Available online at www.sciencedirect.com

ScienceDirect Behavior Therapy 45 (2014) 116 – 125

www.elsevier.com/locate/bt

Cognitive Indicators of Social Anxiety in Youth: A Structural Equation Analysis Brittany M. Rudy Thompson E. Davis, III Louisiana State University Russell A. Matthews Bowling Green State University

Previous studies have demonstrated significant relationships among various cognitive variables such as negative cognition, self-efficacy, and social anxiety. Unfortunately, few studies focus on the role of cognition among youth, and researchers often fail to use domain-specific measures when examining cognitive variables. Therefore, the purpose of the present study was to examine domain-specific cognitive variables (i.e., socially oriented negative self-referent cognition and social self-efficacy) and their relationships to social anxiety in children and adolescents using structural equation modeling techniques. A community sample of children and adolescents (n = 245; 55.9% female; 83.3% Caucasian, 9.4% African American, 2% Asian, 2% Hispanic, 2% “other,” and 1.2% not reported) completed questionnaires assessing social cognition and social anxiety symptomology. Three latent variables were created to examine the constructs of socially oriented negative self-referent cognition (as measured by the SONAS scale), social self-efficacy (as measured by the SEQSS-C), and social anxiety (as measured by the SPAI-C and the Brief SA). The resulting measurement model of latent variables fit the data well. Additionally, consistent with the study hypothesis, results indicated that social self-efficacy likely mediates the relationship between socially oriented negative self-referent cognition and social anxiety, and socially oriented negative self-referent cognition yields significant direct and indirect effects on social anxiety. These findings indicate that socially oriented negative cognitions are Address correspondence to Brittany M. Rudy, Department of Psychology, Louisiana State University, 236 Audubon Hall, Baton Rouge, LA 70803; e-mail: [email protected]. 0005-7894/45/125-134/$1.00/0 © 2013 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.

associated with youth's beliefs about social abilities and the experience of social anxiety. Future directions for research and study limitations, including use of cross-sectional data, are discussed.

Keywords: social anxiety; social self-efficacy; socially oriented negative self-statements; youth

THE MOST EFFICACIOUS METHOD of treatment for social anxiety among children and adolescents is cognitive-behaviorally based treatment (Chambless & Ollendick, 2001; Chambless et al., 1998; Davis, May, & Whiting, 2011; Hodson, McManus, Clark, & Doll, 2008; Ollendick & King, 2004; Silverman & Ollendick, 2008; Silverman, Pina, & Viswesvaran, 2008). Unfortunately, until recently, cognitive variables have been largely overlooked in children (Davis & Ollendick, 2005; Davis et al., 2011), and cognitive components of treatment have been focused more on skills implementation than thought modification (Melfsen et al., 2011). Although exposure is the most critical component of cognitive behavior therapy (CBT; Davis & Ollendick, 2005), it has been posited that children can benefit from cognitive strategies in therapy, given the appropriate circumstances and modifications (Davis et al., 2011; Melfsen et al., 2011; Ronen, 1992). Previous studies have implicated specific cognitive variables (e.g., negative selfreferent cognitions, self-concept, self-efficacy, and coping beliefs) in the presentation of social anxiety, but questions remain as to the specific role that each cognitive variable plays. The purpose of the present

cognitive indicators of social anxiety study was to further explore the relationships among socially oriented cognitive variables and social anxiety among children and adolescents using structural equation modeling (SEM) techniques to better delineate the cognitive process of social anxiety in youth and perhaps inform and enhance cognitivebehavioral treatment practices. Although social phobia is known as a clinical disorder that involves an excessive or unreasonable, marked and persistent fear of social situations that invariably provokes an anxiety response (DSM-IV-TR; American Psychiatric Association, 2000), Schlenker and Leary (1982) defined the continuous construct of “social anxiety,” as a broader and more encompassing construct that results from the prospect or presence of personal evaluation along with a fear of social failure and criticism. Two types of situations, interaction anxiety and performance anxiety, have predominantly been identified as anxiety-provoking and distinguishable areas of social anxiety (Mattick & Clarke, 1998). Whereas each of these areas is distinct, both have demonstrated strong relationships with socially oriented cognitive variables. Negative cognition has been linked to the manifestation of social anxiety among youth and adults (Cieslak, Benight, & Lehman, 2008; Glass & Furlong, 1990; Leary & Atherton, 1986; Muris, 2002). Children are thought to be capable of perspective taking and recursive thought (i.e., thinking about what another person is thinking about) beginning around 7–8 years of age, with full development of more abstract “other” perspective thinking by age 15 years (Selman, 1980; Selman & Jaquette, 1977). Similarly, research regarding theory of mind (TOM), the ability to attribute mental states, beliefs, intents, and so on to oneself and others, suggests that perspective taking and metacognition begin to develop as young as 5–7 years with TOM being more fully developed by the middle teenage years (Wellman, Cross, & Watson, 2001). Negative self-evaluations, as well as perceptions of how others view oneself, are thought to heavily impact one’s own anxiety for social situations and performance in those situations (Clark & Wells, 1995; Leary, 1983). Given this information, onset of childhood social anxiety typically occurs, not surprisingly, between the late childhood and midteenage years (DSM-IV-TR; Davis, Munson, & Tarzca, 2009; Van Roy, Kristensen, Groholt, & Clench-Aas, 2009). Negative self-referent cognition is classified as internal dialogue that is negative in content, state oriented, and can have a lasting emotional impact on disordered and nondisordered youth (Ronan, Kendall, & Rowe, 1994; Sood & Kendall, 2007). General negative self-statements are associated with

