Ethnic differences in social anxiety between individuals of Asian heritage and European heritage: A meta-analytic review

By: Alexander Krieg
Department of Psychology, University of Hawaii at Manoa
Yiyuan Xu
Department of Psychology, University of Hawaii at Manoa;

Acknowledgement:

Social anxiety refers to the experience of intense apprehension and fear related to social interaction or performance (Good & Kleinman, 1985; Mesquita & Frijda, 1992). It is characterized by the persistent fear of social situations in which one’s behavior may be scrutinized by others, an experience that often leads to anticipation of embarrassment or humiliation (American Psychiatric Association, 2000). Symptoms of social anxiety include high physiological arousal, such as shortness of breath and sweating, in feared situations; self-defeating cognitions regarding one’s own social performance and others’ evaluation; as well as avoidance and/or anticipation of feared situations (American Psychiatric Association, 2000; Heimberg, Brozovich, & Rapee, 2010). These symptoms are associated with both occupational and social impairment (Aderka et al., 2012; Wittchen & Beloch, 1996), and carry great cost at both the individual and societal levels.

The experience of social anxiety appears to be relatively common, with about 25% to 33% of adults reporting intense anxiety and/or avoiding certain social situations at some point in their lives (Kessler, McGonagle, & Zhao, 1994; Kessler, Stein, & Berglund, 1998; Ruscio et al., 2008). This prevalence appears to differ between individuals of Asian heritage (AH) and individuals of European heritage (EH), but the findings seem to be mixed (Abe & Zane, 1990; Gordon & Teachman, 2008; Hardin & Leong, 2005; Hsu, 2004; Hsu & Alden, 2007, 2008; Hsu et al., 2012; Lau, Fung, Wang, & Kang, 2009; Lee, Okazaki, & Yoo, 2006; Norasakkunkit & Kalick, 2002, 2009; Norasakkunkit, Kitayama, & Uchida, 2012; Okazaki, 1997, 2000, 2002; Okazaki, Liu, Longworth, & Minn, 2002; Sue, Ino, & Sue, 1983; Sue, Sue, & Ino, 1990). For instance, some epidemiological studies, which used interviews to assess social phobia—a clinically significant subtype of social anxiety—reported a greater 12-month prevalence for social phobia diagnoses among European Americans compared with Asian Americans (3.00% and 2.30%, respectively; Smith et al., 2006). Another social epidemiological study reported the difference being as great as 7.30% (European Americans, 12.60%, and Asian Americans, 5.30%; Asnaani, Richey, Dimaite, Hinton, & Hofmann, 2010). In contrast, studies that relied on self-reports among mostly college students have shown that individuals of AH tended to report higher levels of overall social anxiety symptoms than did their EH counterparts (Hsu et al., 2012; Lau et al., 2009; Norasakkunkit & Kalick, 2009; Okazaki, 1997, 2000, 2002). Therefore, the primary purpose of this study was to review extant findings on the difference in social anxiety between AH and EH individuals using meta-analyses. In addition, we also examined how factors that may be pertinent to distinct experiences of AH and EH individuals, such as independent and interdependent self-construals (Markus & Kitayama, 1991) and acculturation (Berry, Kim, Minde, & Mok, 1987; Zheng & Berry, 1991), may contribute to the AH–EH difference in social anxiety.

The AH–EH Difference in Social Anxiety

Two epidemiological studies found a lower prevalence of social phobia diagnoses in Asian Americans compared with European Americans (Asnaani et al., 2010; Smith et al., 2006). An epidemiological study typically focuses on prevalence rates of certain disorders across various social categories, such as sex, race, generation, and social status (Honjo, 2004). In a study that interviewed a random sample of 24,507 European Americans and 1,332 Asian Americans, Smith et al. (2006) found 12-month prevalence rates of 3.01% and 2.13% for social phobia diagnosis for European Americans and Asian Americans, respectively. Asnaani et al. (2010) reported the 12-month prevalence rate of 12.6% and 5.3%, respectively, for another randomly selected sample of European and Asian Americans. Despite the large sample sizes and the use of diagnostic interviews, there were at least two limitations of these epidemiological studies: They relied on lay interviewers with limited training rather than professional clinicians, and there was no information on either reliability or validity of these diagnoses. As a consequence, the prevalence rates seemed to vary significantly across different studies.

Other studies have used a variety of self-report methods, with samples consisting primarily of college students (Hsu et al., 2012; Lau et al., 2009; Okazaki, 1997). The findings of these studies have shown decent consistency in demonstrating elevated levels of overall social anxiety symptoms among AH individuals compared with EH individuals. For example, in a study that examined a group of 200 Asian American and 200 European American undergraduate students recruited from the registrar’s list, Okazaki (2002) found that Asian Americans scored significantly higher than European Americans on the intensity of social anxiety, as measured by the Social Phobia and Anxiety Inventory (SPAI; Turner, Beidel, Dancu, & Stanley, 1996).

