Journal of Anxiety Disorders 22 (2008) 612621

Reliability and validity of the screen for child anxiety related emotional disorders (SCARED) in Chinese children
Su Linyan*, Wang Kai, Fan Fang, Su Yi, Gao Xueping
Mental Health Institute of the Second Xiangya Hospital, Central South University, Changsha 410011, China Received 13 November 2006; received in revised form 30 May 2007; accepted 31 May 2007

Abstract The present study examined the psychometric properties of the screen for child anxiety related emotional disorders (SCARED) in a large community sample of Chinese children. The 41-item version of the SCARED was administered to 1559 primary and junior high school students (774 boys and 785 girls, mean age 11.8 2.11) in 12 Chinese cities. The SCARED demonstrated moderate to high internal consistency (alpha = 0.430.89) and testretest reliability (intraclass correlation coefcients = 0.460.77 over 2 weeks and 0.240.67 over 12 weeks), moderate parentchild correlation (r = 0.490.59) and good discriminant validity (between anxiety and non-anxiety disorders). The SCARED total score was signicantly correlated with the internalizing factor of the child behavior checklist (0.41). Factor analyses revealed the same ve-factor structure as the original SCARED. These ndings support that the SCARED is a reliable and valid anxiety screening instrument in Chinese children. # 2007 Elsevier Ltd. All rights reserved.
Keywords: Screen for child anxiety related emotional disorders; Anxiety disorder; Psychometrics; Children; China

1. Introduction Anxiety disorders represent one of the most prevalent childhood psychiatric disorders. In epidemiological studies world wide, approximately 520% children and adolescents have been estimated to have an anxiety disorder (Costello & Angold, 1995; Essau, 2000; Newman et al., 1996). The high comorbidity rate of these disorders has also been documented (Essau, 2000; Ginsburg & Silverman, 1996; Kendall, Brady, & Verduin, 2001), with rates for comorbid depression and anxiety ranging from 15.9 to 61.9% (Brady & Kendall, 1992). Childhood anxiety disorders are associated with school
* Corresponding author. Tel.: +86 731 5531781; fax: +86 731 5531781. E-mail address: su-linyan@hotmail.com (L.Y. Su). 0887-6185/$ see front matter # 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.janxdis.2007.05.011

problems, psychosocial difculties, increased risk for substance abuse and other externalizing disorders (Klein, 1994; Strauss, Lahey, Frick, Frame, & Hynd, 1988), and they tend to be stable over time (Beidel, Morris, & Turner, 2004; Bernstein, Hektner, Borchardt, & McMillan, 2001; Cohen, Cohen, & Brook, 1993; Keller et al., 1992). Taken together, all these ndings merit a serious attention to childhood anxiety disorders and indicate the need for reliable and valid screening tools. There are reliable structured diagnostic interviews to diagnose anxiety symptoms in children, but they are difcult to conduct (Silverman & Ollendick, 2005), time consuming and require well-trained interviewers (Klein, 1994; Pavuluri & Birmaher, 2004). Alternatively, clinician-based rating scales can also be used to measure anxiety symptoms, yet several studies have questioned their discriminant validity (Hoehn-Saric,

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621

613

Maisami, & Weigand, 1987; Mattison & Babnato, 1987; Perrin & Last, 1992; Seligman, Ollendick, Langley, & Baldacci, 2004; Strauss et al., 1988) and their efcacy of differentiating individual anxiety disorders (Brady & Kendall, 1992; Klein, 1994). By contrast, self-report scales are an efcient, less expensive and easy way to screen for childhood anxiety disorders (Essau, Muris, & Ederer, 2002; Pavuluri & Birmaher, 2004). The most widely used self-report scales to screen for childhood anxiety disorders include instruments such as the revised childrens manifest anxiety scale (Reynolds & Richmond, 1978), the revised fear survey schedule for children (Ollendick, 1983), the state-trait anxiety inventory for children (Spielberger, 1973) and the social anxiety scale for children-revised (La Greca & Stone, 1993). These questionnaires possess adequate psychometric properties, such as moderate to high internal consistency, moderate testretest reliability and positive concurrent validity and divergent validity (Silverman & Ollendick, 2005), but they do not always ascertain DSM-IV (American Psychiatric Association, 1994) related anxiety symptoms. The screen for child anxiety related emotional disorders (SCARED) was developed by Birmaher et al. (1997) for the screening of DSM-IV childhood anxiety disorders. Exploratory factor analysis of the SCARED revealed ve factors that parallel the DSM-IV classication of anxiety disorders: panic/somatic, generalized anxiety, separation anxiety, social phobia, and school phobia. In a replication study, Birmaher et al. (1999) added three items reecting social phobia to the original version, resulting in a 41-item self-report questionnaire. Since the initial publication of the SCARED, the psychometric properties of this screening instrument have been examined in clinical, community, and primary care samples. The SCARED was found reliable in terms of internal consistency, testretest reliability (Birmaher et al., 1997, 1999; Boyd, Ginsburg, Lambert, Cooley, & Campbell, 2003; Essau et al., 2002; Hale, Raaijmakers, Muris, & Meeus, 2005; Muris, Merckelbach, Schmidt, & Mayer, 1999), and parentchild agreement (Birmaher et al., 1997; Muris et al., 1999; Wren, Bridge, & Birmaher, 2004). The convergent validity of the SCARED was supported because of its signicant correlations with other measures for childhood anxiety disorders (Muris et al., 2002; Muris, Merckelbach, Ollendick, King, & Bogie, 2002), such as the Revised Childrens Manifest Anxiety Scale (Boyd et al., 2003; Muris et al., 1998) and the Spence Childrens Anxiety Scale (Essau et al., 2002). Evidence has also accumulated for the good discriminant validity of the SCARED both between anxiety and other