117

irrational beliefs, fear of negative evaluation, and global behavior ratings of social interactions (Glass & Furlong, 1990), with a greater number of negative self-statements being positively correlated with higher levels of social anxiety in adults (e.g., Glass & Furlong, 1990), children, and adolescents (e.g., Wichmann, Coplan, & Daniels, 2004). However, the content and frequency of negative self-statements tend to differ across different anxiety-provoking situations (King, Mietz, Tinney, & Ollendick, 1995; Prins, 1986; Sood & Kendall, 2007), and therefore, it may be more appropriate to examine the impact of more situationally specific negative self-referent cognition in social anxiety. Rudy and Davis (2013) created a measure designed specifically to address negative self-referent cognition among the two types of social anxiety (performance anxiety and interaction anxiety). Preliminary investigations indicated that situationally constrained (socially oriented) negative cognitions were highly associated with the increased self-report of social anxiety among youth (Rudy & Davis, 2013). Melfsen and colleagues (2011) demonstrated similar findings with a similar assessment measure (i.e., the SAKK; Graf, Gerlach, & Melfsen, 2007) designed for assessment of German youth. Using vignette-based social scenarios, Prinstein, Cheah, and Guyer (2005) found that their parallel construct, critical self-referent attributions were positively associated with increased internalizing symptoms such as depression, social anxiety, and loneliness. Further, Clark and Wells’s (1995) cognitive model of social anxiety indicated that negative social cognitions and self-focused attention play a crucial role in the manifestation of social anxiety, a finding that has been corroborated with youth by multiple researchers (e.g., Hodson et al., 2008; Johnson & Glass, 1989). The construct of self-efficacy has also been established as an important cognitive component of the experience of social anxiety. Bandura (1977, 1997) described self-efficacy as the belief in one’s own capability to do something. This construct is also thought to be best understood as situationally specific in nature (Bandura, 1977, 1997; Leary & Atherton, 1986) but is likely somewhat more stable or trait based than negative self-statements (Bandura, 2001). Socially oriented self-efficacy, or “social self-efficacy,” is best described as a person's belief in his or her ability to accurately execute social situations. For instance, if a person does not believe that he or she can accurately perform a social situation, the lack of confidence is likely to heighten anxiety and decrease the probability of competently handling the situation (Leary & Atherton, 1986). In youth, researchers have repeatedly demonstrated the connection between social self-efficacy and social

118

rudy et al.

anxiety, with social self-efficacy being highly associated with the experience of feelings of social anxiety (Muris, 2002; Smári, Pétursdóttir, & Þorsteinsdóttir, 2001; Spence, Donovan, & Breckman-Toussaint, 1999; Wheeler & Ladd, 1982). Further, social self-efficacy has been determined as distinct among different relationships (i.e., peers, strangers, adults), and may differentially impact anxiety within those relationships (Hannesdóttir & Ollendick, 2007). Specifically how this construct affects social anxiety, however, is poorly understood. To better comprehend the cognitive aspect of social anxiety in youth, it is important to understand associated cognitive variables in the context of one another, rather than just their singular associations with social anxiety. Negative self-referent cognition has been found to be highly associated with selfefficacy (Rudy, Davis, & Matthews, 2012; Rudy, May, Davis, & Matthews, 2013). In referencing the theory of self-reflectiveness, Bandura (2001) discussed that self-reflectiveness is considered to be a multitier process in which thoughts may lead to beliefs, which lead to future actions or agency. Self-reflectiveness influences a person’s beliefs, and in turn, his or her competency and life choices. Taking into consideration the situationally specific nature of these constructs, it is likely that socially oriented negative self-referent cognition (more state based) influences social self-efficacy (more trait based), rather than vice versa, which in turn influences social anxiety. This specific model for variable relationships has been demonstrated in other contexts such as coping and PTSD (see Cieslak et al., 2008). Rudy and colleagues (2012) demonstrated that global self-efficacy mediated the relationship between global negative selfreferent cognition and social anxiety; however, social self-efficacy did not fully mediate the relationship between global negative cognition and social anxiety. The question remains as to whether social selfefficacy was too specific to encompass the effects of broad negative cognitions (i.e., a method effect— constructs poorly matched) or if even socially oriented negative self-statements (a more situationally specific and contextually appropriate construct) are powerful enough to influence social anxiety beyond their effect on social self-efficacy. Melfsen and colleagues (2011) demonstrated that cognitive therapy targeting negative cognitions and coping/efficacy in social situations significantly decreased socially related negative selfevaluations and overall levels of social anxiety as compared with controls; however, coping/efficacy beliefs for social situations remained the same. Also, while significant decreases in social anxiety were demonstrated, only 36% of the sample no longer met criteria for social phobia at treatment completion (Melfsen et al., 2011). Understanding the relationship

