ARTICLE IN PRESS

0005-7967/$ - se

doi:10.1016/j.br

�CorrespondE-mail addr

{Deceased.

Behaviour Research and Therapy 45 (2007) 1181–1191

www.elsevier.com/locate/brat

Social anxiety disorder in childhood and adolescence:Descriptive psychopathology

Patricia A. Raoa,�, Deborah C. Beidela, Samuel M. Turner{, RobertT. Ammermanb, Lori E. Crosbyb, Floyd R. Salleec

aPenn State Hershey College of Medicine, Hershey Medical Center, H073, 500 University Drive, Hershey, PA 17033, USAbDepartment of Pediatrics, Cincinnati Children’s Hospital Medical Center, University of Cincinnati College of Medicine, Cincinnati, OH,

USAcDepartment of Psychiatry, University of Cincinnati College of Medicine, Cincinnati, OH, USA

Received 30 January 2006; received in revised form 14 June 2006; accepted 27 July 2006

Abstract

Although the presentation of social anxiety disorder (SAD) in adults is well documented, less is known about its clinical

manifestation in children and adolescents. To date, most studies have included combined samples of children and adolescents

despite the fact that this age range represents an extensive period of growth and development. This study compares and

contrasts the clinical presentation of SAD among children (ages 7–12) and adolescents (ages 13–17). One hundred and fifty

children (n ¼ 74) and adolescents (n ¼ 76) with a primary diagnosis of SAD participated in the study. The assessment battery

included clinical ratings and behavioral observation as well as parental and self-report. The results indicate that, although the

symptom presentation of children and adolescents with primary SAD shares many features, children tend to present with a

broader pattern of general psychopathology, while adolescents have a more pervasive pattern of social dysfunction and may

be more functionally impaired as a result of their disorder. These findings suggest that interventions for SAD need to

carefully consider clinical presentation of the disorder as it manifests in childhood and adolescence.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Anxiety; Children; Adolescents; Social anxiety disorder

Background/Introduction

Social anxiety disorder (SAD) is characterized by a pervasive pattern of social timidity (AmericanPsychiatric Association [APA], 1994). It is the third most common psychiatric disorder, with a lifetimeprevalence of approximately 13% of the general population (Kashdan & Herbert, 2001). The average age ofonset is mid-adolescence, but the disorder occurs in children as young as age eight (Beidel, Turner, & Morris,1999). Although SAD has been recognized in adult populations for some time, it was not until the fourth

e front matter r 2006 Elsevier Ltd. All rights reserved.

at.2006.07.015

ing author. Tel.: +1717 531 9763.

ess: prao1@hmc.psu.edu (P.A. Rao).

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–11911182

edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; APA, 1994) that the disorderwas formally recognized in children (Beidel & Turner, 1998).

Over the past several decades, considerable research has examined the psychopathology of SAD in adults(e.g., Beidel, Turner, & Dancu, 1985; Heimberg, Holt, Schneier, Spitzer, & Liebowitz, 1993; Turner, Beidel,Dancu, & Keyes, 1986), but to a much lesser extent in adolescents (e.g., Beidel, Turner, Young et al., in press;Essau, Conradt, & Petermann, 1999), and children (Beidel, 1991; Beidel et al., 1999; Spence, Donovan, &Brechman-Toussaint, 1999). The outcome of these investigations has clearly documented that those with SADdiffer from normal control subjects on measures of emotional, occupational, academic and social functioning.One limitation of the extant data is that with very few exceptions (e.g., Beidel, 1991; Beidel et al., 1999), themajority of studies of childhood psychopathology included combined samples of children and adolescents(e.g., Arnold et al., 2003; Spence et al., 1999), which is problematic because this age range represents anextensive and evolving period of human development. Characterized by rapid physical, cognitive, andemotional maturation, it is likely that symptom presentation may vary by age group. Thus, data based on pre-adolescent samples may not be relevant for an adolescent population and vice versa.

In support of this hypothesis, age group differences have been documented in other anxiety disorders. Forexample, among samples of children with generalized anxiety disorder/overanxious disorder (GAD/OAD), olderchildren report more symptoms of the disorder than younger children (Strauss, Lease, Last, & Francis, 1988;Tracey, Chorpita, & Douban, 1997). The different symptom presentation suggests that without the knowledge ofdifferences and attention to age-related symptom variation, GAD might go undetected in younger children.

As noted, the few studies of the psychopathology of childhood SAD have focused on children and very youngadolescents (Beidel et al., 1999; Spence et al., 1999) and have not compared younger versus older groups. In theonly study of which we are aware that compares potential group differences in clinical presentation (Alfano,Beidel, & Turner, 2006), the presence of negative cognitions during social interaction tasks were reported byadolescents, but not children, with SAD. Thus, with respect to the psychopathology of SAD in youth, manyquestions remain unanswered. For example, are children and adolescents with SAD equally impaired by thedisorder? Do comorbid conditions exist in both groups, are they the same disorders and do they exist at equalfrequency? Given their greater physical and cognitive maturity, are adolescents more likely to avoid socialsituations? Thus, despite the obvious need, there has been only minimal attention to potential group variations.