psychiatric disorders and within anxiety disorders (Birmaher et al., 1997, 1999). In general, these studies support the usefulness of the SCARED as a reliable and valid screening tool for childhood anxiety disorders. Previous studies on the psychometric properties of the SCARED have been primarily conducted with Caucasian samples in Western countries (Birmaher et al., 1997, 1999; Essau et al., 2002; Muris et al., 1998, 1999; Wren et al., 2004). This issue poses a signicant limitation in prior studies because the factor structure obtained from Caucasian populations may not be appropriate for other ethnic/racial groups (Carter, Miller, Sbrocco, Suchday, & Lewis, 1999). For example, in a South African community study (Muris, Schmidt, Engelbrecht, & Perold, 2002) and a newly published research on multiethnic primary care population (Wren et al., 2007), the SCARED was found to yield a four-factor structure rather than the ve-factor structure reported in primarily Caucasian samples. Also, in an African-American high school sample, only three factors emerged, including somatic/ panic, social anxiety and generalized anxiety (Boyd et al., 2003). Although in a recent study, the ve-factor SCARED also applied to ethnic minority adolescents not of Dutch descent, further studies on different ethnic groups are warranted (Hale et al., 2005). To our knowledge the SCARED has not been validated with any racial/ethnic group in Asia, such as the Han Chinese. The purpose of the current study was to address this gap in the literature by examining the reliability and the validity of the 41-item SCARED in a large community sample of school children in China. 2. Method 2.1. Subjects and procedures The participants consisted of 1680 Chinese students aged 816 years enrolled in grades 39. These children were recruited from 12 cities which represent different socioeconomic levels and regional backgrounds in China, including Beijing, Tianjin, Changsha, Hangzhou, Jinan, Taiyuan, Dalian, Siping, Liuzhou, Shenzhen, Mianyang and Yichun. From each city, one junior high school and one primary school were selected. From each of these grades, 10 boys and 10 girls were randomly chosen to participate in the study. Of the initial 1680 students, 121 (7.2%) were excluded because of incomplete data. The nal sample consisted of 1559 subjects, including 785 (50.4%) females with a mean age of 11.8 2.1 years old. There were 972 (62%) children (10.4 1.3 years old, range:

614

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621

812 years old) and 587 (38%) adolescents (14.0 1.0 years old, range: 1316 years old). All participants were Han Chinese. There were no signicant differences in the demographics between the children who were included in the study and those who were excluded. After appropriate permissions from the School Boards and after participants consents were obtained, the students completed the SCARED during regular classes hours. Children were allowed to clarify the meaning of some questions, but their responses to the anxiety items were not inuenced by the research assistants. The investigation was carried out in accordance with the latest version of the Declaration of Helsinki and the study design was reviewed by an ethical committee. The investigation was carried out simultaneously in each city and all the questionnaires were returned to the researchers in Changsha for further analysis. To examine the testretest reliability of the SCARED, from the general population, a subgroup of regular school children in Changsha city (n = 33) were retested over a period of 2 weeks and 12 weeks after the initial screening. In order to further examine the SCARED convergent validity, a subgroup of parents, including 139 parents of regular school students and 130 parents of the clinical cases from Changsha city completed the Children Behavior Checklist (CBCL) (Achenbach, 1991). To test parentchild correlation, of the total sample of parents requested to ll out the SCREAD parent version, 1368 parents completed the entire form. The 1368 parents represented one parent per child, including 696 (50.9%) fathers and 672 (49.1%) mothers. We found no signicant rating difference between different genders. To evaluate the discriminant validity of the SCARED, in Changsha, 130 children and adolescents with anxiety (n = 48), pure depressive disorders (n = 30), or attention decit hyperactivity (n = 52) disorders were studied (46 females, 12.0 2.5 years old, range: 816 years old). These children were interviewed with a DSM-IV based interview schedule developed by the researchers. The diagnostic interview was administered by trained clinicians supervised by the rst author and the psychiatric diagnoses were ascertained by this child psychiatrist using the DSM-IV diagnostic criteria. Forty-eight patients had anxiety disorders (24 GAD, 5 SAD, 16 social phobia, 7 school phobia, 10 specic phobia and 3 panic disorder), 15 (31.25%) had two or more anxiety disorders. Thirty had depressive disorders without comorbid anxiety (22 major depressive disorder, 6 dysthymia, 2 depressive disorder not otherwise

