Available online at www.sciencedirect.com Behavior Therapy 42 (2011) 183 – 196 www.elsevier.com/locate/bt The Negative Self-Portrayal Scale: Development, Validation, and Application to Social Anxiety David A. Moscovitch Vanessa Huyder University of Waterloo The Negative Self-Portrayal Scale (NSPS) is a new questionnaire designed to assess the extent to which individuals are concerned that specific self-attributes they view as being deficient will be exposed to scrutiny and evaluation by critical others in social situations. These concerns have been proposed to drive symptoms of social anxiety and account for individual differences in social fears and avoidance behaviors (Moscovitch, 2009). Here, we introduce the NSPS and examine its factor structure and psychometric properties across two large samples of North American undergraduate students with normally distributed symptoms of social anxiety. Exploratory and confirmatory factor analyses supported a 3-factor solution representing concerns about (a) social competence; (b) physical appearance; and (c) signs of anxiety. The NSPS was found to have good internal consistency and test-retest reliability, strong convergent validity, and adequate discriminant validity. In addition, NSPS total scores accounted for a significant proportion of unique variance in self-concealment (i.e., safety) behaviors over and above established symptom measures of social interaction anxiety, social performance anxiety, and depression. Results are discussed in relation to theoretical models of social anxiety and the potential utility of the NSPS for both clinical research and practice. This research was supported, in part, thanks to funding from the Canada Research Chairs Program to David A. Moscovitch. We are grateful to Dr. Erik Woody for his helpful suggestions and assistance with the statistical analyses. Address correspondence to David A. Moscovitch, Ph.D., Canada Research Chair in Mental Health Research, Department of Psychology, University of Waterloo, 200 University Ave. West Waterloo, Ontario, Canada N2L 3G1; e-mail:
[email protected]. 0005-7894/10/183–196/$1.00/0 © 2010 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved. THERE IS A GROWING consensus that the perception of self plays a crucial role in the pathogenesis and persistence of social anxiety (cf. Stopa, 2009a). Although specific conceptualizations of the self differ across contemporary cognitive and interpersonal models of social anxiety (Alden & Taylor, 2004; Clark & Wells, 1995; Hofmann, 2007; Leary & Kowalski, 1995; Rapee & Heimberg, 1997), there is now strong empirical support for the view that symptoms of social anxiety arise from the discrepancy produced by the motivation to convey a desired social impression of oneself to others in combination with the expectation that one will fail to do so (e.g., Schlenker & Leary, 1985). Indeed, numerous studies have shown that highly socially anxious individuals significantly underestimate their social performance abilities as well as other salient social self-attributes, and routinely provide selfratings that fall well below their perception of the standard they believe is required by relevant “audience” evaluators (e.g., for reviews of this literature, see Moscovitch & Hofmann, 2007; Orr & Moscovitch, 2010a; Moscovitch, Orr, Rowa, Gehring Reimer, & Antony, 2009). Under social threat, highly socially anxious and phobic individuals allocate increased attentional resources toward monitoring their internal anxiety cues and other perceived negative features of the self (Rapee & Heimberg, 1997; Spurr & Stopa, 2002). This increase in self-monitoring coincides with the experience of spontaneous, recurrent, intrusive, and often vivid negative mental self-images that are viewed from an observer's perspective (Hackmann, Clark, & McManus, 2000; Hackmann, Surawy, & Clark, 1998). These images tend to be negatively g. and (d) concerns about personality. thereby having the potential to enhance SAD treatment outcomes. & Heimberg. Thus. 2006) and personality characteristics (Mansell & Clark. help facilitate the delivery of CBT for SAD in a more individually customized manner. SPAI. 2008). In the same vein. Palyo. Heinrichs. & Moscovitch. investigators have only recently begun to examine the link between social anxiety and heightened concerns about physical appearance (Hart. in turn.g. as well as the types of subtle avoidance or safety behaviors they tend to use (e. Moscovitch (2009) recently recommended that scientist-practitioners engaged in assessing and treating patients with social anxiety disorder (SAD) within an exposurebased cognitive behavioral therapy (CBT) framework shift their focus in treatment from targeting patients' feared social situations.. Wilson & Rapee. 1999. & Clark. 1997). (c) concerns about physical appearance. Most of the excellent social anxiety self-report measures currently in circulation (e.. . social situations that are endorsed as being anxiety-provoking are those which individuals perceive as having the capacity to expose their perceived self-deficiencies for public consumption. According to Moscovitch (2009). Moscovitch proposed that the feared stimuli in SAD are specific self-attributes that socially anxious individuals perceive as being flawed or contrary to perceived sociocultural norms. 2004 for a review of this literature). we wished to create a novel measure that directly assessed the specific types of self-attributes that Moscovitch (2009) proposed would elicit concerns about self-presentation for socially anxious individuals in anxiety-provoking social situations. SPIN.184 moscovitch & huyder distorted and exaggerated representations of the self. Although numerous psychometrically validated measures of social anxiety are now widely available. SPS. Rapee & Abbott.. 2005). hypothesizing that it might be possible to organize these self-attributes into broad themes that fall across four nonorthogonal dimensions: (a) concerns about social skills and behaviors. Holle. Hurley et al. we follow up the EFA in Sample 2 with confirmatory factor analysis (CFA) of the derived model and plausible competing models (Brown. Thus. & Roemer. subtle avoidance and safety behaviors represent attempts to conceal or prevent the public exposure of those self-attributes. Specifically.. 2006). 2004. no previous studies have tested these predictions empirically. (b) concerns about visible signs of anxiety. Socially anxious people are heterogeneous with respect to the kinds of social situations they fear and avoid (see Hofmann. Sacadura. see Antony. none specifically assesses self-portrayal concerns across the proposed dimensions. As a result. This may. SIAS. Being able to identify and measure such concerns might conceivably both enhance our understanding of symptom heterogeneity in social anxiety and guide case conceptualization of SAD in clinical practice. based on the results from Sample 1. core concerns about self-attribute flaws are proposed to drive the constellation of social anxiety symptoms—subjective distress.. Izgiç. and interpersonal lives of individuals with SAD (Ledley & Heimberg. Fresco. for a review of this literature). Flora. Moscovitch's (2009) model offers a potentially useful. avoidance. these images capture precious emotional and cognitive resources and possibly prevent the encoding of salient positive social and interpersonal cues in the external environment (Clark & Wells. and self-concealment—that are associated with high levels of functional impairment in the emotional. Here. & Kugu. 2006). nevertheless. we examine the factor structure and psychometric properties of the NSPS in two large