among these three variables could have important implications in identifying treatment channels and target variables. As stated before, CBT has been demonstrated as a well-established, best-practice technique for treating social anxiety in youth (e.g., Chambless & Ollendick, 2001; Chambless et al., 1998; Ollendick & King, 2004; Silverman et al., 2008). Although exposure is likely the most critical component of treatment (Chorpita & Daleiden, 2009; Davis & Ollendick, 2005), a basic premise of the utility of CBT lies in the idea that through cognitive restructuring, altering cognitions can change core beliefs (Chorpita, 2007; Kendall, 1993). If, in this way, negative cognitions (i.e., negative self-statements) are suspected theoretically to influence core beliefs (e.g., self-efficacy), targeting these variables more specifically in a stepwise progression as a part of exposure-based CBT may lead to enhanced treatment outcomes.

Present Study The purpose of the present study was to examine the relationships among contextually appropriate, socially oriented cognitive variables and social anxiety among children and adolescents using SEM techniques. Variable modeling using SEM may provide important information about these specific variables and their place among the cognitive trends of social anxiety among youth. Therefore, a measurement model was tested to verify the latent constructs of socially oriented negative self-referent cognition, social self-efficacy, and social anxiety. A structural model was then utilized to test whether the latent construct of social self-efficacy would mediate the relationship between socially oriented negative selfreferent cognition and social anxiety. Bootstrapping was also utilized to determine whether socially oriented negative self-referent cognition would have a significant indirect effect on social anxiety. Based on results of previous contextually similar analyses indicating that global self-efficacy mediated the relationship between global negative self-statements and social anxiety (Rudy & Davis, 2013), it was hypothesized that the measurement and structural models would be well fitted to the data with the structural model indicating that social self-efficacy would mediate the relationship between socially oriented negative self-referent cognition and social anxiety such that these results will help to further define the role of that cognitive variable with respect to the proposed relationships.

Method participants Two hundred forty-five children and adolescents from local area schools (i.e., a community sample)

cognitive indicators of social anxiety participated in the current study, ensuring that the study was adequately powered per suggested parameters for use of SEM (Kline, 2005). In total, 260 participants were recruited for participation in the study; however, 15 participants were excluded from analysis due to missing data (N 10% of items on any of the utilized measures unanswered). Youth from the final sample were between the ages of 8 and 16 years (M = 13.27, SD = 2.14) and were 55.9% female. This age span is noted to be wide in range but was chosen to allow for developmental sensitivity given knowledge of developmental theories of perspective taking and recursive thought (Selman, 1980; Selman & Jaquette, 1977) while reaching a wide range of potentially affected youth (Davis et al., 2009; Van Roy et al., 2009). Further, 72.1% of the sample was above the age of 12 years, increasing the likelihood that the majority of the sample yielded more fully developed perspective-taking skills. Ethnicity was distributed as follows: 83.3% Caucasian, 9.4% African American, 2% Asian, 2% Hispanic, 2% reporting “other” ethnic origin, and 1.2% not reporting ethnic origin. Consistent with Kessler and colleagues’ estimation of social phobia prevalence among community samples (12.1%; 2005), 13.1% of the sample exceeded the clinical cutoff of 18 on the Social Phobia and Anxiety Inventory for Children (SPAI-C; Beidel, Turner & Morris, 1998; see Measures below), and 11.4% of the sample scored at or above a score of 4 on the Brief Social Anxiety Questionnaire (Brief SA; Rudy & Davis, 2013; see Measures below).

measures Demographics A demographic questionnaire was created and used to obtain background information about the child and his or her family. The questionnaire gathered pertinent information such as age, gender, race/ethnicity of the child, and presence of any developmental disabilities. Socially Oriented Negative Self-Referent Cognition A new measure of socially oriented negative selfstatements, the Socially Oriented Negative Anxious Statement Scale (SONAS; Rudy & Davis, 2013) was used to measure participants’ socially oriented negative self-referent cognition. The SONAS is a 16-item self-report questionnaire that assesses two domains of socially oriented negative self-statements. The first factor is conceptualized to tap performance anxiety (α = .84) and the second factor conceptually taps interaction anxiety (α = .80). Items (worded in terms of anxious statements) are rated on a 4-point Likert scale from never to being all the time. The measure was designed for youth ages 8–16 years and demonstrated good psychometric properties. Rudy and