It should be noted that not all self-report studies consistently find the AH–EH difference in social anxiety symptoms. In a study comparing 47 Asian American and 41 European American undergraduate students, Lee et al. (2006) found no significant ethnic difference in social anxiety symptoms (measured using the SPAI) or in the frequency of the experience of social anxiety, as recorded by daily diary entries. Similarly, Gordon and Teachman (2008) found mixed results when surveying 33 Asian American and 39 European American undergraduate students. Although two measures of social anxiety—the Fear of Negative Evaluation Scale (FNES; Watson & Friend, 1969) and the Social Avoidance Distress Scale (SADS; Watson & Friend, 1969)—were administered, Asian Americans scored higher than European Americans only on the FNES scores. Furthermore, no AH–EH differences were found when social anxiety was elicited experimentally (Gordon & Teachman, 2008). Thus, it appears that the robustness of the AH–EH difference in social anxiety may be contingent upon not only the assessment method (i.e., epidemiological interviews vs. self-report questionnaires) but also the specific self-report measure being used.

There is also emerging evidence to suggest that method of assessment may not fully account for the discrepant findings with regard to the AH–EH difference in social anxiety. Okazaki (2000) administered both the self-report and interview forms of the FNES to 39 Asian American and 42 European American undergraduate students, and showed that the AH–EH difference in social anxiety did not vary significantly across different methods of assessment. Hsu (2004) found greater social anxiety symptoms among Chinese Canadians compared with European Canadians, even when a diagnostic interview was used to assess social anxiety. Using a modified version of the Anxiety Disorders Interview Schedule – Version 4 (ADIS-IV; Brown, DiNardo, & Barlow, 1994), Hsu and Alden (2007) showed that Chinese Canadians reported higher social anxiety symptoms, as well as higher social impairment, than did their European Canadian counterparts, suggesting that the discrepant findings on the AH–EH difference in social anxiety may not necessarily be related to methods of assessment.

Self-Construals, Acculturation, and the AH–EH Difference in Social Anxiety

There is also considerable interest in understanding what factors may explain the AH–EH difference in social anxiety. Some studies have shown that the AH–EH difference in social anxiety may be explained by differences in cultural values and norms related to fundamental views of self (Okazaki, 1997). Others have suggested that unique challenges associated with acculturation and being a visible ethnic minority may also account for higher social anxiety among AH than EH individuals (Hsu et al., 2012). In this review, we focused on three explanatory factors that have been most commonly examined in the literature: independent self-construal, interdependent self-construal, and acculturation.

The AH–EH difference in social anxiety may be associated with different cultural views of self in relation to others, that is, lower independent self-construal and higher interdependent self-construal among AH than EH individuals (Okazaki, 1997; Norasakkunkit & Kalick, 2002). Independent self-construal refers to viewing the self as independent and autonomous from others, whereas interdependent self-construal is defined as viewing the self as connected with members of the immediate group and surrounding environment (Markus & Kitayama, 1991). Although one’s independence and autonomy of the self is typically emphasized in Western cultures, adapting to interpersonal and social demands while maintaining social harmony are important socialization goals in many Asian cultures (Markus & Kitayama, 1991; Singelis, 1994). In line with these fundamental differences in cultural values and norms, assertiveness, direct communication, and competitiveness are typically promoted in Western contexts. In contrast, accomplishing a sense of heightened sensitivity toward other’s feelings, opinions, and evaluations, and regulating and suppressing negative emotions, are particularly salient in Asian cultures (Chen & Rubin, 2011; Clark, 2001; Markus & Kitayama, 1991). Moreover, some behavior that appears to be more characteristic of cultural emphasis on interdependent self-construal, such as eye-gaze aversion, nonassertiveness, and more frequent silence (Baker & Edelmann, 2002; Beidel & Turner, 1999), seem to mimic symptoms or behavioral indices of social anxiety as defined by Western theorists (Clark & Wells, 1995), and thus may be misinterpreted as social anxiety.

There is some support for the relations between independent and interdependent self-construal and the AH–EH difference in social anxiety. For instance, Ho and Lau (2011) examined the relation between self-construal and social anxiety symptoms in a sample of 74 European American, 83 U.S.-born Asian American, and 72 foreign-born Asian American undergraduate students. In their analysis, they ran two models, with independent and interdependent self-construal entered separately in each model, along with two additional variables: (a) contrasting foreign-born Asian Americans and European Americans, and (b) contrasting U.S.-born Asian Americans and European Americans. They found that interdependent self-construal was positively related to social anxiety, whereas independent self-construal was negatively associated with social anxiety, and that both types of self-construals fully mediated the AH–EH difference in social anxiety. Likewise, Hong and Woody (2007) tested a mediation model with 251 Korean Canadians and 254 European Canadians recruited from the community. In these models, the ethnic difference in social anxiety was reported separately after the variance from independent and interdependent self-construal was partialed out. Their results indicated that independent self-construal fully mediated the AH–EH difference in social anxiety, whereas interdependent self-construal only partially mediated this ethnic difference.

The AH–EH difference in social anxiety may also be related to the challenges experienced by AH individuals because of their immigrant and/or minority status. One such challenge is acculturative stress, defined as the distress that immigrant and/or minority groups experience when adjusting to life in a new context in which the new cultural values and goals are in conflict with their previously socialized cultural standards (Hsu et al., 2012). When social behaviors in a Western setting create an apparent conflict with Asian values, the resulting acculturative or bicultural stress can lead to elevated symptoms or the experience of social anxiety. For instance, Hsu et al. (2012) showed that monocultural groups (East Asian nationals and European Americans) displayed lower levels of social anxiety and social comparison compared with the bicultural group (Asian Americans). Hsu et al. suggested that the conflict between the two value systems may have accounted for increased wariness and anxiety in social situations among Asian Americans. In a similar vein, Okazaki (1997) found that Asian Americans who were less acculturated to the social norms of American culture were more likely to report higher avoidance of, and distress in, social situations than their European American counterparts. It should be noted, however, that most studies did not directly assess acculturative stress, but rather measured generation status or acculturation-assimilation to American mainstream culture (e.g., language preference, social participation) among AH individuals.