samples of individuals with social anxiety symptoms distributed across the normal spectrum and begin to investigate some of the central. theory-guided heuristic for conceptualizing symptom variations and individual differences in social anxiety. as recommended in the early stages of scale development (e. variations in social fears and safety behaviors are functionally related to underlying differences in self-attribute concerns across the proposed dimensions. Our objective in designing the Negative Self-Portrayal Scale (NSPS) was not to add another questionnaire to this list. LSAS-SR. they are perceived and processed by socially anxious individuals as being accurate. Akyüz. While research on SAD has historically focused on socially anxious individuals' concerns about social skills/behaviors and signs of anxiety (see Moscovitch. etc. On the basis of clinical observations and the burgeoning empirical literature on the central role of the self in social anxiety. Next. Despite our theoretical hypotheses about the factor structure of the NSPS across the proposed dimensions. Rather. vocational. Dogan. theoretical claims of Moscovitch's (2009) model. to targeting more directly the core feared stimuli in SAD. 2001) assess the level or type of social anxiety symptoms that individuals may experience across a variety of social performance and interaction situations. Orsillo. we first present an exploratory factor analysis (EFA) of the NSPS in participant Sample 1. 2009. 2008. McManus. 1995). but previously untested.g. Peters.93) and convergent validity (r = 0.91-0. Heinrichs. internal consistency (Cronbach's alpha of 0. and convergent and discriminant validity (e. provided they were face valid and nonredundant with other items already on the list. Initial Item Generation An initial pool of 62 NSPS items was constructed by the authors.. Assessors were instructed to eliminate any items that seemed redundant or unclear.77 with other measures of social anxiety). Any new items that were suggested by at least one assessor were added to the NSPS item pool. Both scales have been found to have good test-retest reliability (r = 0. Barrios. Given that the SBQ is an unpublished scale with unknown psychometric properties. with descriptive anchors listed as “not at all concerned” to “extremely concerned.90 and 0. Liebowitz Social Anxiety Scale–Self Report Version (LSAS-SR.59-0. resulting in a total of 57 NSPS items.1 Social Phobia Scale (SPS) and Social Interaction Anxiety Scale (SIAS) (Mattick & Clarke. we chose to add items deemed by 185 experienced clinicians to be relevant to socially anxious individuals. the LSAS-SR has been shown to have good test-retest reliability (r = 0. see also McManus et al.negative self-portrayal scale Method participants Two groups of nonoverlapping undergraduate participants (n = 225. the OCI-SV is an 18-item 1 A copy of the SCBI is available upon request from the first author.. Kopper. 1998.” There were no reverse-scored items.. For the fear subscale (LSAS-F). 1998) The SPS and SIAS are companion 20-item selfreport scales that assess social performance and interaction fears. in separate subscales. according to the definitions given to them. Response options per item ranged from 1 to 5. Foa et al. concerns about visible signs of anxiety. & Wells. higher scores represented greater self-portrayal concerns. . urban Canadian university completed a series of online questionnaires in exchange for partial course credit. measures Negative Self-Portrayal Scale (NSPS) Participants were administered the 57-item NSPS. add items they thought were appropriate. across a number of social and performance situations. 2000). This was done using the expert-guided process described above for the NSPS. & Amir. These were then sent via electronic mail to three independent Ph. Butler. among patients with SAD. Respondents were instructed to rate how frequently they typically engage in each of the listed behaviors before or during anxiety-provoking social situations in order to prevent or alleviate their experience of anxiety. & Hofmann.e. 1999) that assesses fear and avoidance. resulting in a total of 53 SCBI items. Each item is rated on a scale from 0 (not at all characteristic or true of me) to 4 (extremely characteristic or true of me). respectively. Kim. each item is rated on a scale of 0 (none) to 3 (severe). Osman. This measure is essentially an expanded version of the unpublished Social Behaviour Questionnaire (SBQ. 2002) Originally adapted from the longer 42-item Obsessive-Compulsive Inventory (Foa. & Chiros. Salkvoskis. Instructions to respondents asked them to rate the degree to which they are concerned about certain aspects of themselves when in anxiety-provoking social situations. 2002.and reverse-scored items. Like the clinician-administered version. 1998. each item is rated on a scale of 0 (never) to 3 (usually). Coles. 2001). safety) behaviors.e. and indicate the dimension in which they thought each self-attribute item belonged.g. correlations with other measures of social anxiety were found to be significantly stronger than correlations with measures of depression) (Baker et al. For the avoidance subscale (LSAS-A). Fresco et al.. Fennell.91). Self-Concealment Behaviors Inventory (SCBI) The SCBI is a self-report measure of self-concealment (i. Following this process.-level scientist-practitioners with expertise in social anxiety research and extensive experience treating individuals with SAD in clinical settings. and concerns about personality). Obsessive Compulsive Inventory–Short Version (OCI-SV.. Gutierrez.D. 2009) dimensions of self-attribute concerns (i. Hackmann. and to distinguish between individuals with and without SAD (Mattick & Clarke. 1995. Baker.83). 2 items were removed and 9 items were added to our original 46-item pool. with a combination of sum.. Following this process. Clark.. n = 316) at a large.. 2009). concerns about physical appearance. 2002) The LSAS-SR is a 24-item self-report version of a well-validated clinician-administered instrument (Heimberg et al. 1998). Any items eliminated or misallocated by more than one assessor were removed from the respective inventory. high levels of internal consistency (Cronbach's alpha indices of 0.95). McManus. concerns about social skills and behaviors. Assessors were given a randomized list of the preliminary items along with an instruction sheet defining the four hypothesized (Moscovitch. 11 items were removed and 6 items were added. Kozak. Osman.3% 4.60–.. When imputing these missing values.86 in clinical samples). Respondents are asked to rate how much each item distresses or bothers them on a scale from 0 (not at all) to 4 (extremely). participants were then asked to complete the NSPS and SCBI in a separate session 2 days after Session 1. One week prior to this session.83 in nonclinical samples and r = . LSAS-SR. we also asked participants to complete the NSPS (but none of the other measures) to enable us to evaluate its test-retest reliability. 1996) The BDI-II is a 21-item self-report instrument for assessing the intensity of depression in diagnosed patients and detecting symptoms of depression in normal populations.93) (Foa et al.6% 4.g. missing data were imputed for all measures using the expectation-maximization (EM) estimation method in SPSS Statistics 18.92 in nonclinical populations. and the correlation of quantitative (scale) variables of missing values. LSAS-SR. and SIAS. Gutierrez. the BDI-II has been found to be significantly more strongly correlated with depression relevant measures. 276 for Group 2) Caucasian Asian East Indian Middle Eastern Other Group 1 Group 2 19.6 years (SD = 3.g. Steer. 