specied); 52 had disruptive disorders without comorbid anxiety (47 attention decit/hyperactivity disorder, 23 oppositional deant disorder and 4 conduct disorder). 2.2. Measures The SCARED (Birmaher et al., 1999) is a 41-item self-report measure designed to screen for DSM-IV anxiety disorders. The SCARED includes 5 factors: somatic/panic (13 items; e.g., When I feel frightened, it is hard to breathe), generalized anxiety (9 items; e.g., I worry about other people liking me), separation anxiety (8 items; e.g., I get scared if I sleep away from home), social phobia (7 items; e.g., I dont like to be with people I dont know well), and school phobia (4 items; e.g., I get headaches when I am at school). The participants rated the items of each factor on a 3-point scale (0 = not true or hardly ever true, 1 = sometimes true, and 2 = true or often true). The SCARED total score, derived by adding the responses of the 41 items, ranges from 0 to 82. The CBCL (Achenbach, 1991) is a 118-item parent measure to assess child behavior and emotional problems. Parents are asked to evaluate whether the behavior is not true for their child (0), somewhat or sometimes true (1), or very true or often true (2), now or during the past six months. As a widely used screening tool, the CBCL was validated for Chinese children by Su, Li, Yang, Luo, & Wan (1997). Parallel to Monga et al. (2000) study, the CBCL was used in this study to examine convergent validity. The English versions of the SCARED and the CBCL were adapted and translated using the back-translation method. The two scales were translated into Chinese by the rst author and then blindly back-translated to English by a professor from the English Training Center of Ministry of Health of the Peoples Republic of China. It was assured that the content of the translated versions were similar to the original English versions by both translators. 2.3. Data analysis Cronbach a coefcients were calculated to evaluate the internal consistency of the SCARED total and subscale scores. Gender and age differences were evaluated using analyses of variance (ANOVA) and effect size statistics. Pearson product moment correlations and intraclass correlation coefcients (ICCs) were used to assess testretest reliability. Parentchild correlations were tested through Pearson product moment correlations. Analyses of variance (ANOVA)

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621 16.15*** 6.75** 9.04**

615 0.012 0.04 Note: CH: Child; AD: Adolescent. ** P < 0.01. *** P < 0.001. 0.79 1.02 0.98 1.34 0.16 9.23** 0.86 1.20 0.85 1.06 0.01 3.23 3.80 3.12 0.08

Cohs d F

14.35 9.65 16.07 11.10 0.17

0.15

0.33

3.64 3.29

4.33 3.49

3. Results 3.1. Age and sex differences

Table 1 Comparison of the total anxiety score and the subscale scores between age groups/gender groups (x s) AD (n = 587)

3.17 2.90

3.03 2.74 3.53 2.73 0.26

3.2.1. Internal consistency The internal consistency of the SCARED total score and subscales was measured by calculating Cronbach a coefcients. The coefcient a values were 0.89 for the total score, 0.70 for somatic/panic, 0.77 for generalized anxiety, 0.67 for separation anxiety, 0.76 for social phobia, and 0.43 for school phobia. For male group, the coefcient a values were 0.90 for the total score and 0.570.85 for the subscales; for female group, the coefcient a values were 0.89 for the total score and 0.440.82 for the subscales. For adolescent group, the coefcient a values were 0.91 for the total score and 0.660.85 for the subscales; for child group, the coefcient a values were 0.89 for the total score and 0.430.79 for the subscales.

Child group

Boy (n = 474)

Girl (n = 498)

Cohs d Boy (n = 300)

3.2. Reliability

Total 13.79 9.76 14.89 9.53 0.11 15.25 11.46 16.93 10.66 score Somatic/ 3.16 3.01 3.17 2.81 0.003 3.58 3.52 3.70 3.033 panic General 3.18 3.04 3.38 2.89 0.068 4.13 3.61 4.53 3.36 anxiety Separation 3.32 2.71 3.73 2.73 0.19 2.67 2.55 3.40 2.87 anxiety Social 3.35 2.65 3.80 2.86 0.16 3.87 3.17 4.34 3.26 phobia School 0.78 1.01 0.81 1.03 0.03 1.00 1.44 0.95 1.24 phobia

Adolescent group

As depicted in Table 1, the adolescent group (aged 1316 years) had signicantly higher scores than the children group (aged 812 years) on the total anxiety score, and with exception of separation anxiety, higher scores on all the SCARED factors (all p-values <0.01). As expected, children scored signicantly higher than adolescents in the separation anxiety factor ( p < 0.01). As for different genders, females reported signicantly higher levels of anxiety on the SCARED total score, separation anxiety, and social phobia subscales than males (all p-values <0.01).

Total by age

Cohs d CH (n = 972)

0.017

0.115

0.15

3.28 2.96

Girl (n = 287)

0.15

3.58 2.76

4.10 3.22

0.17

and effect size statistics were also utilized to test the SCARED discriminant validity. To assess the optimal cutoff score that optimally discriminated between clinically referred children with an anxiety disorder and community children, the receiver operation curve (ROC) method was used (Somoza & Mossman, 1991). Pearson product moment correlations were used to test convergent validity. Replicating the factor analysis methodology of Birmaher et al. (1997, 1999), a principle components factor analysis with varimax rotation was conducted. For conrmatory factor analysis, the model t was evaluated using structural equation model with the following indicators: Chisquare statistic divided by the degree of freedom (x2/ d.f.), the normed t index (NFI), the non-normed t index (NNFI), the comparative t index (CFI), the goodness of t index (GFI) and the root mean square error of approximation (RMSEA).

