovich, F., & Heimberg, R.G. (2008). An analysis of post-event processing in social anxiety
794 disorder. Clinical Psychology Review, 28, 891–903.
795 Brozovich, F., & Heimberg, R. (2011). The relationship of post-event processing to self-evaluation
796 of performance in social anxiety. Behavior Therapy, 42, 224–235.
797 Butler, G., & Mathews, A. (1983). Cognitive processes in anxiety. Advances in Behaviour Research &
798 Therapy, 5(1), 51–62.

Behaviour Change
 Erika S. Penney and Maree J. Abbott

799 Campbell, J.D., & Lavallee, L.F. (1993). Who am I?: The role of self-concept confusion in under-
800 standing the behaviour of people with low-self esteem. In R.F. Baumeister (Ed.), Self-esteem: The
801 puzzle of low self-regard (pp. 3–20). New York: Plenum.
802 Chen, J., Rapee, R.M., & Abbott, M.J. (2013). Mediators of the relationship between social anxiety
803 and post-event rumination. Journal of Anxiety Disorders, 27, 1–8.
804 Clark, D. (2001). A cognitive perspective on social phobia. In W.R. Crozier & L.E. Alden (Eds.),
805 International handbook of social anxiety: Concepts, research and interventions relating to the self and
806 shyness (pp. 405–430). New York: John Wiley & Sons.
807 Clark, D.M, Elhers, A., Hackmann, A., McManus, F., Fennell, M., Grey, N., Waddington, L., &
808 Wild, J. (2006). Cognitive therapy versus exposure and applied relaxation in social phobia: A
809 randomized controlled trial. Journal of Consulting and Clinical Psychology, 74, 568–578.
810 Clark, D., & Wells, A. (1995). A cognitive model of social phobia. In R. Heimberg, M. Liebowitz,
811 D. Hope, & R. Schneier, (Eds.), (Social phobia: Diagnosis, assessment and treatment (pp. 69–93).
812 New York: Guilford Press.
813 Cody, M.W., & Teachman, B.A. (2012). Global and local evaluations of public speaking performance
814 in social anxiety. Behavior Therapy, 42, 601–611.
815 Crum, R., & Pratt, L. (2001). Risk of heavy drinking and alcohol use disorders in social phobia: A
816 prospective analysis. American Journal of Psychiatry, 158, 1693–1700.
817 Dannahy, L., & Stopa, L. (2007). Post-event processing in social anxiety. Behaviour Research and
818 Therapy, 45, 1207–1219.
819 Davidson, J.R.T., Foa, E.B., Huppert, J.D., Keefe, F., Franklin, M., Compton, J., . . . Kishore, G.
820 (2004). Fluoxetine, comprehensive cognitive behavioral therapy, and placebo in generalized
821 social phobia. Archives of General Psychiatry, 61, 1005–1013.
822 Dewit, D.J., Ogborne, A., Offord, D.R., & MacDonald, K.M. (1999). Antecedents of the risk of
823 recovery from DSM–III–R social phobia. Psychological Medicine, 29, 569–582.
824 Edwards, S.L., Rapee, R.M., & Franklin, J. (2003). Postevent rumination and recall bias for a social
825 performance event in high and low socially anxious individuals. Cognitive Therapy and Research,
826 27, 603–617.
827 Fehm, L., Schneider, G., & Hoyer, J. (2007). Is post-event processing specific for social anxiety?
828 Journal of Behavior Therapy and Experimental Psychiatry, 38, 11–22.
829 Feske, U., & Chambless, D.L. (1995). Cognitive behavioral versus exposure only treatment for social
830 phobia: A meta-analysis. Behavior Therapy, 26, 695–720.
831 Furmark, T. (2002). Social phobia: Overview of community surveys. Acta Psychiatrica Scandinavica,
832 105(2), 84–93.
833 Gould, R.A., Buckminstel, S., Pollack, M.H., Otto, M.W., & Yap, L. (1997). Cognitive-behavioral
834 and pharmacological treatment for social phobia: A meta-analysis. Clinical Psychology: Science
835 and Practice, 4, 291–306.
836 Heimberg, R.G., & Becker, R.E. (2002). Cognitive behavioral group therapy for social phobia: Basic
837 mechanisms and clinical strategies. New York: Guilford Press.
838 Heimberg, R.G., Hope, D.A., Dodge, C.S., & Becker, R.E. (1990). DSM-III subtypes of social phobia:
839 Comparison of generalised social phobics and public speaking phobics. Journal of Nervous and
840 Mental Disease, 178, 172–179.
841 Herbert, J.D., Rheingold, A.A., & Goldstein, S.G. (2002). Brief cognitive behavioral group therapy
842 for social anxiety disorder. Cognitive and Behavioral Practice, 9, 1–8.
98
843 Hofmann, S. (2007). Cognitive factors that maintain social anxiety disorder: A comprehensive model
844 and its treatment implications. Cognitive Behaviour Therapy, 36(4), 193–209.
845 Kashdan, T., & Roberts, J. (2007). Social anxiety, depressive symptoms, and post-event rumination:
846 Affective consequences and social contextual influences. Journal of Anxiety Disorders, 21, 284–
847 301.
848 Kessler, R.C., Berglund, P., Demler, O., Jin, R., & Walters, E. (2005). Lifetime prevalence and
849 age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication.
850 Archives of General Psychiatry, 62, 593–602.