2002). Respondents are instructed to choose the statement amongst a group of statements that best applies to how they generally feel. SPS.6% 8. The BDI-II has consistently been shown to have adequate test-retest reliability (r = . I collect things I don't need. results Descriptive Analyses Demographic information (reported age. participants in Group 2 completed all of the counterbalanced measures—BDI-II.g.84).g. Preliminary item-by-item frequency analyses for each measure administered in the study indicated that the total proportion of missing data was low (0% to 2% for the vast majority of items) across all measures and participants in both groups.85 with other measures of OCD).9% 3. 2008). 277 for Group 2) Gender (% female) (N = 225 for Group 1.3% 45. Using an iterative process. OCI-SV. than with measures of anxiety.73 to . F(6..70) with the Beck Depression Inventory.0% 47. the EM method estimates the means.9% 12.35) 72. however. p N .48–. Thus. A MANOVA with gender as the grouping variable was conducted for Groups 1 and 2. p b . and SCBI—in one session.9% 20. Williams. .01).9% Note. SPS. 218) = 1. was Foa and colleagues' (2002) finding that the OCI-SV had a high correlation (r = 0. There were no significant effects of gender on any of the measures. Beck Depression Inventory II (BDI-II. Participants in Group 1 completed two online sessions. Due to the anticipated length and monotony of completing multiple questionnaires in one sitting. I get upset if objects are not arranged properly.. and high internal consistency (Cronbach's alpha indices of 0.8% 32. Barrios. SIAS.186 moscovitch & huyder brief self-report scale that measures concerns associated with obsessive-compulsive disorder (e.76 to 0. etc..3% 10.53–0. The OCI-SV has been found to have excellent test-retest reliability (r = 0.05. gender. we included only the measures that were used in each relevant analysis described below. 281 for Group 2) Ethnicity (N=224 for Group 1. a conceptually distinct measure of depression. Of note.. OCISV.0 (2009). Thus. Each group of statements ranges from 0 (e.67. The first session included administration of the BDI-II. Ns vary for Groups 1 and 2. Due to some missing data that could not be imputed. I do not feel sad) to 3 (e. good convergent and discriminant validity (e.9 years (SD = 3.95 in clinical populations and . & Brown. procedure All questionnaires were administered in online sessions via a standardized web-based automated computer program that was designed to protect participants’ confidentiality and the security of their data. good convergent validity (r = 0. such as the Beck Hopelessness Scale. & Bailey.5% 29. Our experience from administering the questionnaires to Group 1 indicated that all measures could be completed in a reasonable amount of time without imposing unrealistic demands on the participants.82–0. the covariance matrix. and ethnicity) for participants in Groups 1 and 2 are presented in Table 1. NSPS.81 to 0. All measures in both sessions were presented in a randomly counterbalanced order.). Beck.74) 68. and high internal consistency (Cronbach's alpha indices of 0. for Table 1 Demographic Characteristics of Participants Age (N = 220 for Group 1. I am so sad or unhappy that I can't stand it). A MANOVA with ethnicity (Caucasian. and Other) as the grouping variable was conducted on the SIAS.51) 23. SPS. p N .65) 20. with the exception of the one “blushing” item.15) 25.1% of the variance and was labeled concerns about showing signs of anxiety.93 0.21 (22. all of which had corrected item-total correlations greater than or equal to . As indicated above. Wilhelm. Means and (Standard Deviations) of Study Measures Measure NSPS SCBI SPS SIAS LSAS-SR-F LSAS-SR-A BDI-II OCI-SV Group 1 Group 2 Cronbach's α Mean (SD) 1 Cronbach's α Mean (SD) 1 0.90 (10. concerns about social competence. Results in both Groups indicated that relative to Caucasians. which cross-loaded onto two factors but was retained because of its clinical relevance in socially anxious 2 Independent-sample t-tests were also conducted to compare scores across measures for the primary ethnic groups (Caucasian vs.43. LSAS-F.71 (16.32 on any other factor were retained. All retained items with factor loadings of N. indicated that a three-factor solution provided the best fit. 1999.negative self-portrayal scale 187 Table 2 Internal Consistency. SIAS = Social Interaction Anxiety Scale. Asian participants reported higher scores on the SIAS (ts N 2. 1162) = 1.96 0.07 (14. in their recent psychometric study of the Appraisal of Social Concerns Scale (ASC).74) 16. Items were removed that were cross-loading or unreliable. As the hypothesized factors of the NSPS were expected to correlate with each other (Moscovitch.90 0. Table 3 presents the rotated factor loadings from the EFA pattern matrix for each of the 27 NSPS items that were retained in the first group of participants (please also see the notes below Table 3 for the NSPS respondent instructions). and NSPS (ts N 2.45 on a single factor and b. Group 2 scale item characteristics and confirmatory factor analysis.10.90 0. the one blushing item included in the measure for participants in Group 1 did not load cleanly onto one factor. 2009). The three factors accounted for 55. EFA with maximum likelihood extraction was used to examine the underlying structure of the NSPS in the first group of participants. p N . but instead cross-loaded onto both concerns about social competence (.35 (12.29 (15. There were no other significant differences. who. LSAS-SR-A= Liebowitz Social Anxiety Scale – Self Report Version .74) 29.93 0. The results indicated no significant effects of ethnicity on any of the measures. standard deviations.91 0. was labeled concerns about physical appearance.64 (10.05 ).80.972. ps b .64 (23. LSAS-SR-F = Liebowitz Social Anxiety Scale – Self Report Version Fear Subscale.52 (9. 2 Despite the high internal consistency values.93 0. the highest correlation between any two of the retained NSPS items across both samples was r = . This finding was consistent with those of Schultz and colleagues (2006).89 (23. SCBI = Self-Concealment Behaviors Inventory. for Group 2. BDI-II = Beck Depression Inventory II. We chose to retain the blushing item . F (36. for Group 2. An inspection of all eigenvalues greater than 1. SPS = Social Phobia Scale.15. in accordance with Comrey and Lee's (1992) recommendations for factor analyses involving multiple subscales. for Group 1. OCI-SV = Obsessive Compulsive Inventory – Short Version.g.05.88) 31.41 (12.06 (11.05.30) 20.30) 99.4% of the variance.95 0.1% of the total variance.96 2 0. as well as the scree plot.73) Note. Bögels & Reith.Avoidance Subscale.93 0.95 (21.15. Scale Item Characteristics and Factor Analysis Group 1 scale item characteristics and exploratory factor analysis. 274) = . ps b .95 0. Asian. reported that the one ASC blushing item was discarded because it cross-loaded both onto a factor related to social competence and a factor related to social anxiety. and F(36. and F(6. NSPS.933) = 1. Group 1.61) 101. accounting for 5. p N . which accounted for 45. the third factor accounted for 4. East Indian. suggesting non-redundancy between the items.36. 27 items remained.54) 15.47 (9.2 Mean scores.05).13) 11. and Cronbach's alphas for each measure in Groups 1 and 2 are presented in Table 2. Ns = 225 (Group1) and 316 (Group 2).29) and concerns about showings signs of anxiety (. NSPS = Negative Self-Portrayal Scale. LSAS-A.50 (14. Gerlach.89 60.89 57.76 (12. Interpreting the factor loadings. and SCBI scores for Groups 1 and 2. 