The Current Study

Our meta-analytic review focused on three research questions. First, we examined whether, on average, AH individuals tended to have higher social anxiety than EH individuals. Second, we explored whether the AH–EH difference in social anxiety varied as a function of method of assessment (i.e., interviews vs. self-reports) and specific measure of social anxiety. Third, we examined to what extent independent and interdependent self-construals and acculturation accounted for the AH–EH difference in social anxiety.

Method
Literature Search

Literature searches were performed with 17 databases covering the years 1980 to 2013. First, we started with searches based on general keywords pertaining to social anxiety, that is, “social anxi*” (e.g., social anxiety, social anxiety disorder), “social phob*” (e.g., social phobia, social phobic), “fear of negative evaluation,” “social fear*,” “social avoid*,” and “taijin kyofusho,” intersected with “Asia*” (e.g. Asia, Asian, Asian American) and “East*” (e.g., East, Eastern, Easterner). Second, we conducted searches based on measures of social anxiety that were used in the studies identified during the first step, including “SPAI” (Social Phobia and Anxiety Inventory), “SAD*” (Social Avoidance and Distress [Scale]), “FNE*” (Fear of Negative Evaluation [Scale]), “BFNE*” (Brief Fear of Negative Evaluation [Scale]), “TKS” (Taijin Kyofusho Scale), “SPIN*” (Social Phobia Inventory), “LSAS” (Liebowitz Social Anxiety Scale), “SAS-A” (Social Anxiety Scale-Adult), and “ADIS-SP” (Anxiety Disorders Interview Scale-Social Phobia), intersected with “Asia*” (e.g. Asia, Asian, Asian American) and “East*” (e.g., East, Eastern, Easterner). Third, we combined all the keywords used in the first two steps and reconducted searches to cross-validate the results. Of the articles retrieved during this third step, we explored the references for previous literature (“ancestor” search) and examined the works that listed citations of these articles (“descendant” search) in each database. In addition, we performed a general name search for each author who had contributed at least two studies to our sample. These investigators (n = 8) were contacted via e-mail and were asked if they knew of any unpublished studies that should be included in this project (a response rate of 75% [six of eight] resulted in two additional entries of currently unpublished work).

Inclusion and Exclusion Criteria

The following criteria were used to determine the studies’ appropriateness for inclusion in this meta-analysis: (a) the data in the study was accessible in English until August 2013; (b) the study reported a direct and explicit measure of social anxiety, its subtypes (e.g., social phobia), or specific aspects of social anxiety (e.g., fear of negative evaluation); (c) the study included both a group of AH participants and a group of EH participants; and (d) the study contained data (i.e., means and standard deviations, or prevalence rate and standard errors; Cohen, 1988) by which an effect size could be calculated. This was only possible when enough information was available either directly or indirectly (e.g., tables, charts) to determine the direction of the effect; one study (Heinrichs et al., 2006) that did not meet this inclusion criterion was not included in the analysis. Using these criteria, 32 studies, including peer-reviewed articles (n = 25), dissertations (n = 3), and unpublished theses (n = 4), that had examined AH and EH difference in social anxiety were included in our meta-analysis (see Table 1).

Coding System and Coding Procedures

Three raters—the primary investigator and two trained research assistants—used a standardized coding system to code study-level variables for all the studies, including sample characteristics (e.g., region, age, percent foreign born, sampling style) and study characteristics (e.g., design type, measure used; see Table 2). The overall interrater agreement among the three raters was kappa = .80; all disagreements were resolved by discussion.

Additional coding decisions included the following: (a) when the same sample was used in two different studies (i.e., Hsu & Alden, 2007, 2008), we collapsed the information on these participants and the measures reported in the studies into one entry; (b) when one European American group was compared with multiple Asian American groups (i.e., first generation and second generation; n = 8), we combined the groups by weighting and averaging the means and standard deviations.

Meta-Analysis

Computation of effect sizes

We used Cohen’s d as the measure of effect size. Cohen’s d is calculated as the difference between the means of the two groups, divided by their pooled standard deviations. We coded the effect sizes so that a positive d statistic indicated higher social anxiety in AH individuals, and a negative d statistic indicated higher social anxiety in EH individuals. Standard errors were calculated using the following formula to account for sampling error (Hunter & Schmidt, 2004):

Test of homogeneity and random effects model

The homogeneity of the overall effect size was assessed using the homogeneity test statistic Q (Hedges & Olkin, 1985). This statistic was then fitted into a categorical model and tested for homogeneity in each class by generating a chi-square statistic.

Random versus mixed effect model

In the following analyses, when we estimated a model without any moderators, a random effects model was chosen to provide a conservative estimate of the overall effect. In contrast, when we estimated a model with known moderators (see Table 2), a mixed effect model was used in which the variation associated with the coded moderators was partialed out from the random variation that cannot be accounted for by the moderators (Bakermans-Kranenburg, van Ijzendoorn, & Juffer, 2003). All meta-analyses were carried out using the metafor package (Viechtbauer, 2010) on the statistical platform R (R Development Core Team, 2012).