Behaviour Change
 Rumination in Social Anxiety Disorder

851 Kim, E. (2005). The effect of the decreased safety behaviors on anxiety and negative thoughts in
852 social phobics. Anxiety Disorders, 19, 69–86.
853 Kocovski, N.L., & Rector, N.A. (2007). Predictors of Post-Event Rumination Related to Social
854 Anxiety. Cognitive Behaviour Therapy, 36, 112–122.
855 Kocovski, N.L., & Rector, N.A. (2008). Post-event processing in social anxiety disorder: Idiosyncratic
856 priming during the course of CBT. Cognitive Therapy and Research, 32, 23–36.
857 Lampe, L., Slade, T., Issakidis, C., & Andrews, G. (2003). Social phobia in the Australian National
858 Survey of Mental Health and Well-Being (NSMHWB). Psychological Medicine, 33, 637–646.
859 Laposa, J.M., & Rector, N.A. (2011). A prospective examination of predictors of post-event process-
860 ing following videotaped exposures in group cognitive behavioural therapy for individuals with
861 social phobia. Journal of Anxiety Disorders, 25(4), 568–73.
862 Leary, M.R. (1983). A brief version of the Fear of Negative Evaluation Scale. Personality and Social
863 Psychology Bulletin, 9, 371–375.
864 McEvoy, P., & Kingsep, P. (2006). The post-event processing questionnaire in a clinical sample with
865 social phobia. Behaviour Research and Therapy, 44, 1689–1697.
866 McManus, F., Clark, D.M., & Hackmann, A. (2000). Specificity of cognitive bias in social phobia
867 and their role in recovery. Behavioural and Cognitive Psychotherapy, 28, 201–209.
868 Makkar, S., & Grisham, J. (2011). The predictors and contents of post-event processing in social
869 anxiety. Cognitive Therapy & Research, 35, 118–133.
870 Makkar, S., & Grisham, J. (2013). Effects of false feedback on affect, cognition, behaviour, and
871 postevent processing: The mediating role of self-focussed attention. Behavior Therapy, 44, 111–
872 124.
873 Mattick, R.P., & Clarke, J.C. (1998). Development and validation of measures of social phobia
874 scrutiny fear and social interaction anxiety. Behaviour Research and Therapy, 36, 455–470.
875 Mellings, T.M.B., & Alden, L.E. (2000). Cognitive processes in social anxiety: The effects of self-
876 focus, rumination and anticipatory processing. Behaviour Research and Therapy, 38, 243–257.
877 Moscovitch, D.A., & Hofmann, S.G. (2007). When ambiguity hurts: Social standards moder-
878 ate self-appraisals in generalized social phobia. Behaviour Research and Therapy, 45, 1039–
879 1052.
880 Nolen-Hoeksema, S. (1987). Sex differences in unipolar depression: Evidence and theory. Psycholog-
881 ical Bulletin, 101, 259–282.
882 Nolen-Hoeksema, S. (1991). Responses to depression and their effects in the duration of depressive
883 episodes. Journal of Abnormal Psychology, 100, 569–582.
884 Nolen-Hoeksema, S. (2004). The response styles theory. In C. Papageorgiou & A. Wells (Eds.),
885 Depressive rumination: Nature, theory, and treatment (pp. 107–124). Chichester, UK: John Wiley
886 & Sons.
887 Nolen-Hoeksema, S., Parker, L.E., & Larson, J. (1994). Ruminative coping with depressed mood
888 following loss. Journal of Personality and Social Psychology, 67, 92–104.
889 Perini, S., Abbott, M.J., & Rapee, R.M. (2006). Perception of performance as a mediator in the
890 relationship between social anxiety and negative post-event rumination. Cognitive Therapy and
891 Research, 30, 645–659.
892 Perugi, G., Frare, F., Toni, C., Mata, B., & Akiskal, H.S. (2001). Bipolar II and unipolar comorbidity
893 in 153 outpatients with social phobia. Comprehensive Psychiatry, 42, 375–381.
894 Rachman, S., Gruter-Andrew, J., & Shafran, R. (2000). Post-event processing in social anxiety.
895 Behaviour Research and Therapy, 38, 611–617. 99
896 Rapee, R.M., & Abbott, M.J. (2007). Modelling relationships between cognitive variables during
897 and following public speaking in participants with social phobia. Behaviour Research and Therapy,
898 45, 2977–2989.
899 Rapee, R.M., Gaston, J.E., & Abbott, M.J. (2009). Testing the efficacy of theoretically derived
900 improvements in the treatment of social phobia. Journal of Consulting and Clinical Psychology, 77,
901 317–327.

Behaviour Change

View publication stats