1 Based on 27-item NSPS.96 2 0. & Roth. populations (e.93 0.16) 26. the NSPS items were subjected to analysis using an oblique (promax) rotation. we labeled the first factor. 2001).82) 0. Finally. Asian).93) 11. Gruber.96) 23.24). The second factor.. Following this process. Middle Eastern.6% of the variance.92 0.05. 00 .72 .” and “extremely concerned. the four blushing items were highly correlated with each other (r = .41 .79 .93 .82 . Unfashionable 9.32 . only the one “blushing” item was retained. Missing data were handled with the full information maximum likelihood method (Anderson. talking to someone who is a stranger.02 . Each subscale can also be summed separately to obtain data on the relative strength of self-portrayal concerns across dimensions within individuals and across individuals within a larger sample.46 .03 -. Lacking personality 10.” “slightly concerned.e.09 . A 4-factor model was evaluated with CFA. Stupid 19. which we administered to participants in Group 2. Speaking incoherently 20. values for univariate skew were all less than |1.01 -. Sweating 7. Lacking social skills 12.51 .74 .09 .07 -.77 . because of its clinical relevance for socially anxious individuals and our overarching theoretical model (i.53 ..10 . which included the 3-factor structure from Group 1 and a separate fourth factor containing the four blushing items.75 .29 -.44 -.47 .71 . Humorless 23.11 -.59 .02 .07 -.13 -.66 .46 . root-mean-square error of approximation (RMSEA) = .08 . probability of close fit (pclose) = .097. Stuttering 15. as implemented in Amos. Poorly dressed Signs of Anxiety 8. 2009).07 . answering a question in class.” For each item. becoming red in the face.56 .08 -.08 .16 . χ2 (399) = 1579. Fat 26.46 -. The data met the criteria for univariate and multi- variate normality suggested by Kline (1998) and West.05 .11% Note.07 -.e. thus.02 .75 . comparative fit index (CFI) = . .20 -.88) and deemed redundant.188 moscovitch & huyder Table 3 Rotated Factor Loadings from Pattern Matrix for Non-Discarded Items of the Negative Self-Portrayal Scale for Group 1 Item Social Competence 21. Speaking with a trembling voice 25.50 -.63 .49 . In order to explore whether blushing might represent a separate fourth factor.09 .04 .04 . it will become obvious to other people that I am ____.43 .63 .53 .01 . Aloof 3.61 .21 55. Twitching (i. However. Fidgeting 1. Having a bad hair day 2. Losing control of my emotions 4.00 . Blemished (i.06 .” Response options per item range from 1-5. NSPS instructions to respondents read: “Using the scale provided above the following list of items.0 (a structural equation modeling program.95 .09 . etc.37 . Moscovitch.02 .05 -.64 .09 -. Blushing Social Competence Physical Appearance Signs of Anxiety Communalities 1. Interpersonally ineffective 14.73 .17 .10 . for example.e.62 .38 -.12 -.” “moderately concerned.10 .61|. Socially awkward 24.” The NSPS total score is a sum score of all items. This model was evaluated and fit moderately well.19 . Weird-looking 13.58 .76-. 2008).02 -.04 .000. Frozen 6. and Curran (1995).01 -.” “very concerned. Physically unattractive 11. p b .80 .03 .12 . Unable to express myself 18. Arbuckle.60 . giving a speech in front of an audience.01 .51 .13 . Ugly 5. CFA was used to analyze this 30-item NSPS scale for participants in Group 2 using Amos 17.g.06 .67 . and becoming flushed) and added to the NSPS (resulting in a 30-item measure).57| and values for univariate kurtosis were all less than |1.73 .e.97 .80 . respondents are asked to consider the following statement: “In social situations (in which I feel anxious).72 . my facial muscles) 16.51 .83.24 .24 .58 .32 .76 .16 .08 .18 .04 .. my appearance) 22.64 .001.14 . please select the circle under the appropriate number/statement to indicate the degree to which you are concerned about the following aspects of yourself when you are in anxiety-provoking social situations (e.55 . with descriptive anchors including “not at all concerned. Boring Physical Appearance 27.66 . becoming hot in the face..13 -. 1957).). Finch.50 .58 ..846. Reserved 17.05 . three new blushing items were constructed (i.07 -. .96). consisting of 27 items loading onto the three factors found in Group 1. The one “blushing” item was constrained to load onto the concerns about signs of FIGURE 1 Confirmatory factor analysis of the NSPS 3-factor model for Group 2. p b . Thus.001. indicating that they were not distinct factors and would be more accurately represented as one factor (concerns about social competence).001 (2-tailed). The model demonstrated good fit. All factor loadings had a p-value b . CFI = . pclose = .negative self-portrayal scale CFA was then conducted to test the original Moscovitch (2009) 4-factor model with the remaining 27 NSPS items.000. two of the factors—concerns about social skills and behaviors and concerns about personality—were highly corre- 189 lated (r = .898. χ2(318) = 948.079. RMSEA = . However.52. Items are labelled with I = item and the corresponding number from the NSPS items shown in Table 3. a new model was evaluated via CFA. 96 (n = 225) and . BDI-II.888. and SIAS scores (r =.001. Tolerance values were .e.88 (n = 316) in Group 2. n = 316).27. and (c) Signs of Anxiety. SPS. 3.96. and SIAS scores (r = .083. and .93 (n = 316). n = 225). with Cronbach's alpha coefficients for the social competence. and physical appearance) are reported in Table 4. For Groups 1 and 2. This 3-factor model fit well. Cronbach's alphas for the 27-item NSPS in Groups 1 and 2 were both .94 (n = 316). Though not measured in Group 1. consisting of 27 items loading onto the three factors found in Group 1 and the one “blushing” item constrained to load onto the social competence factor. CFI = .080. pclose = . LSAS-SR-A. The SCBI also demonstrated excellent internal consistency.000. We were. for comparative purposes. Convergent and Discriminant Validity As depicted in Table 5. . and signs of anxiety factors. and .73) than with measures of theoretically distinct constructs such as OCD (r = . the Browne-Cudeck Criterion (BCC) and expected cross-validation index (ECVI) values were inspected for each of these models.05. Cronbach's alphas for the 53-item SCBI in Groups 1 and 2. Although the NSPS in Group 2 was significantly more strongly correlated with measures of social anxiety than with the theoretically distinct construct of depression. interested in examining the hypothesis that the NSPS would be a useful measure that would help account for the wide-ranging use of self-concealment behaviors. the test-retest reliability of the 27-item NSPS in Group 2 was satisfactory (r = . NSPS total scores for Groups 1 and 2 demonstrated significantly higher zero-order correlations with measures of social anxiety.50-.40–. we predicted that self-concerns (even in the unselected samples in the present study) would account for a significant amount of variance in self-concealment behaviors over and above the influence of symptoms of social interaction anxiety. n = 225 and 316) and BDI-II (r = . of . and NSPS scores) entered in four separate steps.01. including both the OCI-SV (r = .43-. Finally. Thus. The three factors are labeled: (a) Social Competence.. including LSAS-SR-F. *p b .66) and ECVI (3. therefore. n = 316) over a 1-week interval.65) (pdifference N . Each item of the scale is identified by number. Relative to the 3-factor model with blushing loading onto the social competence factor. as measured by BDI-II depression scores (r = .91 (n = 225). and . LSAS-SR-A.63 vs. an alternative 3-factor model was tested via CFA.05.61) (pdifference b . safety) behaviors are functionally related to one another.58. signs of anxiety. and depression. 1 with standardized estimates. p b . CFI = . 