Knowledge of differences in clinical presentation is important for developing interventions and evaluatingtreatment outcome as well. For example, the paucity of negative cognitions among children and adolescentsobserved by Alfano et al. (2006) may help to explain the observation by Spence and colleagues (Spence,Donovan, & Brechman-Toussaint, 2000) that a subset of children participating in a cognitive-behavioraltreatment program were unable to participate in its cognitive-restructuring component. To summarize,although the extant literature suggests that SAD can be identified in both children and adolescents (e.g., Beidel& Turner, 1998), it remains unclear whether the disorder presents differently in younger versus older youth.Identifying potential age differences in clinical presentation is not simply an academic exercise. Suchinformation is important in order to accurately identify and treat those who suffer from SAD. Identification ofpotentially age-specific differences in clinical presentation could allow for the early initiation of treatment andthereby possibly avoid the resultant functional impairment (Alnaes & Torgersen, 1999; Davidson, Hughes, &George, 1993; Keller, 2003). Furthermore, elucidating its expression in different groups could allow for thedevelopment of more targeted and therefore, efficacious treatment strategies. The purpose of this investigationwas to compare and contrast the clinical presentation of SAD among children and adolescents. Data werecompared across a range of variables including diagnostic interviews, self and parental report and behavioralassessment. It was expected that although both groups would equally manifest the core features of SAD,group differences might emerge in secondary clinical features.

Method

Participants

Participants consisted of 150 children and adolescents and their parents who sought treatment in a researchproject investigating the comparative efficacy of two interventions for SAD. Children and adolescents were

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–1191 1183

referred by local clinicians or were recruited through announcements placed in local newspapers andbroadcast on local radio stations, literature sent to school counselors, and flyers posted in public libraries, areachurches and community centers. All participants met DSM-IV criteria for a primary diagnosis of SAD.Participants ranged in age from 7 to 17 years (M ¼ 12.1 years; SD ¼ 2.7 years); there were 72 males (48.0%)and 78 females (52.0%). Ninety-two of the subjects were Caucasian (61.3%), 35 were African American(23.3%), 9 were Asian American (6.0%), 7 were Hispanic (4.7%), 1 was Indian Subcontinent (.7%), 1 wasAmerican Indian/Alaskan (.7%), and 3 did not endorse any specific racial category (2.0%). Eighty-five percentof the sample came from families classified in the middle three socioeconomic categories, as identified by theHollingshead Index of Social Position (Hollingshead, 1957). The mean IQ total score for the sample was106.24 (SD ¼ 16.8) based on scores on the Block Design and Vocabulary subsections of the WISC-III(Wechsler, 1991).

Participants were classified as children (n ¼ 74), ranging in age from 7 to 12 years (M ¼ 9.8 years, SD ¼ 1.5years), or adolescents (n ¼ 76), ranging in age from 13 to 17 years (M ¼ 14.4 years, SD ¼ 1.1 years).Chi-square analyses indicated that there were no significant differences on any demographic variables(gender, race or socioeconomic status). Independent sample t-tests revealed no significant differences in IQ(Ms ¼ 106.5 and 105.3, respectively, for children and adolescents).

Assessment

Diagnostic interview

Children, adolescents and their parents were interviewed by a doctoral-level clinician using the AnxietyDisorders Interview Schedule-Child/Parent Version (ADIS-C/P; Silverman & Albano, 1996). The ADIS-C/Pis a semi-structured interview that allows for the assessment of Axis I disorders and uses an 8-point clinicianseverity rating (CSR) to assess the severity of any assigned diagnosis (with higher scores indicating greaterseverity). The participant and parent were interviewed separately by the same interviewer, who usedinformation from both informants to determine the final diagnosis and assign a severity rating. In those rareinstances when parent and participant provided conflicting information, the case was reviewed by the secondand third author who used not only the diagnostic interview information but also the additional assessmentdata to determine final diagnostic status.

To establish inter-rater reliability for assigned diagnoses and CSR, 20% of the interviews, chosen atrandom, were videotaped and subsequently rated by a second doctoral level clinician. The Kappa coefficientfor SAD diagnosis was k ¼ :78, although there was only one case of diagnostic disagreement. Chronbach’salpha for SAD severity rating was a ¼ .84.

In addition to establishing the presence of a diagnosis of SAD and its severity, the SAD section of theADIS-C/P was used to examine fear and anxiety for 20 different social situations. Fear ratings are assigned foreach situation based on an 8-point scale, with 1 indicating no distress and 8 indicating severe distress.Avoidance of these social situations was coded dichotomously. Fear and avoidance ratings were collectedfrom parent and child. For the purposes of this study, only participant ratings are presented.