Cohs d F

3.37 2.90 0.01

0.15

3.61 2.79 0.20 3.07 2.67 11.87 ** 0.18

Total by gender

14.36 10.47 15.62 9.99 0.12

Girl (n = 785)

3.33 3.22

3.55 3.30

Boy (n = 774)

40.38 ***

10.75 **

8.77**

11.93 **

3.55 2.87

4.00 3.02 0.15

616

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621

Table 2 Comparison of the testretest reliability of total anxiety score and the subscale scores in test 1, test 2 and test 3 (n = 33) Test 1 (x s) Total score Somatic/panic General anxiety Separation anxiety Social phobia School phobia 16.06 10.09 3.53 3.61 3.27 2.48 4.18 3.34 4.36 2.88 0.79 0.93 Test 2 (x s) 12.45 10.23 2.58 2.93 2.82 3.42 2.88 3.14 3.58 2.93 0.61 0.83 Test 3 (x s) 11.55 9.92 2.64 3.28 2.39 3.13 2.88 3.24 3.06 2.86 0.57 0.83 r12 0.61 0.82 0.52 0.51 0.55 0.66 r13 0.57 0.69 0.50 0.55 0.29 0.53 ICC12 0.57 0.77 0.50 0.46 0.53 0.65 ICC13 0.50 0.67 0.46 0.51 0.24 0.52

Note: Test 1: the rst test; Test 2: the re-test after 2 weeks; Test 3: the re-test after 3 months.

3.2.2. Testretest reliability As shown in Table 2, the 2-week and 12-week Pearson correlation coefcients (r) were 0.61 and 0.57 for the total score, 0.82 and 0.69 for somatic/panic, 0.52 and 0.50 for generalized anxiety, 0.51 and 0.55 for separation anxiety, 0.55 and 0.29 for social phobia, and 0.66 and 0.53 for school phobia, respectively. Intraclass correlation coefcients (ICCs) were 0.57 and 0.50 for the total score, 0.77 and 0.67 for somatic/panic, 0.50 and 0.46 for generalized anxiety, 0.46 and 0.51 for separation anxiety, 0.53 and 0.24 for social phobia, and 0.65 and 0.52 for school phobia, respectively. Except for social phobia, the testretest reliability of the total scale and the other subscales was acceptable. 3.2.3. Parentchild correlations The parent and child total and each factor scores were signicantly correlated with rs between 0.49 for somatic/ panic factor and 0.59 for the total score (all p-values <0.01). Children reported more anxiety symptoms than parents on the SCARED total score (children: 15.0 10.25, parents: 11.5 8.74; p < 0.0001) and all the ve factors (children: 3.35 3.06, parents: 2.09 2.38 for somatic/panic; children: 3.68 3.21, parents: 2.56 2.65 for generalized anxiety; children: 3.34 2.74, parents: 2.96 2.52 for separation anxiety; children: 3.77 2.95, parents: 3.41 2.92 for social

phobia; children: 0.80 1.15, parents: 0.52 0.85 for school phobia) (all p-values <0.01). The parents scores were signicantly higher correlated with the scores of the adolescents than with the children scores (total score: r = 0.67 versus 0.52; somatic/panic: 0.57 versus 0.42; generalized anxiety: 0.56 versus 0.46; separation anxiety: 0.64 versus 0.51; social phobia: 0.63 versus 0.51; school phobia: 0.54 versus 0.47, respectively (all p-values 0.04). Girls showed signicantly higher parentchild correlations than boys on the social phobia factor (r = 0.62 versus 0.50, p = 0.0006). 3.3. Validity 3.3.1. Discriminant validity As depicted in Table 3, the total score and the scores for each of the ve subscales were signicantly higher in children with anxiety disorders (n = 48) than nonreferred school children (n = 1559) and children with ADHD (n = 52) (all p-values <0.001). The total anxiety and the panic/somatic, separation anxiety and social phobia subscales signicantly discriminated anxious children from depressed children (n = 30) (all p-values <0.001). Except for the difference in school phobia between anxious and depressed children (d = 0.19), the differences in scores between the anxiety group and

Table 3 Comparison of children with anxiety disorders, the school sample, and children with other psychiatric disorders (means S.D.) Anxiety (n = 48) Total score Somatic/panic General anxiety Separation anxiety Social phobia School phobia 30.27 9.94a,b,c 7.42 3.99a,b,c 8.46 3.72a,c 5.90 2.87a,b,c 5.98 3.14a,b,c 2.31 2.07a,c School sample (n = 1559) 15.00 10.25 3.35 3.06 3.68 3.21 3.34 2.74 3.77 2.95 0.86 1.13 Pure depression (n = 30) 24.67 10.43 6.10 4.40 7.53 3.29 3.73 2.26 4.60 3.37 2.70 2.10 ADHD (n = 52) 21.00 11.99 5.35 4.57 5.06 2.89 4.58 2.96 4.58 2.96 1.44 1.75 F 46.58 37.84 49.54 16.47 10.21 46.99 Cohs da 1.5 1.14 1.38 0.91 0.72 0.87 Cohs d b 0.55 0.31 0.26 0.84 0.42 0.19 Cohs dc 0.84 0.48 1.02 0.45 0.46 0.45

Note: p < 0.001 for all. ADHD: Attention decit hyperactivity disorder. a Comparison with school sample. b Comparison with pure depression. c Comparison with ADHD.