119

Davis (2013) reported good construct validity— convergent and divergent—when compared with similar measures of negative cognition in youth (i.e., NASSQ; Ronan et al., 1994; and the CATS; Schniering & Rapee, 2002) and subscales of those measures. The SONAS was also found to exhibit good criterion-related validity (i.e., was appropriately associated with increased report of social anxiety) and appropriate demographic analysis results with expected gender differences (fewer thoughts reported by males—consistent with previous literature—see Schniering & Rapee, 2002), and no significant age or ethnicity effects (Rudy & Davis, 2013). Guidelines indicate that alpha coefficients (i.e., Cronbach’s alpha) below .70 indicate unacceptable internal consistency, coefficients between .70 and .79 indicate fair internal consistency, coefficients between .80 and .89 indicate good internal consistency, and coefficients above .90 indicate excellent internal consistency (Ciccheti, 1994; Ciccheti & Sparrow, 1990). For the current sample, the performance (α = .84), interaction (α = .80), and total scales (α = .88) demonstrated good internal consistency as measured by Cronbach’s alpha. Social Anxiety For the purpose of this study, social anxiety was indexed in terms of two independent but related factors assessed by separate anxiety-based measures. The first measure, the Social Phobia and Anxiety Inventory for Children (SPAI-C; Beidel et al., 1998) was used to assess the social anxiety. The SPAI-C is a 26-item self-report questionnaire designed for children and adolescents ages 7–16 years. The measure assessed the physical, cognitive, and avoidant domains of social phobia. Items are rated on a 3-point Likert scale with 0 being never or hardly ever and 2 being always or almost always. Within the normative sample, the mean for nonsocially anxious children was 13.74 (SD = 8.5) and the mean for socially anxious children was 21.8 (SD = 8.4) with the clinical cutoff for social phobia being 18 (Beidel & Morris, 1995; Beidel et al., 1998). The inventory has previously demonstrated good internal consistency and test–retest reliability (α = .95, r = .86; Beidel et al., 1998). For the current sample, the measure yielded good internal consistency (α = .95). In order to be consistent with how this measure is scored (i.e., the measure is not composed of separate subscales; Beidel et al., 1998), the SPAI-C was indexed in terms of a total overall score. Additionally, a brief social anxiety questionnaire, the Brief SA Questionnaire (Rudy & Davis, 2013), was assessed as a second indicator of social anxiety. This measure addressed self-endorsed feelings of

120

rudy et al.

interaction and performance anxiety, interference, and distress in a DSM-IV-TR (American Psychiatric Association, 2000) checklist format. Participants were asked to provide “yes” or “no” responses to appropriately worded DSM-IV-TR criteria (seven items) for social phobia (American Psychiatric Association, 2000). The measure was scored additively with a score above 4 warranting clinical attention. The measure yielded an acceptable internal consistency (α = .78). Self-Efficacy The Self-Efficacy Questionnaire for Social Skills for Children (Ollendick & Schmidt, 1987) assessed children’s self-efficacy specifically for social situations by inquiring if the child thinks that he or she could complete the social task being asked of him or her. The scale is a 10-item self-report measure that is rated on a 5-point Likert scale with 1 being not sure at all and 5 being really sure. The scale has high internal consistency and test–retest reliability (r = .87, r = .75; Ollendick & Schmidt). For the purpose of this study, each of the 10 items was rated for peers, strangers, and adults separately, creating a total of 30 items to be used in a summed total score. Internal consistency was excellent for the total scale (α = .92), good for the peers subscale (α = .87) and strangers subscale (α = .86), and adequate for the adult subscale (α = .79).

procedure Permission was obtained from the university’s Institutional Review Board to recruit and gather information from participants. Participants were recruited from local area (urban) elementary, middle, and high schools in the southeastern United States. An informational e-mail was used to inform teachers of the study, and a promotional letter was sent home to inform parents of the study and to ask permission for their child’s participation. Parents were required to sign and return informed consent forms and demographic forms before participants were eligible for participation. Demographic questionnaires were screened for developmental disabilities and non-English-speaking families (exclusion criteria). Participants whose parents signed and returned the consent forms and demographic information completed questionnaire packets during an arranged assembly. Assent forms were signed by participants to ensure their agreement to participate as well as their understanding of the study and limits of confidentiality before questionnaire packets were completed. Multiple experimenters were available during the assembly to answer any questions from the participants. A debriefing sheet was sent home with each partici-

pant upon measure completion containing information about the study purpose as well as investigator contact information for any further questions, concerns, or results inquiry.

Results preliminary analysis Given that the Brief SA Questionnaire is a relative new measure, a preliminary confirmatory factor analysis was conducted on the measure using Mplus (v6.0 with WLSMV estimation; Muthén & Muthén, 1998– 2012). Because it is assumed that each of the seven dichotomous items is related to a continuous higherorder latent factor, a tetrachoric correlation matrix was used. The single-factor model fit the data well: χ 2(14) = 10.07, p N .05, CFI = 1.00, RMSEA = .00. model testing Structural equation modeling techniques (Mplus, v6.0 with maximum likelihood estimation; Muthén & Muthén, 1998–2012) were utilized to examine potential predictive models. Also, the use of latent models allowed us to model the construct indicators to ensure adequate measurement of core constructs, as well as model measurement error. Four measures of model fit, chi-square (χ 2), comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR) were used to determine the fit of the tested models. When evaluating model “fit,” model complexity, the theoretical underpinnings of the model, and the interpretability of estimates must be considered (Hu & Bentler, 1999; Marsh, Hau, & Wen, 2004). Accordingly, we considered several key issues in our evaluation of model fit. First, modification indices were examined to ensure the model did not systematically deviate from the data. A nonsignificant χ 2 indicates good model fit; however, χ 2 is sensitive to sample size. A CFI value of .95 or higher and a SRMR value of .08 or lower is indicative of good model fit (Hu & Bentler, 1999). Next, model fit also was assessed in terms of congruence with theoretical propositions; if at least two thirds of proposed hypotheses were supported, the model would be classified as having good fit. Finally, the interpretability of model parameters was assessed (i.e., no negative variance estimates). First, the measurement model was assessed. The three core constructs (i.e., socially oriented negative self-referent cognition, social self-efficacy, and social anxiety) were each represented as a latent construct. For socially oriented negative self-referent cognition, the two subscales from this measure were each modeled as latent variables, with items from those subscales serving as indicators (see Figure 1). For social self-efficacy the three subscales from the

cognitive indicators of social anxiety

121

FIGURE 1 Social self-efficacy was found to mediate the relationship between socially oriented negative self-referent cognition and social anxiety in youth, with socially oriented negative self-referent cognition yielding significant standardized direct and indirect effects on social anxiety in this proposed structural model.