Functional impairment

Rated by the same doctoral-level clinicians who completed the ADIS-C/P, overall functioning wasdetermined using the Children’s Global Assessment Scale (K-GAS; Shaffer et al., 1983). The K-GAS is a9-point scale assessing the child’s overall psychological, social, and school functioning, with 1 indicating severeimpairment, and 9 indicating good functioning in all areas. Twenty percent of the diagnostic interviews werevideotaped and rated by a second interviewer. Chronbach alpha inter-rater reliability coefficient was a ¼ :78.

Self-report inventories

Participants completed four self-report inventories: (a) the Loneliness Scale (LS; Asher & Wheeler, 1985),which assesses subjective feelings of loneliness and social isolation, (b) the Eysenck Personality Questionnaire-

Junior (EPQJ; Eysenck & Eysenck, 1975) which assesses extraversion and neuroticism, (c) the Social Phobia

and Anxiety Inventory for Children (SPAI-C; Beidel, Turner, & Morris, 1995), which assesses social anxiety ina broad range of social situations, and (d) the Multidimensional Anxiety Scale for Children (MASC; March &

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–11911184

Parker, 1999), which assesses four dimensions of anxiety (physical symptoms, harm avoidance, social anxiety,and overall anxiety).

Parent inventories

Parents completed the Child Behavior Checklist (CBCL; Achenbach, 1991); the Internalizing, Externalizingand Total Behavior Problems scores were included in this analysis.

Behavioral assessments

The behavioral assessment was conducted during a second clinic visit. Twenty-two children and 18adolescents did not return following the first visit and thus, did not participate in the behavioral assessment.There were no demographic differences between those who did or did not complete the behavioral assessment,nor were there significant differences on the ADIS-C/P CSR (p4.05). Inadequate or damaged videotaperesulted in observer scorable behavioral videotaped assessments for 48 children and 57 adolescents. Thebehavioral assessment consisted of two tasks: (a) role-play and (b) read-aloud. The role-play task consisted ofscenarios where the participant interacted with a same-aged peer trained to give friendly but neutral responses.Scene content included: (a) starting a conversation with an unfamiliar peer, (b) offering to help another peer,(c) giving a compliment, (d) receiving a compliment, and (e) responding assertively to a peer’s inappropriatebehavior. During the read aloud task, each participant read aloud for 10min (either Jack and the Beanstalk forchildren age 12 or younger or The Ransom of Red Chief for those 13 and older). The audience consisted of anadult experimenter and a same-aged peer. Order of task presentation was determined by random assignment.

The behavioral assessment was videotaped and coded by independent raters on the following dimensions:(a) speech latency (time to respond to a verbal prompt), (b) anxiety (observable signs of distress) and(c) effectiveness (quality of the social interaction/public performance). Speech latency was timed in seconds.Anxiety and effectiveness were rated using 4-point Likert-type scales. Anxiety ratings ranged from 1 (not at allanxious) to 4 (severely anxious) and social effectiveness ratings ranged from 1 (not effective at all) to 4 (veryeffective). In addition, participants rated their own level of distress during the two tasks on a 5-point ratingscale ranging from 1 (not at all nervous) to 5 (very nervous).

Twenty-five percent of the assessments, chosen at random, were coded by a second rater to establish inter-rater reliability. Chronbach alpha interrater coefficients for anxiety and effectiveness across the two behavioraltasks ranged between .84 and .96. Chronbach alpha interrater reliability for latency was a ¼ .90.

Results

Data were grouped conceptually according to the following dimensions of the clinical syndrome: socialanxiety and distress, other aspects of psychopathology, social skills and performance.

Social anxiety and distress

Age-related differences in degree of distress were examined for each of the 20 social situations listed in theADIS-C/P (see Table 1). Across all situations, adolescents’ fear ratings were significantly higher than thechildren’s (Hotellings t2 [F ¼ 39:74, df ¼ 19,131] po.001). When examined individually, there were 10 (outof 20) situations where there was a higher percentage of adolescents who endorsed at least moderate distress.These situations included ‘‘attending social activities,’’ ‘‘working/playing with a group,’’ ‘‘asking the teacher aquestion,’’ ‘‘participating in gym class,’’ ‘‘walking in the hallways,’’ ‘‘inviting a friend to get together,’’‘‘writing on the chalkboard,’’ ‘‘eating in front of others,’’ ‘‘dating,’’ and ‘‘answering or talking on the phone’’(see Table 1). More frequent concerns by adolescents for the item ‘‘dating’’ clearly reflects the age-specificnature of this task; the other situations are commonly encountered by all youth and suggest that a significantlylarger proportion of adolescents than children are severely affected.