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621 Table 4 Factor analyses of the SCARED in Chinese children (N = 1559) Item 33. I worry about what is going to happen in the future 37. I worry about things that have already happened 31. I worry that something bad might happen to my parents 20. I have nightmares about something bad happening to me 30. I am afraid of having anxiety (or panic) attacks 14. I worry about being as good as other kids 16. I have nightmares about something bad happening to my parents 23. I am a worrier 28. People tell me that I worry too much 7. I am nervous 6. When I get frightened, I feel like passing out 24. I get really frightened for no reason at all 15. When I get frightened, I feel liking things are not real 5. I worry about other people liking me 21. I worry about things working out for me 12. When I get frightened, I feel like I am going crazy 32. I feel shy with people I dont know well 10. I feel nervous with people I dont know well 26. It is hard for me to talk with people I dont know well 3. I dont like to be with people I dont know well 41. I am shy 40. I feel nervous about going to parties, dances, or any place where there will be people that I dont know well 39. I feel nervous when I am with other children or adults and I have to do something while they watch me (for example: read aloud, speak, play a game, play a sport) 35. I worry about how well I do things 18. When I get frightened, my heart beats fast 22. When I get frightened, I sweat a lot 9. People tell me I look nervous 27. When I get frightened, I feel like I am choking 1. When I feel frightened, it is hard to breathe 34. When I get frightened, I feel like throwing up 38. When I get frightened, I feel dizzy 13. I worry about sleeping alone 25. I am afraid to be alone in the house 8. I follow my mother or father wherever they go 29. I dont like to be away from my family 4. I get scared if I sleep away from home 36. I am scared to go to school 17. I worry about going to school 2. I get headaches when I am at school 19. I get shaky 11. I get stomachaches at school Factor 1 0.629 0.625 0.556 0.527 0.503 0.479 0.473 0.464 0.457 0.446 0.441 0.437 0.437 0.432 0.408 0.333 0.197 8.019E02 8.546E02 2.680E02 0.262 0.210 0.178 Factor 2 0.244 0.171 0.135 6.101E02 0.263 0.393 2.854E02 0.236 0.197 0.173 6.585E02 2.467E02 0.153 0.248 0.227 1.071E02 0.675 0.591 0.588 0.575 0.575 0.562 0.527 Factor 3 5.200E02 6.844E02 2.574E03 0.174 0.251 7.722E03 0.120 4.340E02 0.206 0.186 0.348 0.269 0.334 1.257E02 0.140 0.330 8.840E02 0.147 0.182 7.781E03 6.075E02 0.104 0.129 Factor 4 7.072E02 1.196E02 0.268 0.172 0.153 7.476E02 0.159 5.255E02 3.403E02 8.199E02 2.328E02 0.172 7.058E02 3.153E02 7.473E02 3.917E02 3.371E02 0.188 3.124E02 9.867E02 3.372E02 6.633E02 0.107 Factor 5 0.187 0.113

617

3.691E02 0.147 4.234E02 0.109 2.023E03 0.187 4.979E02 0.206 8.693E02 0.132 1.484E02 0.143 3.092E02 8.923E02 7.651E02 6.892E02 6.297E02 7.223E02 0.123 0.113 9.544E02

0.424 0.227 0.117 0.154 0.164 6.422E02 0.105 0.103 0.123 8.792E02 3.573E02 0.120 0.133 8.472E02 0.110 0.188 0.224 0.124

0.479 0.373 0.309 0.241 0.120 0.216 2.445E03 4.756E02 2.453E02 9.745E02 0.143 0.343 0.335 8.220E02 0.131 5.956E02 0.147 6.591E02

4.445E02 0.182 0.299 0.236 0.629 0.589 0.581 0.486 5.963E03 5.271E02 0.110 5.073E02 0.103 6.049E02 3.618E02 0.115 0.274 0.149

1.595E02 9.300E02 9.130E02 5.780E02 2.059E03 4.938E03 8.388E02 0.196 0.702 0.696 0.646 0.415 0.355 3.835E02 6.942E02 0.182 2.750E02 0.251

0.120 3.728E02 0.183 0.154 3.367E02 3.102E02 0.138 0.226 1.236E02 4.104E02 0.105 4.981E02 6.390E02 0.761 0.725 0.504 0.357 0.321

Note: SCARED: Screen for child anxiety related emotional disorders.

other groups were modest to strong effect size (d = 0.311.5). The ROC method was used to assess whether any specic cutoff scores signicantly discriminate between anxious and community children. The optimal cutoff

point is determined by plotting sensitivity versus 1specicity and examining the point of maximum deviation from chance. A total cutoff point score of 25 on the child SCARED resulted in the optimal sensitivity (79%) and specicity (82%) separating