Self-Construals and Social Anxiety: Meta-Analytic Structural Equation Modeling (MASEM)

According to Cheung (2014a), the purpose of MASEM is to synthesize correlation (or covariance) matrices using meta-analytic techniques, and then to fit a structural equation model on the pooled correlation (or covariance) matrix. This is typically done in two stages (Cheung & Chan, 2005, 2009). In the first stage, the correlation or covariance matrices are tested for homogeneity and then combined to form a pooled matrix. In the second stage, the pooled matrix is used as the observed correlation or covariance matrix and incorporated into a structural equation modeling analysis. A random effects model will be used in the second stage if heterogeneity among correlation matrices is found in the first stage. This is a useful tool for fitting complex models to data aggregated from descriptive statistics across many studies. Given the limited information on means and standard deviations in prior studies, the MASEM analyses, conducted with the metaSEM package (Cheung, 2014b), were based on a pooled correlation matrix.

Following previous studies (e.g., Ho & Lau, 2011; Hong & Woody, 2007; Okazaki, 1997), we tested a mediation model, which assumes that the AH–EH difference in self-construals may at least partially explain (i.e., mediate) the AH–EH difference in social anxiety (see Figure 1). Correlation matrices were pooled from studies that included measures of independent and interdependent self-construal as well as a measure of social anxiety. Ethnicity (AH coded as “1”; EH coded as “0”) was also included in each of the correlation matrices by calculating the Cohen’s d statistic from group means and standard deviations and converting it to an equivalent correlation. Heterogeneity was measured via the homogeneity test statistic Q (Hedges & Olkin, 1985). The final correlation matrix was fitted to a SEM-based mediation model in order to test the mediating effect of self-construals (see details in the Results section).

Acculturation and Social Anxiety: Meta-Regression Analyses

According to Sutton and Higgins (2008), a meta-regression is “a combination of meta-analytic principles (of combining results from multiple studies with due attention to within-study precision and among-study variation) with regression ideas (of predicting study effects using study-level covariates)” (p. 629). Using the R package metafor (Viechtbauer, 2010), we adopted a random effects model in carrying out the meta-regression analysis. A random effects model can help reveal heterogeneity among studies and accordingly identify potentially important covariate variables (i.e., sample composition or features) that require additional analyses, based on subgroups. Because acculturation measures were only used for AH individuals, Cohen’s f2, which can be calculated based on Pearson’s correlations reported in previous studies, was used to estimate variance in social anxiety that can be accounted by acculturation.

Results
Descriptive Statistics for the Study-Level Variables

The current sample includes 32 studies from the United States, Canada, and Australia, with a wide variety of demographic and methodological differences (see Table 2). These study-level variables were coded and entered into the analysis to estimate their relevant influence on the overall effect size. The homogeneity measure of our mixed effects model indicated significant residual heterogeneity among the effect sizes, QE (18) = 2484.50 (p = .00), with the overall effect of study-level variables being significant, QM (13) = 39.83, p = .00. Only one individual study-level variable had a significant effect on the overall effect size: A higher percentage of individuals with Japanese ancestry was significantly associated with lowered effect size (Cohen’s d = −.20, p = .01). Table 3 provides information on all other nonsignificant study-level variables.

The AH–EH Difference in Social Anxiety

Of the 32 independent effect sizes, 28 (87.50%) were positive in direction and 27 (96.43%) were statistically significant. Using a random effects model, the overall effect size across all studies was d = .36 (95% confidence interval [CI] [.27, .45]), indicating that AH individuals tended to report higher social anxiety than EH individuals. The trim and fill statistic (Duval & Tweedie, 2000) was .29 (95% CI [.20, .39]), which indicated that even with a more balanced distribution of studies on both sides of the overall effect, the overall effect would remain statistically significant.

In order to provide the most accurate estimate of the overall effect in the population, one effect size was calculated for each independent sample. Effect sizes of two studies by Hsu and Alden (2007, 2008), which were based on data from the same sample, were combined as one independent estimate. As recommended by Hedges and Olkin (1985), when multiple measures of social anxiety were used in a single study, effect sizes were averaged across measures within this study to attain an aggregated effect size, so that a single sample would not be overrepresented multiple times in the calculation of an overall effect. Seventeen studies with two or more measures of social anxiety were averaged in their respective categories.

Assessment Method and Specific Measure of Social Anxiety

We also examined whether the AH–EH difference in social anxiety was moderated by method of assessment (interview vs. self-report) or specific self-report measure. When more than one measure of social anxiety was used in a single study, effect sizes calculated for each individual measure were treated as independent entries, rather than being aggregated, in estimating the moderating effect. The results showed that neither moderator had a significant effect: The Cohen’s d estimates were .02 (p = .91) and .01 (p = .58), respectively.

We further explored the relation between method of assessment and the AH–EH difference in social anxiety by carrying out a follow-up meta-analysis that only included the studies using interview measures of social anxiety (n = 5). Three effect sizes were positive and significant. Using a mixed effects model, the effect size across the five studies was d = .37 (95% CI [−.09, .82]), which is very similar to the overall effect found with the entire sample of studies.