In addition to endorsing moderate to severe distress, a substantial percentage of children and adolescentsavoid these social situations. Across all settings, adolescents had higher ratings of avoidance than did children(Hotelling’s t2(df ¼ 19,131) ¼ 18.16, po.001) and there were 8 individual situations (40% of the total), wherea significantly larger percentage of adolescents than children endorsed at least moderate avoidance including

ARTICLE IN PRESS

Table 1

Percentage of subjects endorsing moderate to severe levels of fear in social situations

Item Children (n ¼ 74) Adolescents (n ¼ 76) w2

Oral reports/reading aloud 82.4 88.2 .98

Taking tests 39.2 51.3 2.23

Attending parties/dances/social activities 58.1 81.6 9.84**

Working/playing with a group 44.6 61.8 4.48*

Asking teacher a question 70.3 86.8 6.14**

Answering questions in class 67.6 75.0 1.01

Participating in gym class 25.7 52.6 11.42***

Walking in hallways 13.5 43.4 16.40***

Initiating/joining conversations 82.4 90.8 2.27

Speaking to new/unfamiliar people 87.8 89.5 .10

Speaking to adults 81.1 82.9 .08

Inviting a friend to a get together 41.9 63.2 6.80**

Attending meetings 54.1 56.6 .10

Dating 8.1 47.4 28.67***

Eating in front of others 16.2 34.2 6.42**

Using public bathrooms 17.6 30.3 3.31

Writing in front of others on the chalkboard 50.0 67.1 4.52*

Answering/talking on telephone 33.8 52.6 5.42*

Musical or athletic performance 81.6 81.6 .01

Having a picture taken 27.0 40.8 3.17

*po.05; **po.01; ***po.001.

P.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–1191 1185

‘‘asking the teacher a question,’’ ‘‘attending social activities,’’ ‘‘inviting a friend to a get together,’’ ‘‘writing onthe chalkboard,’’ ‘‘working with a group,’’ ‘‘eating in front of others,’’ ‘‘walking in the hallways,’’ and‘‘dating’’ (see Table 2 for percentages). Again, ‘‘dating’’ is obviously more relevant to adolescent functioning;other situations occur with equal frequency and indicate more frequent avoidance by adolescents (see Table 2).

Overall, the ADIS-C/P CSR for SAD indicated moderate severity for both groups (5.91 for children versus6.15 for adolescents) and the scores were not significantly different. Additionally, the groups were notsignificantly different with respect to the percentage who were assigned the specific subtype (11% for childrenand 8% for adolescents) or on K-GAS ratings (5.67 for children and 5.57 for adolescents), which weremoderate for both groups.

The groups were not significantly different on the SPAI-C [tð139Þ ¼ �1:69, p4.05; see Table 3] or on theEPQJ Extraversion subscale [tð125Þ ¼ 1:72, p4.05, see Table 3]. In contrast, adolescents scored significantlyhigher on the MASC Social Anxiety subscale [tð139Þ ¼ �2:17, po.05], and on the LS (tð139Þ ¼ �3:21,pp:005).

Other measures of psychopathology

Eighty-two participants (54.6%) met criteria for a secondary Axis I diagnosis and the rate was higher (butnot significantly) for children (62.2%) than adolescents (47.4%; w2 ¼ 3:31; p4.05). The most commonsecondary diagnosis for both groups was generalized anxiety disorder (27.0% for children and 26.3% foradolescents). The groups differed significantly only on the diagnosis of Separation Anxiety Disorder, where asignificantly higher percentage of children (14.9%) than adolescents (3.9%) were assigned this diagnosis(w2 ¼ 5:28; po.02). Rates of comorbidity are depicted in Table 4.

With respect to dimensional aspects of psychopathology, the groups did not differ on the EPQJ Neuroticism

subscale [tð137Þ ¼ �:01, p4.05], the MASC Physical Symptoms [tð139Þ ¼ �:15, p4.05] or the MASC Total

Anxiety [tð139Þ ¼ :88, p4.05] subscales. However, children scored significantly higher than adolescents on theMASC Harm Avoidance [tð139Þ ¼ 2:33, po.025] and Separation Anxiety subscales [tð142Þ ¼ 4:37, po.001].There were no significant group differences on the CBCL Internalizing [tð125Þ ¼ 1:00, p4.05], Externalizing

[tð125Þ ¼ �:45, p4.05], or Total Behavior Problems [tð125Þ ¼ �:68, p4.05] subscales.

ARTICLE IN PRESS

Table 2

Percentage of subjects endorsing avoidance of social situations

Situation Children (n ¼ 74) Adolescents (n ¼ 76) w2

Speaking to unfamiliar people 71.6 73.7 .08

Initiating/joining conversations 66.2 75.0 1.40

Asking teacher a question 55.4 73.7 5.48**

Giving oral reports/ presentations 59.5 71.1 2.23

Speaking to adults 59.5 67.0 .94

Attending parties/dances or other social activities 35.1 65.8 14.10***

Answering questions in class 56.8 63.2 .64

Musical or athletic performance 48.6 55.3 .66

Inviting friend to a get together 32.4 53.9 7.07***

Writing on chalkboard in front of others 33.8 50.0 4.05*

Attending meetings 33.8 44.7 1.89

Working/playing with a group 27.0 42.1 3.76*

Dating 5.4 35.5 20.75***

Answering/talking on telephone 29.7 32.9 .18

Participating in gym class 18.9 31.6 3.17

Eating in front of others 6.8 28.9 12.51***

Walking in hallways 5.4 27.6 13.34***

Having a picture taken 21.6 22.4 .01

Using public bathrooms 12.2 21.1 2.13

Taking tests 13.5 19.7 1.05

*po.05; **po.02; ***po.001.