618

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621

children with anxiety disorders (n = 48) from community children (n = 1559). 3.3.2. Convergent validity Based on the school (n = 139) and clinical (n = 130) sample in Changsha city, the childrens SCARED total score correlated signicantly better with the parents CBCLs internalizing score (r = 0.41) than with the externalizing (r = 0.18) ( p < 0.01). 3.4. Factor analyses Several factor solutions were analyzed. Only those solutions with factors with eigenvalues >1 and which were clinically sound were chosen. A principal components analysis with varimax rotation was conducted to assess the correlated matrix of the SCAREDs 41 items. In this approach, the t of the factor structure was evaluated by multiple criteria: item loadings above 0.30; more than three items loading on the retained factors; and conceptual meaning of the emerging factor structure. As shown in Table 4, 40 items were retained which accounted for 36.8% of the explained variance. Item 9 did not load on any of the factors with item loadings under 0.30. The ve factors were generally corresponding to the SCAREDs original factors. However, only four somatic items from Birmaher et al. (1999) somatic/panic subscale load on the original factor, while all of the panic items together with the remaining somatic items load on other factors, especially on the generalized anxiety subscale (items 30, 6, 24, 15 and 12). The ve factors were positively and signicantly correlated with the SCARED total score ( p-values <0.001). The internal consistency of these factors ranged from 0.58 to 0.84. The 2-week testretest reliability for these factors ranged from 0.51 to 0.63 and the 12-week testretest reliability from 0.38 to 0.52. Conrmatory factor analysis with LISREL8.51 soft reskog & So rbom, 1996) revealed that this ware (Jo modied 5-factor, 40-item SCARED model was an acceptable model for this Chinese sample (x2/d.f. = 3.8, NFI = 0.87, NNFI = 0.91, CFI = 0.93, GFI = 0.92, RMSEA = 0.029). 4. Discussion This study examined the psychometric properties of the SCARED in a large school sample of children and adolescents in China. Our results indicate that, similar to other clinical and community studies across the world (Birmaher et al., 1999; Essau et al., 2002;

Muris et al., 1999), the SCARED has appropriate psychometric properties and is a useful instrument to screen for DSM anxiety disorders in Chinese youth. 4.1. Sex and age effects Similar to other studies (Cohen et al., 1993; Compton, Nelson, & March, 2000; Craske, 1997; Hale et al., 2005) females had signicantly higher anxiety scores on the total anxiety, separation anxiety, and social phobia. Concurring with prior publications, children had higher scores on separation anxiety than the adolescents (Bernstein & Layne, 2004; Birmaher et al., 1997; Hale et al., 2005). 4.2. Reliability This study supported the previous ndings of the SCAREDs high internal consistency and moderate testretest reliability (Birmaher et al., 1997, 1999; Boyd et al., 2003). Comparable with the parentchild correlations for anxiety symptoms reported in the western literature (Birmaher et al., 1997; Wren et al., 2004), this study reported higher parentchild correlations with a correlation of 0.59 for the total anxiety score and 0.670.89 for the ve-factors. Moreover, we replicated the nding that adolescents and females show signicantly higher parentchild correlations than children and males (Birmaher et al., 1999; Wren et al., 2004). 4.3. Validity The results of the present study also supported the discriminant validity of the SCARED. The total score and the ve subscales signicantly differentiated children with anxiety disorders from school children and children with ADHD. Consistent with prior studies (Birmaher et al., 1997, 1999; Monga et al., 2000), the SCARED signicantly discriminated between anxious and pure depressed children on the total score and panic/somatic, separation anxiety and social phobia subscales. Noteworthy, consistent with Birmaher et al. (1999), a cutoff point of 25 on the child SCARED yielded optimal sensitivity (79%) and specicity (82%). Finally, parallel to Monga et al. (2000) study, using the CBCL as the comparison instrument, the SCARED showed good convergent and divergent validity. The SCARED total score was found to correlate signicantly better with the internalizing than with the externalizing factors of the CBCL.

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621

619

4.4. Factor analyses In this study, the SCARED showed a ve-factor structure with most items corresponding to the items included in the ve anxiety factors the original scale measured (Birmaher et al., 1997, 1999). However, Item 9 (people tell me I look nervous) did not load on any of the factors, which may be explained by the indirectness of Chinese culture. Chinese people tend to rely on indirect, more complex methods of communication like indenite comments while Westerners tend to adopt direct and simple methods. Also in contrast to the study by Birmaher and colleagues, only four somatic items loaded on the original somatic/panic factor. These discrepancies may be accounted by differences in the sample composition and culture factors. For example, in contrast with other studies (Birmaher et al., 1997, 1999; Hale et al., 2005), most of the subjects included in our study were young children (62%). Also, it is possible that somatic/panic symptoms may differ in presentation in Chinese youths. Somatic/panic symptoms in Chinese children often represent physical symptoms of the underlying anxiety factors including generalized anxiety, separation anxiety, social phobia and school phobia. Interestingly, African Americans appear to have different characteristics in the frequency and distribution of their somatic/panic symptoms when compared with Caucasians (Boyd et al., 2003; Carter et al., 1999; Guarnaccia, 1997). In Wren et al. (2007) study, the researchers also reported a large variation in factor structure across ethnic groups in the areas of somatic/ panic. Thus, it is possible that somatic/panic symptoms in particular are sensitive to cultural issues. 5. Limitations This study only evaluated a small sample of children for the testretest reliability and ndings need to be replicated with larger samples. Secondly, only school students living in urban environments were included. Finally, further studies including larger samples of children with various anxiety disorders and other psychopathologies are warranted to evaluate the usefulness of the SCARED in clinical populations. 6. Clinical implications The SCARED appears to be a reliable and valid screening scale for anxiety disorders in Chinese children and adolescents. With limited mental health resources, the inadequate clinical attention dedicated to children with internalizing disorders has recently been

recognized as a key area of concern in China. Because of the inexpensive cost, convenient administration, and the ease of standardization, the SCARED seems to be an ideal and valuable instrument for identifying childhood anxiety disorders in Chinese children and adolescents.