Similarly, we also conducted a follow-up meta-analysis that only included the studies using self-report measures of social anxiety (n = 30). Among the 47 independent effect sizes, 41 were positive and 40 were statistically significant. The effect size across all specific measures was d = .37 (95% CI [.28, .46]). For the five most commonly used measures of social anxiety—SADS, FNES, SPS, SIAS, and SPAI (see Table 2)—we found average effect sizes of .47, .19, .36, .50, and .45, respectively.

Self-Construal and the AH–EH Difference in Social Anxiety

To examine how self-construal may account for the AH–EH difference in social anxiety, we examined the roles of independent and interdependent self-construals as mediators in the relation between ethnicity (AH vs. EH, dummy coded) and social anxiety using a two-stage MASEM analysis. We first examined heterogeneity of correlational matrices from the seven studies (n = 2,052) that included ethnicity, independent self-construal, interdependent self-construal, and social anxiety. The Q test (Hedges & Olkin, 1985) indicated that these correlation matrices were heterogeneous, Q (33) = 145.82, p = .00, which was expected, given that no study-level variables were taken into account. According to Cheung (2014a), significant heterogeneity in this initial step does not preclude us to continuing to specify structural equation models, but instead indicates between-study variation, making the use of a random effects model preferable. Thus, all the correlation matrices were synthesized into a pooled correlation matrix and analyzed using a random effects model. The synthesized matrix was fitted to a structural equation model that included mediation of independent and/or interdependent self-construals in the relation between ethnicity and social anxiety.

Given that independent and interdependent self-construals may not fully mediate the relations between ethnicity and social anxiety, six competing mediation models were tested and compared with each other based on model fit indices and chi-square differences (see Table 4): three full mediation models (1A through 1C) and three partial mediation models (2A through 2C). We specified three variations of full mediation models without the direct effect from ethnicity to social anxiety: Model 1A, with the mediation of both independent and interdependent self-construals; Model 1B, with the mediation of independent self-construal only; and Model 1C, with the mediation of interdependent self-construal only (see Table 4). In a similar vein, we specified three variations of partial mediation model with the direct effect from ethnicity to social anxiety: Model 2A, with the partial mediation of both independent and interdependent self-construals; Model 2B, with the partial mediation of independent self-construal only; and Model 2C, with the partial mediation of interdependent self-construal only (see Table 4).

As shown in Table 4, Model 2A, that is, the partial mediation model with both independent and interdependent self-construal as mediators, had better model fit than the other five alternative models. However, the path from interdependent self-construal to social anxiety was not significant, resulting in a nonsignificant, indirect (i.e., mediating) effect of interdependent self-construal (β = .03, p = .06). When the path from interdependent self-construal to social anxiety was fixed to zero in Model 2B (see Figure 1), the model still reasonably fit the data, χ2(3) = 5.58, p = .06; Comparative Fit Index (CFI) = .99; Tucker-Lewis Index (TLI) = .98; root mean square error of approximation (RMSEA) = .03. Moreover, the direct effect of ethnicity on social anxiety decreased from .26 to .20 when independent self-construal was included in the model (Sobel test = 3.38, p = .00). Taken together, Model 2B appears to be the most parsimonious model that fit the data reasonably well. In addition, it appears that the mediating role of independent self-construal may be more salient than that of interdependent self-construal.

The differences in self-construals were expected not only between AH and EH individuals but also between Asian Nationals (people born, raised, and currently residing in an Asian country) and EH individuals. Therefore, to further understand the role of self-construals in ethnic differences in social anxiety, we performed an additional meta-analysis that included 11 studies that also compared social anxiety between Asian Nationals and EH individuals. All 11 studies found higher social anxiety among Asian nationals, of which nine were significant. Similar to the results of main meta-analysis, the overall effect size for these 11 studies (d = .31 (95% CI [.16, .48]) was positive, but was slightly lower.

Acculturation and Social Anxiety

To examine to what extent acculturation may account for variance in social anxiety among AH individuals, a meta-regression analysis was conducted with four studies that included a measure of acculturation (see details in the Method section). The results from our mixed effects model showed that acculturation did not account for significant variance in social anxiety, f2 = .04, p = .85. The effect sizes across the four studies were relatively homogenous, QE (29) = .02, p = .99.

Given the limited number of studies that had measured acculturation, we conducted a follow-up meta-analysis that included studies that specified generational status among AH-heritage individuals. Given that AH individuals of different generations may vary in their acculturation experiences, if acculturative stress is associated with social anxiety, foreign-born (first-generation) AH individuals may differ from U.S.-born (second-generation or greater) AH individuals in their experience of social anxiety. This meta-analysis included eight studies and compared social anxiety between foreign-born (first generation) and U.S.-born (second generation or higher) AH individuals. The results showed that five studies found higher social anxiety among foreign-born AH individuals. The overall effect size was d = −.21 (95% CI [−1.96, 1.55]). Given the wide range of the 95% CI, there was no conclusive evidence to suggest a difference in social anxiety between foreign-born and U.S.-born AH individuals.

Discussion

Ethnic difference in social anxiety between AH and EH individuals has received considerable attention in the last 30 years. However, it was unclear whether this difference was robust, how it varied across assessment method and specific measure of social anxiety, and to what extent factors such as self-construal and acculturation may account for the AH–EH difference in social anxiety. To address these three questions, the current meta-analysis reviewed the extant findings of AH–EH difference in social anxiety from a wide range of studies with various methods and samples (see Table 1).