Table 3

Group means, standard deviations, and t-scores on the self-report measures

Children (n ¼ 74) Adolescents (n ¼ 76) t-scores

Mean (SD) Mean (SD)

Child measures

Loneliness scale 36.6 (11.2) 43.0 (12.6) �3.14**

EPQJ

Extraversion 12.5 (4.7) 11.2 (5.1) 1.50

Neuroticism 10.6 (5.1) 10.6 (5.4) �.01

MASC

Physical symptoms 12.1 (7.6) 12.3 (7.6) �.15

Harm avoidance 17.4 (5.1) 15.5 (4.6) 2.33*

Social anxiety 15.8 (7.2) 18.1 (6.7) �1.94*

Separation anxiety 11.7 (5.0) 8.2 (4.5) 4.37***

Total anxiety 57.0 (20.1) 54.2 (17.7) .88

SPAIC 23.8 (10.5) 26.5 (11.9) �1.42

Parent measures

Child behavior check list

Internalizing behavior 67.0 (10.8) 65.2 (9.9) 1.00

Externalizing behavior 46.9 (10.6) 47.7 (9.7) �.45

Total 58.6 (11.0) 57.3 (9.9) .68

*po.05; **po.01; ***po.001.

P.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–11911186

Behavioral assessment of social and performance skill

Children had significantly longer speech latencies than adolescents during the role play task (po.01).Furthermore, children were rated as significantly more anxious (po.05) as well as significantly less skilled

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–1191 1187

(po.01). Similarly, during the read-aloud task, children were rated as significantly more anxious (po.05) andsignificantly less skilled (po.05) than adolescents. There were no group differences on self-ratings of anxietyduring either the role-play or read-aloud tasks (p4.05). See Table 5 for means, standard deviations and t-scores.

Effects of gender

Because girls often report higher anxiety than boys (cf. Beidel & Turner, 2005), self-report data wereanalyzed for potential gender differences. There were gender effects only for the SPAI-C, where girls hadhigher scores than boys (Ms ¼ 28.11 and 22.15, respectively) [tð139Þ ¼ �3:23; p ¼ :002]. However, thisdifference was accounted for by the adolescent group, with girls scoring significantly higher than boys(Ms ¼ 30.73 and 22.50, respectively) [tð70Þ ¼ �3:10, p ¼ :003].

Discussion

Previous investigations (e.g., Beidel et al., 1999; Beidel et al., in press; Spence et al., 1999) clearly indicatethat both children and adolescents with SAD suffer significant emotional, academic and social impairment

Table 4

Percentage of subjects diagnosed with secondary diagnoses

Secondary diagnosis Children (n ¼ 74) Adolescents (n ¼ 76)

Any secondary disorder 62.0 47.4

Generalized anxiety disorder 27.0 26.3

Specific phobia 18.9 10.5

Attention-deficit/hyperactivity disorder 8.1 7.9

Depressive disordera 2.7 5.3

Obsessive-compulsive disorder 2.7 —

Oppositional defiant disorder — 2.6

Separation anxiety disorder 14.9 3.9*

Selective mutism 9.5 2.6

Reading disorder 1.4 —

Mixed receptive-expressive language disorder 1.4 —

*po.025.aFor Adolescents, Depressive Disorder combined three categories: (a) single episode, (b) recurrent, severe, and (c) dysthymic disorder.

In children, only dysthymic disorder was endorsed as a Depressive Disorder.

Table 5

Group means and standard deviations on the behavioral assessments

Children (n ¼ 48) Adolescents (n ¼ 57) t-values

Observer ratings

Role-play speech latencya 3.9 (2.5) 2.6 (2.3) 2.71**

Role-play anxietyb 2.8 (.85) 2.4 (.80) 2.05*

Role-play effectivenessc 1.8 (.68) 2.2 (.80) �2.73**

Read-aloud anxietyb 2.2 (.80) 1.9 (.64) 2.04*

Read-aloud effectivenessc 2.4 (.79) 2.7 (.70) �2.02*

Self report ratings

Role-play anxietyb 2.3 (1.0) 2.5 (1.1) �.74

Read-aloud anxietyb 2.2 (1.1) 2.4 (1.1) �.60

*po.05; **po.01.aMean scores are indicated in seconds.bLower scores indicate less anxiety.cLower scores indicate less skill.

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–11911188

when compared to those with no disorder. This investigation differs from previous studies, however,in that it is the first to compare and contrast the clinical presentation of SAD in children and adolescents.The results indicate that although its basic form appears consistent across groups, there are differencesas well.