2004). p b . Internal Consistency and Test-Retest Reliability The NSPS demonstrated excellent internal consistency. 1 (social competence. respectively. Construct Validity Moscovitch (2009) theorized that self-portrayal concerns and self-concealment (i. χ2 (321) = 1018.93 (n = 225).001.75. and is illustrated in Fig. Intercorrelations between the three NSPS factors modeled in Fig. we conducted a hierarchical multiple regression analysis with SCBI scores as the criterion variable and four predictor variables (SIAS. This model did not fit as well. Ns = 225 (Group 1) and 316 (Group 2). the NSPS in Group 1 was not more strongly correlated with other measures of social anxiety than with the BDIII (r = . SPS. 1 could not be tested because they are not nested models. pclose = .190 moscovitch & huyder anxiety factor. χ2(321) = 974. social performance anxiety.71–.001. n = 225 and 316). the 53-item version of the SCBI total score demonstrated significantly higher zeroorder correlations with measures of social anxiety. RMSEA = .000.001.63–. such that self-concealment is used in a strategic manner by socially anxious individuals to prevent others from observing their negative self-attributes in social situations.77) values and was. were .001 (2-tailed).97 vs.45) (pdifference b . SPS. Table 4 Intercorrelations Between the Factors of the Negative Self-Portrayal Scale Note. therefore.77) than with measures of theoretically distinct constructs. Each factor also demonstrated strong internal consistency. To this end. retained. . the 3-factor model with blushing loading on the signs of anxiety factor (displayed in Fig. respectively. (b) Physical Appearance.895. according to the 27item measure presented in Table 3. However. 1202.95 (n = 316).53) (pdifference b . over and above established symptom measures of social anxiety and depression. with smaller values indicating better model fit (Wicherts & Dolan.85 (n = 225) in Group 1. including LSAS-SR-F. 1) demonstrated lower BCC (1158. n = 225 and 316). Group 1 correlations are in the gray-shaded region. RMSEA = .47) (p-values b. physical appearance. The significance of the difference in model fit between this model and the one displayed in Fig. 024 for Step 4 (p b .001 (2-tailed). LSAS-SR-F = Liebowitz Social Anxiety Scale – Self Report Version Fear Subscale.67 ⁎ .77 ⁎ . SPS = Social Phobia Scale. OCI-SV = Obsessive Compulsive Inventory – Short Version. Group 1: N = 225. NSPS = Negative SelfPortrayal Scale.49 ⁎ .56 ⁎ .516 for Step 1 (p b .10 0.12 0. ⁎⁎ p b . BDI-II = Beck Depression Inventory II.54 ⁎ .10 0.60 ⁎ . ⁎ p b .10 0.001).42 ⁎⁎ 0.68 ⁎ .23 ⁎⁎ 0.001).55 0. SIAS = Social Interaction Anxiety Scale.001).36 ⁎⁎ 0.60 ⁎ .05 0.35 ⁎ .12 0.61 ⁎ .77 ⁎ .11 0.565 for Step 1 (p b .52 ⁎ .72 ⁎ .69 ⁎ .69 ⁎ .15 ⁎ 0. Group 2: N = 316.57 0. and Other Self-Report Measures Measures SCBI Group 1 Group 2 SPS Group 1 Group 2 SIAS Group 1 Group 2 LSAS-SR-F Group 1 Group 2 LSAS-SR-A Group 1 Group 2 BDI-II Group 1 Group 2 OCI-SV Group 1 Group 2 NSPS Total NSPS Social Competence NSPS Physical Appearance NSPS Signs of Anxiety SCBI .72 ⁎⁎ 0.05).75 ⁎⁎ 0.62 ⁎ .55 ⁎ .72 ⁎ .45 ⁎ .05.55 0.41 ⁎⁎ 0.76 ⁎ .66 ⁎ .71 ⁎ . SIAS = Social Interaction Anxiety Scale.41 0.24 ⁎⁎ Note.53 ⁎ .06 0. .52 ⁎ . R2 = .191 negative self-portrayal scale Table 5 Zero-Order Correlations Between the NSPS Total Score.05 0.08 0.050 for Step 4 (p b .54 0.63 0. ⁎ p b .58 ⁎ .76 ⁎ .59 ⁎ .001). NSPS = Negative Self-Portrayal Scale.10 0.08 0.66 ⁎ .07 0.003 for Step 3 (p N .67 ⁎ .25 0.Avoidance Subscale.40 ⁎ .57 ⁎ .07 0.73 ⁎ .07 0.43 ⁎ Note.62 ⁎ .83 0.63 ⁎ .75 ⁎ .76 ⁎ .71 ⁎ .58 ⁎ .16 0. NSPS Subscales.71 ⁎ . R2change = .44 ⁎ .08 0.46 0.09 0.37 ⁎ .73 ⁎ .50 ⁎ .64 ⁎ .31 ⁎⁎ 0.03 0. All correlations for Groups 1 and 2 are based on the 27-item NSPS.087 for Step 2 (p b .39 0.035 for Step 3 (p b .55 ⁎ .37 ⁎ . SCBI = Self-Concealment Behaviors Inventory.73 ⁎ .001).58 ⁎ .45 0.24 0. SPS = Social Phobia Scale.07 0.75 ⁎ .32 ⁎⁎ 0.34 ⁎⁎ 0. Table 6 Summary of Hierarchical Regression Analysis for Variables Predicting Levels of Social Self-Concealment on the SCBI Predictor Step 1 SIAS Step 2 SIAS SPS Step 3 SIAS SPS BDI Step 4 SIAS SPS BDI NSPS B SE B ß Group 1 Group 2 Group 1 Group 2 Group 1 Group 2 1.28 ⁎⁎ 0.65 ⁎ .32 ⁎⁎ 0.47 ⁎ .41 ⁎⁎ 0. R2 = . SCBI = Self-Concealment Behaviors Inventory.10 ⁎ 0.65 ⁎ . SCBI. R2change = . LSAS-SR-A= Liebowitz Social Anxiety Scale – Self Report Version .53 ⁎⁎ 0.45 ⁎ .001).07 0.51 ⁎ .108 for Step 2 (p b .61 ⁎ .09 0.48 ⁎ .56 ⁎ . R2change = .33 ⁎ .001). BDI-II = Beck Depression Inventory II.64 ⁎ – – .04 1.44 0.09 0.68 ⁎ .65 ⁎ .75 ⁎ . R2change = .22 0.42 ⁎⁎ 0.65 0.001.44 0. R2change = .07 0.73 ⁎ .58 ⁎ .62 ⁎ .68 ⁎ .02 0. R2change = .59 0. Ns = 225 (Group 1) and 316 (Group 2).50 ⁎ . SIAS scores were entered in step 1. social interaction anxiety.4%– 5.95. . SPS and BDI-II scores).97). Alongside concerns about social competence and in support of Moscovitch's (2009) theory. Whereas psychologists might distinguish between social skills and behaviors..92 PA. .3 Discussion The results of the present study suggest that the NSPS is a psychometrically strong and potentially useful instrument for measuring concerns that individuals have about exposing specific negative self-attributes to critical others in anxiety-provoking social situations. Any differences that emerged between ethnic groups were not found consistently across the two samples and may have been affected by sample size considerations. Asians = . and self-concealment behaviors amongst socially anxious individuals (Moscovitch. Thus.86 SigAnx.74.93 SC. The full set of results is available by request from the first author.95) and Group 2 (Caucasians = . Asians = .95) and Group 2 (Caucasians = .93 SC.g. reliable. and personality.95 SC. It was surprising that the blushing item crossloaded relatively weakly onto both the signs of anxiety and social competence factors in the EFA.93 SocCom. n = 132 in Group 2) and Asian (n = 67 in Group 1.96.75 is consistent with our expectations that the NSPS reflects relatively stable . avoidance. which is well above the minimum criteria of . and depression symptoms each accounted significantly and uniquely for variance in SCBI scores in Steps 1. and NSPS total scores in Step 4. For Group 2. A test-retest correlation of . It enables the specific assessment of the core negative features of the self that are thought to underlie self-presentational concerns that drive social anxiety (e. The NSPS represents a novel assessment tool in the social anxiety field that goes beyond measuring the level or types of social 3 We also analyzed the psychometric properties of the NSPS for the Caucasian (n = 102 in Group 1. on the other. .94 PA. Asians = . As shown in Table 6. Asians = . West. Asians = .. but we chose to retain the blushing item for the CFA due to its empirically supported relevance in social anxiety research and practice. only social interaction and performance anxiety (but not depression symptoms) accounted significantly and uniquely for variance in SCBI scores.84 SA.10 (Cohen. BDI-II scores in Step 3. which we labeled concerns about social competence. Moreover. the addition of NSPS scores to the regression equation in Step 4 accounted for a significant proportion (2. 