Overall, the results of meta-analysis indicated that, with the exception of a few epidemiological studies, AH individuals tended to have higher social anxiety than their EH counterparts, and this difference was modest, yet robust, across most study-level variables. These findings correspond to the AH–EH difference in related constructs such as shyness (Paulhus, 2002), and the personality traits of extraversion and neuroticism (e.g., Iwawaki, Eysenck, & Eysneck, 1980; Loo & Shiomi, 1982), which Kotov, Gamez, Schmidt, and Watson (2010) demonstrated, using a meta-analysis, to be related to social anxiety.

The results of meta-analysis also indicated that method of assessment (interview or self-report) or specific measure of social anxiety did not moderate the AH–EH difference in social anxiety, providing further evidence for the robustness of this ethnic difference. Although a few studies that incorporated a diagnostic interview did find a higher prevalence of social phobia diagnosis among European than Asian Americans (Hsu & Alden, 2007), the inclusion of studies that used diagnostic interviews (vs. self-reports) did not seem to change the overall direction and magnitude of the AH–EH difference in social anxiety (i.e., elevated social anxiety among AH than EH individuals) in the meta-analysis. In addition, when we reran the meta-analysis without these epidemiological studies, the overall direction and magnitude of the effect size remained similar. It should be noted, however, that when we only examined studies that incorporated a diagnostic interview (n = 5), the 95% CIs included the null hypothesis, indicating that there was no significant AH–EH difference in social anxiety in these epidemiological studies.

There are several plausible explanations for the divergence between self-report and epidemiological studies. First, although, for the purpose of meta-analysis, the prevalence rates of social phobia diagnosis can be statistically transformed to a metric that is comparable with mean ratings of social anxiety symptoms (Cohen, 1988, pp. 181–182), epidemiological studies focused on categorical diagnosis of social phobia rather than social anxiety symptoms of a continuous nature. It is plausible that, despite the elevated mean levels of social anxiety symptoms, AH individuals are less likely than their EH counterparts to exhibit severe forms of social anxiety and thus are less likely to be diagnosed with social phobia. Second, some aspects of diagnostic interviews are not present in self-report studies, such as rapport between the interviewer and interviewee and the tone in which the questions were asked. Perhaps the most salience difference between diagnostic interviews and self-reports is the distinct nature of interpersonal interaction in these two types of studies. The dynamic process of having a face-to-face interview may lead to some degree of disclosure bias, particularly among AH individuals. AH individuals may hesitate to disclose experiences of social anxiety to the interviewer, possibly because of being concerned about social desirability or “saving face” in front of strangers (Chan, 1998). Finally, it is also possible that compared with self-report questionnaires, diagnostic interviews are simply more accurate in assessing social anxiety. Clearly, these various possibilities need to be explored in future studies.

Given the relative robustness of the AH–EH difference in social anxiety particularly among self-report studies, it was important to understand what factors might account for this ethnic difference. The results of the MASEM analysis showed that independent self-construal partially mediated the ethnic difference in social anxiety, that is, lower independent self-construal among AH individuals may partly explain why, on average, they report higher social anxiety than EH individuals. These results are consistent with findings of an experimental study conducted by Norasakkunkit and Kalick (Norasakkunkit & Kalick, 2009), in which independent self-construal was primed and examined in relation to social anxiety among 127 Japanese nationals and 126 European Americans. Norasakkunkit and Kalick found that, for both groups, participants who were primed for independent self-construal reported significantly lower levels of social anxiety than those who were not primed for independent self-construal. In addition, in their 2 (priming) × 2 (ethnicity) ANOVA analyses, Norasakkunkit and Kalick failed to find significant interaction between ethnicity and priming, suggesting that the success of independent self-construal priming did not vary across ethnic groups. Thus, it seems likely that the relation between independent self-construal and social anxiety may be similar for both AH and EH individuals.

Unlike independent self-construal, the mediation of interdependent self-construal was barely significant. The roles of independent and interdependent self-construals in the AH–EH difference in social anxiety were typically examined separately in previous studies (e.g., Ho & Lau, 2011). In contrast, our MASEM analyses integrated both self-construals in one model, and estimated unique contribution of one type of self-construal while controlling for another. The results of MASEM analyses suggest that although greater interdependent self-construal is often associated with stronger emphasis on social relationships and social evaluations, it does not necessarily lead to higher social anxiety symptoms among AH individuals, providing some initial evidence to depathologize interdependent self-construal that characterizes many Asian cultures. The higher social anxiety symptoms among AH individuals may depend to a greater extent on independent (or lack thereof) than interdependent self-construal.

The results of MASEM analyses also showed that, even with the inclusion of mediation effects of both types of self-construals, the direct effect from ethnicity to social anxiety remained significant, suggesting that there is unexplained variance in this ethnic difference in social anxiety that needs to be further explored. For instance, AH and EH individuals may differ in how they interpret questions on social anxiety and may respond to them differently. There is limited evidence with regard to whether measures of social anxiety had equivalent measurement properties (i.e., measurement invariance; Little, 1997) for both AH and EH individuals. Some researchers have questioned that social anxiety questionnaires often include behavior that may be relatively “normal” in some Asian cultures, but may be perceived as pathological from a Western mental health standard (Norasakkunkit et al., 2009). Some items in these questionnaires may have different cultural meanings for AH and EH individuals, and thus may not function the same way for the two groups. In addition, a reference-group effect may also confound invariance of social anxiety measures. Heine, Lehman, Peng, and Greenholtz (2002) argued that members of a cultural group often compare themselves with members of the same group when giving a decontextualized subjective self-rating, resulting in shifting standards when group comparisons are made.