With respect to similarities, broad measures of social anxiety (SPAI-C, CSR, extraversion) indicatemoderate and equivalent levels of social distress for both children and adolescents. In fact, over 90%of both groups met criteria for the generalized subtype, a percentage higher than what usually is foundamong adult populations (Turner, Beidel, & Townsley, 1992). Furthermore, based on their K-GAS ratings,both groups experienced equal and moderate impairment as a result of SAD. Thus, although more adolescentsmay endorse significant anxiety across a broader range of situations, the disorder’s impact is equal forboth groups.

With respect to group differences, adolescents endorsed a broader pattern of fear and avoidance whencompared to their younger counterparts. Adolescence is a period when engagement with peers andestablishment of friendships is emphasized and encouraged by parents and other adults. For example, inaddition to general social expectations regarding friendships and age-appropriate activities, dating andheterosocial interactions are also expected adolescent activities. When children are young, interactions withpeers are usually arranged by parents (play dates, dancing lessons, soccer team membership) and the choice toactively avoid these interactions is limited. However, with cognitive and physical maturity, parents are lesslikely to arrange these interactions as that responsibility gradually shifts toward the adolescent (Hartup, 1989;Hartup & Stevens, 1999) thereby allowing opportunities for avoidance. Thus, for adolescents with SAD,physical and cognitive maturity, as well as cultural expectations for independence in social encounters, mayresult in a broader pattern of social avoidance.

The consequences of increased social avoidance include fewer friendships and a stronger feeling of socialisolation. Indeed, the literature addressing the significance of friendships reveals that people with a satisfyingsocial network have a greater sense of well-being than those without ongoing social relationships (Hartup &Stevens, 1999). Thus, this broader pattern of social avoidance in socially phobic adolescents likely accounts fortheir higher scores on the LS.

Adolescents also reported significantly higher scores on the MASC Social Anxiety subscale than didchildren. Specifically, in contrast to the SPAI-C, where items assess anxiety across various social situations,the MASC Social Anxiety subscale (on its surface) appears to more directly tap the ‘‘cognitive’’ aspect of SAD(e.g., I worry about other people laughing at me, I worry about getting called on in class, I worry about whatother people think of me). Thus, the higher scores of adolescents on the Social Anxiety subscale of the MASCmay reflect a cognitive worry component, a dimension that is more likely to reflect the adolescent’s cognitivematurity. A large community study of adolescents in the Netherlands (Westenberg, Drewes, Goedhart,Siebelink, & Treffers, 2004) that found a direct link between cognitive maturity and an increase in socialevaluative fears during adolescence, provides further support for this hypothesis.

In contrast, children with SAD had a broader pattern of general psychopathology than adolescents. Theywere more likely to worry about doing things correctly and about being separated from their parents. Thesedata are consistent with the general understanding of fears and anxiety in children, where prevalence rates forfears and phobias are higher among younger children than adolescents (Beidel & Turner, 2005) with theexception of SAD. With respect to comorbid disorders, the types of co-existing conditions found among thissample are consistent with those of Beidel et al. (1999) whereas the pattern among adolescents is quite differentfrom those among an epidemiological sample of German adolescents, where 41% had a comorbidsomatoform disorder, 29% had a depressive disorder, and 24% had a substance abuse disorder (Essau et al.,1999). However, this latter study differs from the current investigation on several relevant factors including thediagnostic instrument and assessment strategy and the fact that in the epidemiological sample, there was noattempt to determine whether SAD was the primary diagnosis as was the case for the current investigation.The high rate of comorbid somatoform disorder reported by Essau et al. (1999) is inconsistent with any otherinvestigations of the psychopathology of SAD in other reported populations.

Furthermore, although both children and adolescents with SAD have significantly poorer social skills thanyouth with no disorder (Beidel et al., 1999; Beidel et al., in press; Spence et al., 1999), when compared directly,children with SAD displayed significantly less skill and more anxiety than their adolescent counterparts. The

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–1191 1189

behavioral assessment used in this investigation was a structured task, requiring responses to one sentenceprompts provided by a same age peer. Although its discriminative validity for both age groups has beenpreviously demonstrated, it may be that despite their higher level of fear and avoidance, adolescents with SADhave sufficient skills to respond minimally to these very brief and structured interactions. In fact, unstructuredsocial interactions are described as much more challenging for adolescents with SAD (Beidel & Turner, 2005)and future investigations may need to consider the specific conversational format most likely to reveal theextent of skill deficits in various age groups.

The outcome of this investigation has significant implications for the conceptualization of treatment forSAD in youth. First, both groups demonstrated social skills deficits (although children were more impaired)suggesting that both cohorts will require social skills training. Second, both groups endorsed equivalent levelsof social distress indicating that exposure will be necessary for both groups. However, because the pattern ofdistress and avoidance was more pervasive for adolescents, clinicians will need to design clinic treatmentsessions, generalization sessions, and homework assignments directed at a broader range of social situations.Furthermore, the data from this investigation, in conjunction with earlier outcome data (Alfano, Beidel, &Turner, 2002; Alfano et al., 2006; Spence et al., 2000; Westenberg et al., 2004) suggest that cognitiveinterventions might be appropriate for a subset of adolescents. Thus, consistent with their clinicalpresentation, treatment components also might vary based on age.