2009a) and to account for observed heterogeneity in situational fears. (5) NSPS convergent/discriminant and construct validity: Space limitations prevent us from reproducing the correlational matrices and regression results here. respondents may be indicating that concerns about appearing “socially awkward” or “speaking incoherently” (both originally social skills/behaviors items) have the same negative social impact as appearing “aloof” or “boring” (both originally personality items).84 SA) and Group 2 (Caucasians = . This 3-factor solution was associated with good model fit in the subsequent CFA. on one hand. social performance anxiety. Empirical evidence obtained across two independent undergraduate samples in the present study indicated that NSPS items developed to tap concerns about social skills and concerns about personality loaded jointly onto a single factor. the emergent factor structure of the NSPS appears to be valid. 2 and 3 for Group 1.0%) of the variance (in the use of self-concealment behaviors over and above the contributions of SIAS.33 for all predictor variables. the development of this measure represents a crucial first step in the process of testing Moscovitch's model of social anxiety. n = 90 in Group 2) groups separately: (1) NSPS total score internal consistency: Group 1 (Caucasians = . & Aiken. respectively. anxiety symptoms respondents experience across different situations. Cohen. exhibiting flaws in social skills/behaviors or flaws in personality leads essentially to the same concern: appearing socially incompetent. it is likely that NSPS respondents simply considered social skill/behaviors to be observable indices of personality. which were conducted for the Caucasian and Asian groups separately across both samples. While the observed factors of the NSPS across both groups in the present study did not map perfectly onto the predicted dimensions. Leary & Kowalski.91 PhyApp. (2) NSPS internal consistency for the 3 factors: Group 1 (Caucasians = . Asians = . Importantly.95. 1995. we found reasonably strong model fit when we constrained the blushing item to load only onto the signs of anxiety factor. The overall pattern of results between the groups was similar. The total Ns were 225 and 316 for Groups 1 and 2. Future research is needed to replicate this finding. Asians = .96). making these post-hoc analyses difficult to interpret. in other words. . (4) SCBI internal consistency: Group 1 (Caucasians = . . concerns about signs of anxiety and concerns about physical appearance emerged as separate (nonorthogonal) factors in the EFA. SPS scores in Step 2.75). fairly stable over a 1-week period. 2003). 2009). and conceptually sound.96. . . In the subsequent CFA. (3) NSPS test-retest reliability: Group 2 (Caucasians = . for both Groups 1 and 2.94 PA. .91 SA).192 moscovitch & huyder examined prior to this analysis to assess for the potential impact of multicolinearity and were found to be no less than . Stopa. g.g. 2005). the values of another culture might dictate that social competence is reflected in one's ability to remain silent during social gatherings. panic. Dozois & Frewen. Chavira. We assessed the discriminant validity of the NSPS in relation to the BDI-II and the OCI-SV. it would be prudent to replicate our study with paper questionnaires. & In-Albon. Kemp. but it will be important in future studies to determine how the NSPS correlates with measures of constructs—for example. Thus. university-educated. Indeed. no doubt. what might appear socially incompetent according to the norms of one society might appear quite competent according to the norms of another. the overall mean levels of these symptoms were quite low. worry. Since all questionnaires were administered online. the observed correlations in our study between the NSPS and measures of both social anxiety and depression were quite robust. Ramel.. female. Suvak..g. Peters. Abbott.. Moscovitch. Although we hypothesized that the NSPS would correlate more strongly with measures of social anxiety than with measures of depression. thus.g. Therefore.. Although replication of our results is clearly required in both clinical samples of SAD patients and community samples of healthy and anxious . whereas the values of one culture might dictate that its members make humorous remarks or witty conversation during social gatherings in order to appear socially competent. it is likely that the NSPS is a culturally bound measure. it is impossible to know whether and how the NSPS findings reported here would generalize from the present participants to clinical samples. it is crucial to note that. & Gaston. Our results suggested that the SCBI is a valid and reliable measure. because individuals’ self-portrayal concerns depend on the standards and norms for social behavior and because such standards and norms vary considerably across cultures (e. Moreover. In other words. Bryson. or higher-order dimensions (e. it would be both interesting and important to examine the cross-cultural applicability of the NSPS in future research. In addition. The new measure demonstrated adequate concurrent validity across both samples and explained significant variance in the use of self-concealment behaviors over and above that accounted for by social interaction anxiety. and socioeconomically advantaged individuals. social performance anxiety. no validated. 2005). Winzelberg. Rapee. More research is required to better understand the unique cognitive features of social anxiety and depression and whether the NSPS is capable of distinguishing between them. Markus & Kitayama. 2006. A possible reason for this finding is that the NSPS and BDI-II capture overlapping components of the self-critical cognitive style that is characteristic of—and may serve as a common cognitive diathesis underlying—both syndromes (see Ingram. Another possibility is that the NSPS might have good discriminant validity in relation to some measures (e. Notwithstanding the recent publication of the Subtle Avoidance Frequency Examination (SAFE. our use of multiple sessions for completion of the study measures may have introduced additional variability into the study associated with participants’ fluctuating circumstances and moods across testing sessions. despite the wide range of reported symptoms of anxiety and depression in our study. OCISV) but poor discriminant validity in relation to others (e. and practice listening respectfully to the elders of the community.. and depression symptoms. Similarly. with high Cronbach's alphas across both samples and adequate convergent and discriminant validity. we cannot infer that the results would generalize to samples of individuals with different compositions of demographic characteristics. 1991).negative self-portrayal scale concerns about self-attributes that ought not to change significantly over short time intervals in the absence of treatment.g. it is possible that participants felt less accountable to answer questions as carefully as participants who 193 complete pencil-and-paper measures in the laboratory might feel. Cuming. there are. BDI-II).. although previous studies comparing online and paper administrations of questionnaires have shown that online measures are highly reliable (e. variations across cultures in standards of physical appearance and norms associated with showing signs of anxiety. The present study was not without its limitations. Osborne. published measure of safety behaviors existed when our data were collected. Furthermore. Luce. Das. & Scher. Brown & Barlow. Hofmann. we adapted the SCBI from a previously unpublished measure of safety behaviors and used it as our measure of self-concealment behaviors in the present study.g. 2009). 