Similarly, the measurement issue may also apply to the use of diagnostic interviews to assess social anxiety in AH and EH individuals. Given that many Asian cultures stress hierarchical relationships and sensitivity to power differentials, a clinical interview with an adult stranger who may be in a position of authority may warrant deferential behavior and misguide the interviewer in certain areas of the assessment (Yee, Huang, & Lew, 1998). Thus, there is a clear need to explore whether measures and diagnoses of social anxiety function in the same way for AH and EH individuals.

The higher social anxiety among AH individuals may also be related to challenges experienced by AH individuals, such as acculturation, that are associated with being members of immigrant and/or minority groups. However, the results of meta-regression showed that acculturation was not significantly associated with within-AH group differences in social anxiety. In addition, our follow-up meta-analysis indicated that the differences in social anxiety between foreign-born AH individuals (who may experience more stress associated with acculturation than their U.S.-born counterparts) and U.S.-born AH individuals were not significantly different. Another follow-up analysis found that the differences in social anxiety between Asian Nationals and EH individuals were similar to what was found between AH and EH individuals in the main meta-analyses. Because Asian Nationals and AH individuals may theoretically share some cultural characteristics (while not sharing similar experiences of acculturation), it is plausible to speculate that self-construal, particularly lack of independent self-construal, may partly account for the AH–EH difference in social anxiety as well as the difference in social anxiety between some Asian Nationals and EH individuals.

It should be noted that some measurement problems may have contributed to the nonsignificant influence of acculturation on social anxiety among AH individuals. First, acculturative stress was not directly measured; instead, acculturation was used as a proxy for acculturative stress, which may have underestimated explained variance of social anxiety. Second, existing measures of acculturation often fail to capture multidimensional and dynamic nature of acculturative process (see Yoon et al., 2013, for a review). Unilinear measures of acculturation treat acculturation as a unidimensional process ranging from orientation to one’s natal culture to host culture. In contrast, bilinear measures of acculturation (Berry, 1993; Stephenson, 2000) assume that individuals may adopt a bicultural acculturation style that maintains aspects of one’s natal culture while selectively acquiring some of the host culture (Berry, 1993). Thus, one’s orientation to the host culture, such as the use of English and participation in American cultural activities, is not necessarily in conflict with his or her orientation to natal culture. Unfortunately, the measures of acculturation used in the studies reviewed here were all unilinear, thus limiting explanatory power of acculturation in understanding elevated social anxiety symptoms among AH individuals.

Caveats and Future Directions

Several caveats need to be mentioned. First, the majority of the reviewed studies relied on self-reports of “feeling socially anxious,” which, taken together, was more indicative of overall social anxiety symptoms rather than specific subtypes. Although self-reports have been widely used in clinical settings (e.g., De Los Reyes et al., 2012), other indicators of social anxiety, such as whether an individual is “acting anxious” (observed behavior) or experiencing negative repercussions resulting from the feelings of social anxiety (outcome measures), provide important information that can complement the self-report data on social anxiety. For instance, Okazaki et al. (2002) compared Asian American and European American microlevel behaviors (e.g., fidgeting, gaze avoidance) using a 3-min performance task. Although there were ethnic differences in trait measures of social anxiety, as well as emotion ratings before and after the task, there were no group differences in behavioral indicators of social anxiety. These findings suggested that AH–EH differences may not be consistent across all domains; the AH–EH differences we found here may be related to some factors that uniquely impact self-report, but not necessarily observable, behaviors (e.g., response biases).

Other studies have examined the impact of elevated social anxiety among AH individuals on outcome variables related to general psychological well-being, but the findings were mixed. For instance, Abe and Zane (1990) found that foreign-born Asian Americans reported higher levels of social anxiety and psychological maladjustment than U.S.-born Asian Americans, who, in turn, reported higher social anxiety and psychological maladjustment than European Americans, after controlling for social desirability, self-consciousness, other-directedness, and extraversion. In contrast, Ho and Lau (2011) showed that although higher scores on social anxiety predicted lower psychological well-being for foreign-born Asian Americans, U.S.-born Asian Americans, and European Americans, group membership was a nonsignificant predictor of psychological well-being. These inconsistent results demonstrate a need to better understand (a) the relation between social anxiety and psychological well-being within each group, and (b) whether the AH–EH difference in social anxiety is robust across various indicators of social anxiety, and whether social anxiety among AH individuals fits in the nomological network in the same way as it does for EH individuals.

Second, although the current review focused on between-group comparisons in AH and EH individuals, the within-group variations among distinct Asian heritage groups should not be overlooked. The category “Asian” encompasses more than 30 distinct ethnic groups, each with their own unique values, beliefs, culture, language, political and economic conditions, and ethnic, as well as immigration, history (Chin & Kameoka, 2006). To tentatively explore these variations, we conducted follow-up meta-analyses in which individual Asian heritage group was entered as a moderator. The results showed that only Japanese ethnicity significantly moderated the AH–EH ethnic difference in social anxiety. That is, individuals of Japanese heritage, unlike other Asian heritage groups, did not differ as much from EH group in social anxiety, perhaps because of Japanese Americans’ unique history of immigration and acculturation (Y. Xie & Goyette, 2004). Future studies should be cautious when combining multiple Asian ethnicities and comparing them with EH individuals.