This study is not without its limitations. First, all of the participants in this investigation were seekingtreatment for SAD and thus, may not represent the ‘‘typical’’ child or adolescent with this disorder.However, they are representative of treatment seeking samples and represent the largest sample to date ofcarefully diagnosed children and adolescents with SAD. Although it is possible that community samplesmay differ from clinical samples on various dimensions of psychopathology, such investigations are yetto be conducted and thus, it is difficult to speculate whether or how the results from this sample may differfrom those with the disorder who do not seek treatment. Second, although the overall sample size wasquite large, the number of children representing racial/ethnic minorities did not allow for comparisons basedon this variable. Although previous investigations indicated a lack of differences in the clinical presentation ofSAD across Caucasian and African-American children (Ferrell, Beidel, & Turner, 2004), further studiesincluding a broader range of ethnicity are necessary. Third, although this study represents the largestinvestigation to date examining the psychopathology of youth with primary SAD, the sample size dictated aneed for a categorical approach. Future investigations using much larger samples across the same age rangemay be able to identify more precise ages (e.g., 9–10 versus 13–14) at which symptom presentation changessignificantly. Although assignment of 12-year olds to the adolescent group did not change the outcome in apreliminary analysis of these data, future research using even larger sample sizes (perhaps a communitysample) may allow for a more extensive developmental approach consistent with the developmentalliterature in normal fear (Gullone, 2000; Gullone & King, 1997; Gullone, King, & Ollendick, 2001; Ollendick& Hirshfeld-Becker, 2002).

In summary, the results of this investigation indicate that the symptom presentation of children andadolescents with primary SAD share core features but differ in the extent of their social dysfunction, theirdegree of functional impairment, their pattern of comorbidity and their degree of social skills. Futureinvestigations may include an adult sample, thus providing even further information regarding the clinicalpresentation of SAD across various stages of development.

Acknowledgment

This manuscript was supported by NIMH grant 53703 to the second, third and sixth author.

References

Achenbach, T. (1991). Manual for the child behavioral checklists/4– 18 and 1991 profile. Burlington: University of Vermont.

Alfano, C. A., Beidel, D. C., & Turner, S. M. (2002). Considering cognition in childhood anxiety disorders: Conceptual, methodological

and developmental considerations. Clinical Psychology Review, 22, 1209–1238.

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–11911190

Alfano, C. A., Beidel, D. C., & Turner, S. M. (2006). Cognitive correlates of social phobia among children and adolescents. Journal of

Abnormal Child Psychology, 34, 182–194.

Alnaes, R., & Torgersen, S. (1999). A 6-year follow-up study of anxiety disorders in psychiatric outpatients: Development and continuity

with personality disorders and personality trait as predictors. Journal of Psychiatry, 53, 409–416.

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed). Washington, DC: Author.

Arnold, P., Banerjee, P., Bhandari, R., Lorch, E., Ivey, J., Rose, M., et al. (2003). Childhood anxiety disorders and developmental issues in

anxiety. Current Psychiatry Reports, 5, 252–265.

Asher, S., & Wheeler, V. (1985). Children’s loneliness: A comparison of rejected and neglected peer status. Journal of Consulting and

Clinical Psychology, 53, 500–505.

Beidel, D. C. (1991). Social phobia and overanxious disorder in school-age children. Journal of the American Academy of Child and

Adolescent Psychiatry, 30, 545–552.

Beidel, D. C., & Turner, S. M. (1998). Shy children, phobic adults: Nature and treatment of social phobia. American Psychological

Association: Washington, DC.

Beidel, D. C., & Turner, S. M. (2005). Childhood anxiety disorders. New York: Routledge Taylor & Francis Group.

Beidel, D. C., Turner, S. M., & Dancu, C. V. (1985). Physiological, cognitive, and behavioral aspects of social anxiety. Behaviour Research

and Therapy, 23, 109–117.

Beidel, D. C., Turner, S. M., & Morris, T. L. (1995). A new inventory to assess childhood social anxiety and phobia: The social phobia and

anxiety inventory for children. Psychological Assessment, 7, 73–79.

Beidel, D. C., Turner, S. M., & Morris, T. L. (1999). Psychopathology of childhood social phobia. Journal of the American Academy of

Child and Adolescent Psychiatry, 38, 643–650.

Beidel, D. C., Turner, S. M., Young, B., Ammerman, R. T., Sallee, F. R., & Crosby, L. E. (in press). Psychopathology of adolescent social

phobia. Journal of Psychopathology and Behavioral Assessment, in press.