2007). As highlighted by some of the reported differences between Asian and Caucasian participants in the present study. & Taylor. Consequently. constrain the desire to speak out. previous research has highlighted the central role of negative self-related cognitions in the phenomenology of both social anxiety and depression (e. because our unselected samples of participants were comprised predominantly of young. which in turn would dictate cross-cultural differences in the relevance of and responses to the present version of the NSPS. 2009)—that might be even more similar to social anxiety both conceptually and clinically than depression or OCD per se. thus. even though this distinction is not captured in the factor structure of the NSPS. A. A. A cognitive model of social phobia. M. Dordrecht. 2009). 2006). H.. the preliminary validation of the NSPS enables us to renew our call for a paradigm shift in treating SAD (e. E. & Rapee. Journal of Psychopathology and Behavioral Assessment. Baker. Bögels. J. T. and treatment (pp. A. M. see also Moscovitch. 2009b) and. Conversely. L. A. & Reith. 21. N. T.. (1995).. West. (2008) Amos 17.. Hillsdale. but also how much they might overestimate the probability and social costs associated with revealing each self-attribute to critical others. (2001). S..). Hackmann. The Liebowitz social anxiety scale as a self-report instrument: a preliminary psychometric analysis. Blushing and physiological arousability in social phobia. Social behaviour questionnaire. therefore.0 users guide. J.194 moscovitch & huyder control participants (an endeavor that is currently in progress in our laboratory). (2001). L. M. Practitioner's guide to empirically based measures of anxiety: AABT clinical assessment series.. & Swinson. (2003). New York: Oxford University Press. We hope to track how such probability and cost estimates might differ across the NSPS dimensions and change as a result of treatment. A first course in factor analysis (2nd ed. H.. Anderson. & F. Clark. Cohen. Confirmatory factor analysis for applied research. 69–93). Dozois. A. but it is not sufficient. (1992).. and Sweating Questionnaire. J. S. Ehlers. Oxford guide to behavioural experiments in cognitive therapy. (2004). Journal of the American Statistical Association. New York: The Guilford Press. A self-report measure of subtle avoidance and safety behaviors relevant to social anxiety: Development and psychometric properties. Schneier (Eds.. Traditional conceptualizations of exposure-based CBT for SAD (e. Clark. E. Kemp... (2004). 51–66. (1999). R. R. H.). G.. & Wells. L. S.. Bennett-Levy. S. Orsillo. Rodebaugh & Rapee. 2009) are guided by a therapeutic mindset that is characterized by the question: “What types of feared social situations must be confronted in the service of anxiety reduction?” Confronting social situations is indeed necessary in the treatment of social anxiety. Cuming. & Aiken.g. Abbott. we are currently working on adapting the current version of the NSPS to investigate in future studies not only how concerned socially anxious individuals are about publicly exposing feared negative self-attributes. L. might require therapeutic intervention with a different set of intervention strategies. IL. (Eds.. & Gaston. Clark. F. L.. A. It might be worthwhile to remain attuned to the potentially useful distinction between concerns about social skills/behaviors and concerns about personality in the context of SAD. Heinrichs. M.. W. Moscovitch. 110. S. Finally. Butler. . the preliminary validation of the NSPS enables us to envision its future use as a clinical tool to guide the assessment and treatment of SAD.. 90. Arbuckle. Unpublished manuscript.. 40.. (2006). L.. M. T. Antony. 101–109. Liebowitz. L. Social phobia: Diagnosis. M.. D. Comrey.. & Wells. J. Netherlands: Kluwer Academic Publishers. M. Journal of Affective Disorders. W. Hackmann. A. J. Specificity of cognitive structure in depression and social phobia: A comparison of interpersonal and achievement content. Antony... & Westbrook. 2009) and similar types of behavioral experiments (Bennett-Levy et al. Behaviour Research and Therapy. Kim. M. P. concerns about signs of anxiety or personality flaws in the context of SAD may be more resistant to change with video feedback (see Orr & Moscovitch. 857–882. 879–883. 52. Peters.. (2009). (1957)..g. Butler. Gruber. F.... Brown. (2006). New York: Guilford Press.. & Hofmann. & Frewen. M. B. 1986. R. Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed. 247–258.. Stopa. and how can new learning about the self be promoted most effectively during therapy in the service of anxiety reduction?” Future research is required to support this testable prediction. A. S.. 2004) that are designed to improve the self-other discrepancy in imagined behavioral deficits (e. Fennell. concerns about social skills/behaviors that are endorsed by patients with SAD tend to be most responsive to interventions such as video feedback (e. J. Interpersonal processes in social phobia. T. M. 2010b. Journal of Anxiety Disorders. & Lee. Chicago. Foa & McNally. 200–203. Trembling. & Taylor.g. Validity of two questionnaires to assess social fears: The Dutch Social Phobia and Anxiety Inventory and the Blushing.. 2009) will be more effectively facilitated with both an a priori exploration of patients’ self-portrayal concerns and an overarching therapeutic mindset that is guided by the following question: “What about the self does this patient perceive as being flawed. Heimberg.). D.. D. Mueller. G.. (2002). Fennell. McManus. Journal of Abnormal Psychology.). assessment. & Roth. 24. D. Cohen.. NJ: Lawrence Erlbaum Associates. Harvey. Mahwah: Lawrence Erlbaum Associates. (1995). We hypothesize that the extinction of learned fear in patients with SAD vis-à-vis the principles of emotional processing theory (Foa & Kozak. C. In R. Gerlach. A. S. J.g. W. In our clinical experience. M.. In closing. G. K. Wilhelm.. References Alden. Heimberg. G. & Roemer. We hypothesize that clinical outcomes in SAD will be improved by practitioners’ assessment of the nature of their patients’ self-portrayal concerns and the subsequent selection of specific interventions to target those concerns across the dimensions of the NSPS. D. Maximum likelihood estimates for a multivariate normal distribution when some observations are missing. M. Hope. A. N. McManus et al. L. 701–715. 1996. P. G. 23. 2000. Clinical Psychology Review. Rapee. A.. Winzelberg.). The nature and expression of social phobia: Toward a new classification. Stein (Eds. 1183–1192. J. 83–102. S. P. D..). K. S. 455–470.. Heimberg (Eds. F. 14. N.. 667–683. The obsessivecompulsive inventory: Development and validation of a short version.. D.. and motivation. A. C. & Kugu. G. A. In R... Development and examination of the Social Appearance Anxiety Scale.. Sacadura. Emotional processing of fear: Exposure to corrective information. D. Fresco. & Clarke. E. 49. 38. Horner. Fennell. (2001). 47. (2009). R. D. New York: The Guilford Press. 1039–1052. D. R.. 123–134. F. & Clark. R. (2002). M. (2004).. Palyo. Kichic. Assessment. 29. 16. Behaviour Research and Therapy. assessment. The Liebowitz Social Anxiety Scale: A comparison of the psychometric properties of self-report and clinician-administered formats. (2010b). B.. G. Psychometric properties of the Liebowitz Social Anxiety Scale.. M. Liebowitz. E. Luce. 193–209. D. New York: The Guilford Press. E. Social anxiety and self-impression: cognitive preparation enhances the effects of video feedback following a stressful social task. Vandenberg. S. D. issues. Hofmann. G. M. Social phobia: Diagnosis. (1997). Psychological Review. 