Acknowledgements This work was supported by the National Natural Science Fund of China (30370521, Su). In addition, the authors thank Dr. Boris Birmaher and the two reviewers for their valuable suggestions in preparing this manuscript.

References
Achenbach, T. M. (1991). Manual for the child behavior checklist/418 and 1991 prole. Burlington, VT: University of Vermont Department of Psychiatry. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (DSM-IV) (4th ed.). Washington, DC: American Psychiatric Association. Beidel, D. C., Morris, T. L., & Turner, M. W. (2004). Social phobia. In: T. L. Morris & J. S. March (Eds.), Anxiety disorders in children and adolescents (2nd ed., pp. 141163). New York: Guilford. Bernstein, G. A., Hektner, J. M., Borchardt, C. M., & McMillan, M. H. (2001). Treatment of school refusal: One year follow-up. Journal of the American Academy of Child & Adolescent Psychiatry, 29, 773781. Bernstein, G. A., & Layne, A. E. (2004). Separation anxiety disorder and generalized anxiety disorder. In: J. M. Wiener & M. K. Dulcan (Eds.), Textbook of child and adolescent psychiatry (3rd ed., pp. 557589). Washington, DC: American Psychiatric Publishing. Birmaher, B., Brent, D. A., Chiappetta, L., Bridge, J., Monga, S., & Baugher, M. (1999). Psychometric properties of the screen for child anxiety related emotional disorders (SCARED): A replication study. Journal of the American Academy of Child & Adolescent Psychiatry, 38, 12301236. Birmaher, B., Khetarpal, S., Brent, D., Cully, M., Balach, L., Kaufman, J., et al. (1997). The Screen for child anxiety related emotional disorders (SCARED): Scale construction and psychometric characteristics. Journal of the American Academy of Child & Adolescent Psychiatry, 36, 545553. Boyd, R. C., Ginsburg, G. S., Lambert, S. F., Cooley, M. R., & Campbell, K. D. M. (2003). Screen for child anxiety related emotional disorders (SCARED): psychometric properties in an African American parochial high school sample. Journal of the American Academy of Child & Adolescent Psychiatry, 42, 1188 1196. Brady, E. U., & Kendall, P. C. (1992). Comorbidity of anxiety and depression in children and adolescents. Psychological Bulletin, 3, 244255. Carter, M. M., Miller, O., Sbrocco, T., Suchday, S., & Lewis, E. L. (1999). Factor structure of the anxiety sensitivity index among African American college students. Psychological Assessment, 11, 525533.

620

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621 Monga, S., Birmaher, B., Chiappetta, L., Chiappetta, L., Brent, D., Kaufman, J., et al. (2000). Screen for child anxiety-related emotional disorders (SCARED): convergent and divergent validity. Depress Anxiety, 12, 8591. Muris, P., Merckelbach, H., Mayer, B., van Brakel, A., Thissen, S., Moulaert, V., et al. (1998). The Screen for child anxiety related emotional disorders (SCARED) and traditional childhood anxiety measures. Journal of Behavior Therapy and Experimental Psychiatry, 29, 327339. Muris, P., Merckelbach, H., Ollendick, T., King, N., & Bogie, N. (2002a). Three traditional and three new childhood anxiety questionnaires: their reliability and validity in a normal adolescent sample. Behavior Research and Therapy, 40, 753772. Muris, P., Merckelbach, H., Ollendick, T., King, N., Meesters, C., & van Kessel, C. (2002b). What is the revised fear survey schedule for children measuring? Behavior Research and Therapy, 40, 13171326. Muris, P., Merckelbach, H., Schmidt, H., & Mayer, B. (1999). The revised version of the screen for child anxiety related emotional disorders (SCARED-R): factor structure in normal children. Personality and Individual Differences, 26, 99112. Muris, P., Schmidt, H., Engelbrecht, P., & Perold, M. (2002c). DSMIV-dened anxiety disorder symptoms in South African children. Journal of the American Academy of Child & Adolescent Psychiatry, 41, 13601368. Newman, D. L., Moftt, T. E., Caspi, A., Magdol, L., Silva, P. A., & Stanton, W. R. (1996). Psychiatric disorder in a birth conhort of young adults: prevalence, comorbidity, clinical signicance, and new case incidence from ages 11 to 21. Journal of Consulting and Clinical Psychology, 64, 552562. Ollendick, T. H. (1983). Reliability and validity of the revised fear survey schedule for children (FSSC-R). Behavior Research and Therapy, 21, 685692. Pavuluri, M., & Birmaher, B. (2004). A practical guide to using ratings of depression and anxiety in child psychiatric practice. Current Psychiatry Reports, 6, 108116. Perrin, S., & Last, C. G. (1992). Do childhood anxiety measures measure anxiety? Journal of Abnormal Child Psychology, 20, 567578. Reynolds, C. R., & Richmond, B. O. (1978). What I think and feel: A revised measure of childrens manifest anxiety. Journal of Abnormal Child Psychology, 6, 271280. Seligman, L. D., Ollendick, T. H., Langley, A. K., & Baldacci, H. B. (2004). The utility of measures of child and adolescent anxiety: A meta-analytic review of the revised childrens manifest anxiety scale, the state-trait anxiety inventory for children, and the child behavior checklist. Journal of Clinical Child Psychology, 33, 557 565. Silverman, W. K., & Ollendick, T. H. (2005). Evidence-based assessment of anxiety and its disorders in children and adolescents. Journal of Clinical Child and Adolescent Psychology, 34, 380411. Somoza, E., & Mossman, D. (1991). ROC curves and the binormal assumption. Journal of Neuropsychiatry and Clinical Neurosciences, 3, 436439. Spielberger, C. D. (1973). State-trait anxiety scale for children: preliminary manual. Palo Alto, CA: Consulting Psychologists Press. Strauss, C. C., Lahey, B. B., Frick, P., Frame, C. L., & Hynd, G. W. (1988). Peer social status of children with anxiety disorders. Journal of Consulting and Clinical Psychology, 56, 137141. Su, L. Y., Li, X. R., Yang, Z. W., Luo, X. R., & Wan, G. B. (1997). The application of child behavior checklist and teachers report