measure were used as indicators (Note: we did not model this construct at the item level given concerns of model complexity relative to sample size; Brown, 2006). Finally, for social anxiety two separate measures of social anxiety (i.e., the SPAI-C and the Brief SA questionnaire) were set indicators of the construct. Because of the manner in which the SPAI-C is scored, it was not possible to model this construct at the item level. And, again in light of model complexity concerns, the Brief SA questionnaire was treated as a summated indicator variable (i.e., the construct was not indexed at the item level). These two constructs correlated at .66 (p b .01), suggesting strong convergent validity. Consistent with standard SEM practices, the three latent constructs (socially oriented negative selfreferent cognition, social self-efficacy, and social anxiety) were then set free to correlate with one another. Although the χ 2 and the CFI did not meet standard cutoff levels χ 2 (184) = 303.51, p b .01, SRMR = .06, CFI = .93, RMSEA = .05, all items/ indicators loaded strongly (.48–.88) on their respective factors (Brown, 2006). Further, based on review of the standardized residuals and modification indices, there was no indication that any of the indicators cross-loaded on another construct. Collectively then, it appears the measurement model fit the data adequately. It should be noted that given the relatively large age range included in our sample, age was initially included as a covariate in our analyses. However,

for reasons of parsimony it was removed from the reported model given inclusion of age as a covariate did not influence the nature of our results; the pattern of parameter was nearly identical when age was included versus when it was excluded. To examine issues of discriminate validity among the constructs, a second measurement model was calculated wherein all indicators were set to load on a single latent composite factor. This model fit the data poorly: χ 2 (189) = 684.26, p b .01, SRMR = .09, CFI = .72, RMSEA = .10. These results suggest that the three latent constructs represented in the measurement model are distinct constructs from one another. Next, we tested the structural model (see Figure 1) to examine whether social self-efficacy mediates the relationship between socially oriented negative selfstatements and social anxiety. This structural model fit the data adequately: χ 2 (184) = 303.51 p b .01, SRMR = .06, CFI = .93, RMSEA = .05. Parameter estimates indicated significant standardized direct effects of socially oriented negative self-statements on social self-efficacy (β = –.46, p b .01) and social anxiety (β = .67, p b .01), as well as a significant standardized direct effect of social self-efficacy on social anxiety (β = –.36, p b .01). Bootstrapping analyses (see Kline, 2005; Preacher & Hayes, 2004; Ullman, 2007) were conducted to test for indirect effects. Maximum likelihood bootstrapping was used to estimate standard errors and confidence intervals (95%) for all relevant effects (5,000 samples were drawn). Socially oriented negative self-statements had

122

rudy et al.

a significant standardized indirect effect (β = .17, p b .01) on social anxiety, confirming that social self-efficacy mediates the relationship between socially oriented negative self-statements and social anxiety.

alternative model testing Although our model was theoretically driven and supported by prior research, the possibility of alternative models that fit equally well exists. As such, we tested an alternative model wherein we examined whether the mediation process actually works better in reverse. Specifically, we tested a model where social anxiety was set to predict negative self-statements with social self-efficacy again acting as a mediator. Although this model fit the data as well as the conceptual model, χ 2 (184) = 303.51, p b .01, SRMR = .06, CFI = .93, RMSEA = .06, the direct effect from social self-efficacy to negative selfstatements was not significant in this model (β = .17, p N .05). These results provide additional support for the ordering of effects originally proposed.

Discussion The aim of the present study was to examine domain-specific cognitive variables as related to social anxiety in youth via theoretical modeling using SEM techniques. Our structural model revealed that social self-efficacy likely mediates the relationship between socially oriented negative self-referent cognition and social anxiety, with socially oriented negative selfreferent cognition also yielding significant direct and indirect effects on social anxiety. The model preliminarily suggests that it is likely that the relationship between socially oriented negative self-referent cognition and social anxiety exists beyond the association between socially oriented negative self-referent cognition and social self-efficacy. These findings emphasize the importance of domain-specific negative cognition and confirm previous findings even with the use of more contextually appropriate variables (see Rudy et al., 2012). Understanding this new model is critical in delineating the cognitive aspect of social anxiety presentation in youth. How children view themselves with regard to social situations and their views of their social abilities each uniquely contributes to the amount of social anxiety experienced. Previous research has focused on ability beliefs as the primary cognitive influence on social anxiety in youth (Muris, 2002; Smári et al., 2001; Spence Donovan et al., 1999; Wheeler & Ladd, 1982) as well as performance and skills deficits (e.g., Cartwright-Hatton, Hodges, & Porter, 2003). It is possible, however, that socially specific negative self-views and self-related thoughts play a unique role in the cognitive process of social anxiety beyond ability beliefs and, therefore,