48. M... Clark. (2010a). & Scher. In M. 485–496. I. R. Social phobia among university students and its relation to selfesteem and body image. B.. & Kowalski. Crozier. G. D. & Goetz. S. B. M. D. In R. Reliability of self-report: Paper versus online administration. K. Cognitive factors that maintain social anxiety disorder: A comprehensive model and its treatment implications.. Dogan.. R. F.. Holle.... A. C. C. When ambiguity hurts: social standards moderate self-appraisals in generalized social phobia. Brannick. A new model to facilitate individualized case conceptualization and treatment of social phobia: An examination and reaction to Moscovitch's model. Handbook of anxiety and anxiety disorders (pp. M.. Hackmann.. Akyüz. Clinical Psychology Review Special Issue: Social Phobia and Social Anxiety.. S. W. Antony. Cognitive Behaviour Therapy. M. J. (1999). D.. C. W. Gehring Reimer. Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. (1998). Moscovitch. S. & Heimberg. Learning to re-appraise the self during video feedback for social anxiety: Does depth of processing matter? Behaviour Research and Therapy. 98. (1998). E. Safren. Langer. & Liebowitz. R. & Clark. 16. Das.. Hirsch.. (2005). Cognitive and Behavioral Practice. 48–59. Clark... Behaviour Research and Therapy. B. J. F. & Antony. R. Waddington. New York: John Wiley & Sons. M. M. 26. J.. D.. Improving outcomes and preventing relapse in cognitive-behavioral therapy (pp. Social anxiety and depression. & Salkovskis. N. Gutierrez. Antony. 329–343). Hofmann. In M.. Clark. McManus. A.. (1995). G. Social anxiety disorder.. Grey. & Clark. M. Salkovskis. Brown... Psychological Assessment. 1025–1035. Psychometric properties of the Beck Depression Inventory-II in nonclinical adolescent samples. 496–503. G. J. Hackmann. Stein. A. 1384–1389. D. (2007). Behaviour Research and Therapy. 224–253. S. (2009).. What is the core fear in social phobia?: A new model to facilitate individualized case conceptualization and treatment. Exposure-based treatments for anxiety disorders: Theory and process. Wild. Fresco. How do I appear to others? Social anxiety and processing of the observable self. R. Current controversies in the anxiety disorders (pp. Mediation of changes in anxiety and depression during treatment of social phobia. 37. R. Moscovitch. G. (2009). 39. Hart. (2008). P. Ledley. G. H. D. & M. M.. Behaviour Research and Therapy. Hurley. & Kozak. Suvak. Submitted for publication. Moscovitch.. The self-presentation model of social phobia.. R. 18. Schneier. & Bailey. J. D. 31.. Orr. McManus. H. M. B. M. 728–737. 3–12. Hofmann.. Bryson. R. J. R... M. Psychological Bulletin. Behaviour Research and Therapy. D. Foa. Ingram. (1986). The essential handbook of social anxiety for clinicians (pp. Flora. Mansell. Heimberg. B. M. Chavira. (2009). 769–797. Leiberg. E.. 419–434. Juster. 73. P. T. G. M. Heimberg. G. Exploratory and confirmatory factor analysis: Guidelines. M. E. Heinrichs... A. N. Surawy. & McNally. J. J. & Heimberg. Ramel. C. M. The validation of a new obsessivecompulsive disorder scale: The Obsessive-Compulsive Inventory. L. & Moscovitch. New York: Guilford Press... Osborne. Psychological Medicine. 10. R. D. O. Journal of Anxiety Disorders. M.... (2008). D. T. Orr. & L. K. 241–264). P. E. Mechanisms of change in exposure therapy.. Huppert. 945–952. Hajcak. A.. (2008). 45. B. C. Harvey. (2007). Mattick. R. 461–475). New York: Guilford Press. J. G... A. A. M.. 23. Moscovitch. and treatment (pp. Why social anxiety persists: An experimental investigation of the role of safety behaviours as a maintaining factor. Seers. E.. A. (1991). 206–214. 135–141. N. M. Computers in Human Behavior.. Psychological Assessment. E. Orr. D. M. H. Cognitive and Behavioral Practice. & Taylor. D. S. L. Rapee (Ed.. Osman. Foa. Behavioral and Cognitive Psychotherapy. A. 23. . S. R.. A.. (1999). & Kitayama. In the absence of rose-colored glasses: Ratings of self-attributes and their differential certainty and importance across multiple dimensions in social phobia. A. (2004).. (1996). M. S. D. S. Coles. (2000). & F.. 147–161. J. A.. Barrios. M. Principles and practice of structural equation modeling. Schriesheim. (2007). Alden (Eds. A. (1998). P. A. (2005). Markus. S. Hackmann. E. X. M. 38. J. & Manley.. Canadian Journal of Psychiatry.). (2009). Liebowitz. R. R. Journal of Organizational Behaivor.. Ehlers.. Coles. R. (2000).. W. Journal of Consulting and Clinical Psychology. 94–112).). M. 24. In R. Kozak... Behaviour Research and Therapy. C. J. D. M.. 64. Culture and the self: Implications for cognition. & R. Hami. R. E.negative self-portrayal scale Foa.. T.. & Moscovitch.. Kline. C. Ledley. D. 20–35. F..... M. W. Heimberg. Hope. R. 99. 36. R. G. A demonstration of the efficacy of two of the components of cognitive therapy for social phobia. Antony. 38–76). M. S. D. M. R. Journal of Clinical Psychology. 66–70. A. 36. K. 601–610. Seeing yourself through others' eyes: A study of spontaneously occurring images in social phobia. Schneier (Eds. R. Psychological Medicine. Liness. G. (2005). R.. & Amir.). Izgiç. A. 15. Moscovitch. T. E. Recurrent images and early memories in social phobia. E. J. Foa. Heimberg. & Williams. Concerns about physical appearance in social anxiety impede the therapeutic effects of video feedback. New York: Oxford University Press. R. 199–212. & McManus. 195 Leary.. B. Williams.. G. Scandura. Rowa. & Rapee.. Journal of Behavior Therapy and Experimental Psychiatry. B. emotion. F.. & Hofmann. & Moscovitch. A. A. A. (1998). and alternatives. & In-Albon. E. 630–634. M.. A. & Swinson.. (2004). the Social Phobia Scale (SPS) and the Social Interaction Anxiety Scale (SIAS). & Telch. (1985). R. & Rapee. L. A. Discriminant validity of the Social Phobia and Anxiety Inventory (SPAI). B. Wilson. & Heimberg. The Appraisal of Social Concerns Scale: Psychometric valida- tion with a clinical sample of patients with social anxiety disorder. M. (1997). R. Schlenker.. (2006). issues. M. M. Clinical Psychology Review. Rodebaugh. Reconceptualizing the self. 38. T. P.). R.196 moscovitch & huyder Osman... & Abbott.. (2006). 20. Structural equation modeling: Concepts. 37. & Dolan. 393–405. J. Why is the self important in understanding and treating social phobia? Cognitive Behaviour Therapy. 44. Those who think they look worst respond best: Self-observer discrepancy predicts response to video feedback following a speech task. K. 48–54. & Chiros. L. X. R E C E I V E D : August 31. 741–756. R. M.. V. Stopa. B. Journal of Language and Social Psychology. R. Cognitive and Behavioral Practice.. C. Hoyle (Ed. Behaviour Research and Therapy. Heimberg. 705–715. 943–950. T. 22. M. R. J. 142–148. West.. A cognitivebehavioral model of anxiety in social phobia. S. Behaviour Research and Therapy. Spurr. P. (1995). Stopa. Schneier. R. Journal of Psychopathology and Behavioral Assessment. 2009 A C C E P T E D : April 27. & Rapee. 35. A. 2010 Available online 16 December 2010 . Structural Equation Modeling. G. H. M. Mental representation of observable attributes in people with social phobia. J. 29. Barrios. (2000). (2006). 38. Cognitive Therapy and Research. J. & Leary. CA: Sage. J. R. R. 113–126. 171–192. Rapee... Self-concept certainty in social phobia. and applications (pp. & Stopa. 16. (2002). M. 947–975. F. G. Liebowitz. Self-focused attention in social phobia and social anxiety. J. Kopper. Wicherts. L. G. M. Behaviour Research and Therapy. 56–75). L. 4. Structural equation models with nonnormal variables: Problems and remedies. Finch. T. Behavior Therapy.. C.. 37. (2009b). J. & Curran. Rodebaugh. L.. (1998).. Social anxiety and communication about the self.. M. 45–50.. Gutierrez. F. In R. F.. (2009a)... 113–136. Schultz. 11. Rapee. Journal of Behavior Therapy and Experimental Psychiatry. R. M. L. (2006). Peters. The Social Phobia and Social Interaction Anxiety Scales: Evaluation of psychometric properties. 249–264. L. E. Newbury Park. M. A cautionary note on the use of information fit indexes in covariance structure modeling with means.