Davidson, J. R. T., Hughes, D. L., & George, L. K. (1993). The epidemiology of social phobia: Findings from the Duke epidemiological

catchment Area study. Psychological Medicine, 23, 709–718.

Essau, C. A., Conradt, J., & Petermann, F. (1999). Frequency and comorbidity of social phobia fears in adolescents. Behaviour Research

and Therapy, 37, 831–843.

Eysenck, H., & Eysenck, S. (1975). Eysenck personality questionnaire (junior & adult) manual. San Diego, CA: Educational and Industrial

Testing Service.

Ferrell, C., Beidel, D. C., & Turner, S. M. (2004). Assessment and treatment outcome of African-American and Caucasian socially phobic

children: A cross cultural comparison. Journal of Clinical Child Psychology, 33, 260–268.

Gullone, E. (2000). The development of normal fear: A century of research. Clinical Psychology Review, 20, 429–451.

Gullone, E., & King, N. J. (1997). Three-year follow-up of normal fear in children and adolescents aged 7 to 18 years. Journal of

Developmental Psychology, 15, 97–111.

Gullone, E., King, N. J., & Ollendick, T. H. (2001). Self-reported anxiety in children and adolescents: A three-year follow-up study.

Journal of Genetic Psychology, 162, 5–19.

Hartup, W. W. (1989). Social relationships and their developmental significance. American Psychologist, 44, 120–126.

Hartup, W. W., & Stevens, N. (1999). Friendships and adaptation across the life span. Current Directions in Psychological Science, 8,

76–79.

Heimberg, R. G., Holt, C. S., Schneier, F. R., Spitzer, R. L., & Liebowitz, M. R. (1993). The issue of subtypes in the diagnosis of social

phobia. Journal of Anxiety Disorders, 7, 249–269.

Hollingshead, A. B. (1957). Two factor index of social position. New Haven: Author.

Kashdan, T. B., & Herbert, J. D. (2001). Social anxiety disorder in childhood and adolescence: Current status and future directions.

Clincial Child and Family Psychology Review, 4, 37–61.

Keller, M. B. (2003). The lifelong course of social anxiety disorder: A clinical perspective. Acta Psychiatrica Scandinavia, 108(Suppl. 417),

85–94.

March, J. S., & Parker, J. (1999). The multidimensional anxiety scale for children (MASC). In M. E. Maruish (Ed.), Use of psychological

testing for treatment planning and outcomes assessment (2nd ed., pp. 299–322). Mahwah, NJ: Lawrence Erlbaum Associates.

Ollendick, T. H., & Hirshfeld-Becker (2002). The developmental psychopathology of social anxiety disorder. Biological Psychiatry, 51,

44–58.

Shaffer, D., Gould, M. S., Brasic, J., Abrosini, P., Fisher, P., Bird, H., et al. (1983). A children’s global assessment scale (K-GAS).

Archives of General Psychiatry, 40, 1228–1231.

Silverman, W. K., & Albano, A. M. (1996). The anxiety disorders interview schedule for DSM-IV—Child and parent versions. San Antonio,

TX: Psychological Corporation.

Spence, S. H., Donovan, C., & Brechman-Toussaint, M. (1999). Social skills, social outcomes, and cognitive features of childhood social

phobia. Journal of Abnormal Psychology, 108, 211–221.

Spence, S. H., Donovan, C., & Brechman-Toussaint, M. (2000). The treatment of childhood social phobia: The effectiveness of social skills

training-based, cognitive-behavioural intervention, with and without parental involvement. Journal of Child Psychology & Psychiatry,

41, 713–726.

Strauss, C. C., Lease, C. A., Last, C. G., & Francis, G. (1988). Overanxious disorder: An examination of developmental disorders. Journal

of Abnormal Child Psychology, 16, 433–443.

Tracey, S. A., Chorpita, B. F., & Douban, J. (1997). Empirical evaluation of DSM-IV generalized anxiety disorder criteria in children and

adolescents. Journal of Clinical Child Psychology, 26, 65–73.

ARTICLE IN PRESSP.A. Rao et al. / Behaviour Research and Therapy 45 (2007) 1181–1191 1191

Turner, S. M., Beidel, D. C., Dancu, C. V., & Keyes, D. J. (1986). Psychopathology of social phobia and comparison to avoidant

personality disorder. Journal of Abnormal Psychology, 95, 389–394.

Turner, S. M., Beidel, D. C., & Townsley, R. M. (1992). Social phobia: A comparison of specific and generalized subtypes and avoidant

personality disorder. Journal of Abnormal Psychology, 101, 326–331.

Wechsler, D. (1991). Wechsler intelligence scale for children (3rd ed). San Antonio, TX: The Psychological Corporation.

Westenberg, P. M., Drewes, M. J., Goedhart, A. W., Siebelink, B. M., & Treffers, P. D. A. (2004). A developmental analysis of self-

reported fears in late childhood through mid-adolescence: Social-evaluative fears on the rise? Journal of Child Psychology and

Psychiatry, 45, 481–495.