should be incorporated into the cognitive framework of social anxiety in youth. Further, alternative model testing supported the proposed ordering of effects. Melfsen and colleagues (2011) demonstrated that within a developmentally appropriate context, cognitive therapy targeting negative cognitions pertaining to social situations can significantly decrease socially related negative self-evaluations and overall levels of social anxiety. Given that a basic premise of CBT is the alteration of core beliefs via cognitive restructuring (Chorpita, 2007; Kendall, 1993), understanding the mechanisms through which cognitive variables are associated with and potentially influence social anxiety is important. The present study highlighted socially oriented negative self-referent cognition as a potentially powerful and important cognitive variable that may influence social anxiety both independently and through its effects on core beliefs and efficacy for social situations. Unfortunately, cognition is often overlooked in treatment of children and adolescents (Davis et al., 2011), particularly with respect to thought modification (Melfsen et al., 2011). Cognitive therapy alone is likely not sufficient for remission in the majority of youth (Melfsen et al., 2011), but perhaps given their importance in the cognitive framework of social anxiety, specifically targeting socially oriented negative cognitions and monitoring these negative self-evaluations more closely as a part of exposure-based CBT may lead to enhanced treatment outcomes. That said, more research is needed to longitudinally examine these constructs within the context of treatment and confirm ordering of effects and the possible predictive nature of these variables.

limitations As with all studies, this study is not without limitations. Only 13.1% of the study sample exhibited clinical elevations of social anxiety on the SPAI-C. This proportion is comparable to Kessler and colleagues’ (12.1%; 2005) community sample estimations; however, more research is needed to determine whether the relationships demonstrated in this study would remain stable in a clinical sample. Further, the sample was 83.3% Caucasian, 72.1% of the sample was above the age of 12, and approximately 76% of participants came from a tuitionrequired school with primarily middle- to upper-class students. Given that variables such as age, ethnicity, and socioeconomic status (SES) were restricted within the sample, generalizability to other populations (e.g., lower SES, minority, younger children) may be limited. Additionally, the study design was cross-sectional in nature. Whereas SEM techniques were explicitly

cognitive indicators of social anxiety chosen to model variable relationships and maximize the variance accounted for when analyzing data (Kline, 2005; Ullman, 2007), and alternative model testing was utilized to help confirm order of effects, longitudinal data are necessary to fully examine predictive relationships and directional effects of cognitive variables such as socially oriented negative self-referent cognition and social self-efficacy on social anxiety. Conclusions drawn from cross-sectional data warrant some concern, cannot be directional, and should be tentative and interpreted with caution (Cole & Maxwell, 2003). Future studies should focus on collection of data at multiple time points as well as within the context of progress monitoring throughout treatment. It should be noted that, like most measures based on dichotomous indicators, the Brief SA questionnaire did demonstrate some non-normality, in terms of skewness. As noted by Kline (2005), maximum likelihood estimation is relatively robust against non-normality. However, non-normality can inflate a model’s χ 2 value. Researchers looking to extend the results from the present study should consider collecting larger samples or avoid using dichotomous measures in order to address this concern (Kline, 2005). More broadly, as noted within the measurement model, while it is recommended that all indicators be modeled at the same level for consistency and standardized interpretation of latent constructs (Kline, 2005), some constructs were modeled at the item level, and some were modeled using computed subscale indicators. In certain instances, such as the SPAI-C, this was necessary because the measure total score is computed based on a multiplicative formula. Further, and more practically, with our relatively small sample size (N = 245), modeling all measures at the item level would not have been consistent with SEM best practices in terms of sample size to indicator ratios and would likely have resulted in unstable parameter estimates and inadequate power (Brown, 2006). Ideally, and consistent with SEM best practices, all constructs would have been modeled at the item level. Further, for our social anxiety latent construct, only two indicators were included, when it is generally recommended that three indicators be used to assess a latent construct. Having only two indicators for a latent factor can lead to estimation problems (Kline, 2005). These are issues researchers should consider in seeking to extend present study findings.

Conclusion In conclusion, an examination of domain-specific cognitive variables as related to social anxiety in youth using SEM techniques yielded significant

123

results with social self-efficacy mediating the relationship between socially oriented negative self-referent cognition and social anxiety among children and early adolescents. Socially oriented negative cognitions were found to be associated with beliefs about social abilities and, subsequently, level of social anxiety. Additionally, socially oriented negative cognition retained a unique contribution to the amount of social anxiety present, asserting its importance in the cognitive framework of social anxiety presentation among youth. Future research should focus on replicating the association between these cognitive variables and social anxiety in youth in clinical samples longitudinally within the context of progress monitoring and treatment. Conflict of Interest Statement The authors declare that there are no conflicts of interest.