Cohen, P., Cohen, J., & Brook, J. (1993b). An epidemiological study of disorders in late childhood and adolescence: II. Persistence of disorders. Journal of Paediatric Psychology, 26, 367374. Cohen, P., Cohen, J., Kasen, S., Velez, C. N., Hartmark, C., Johnson, J., et al. (1993a). An epidemiological study of disorders in late childhood and adolescence: I. Age and gender specic prevalence. Journal of Child Psychology and Psychiatry, 34, 851867. Compton, S. N., Nelson, A. H., & March, J. S. (2000). Social phobia and separation anxiety symptoms in community and clinical samples of children and adolescents. Journal of the American Academy of Child & Adolescent Psychiatry, 39, 10401046. Costello, E. J., & Angold, A. (1995). Epidemiology. In: J. S. March (Ed.), Anxiety disorders in children and adolescent (pp. 109124). New York: Guilford. Craske, M. G. (1997). Fear and anxiety in children and adolescents. Bulletin of the Menninger Clinic, 61(Suppl. A), A4A36. Essau, C. A. (2000). Frequency, comorbidity and psychosocial impairment of anxiety disorders in German adolescents. Journal of Anxiety Disorders, 14, 263279. Essau, C. A., Muris, P., & Ederer, E. M. (2002). Reliability and validity of the Spence childrens anxiety scale and the screen for child anxiety related emotional disorders in German children. Journal of Behavior Therapy and Experimental Psychiatry, 33, 118. Ginsburg, G. S., & Silverman, W. K. (1996). Phobic and anxiety disorders in Hispanic and Caucasian youth. Journal of Anxiety Disorders, 10, 517528. Guarnaccia, P. J. (1997). A cross-cultural perspective on anxiety disorders. In: S. Friedman (Ed.), Cultural Issues in the treatment of anxiety (pp. 320). New York: Guilford. Hale, W. W., Raaijmakers, Q., Muris, P., & Meeus, W. (2005). Psychometric properties of the screen for child anxiety related emotional disorders (SCARED) in the generalized adolescent population. Journal of the American Academy of Child & Adolescent Psychiatry, 44, 283290. Hoehn-Saric, E., Maisami, M., & Weigand, D. (1987). Measurement of anxiety in children and adolescents using semistructured interviews. Journal of the American Academy of Child & Adolescent Psychiatry, 26, 541545. reskog, K. G., & So rbom, D. (1996). Lisrel 8: Users reference Jo guide. Chicago, IL: Scientic Software International. Keller, M. B., Lavori, P. W., Wunder, J., Beardslee, W. R., Schwartz, C. E., & Roth, J. (1992). Chronic course of anxiety disorders in children and adolescents. Journal of the American Academy of Child & Adolescent Psychiatry, 31, 595599. Kendall, P. C., Brady, E. U., & Verduin, T. L. (2001). Comorbidity in childhood anxiety disorders and treatment outcome. Journal of the American Academy of Child & Adolescent Psychiatry, 40, 787 794. Klein, R. G. (1994). Anxiety disorders. In: M. Rutter, E. Taylor, & L. Hersov (Eds.), Child and adolescent psychiatry: modern approaches (pp. 351374). Oxford, England: Blackwell Scientic Publications. La Greca, A. M., & Stone, W. L. (1993). Social anxiety scale for children-revised: factor structure and concurrent validity. Journal of Clinical Child Psychology, 22, 1727. Mattison, R. E., & Babnato, S. J. (1987). Empirical measurement of overanxious disorder in boys 8 to 12 years old. Journal of the American Academy of Child & Adolescent Psychiatry, 26, 536540.

L.Y. Su et al. / Journal of Anxiety Disorders 22 (2008) 612621 form in epidemiological study of child mental health: Choosing an optimal cut-off point. Chinese Mental Health Journal, 11, 3740. Wren, F. J., Berg, E. A., Heiden, L. A., Kinnamon, C. J., Ohlson, L. A., Bridge, J. A., et al. (2007). Childhood anxiety in a diverse primary care population: Parentchild reports, ethnicity and SCARED

621

factor structure. Journal of the American Academy of Child & Adolescent Psychiatry, 46, 332340. Wren, F. J., Bridge, J. A., & Birmaher, B. (2004). Screening for childhood anxiety symptoms in primary care: Integrating child and parent reports. Journal of the American Academy of Child & Adolescent Psychiatry, 43, 13641371.