References American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191–215. Bandura, A. (1997). Self-efficacy: The exercise of control. New York, NY: Freeman. Bandura, A. (2001). Social cognitive theory: An agenic perspective. Annual Review of Psychology, 52, 1–26. Beidel, D. C., & Morris, T. L. (1995). Social phobia. In J. S. March (Ed.), Anxiety disorders in children and adolescents (pp. 181–211). New York, NY: Guilford Press. Beidel, D. C., Turner, S. M., & Morris, T. L. (1998). Social Phobia and Anxiety Inventory for Children. Toronto, Canada: Multi-Health Systems. Brown, T. A. (2006). Confirmatory factor analysis for applied research. Is it multiplicative, or additive or something else? New York, NY: Guilford Press. Cartwright-Hatton, S., Hodges, L., & Porter, J. (2003). Social anxiety in childhood: The relationship with self and observer rated social skills. Journal of Child Psychology and Psychiatry, 44, 737–742. Chambless, D. L., Baker, M. J., Baucom, D. H., Beutler, L. E., Calhoun, K. S., Crits-Christoph, P., ... Woody, S. (1998). Update on empirically validated therapies, II. Clinical Psychologist, 51, 3–16. Chambless, D. L., & Ollendick, T. H. (2001). Empirically supported psychological interventions: Controversies and evidence. Annual Review of Psychology, 52, 685–716. Ciccheti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating standardized assessment instruments in psychology. Psychological Assessment, 6, 184–190. Ciccheti, D.V. & Sparrow, S. S. (1990). Assessment of adaptive behavior in young children. In J. J. Johnson & J. Goldman (Eds.), Developmental assessment in clinical child psychology: A handbook (pp. 173–196). New York: Pergamon Press. Cieslak, R., Benight, C. C., & Lehman, V. C. (2008). Coping self-efficacy mediates the effects of negative cognitions on posttraumatic distress. Behaviour Research and Therapy, 46, 788–798. Chorpita, B. F. (2007). Modular cognitive-behavioral therapy for childhood anxiety disorders. New York, NY: Guilford Press.

124

rudy et al.

Chorpita, B. F., & Daleiden, E. L. (2009). Mapping evidencebased treatments for children and adolescents: Application of the distillation and matching model to 615 treatments. Journal of Counseling and Clinical Psychology, 77(3), 566–579. Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In R. Heimberg, M. Liebowitz, D. A. Hope, & F. R. Schneier (Eds.), Social phobia: Diagnosis, assessment and treatment (pp. 69–93). New York, NY: Guilford Press. Cole, D. A., & Maxwell, S. E. (2003). Testing mediational models with longitudinal data: Questions and tips in the use of structural equation modeling. Journal of Abnormal Psychology, 112, 558–577. Davis, T. E., III, May, A. C., & Whiting, S. (2011). Evidencebased treatment of anxiety and phobia in children and adolescents: Current status and effects on the emotional response. Clinical Psychology Review, 31, 592–602. Davis, T. E., III, Munson, M., & Tarzca, E. (2009). Anxiety disorders and phobias. In J. Matson (Ed.), Social behavior and social skills in children (pp. 219–244). New York, NY: Springer Science and Business Media. Davis, T. E., III, & Ollendick, T. H. (2005). A critical review of empirically supported treatments for specific phobia in children: Do efficacious treatments address the components of a phobic response? Clinical Psychology: Science and Practice, 12, 144–160. Glass, C. R., & Furlong, M. (1990). Cognitive assessment of social anxiety: Affective and behavioral correlates. Cognitive Therapy and Research, 14, 365–384. Graf, A., Gerlach, A. L., & Melfsen, S. (2007). Fragebogen zur Erfassung sozial ängstlicher Kognitionen bei Kindern und Jugendlichen. Zeitschrift für Kinder- und Jugendpsychiatrie und Psychotherapie, 35, 257–264. Hannesdóttir, D. K., & Ollendick, T. H. (2007). Social cognition and social anxiety among icelandic school children. Child and Family Behavior Therapy, 29, 43–58. Hodson, K., McManus, F., Clark, D. M., & Doll, H. (2008). Can Clark and Wells’ (1995) model of social phobia be applied to young people? Behavioural and Cognitive Psychotherapy, 36, 449–461. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. Johnson, R. L., & Glass, C. R. (1989). Heterosocial anxiety and direction of attention in high school boys. Cognitive Therapy and Research, 13, 509–526. Kendall, P. C. (1993). Cognitive-behavioral therapies with youth: Guiding theory, current status, and emerging developments. Journal of Consulting and Clinical Psychology, 61, 235–247. Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM–IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62, 593–602. King, N. J., Mietz, A., Tinney, L., & Ollendick, T. H. (1995). Psychopathology and cognition in adolescents experiencing severe test anxiety. Journal of Clinical Child Psychology, 24, 49–54. Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.). New York, NY: Guilford Press. Leary, M. R. (1983). Social anxiousness: The construct and measurement. Journal of Personality Assessment, 47, 66–75. Leary, M. R., & Atherton, S. C. (1986). Self-efficacy, social anxiety, and inhibition in interpersonal encounters. Journal of Social and Clinical Psychology, 4, 256–267. Marsh, H. W., Hau, K. -T., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis-testing approaches to

setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling, 11, 320–341. Mattick, R. P., & Clarke, J. C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behavior Research and Therapy, 36, 455–470. Melfsen, S., Kühnemund, M., Schwieger, J., Warnke, A., Stadler, C., Poustka, F., & Stangier, U. (2011). Cognitive behavioral therapy of socially phobic children focusing on cognition: A randomized waitlist control study. Child and Adolescent Psychiatry and Mental Health, 5, 5. Muris, P. (2002). Relationships between self-efficacy and symptoms of anxiety disorders and depression in a normal adolescent sample. Personality and Individual Differences, 32, 337–348. Muthén, L. K., & Muthén, B. O. (1998–2012). Mplus user’s guide (7th ed.). Los Angeles, CA: Author. Ollendick, T. H., & King, N. J. (2004). Empirically supported treatments for children and adolescents: Advances toward evidence-based practice. In P. M. Barrett & T.H. Ollendick (Eds.), Handbook of interventions that work with children and adolescents: Prevention and treatment (pp. 3–25). West Sussex, UK: Wiley. Ollendick, T. H., & Schmidt, C. R. (1987). Social learning constructs in the prediction of peer interaction. Journal of Clinical Child Psychology, 16, 80–87. Preacher, K., & Hayes, A. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, and Computers, 36, 717–731. Prins, P. J. (1986). Children’s self-speech and self-regulation during a fear-provoking behavioral test. Behaviour Research and Therapy, 24, 181–191. Prinstein, M. J., Cheah, C. S. L., & Guyer, J. E. (2005). Peer victimization, cue interpretation, and internalizing symptoms: Preliminary concurrent and longitudinal findings for children and adolescents. Journal of Clinical Child and Adolescent Psychology, 34, 11–24. Ronan, K. R., Kendall, P. C., & Rowe, M. (1994). Negative affectivity in children: Developmental and validation of a self-statement questionnaire. Cognitive Therapy and Research, 18, 509–528. Ronen, T. (1992). Cognitive therapy with young children. Child Psychiatry and Human Development, 23, 19–30. Rudy, B. M., & Davis, T. E., III. (2013). Socially oriented negative self-referent cognition in youth: The development and validation of a measure. Unpublished manuscript. Rudy, B. M., Davis, T. E., III, & Matthews, R. A. (2012). The relationship among self-efficacy, negative self-referent cognitions, and social anxiety in children: A multiple mediator model. Behavior Therapy, 43, 619–628. Rudy, B. M., May, A. C., Davis, T. E., III, & Matthews, R. A. (2013). Youth’s negative self-statements as related to social self-efficacy among differing relationships. Journal of Psychopathology and Behavioral Assessment, 35, 106–112. Schlenker, B. R., & Leary, M. R. (1982). Social anxiety and self-presentation: A conceptualization model. Psychological Bulletin, 92, 641–669. Selman, R. L. (1980). The growth of interpersonal understanding. New York, NY: Academic Press. Selman, R. L., & Jaquette, D. (1977). The development of interpersonal awareness. Cambridge: MA: Harvard–Judge Baker Social Reasoning Project. Silverman, W. K., & Ollendick, T. H. (2008). Child and adolescent anxiety disorders. In J. Hunsley & E.J. Mash (Eds.), A guide to assessments that work (pp. 181–206). New York, NY: Oxford University Press.

cognitive indicators of social anxiety Silverman, W. K., Pina, A. A., & Viswesvaran, C. (2008). Evidence-based psychosocial treatments for phobic and anxiety disorders in children and adolescents. Journal of Clinical Child and Adolescent Psychology, 37, 105–130. Smári, J., Pétursdóttir, G., & Þorsteinsdóttir, V. (2001). Social anxiety and depression in adolescents in relation to perceived competence and situational appraisal. Journal of Adolescence, 24, 199–207. Sood, E. D., & Kendall, P. C. (2007). Assessing anxious self-talk in youth: The Negative Affectivity Self-Statement Questionnaire— anxiety scale. Cognitive Therapy and Research, 31, 603–618. Spence, S. H., Donovan, C., & Breckman-Toussaint, M. (1999). Social skills, social outcomes, and cognitive features of childhood social phobia. Journal of Abnormal Psychology, 108, 211–221. Ullman, J. R. (2007). Structural equation modeling. In B. G. Tabachncick & L. S. Fidell (Eds.), Using multivariate statistics (5th ed., pp. 676–781). Upper Saddle River, NJ: Pearson Education.

125

Van Roy, B., Kristensen, H., Groholt, B., & Clench-Aas, J. (2009). Prevalence and characteristics of significant social anxiety in children aged 8–13 years: A Norwegian crosssectional population study. Social Psychiatry Psychiatric Epidemiology, 44, 407–415. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory-of-mind development: The truth about false belief. Child Development, 72, 655–684. Wheeler, V. A., & Ladd, G. W. (1982). Assessment of children’s self-efficacy for social interactions with peers. Developmental Psychology, 18, 795–805. Wichmann, C., Coplan, R. J., & Daniels, T. (2004). The social cognitions of socially withdrawn children. Social Development, 13, 377–392.

R E C E I V E D : January 2, 2013 A C C E P T E D : September 12, 2013 Available online 21 September 2013

Cognitive indicators of social anxiety in youth: a structural equation analysis.

Previous studies have demonstrated significant relationships among various cognitive variables such as negative cognition, self-efficacy, and social a...
341KB Sizes 1 Downloads 0 Views