Third, the current review relied on meta-analysis, which is itself limited in at least two ways. The inclusion of studies in meta-analysis is susceptible to authors’ subjective bias and publication bias. In addition, although some degree of heterogeneity among individual studies can be accounted for by the effects of moderators, the unexplained variance often raises questions concerning whether the studies can be equally compared. These limitations of meta-analysis must be taken into account when interpreting our findings.

Nevertheless, the results of this meta-analytical review demonstrated the relative robustness of the AH–EH difference in self-reported social anxiety, and highlighted the important role that self-construals, particularly independent self-construal, may play in this ethnic difference. The results of this study may help clinicians better understand ethnic differences in subclinical social anxiety and develop more culturally sensitive treatment goals that do not pathologize values and beliefs related to Asian cultures. Future studies should focus on identifying psychological processes that underlie the relation between lower independent self-construal and higher social anxiety among AH individuals.

Footnotes

1  We used the terms “Asian heritage” (AH) and “European heritage” (EH) to refer to people residing in North America or Australia with either Asian or European ancestry. Thus, our review includes samples of not only Asian/European Americans but also Asian/European Canadians and Asian/European Australians.

2  The current study focused on overall social anxiety symptoms, rather than distinguishing its specific subtypes. However, some studies included in our meta-analysis examined social phobia (one subtype of social anxiety) or specific aspects of social anxiety (e.g., fear of negative social evaluation).

3  Taijin kyofusho refers to a culturally specific form of social phobia observed in East-Asian cultures (see Kleinknecht, Dinnel, Kleinknecht, Hiruma, & Harada, 1997). This form of social anxiety focuses more on the fear of offending others rather than causing embarrassment to oneself. Thus, it has been referred to as “allocentric social anxiety” or “social anxiety: offensive subtype.” Given its relevance to the current study, the term was included as a keyword in our literature search.

4  The prevalence rates of social phobia reported in social epidemiological studies can be statistically transformed to Cohen’s d (Cohen, 1988).

5  Because social epidemiological studies were qualitatively different than the rest, we reran the meta-analysis without these two, and found an overall effect of d = .39 (95% CI [.32, .48]), nearly identical to the effect we found when they were included.

6  Some studies included multiple measures of social anxiety, which made the total number of effect sizes greater than the total number of studies.

7  In addition to the reported Cochran’s Q test, we also performed a permutation test, which has been shown to be more conservative and more conducive to meta-regressions with a single predictor variable and low number of observations (Gagnier, Moher, Boone, Bombardier, & Beyene, 2012). We found the same results with only a slightly elevated (but still significant) p value compared with Cochran’s Q.

8  The wide range of the confidence interval was likely because of the inclusion of two studies, one with a large positive effect size (Okazaki et al., 2002) and one with a large negative effect size (Breslau & Chang, 2006). We reran the meta-analysis without these two “outlier” studies, with the following results: d = .15 (95% CI [−.11, .41]). These results still included the null hypothesis and demonstrated no significant generational difference.

9  We also tested a just-identified model (df = 0) that only estimated the mediation of interdependent self-construal. Although with the exclusion of independent self-construal model, fit cannot be estimated, the indirect (mediating) effect of interdependent self-construal was significant (β = .03; p = .03).

References

References marked with an asterisk indicate studies included in the meta-analysis.

*Abe, J., & Zane, N. (1990). Psychological maladjustment among Asian and White American college students: Controlling for confounds. Journal of Counseling Psychology, 37, 437–444. doi:10.1037/0022-0167.37.4.437

Aderka, I. M., Hofmann, S. G., Nickerson, A., Hermesh, H., Gilboa-Schechtman, E., & Marom, S. (2012). Functional impairment in social anxiety disorder. Journal of Anxiety Disorders, 26, 393–400. doi:10.1016/j.janxdis.2012.01.003

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

*Asnaani, A., Richey, J. A., Dimaite, R., Hinton, D. E., & Hofmann, S. G. (2010). A cross-ethnic comparison of lifetime prevalence rates of anxiety disorders. Journal of Nervous and Mental Disease, 198, 551–555. doi:10.1097/NMD.0b013e3181ea169f

Baker, S. R., & Edelmann, R. J. (2002). Is social phobia related to lack of social skills? Duration of skill-related behaviors and ratings of behavioral adequacy. British Journal of Clinical Psychology, 41, 243–257. doi:10.1348/014466502760379118

Bakermans-Kranenburg, M. J., van Ijzendoorn, M. H., & Juffer, F. (2003). Less is more: Meta-analyses of sensitivity and attachment interventions in early childhood. Psychological Bulletin, 129, 195–215. doi:10.1037/0033-2909.129.2.195

Beidel, D. C., & Turner, S. M. (1999). The natural course of shyness and related syndromes. In L. A.Schmidt & J.Schulkin (Eds.), Extreme fear, shyness, and social phobia: Origins, biological mechanisms, and clinical outcomes (pp. 203–223). New York, NY: Oxford University Press. doi:10.1093/acprof:oso/9780195118872.003.0012

*Berger, L., & Zane, N. (2011). Stigmatized attitudes toward the mentally ill among young adults in college. Unpublished manuscript, Citrus Heights, CA: University of California, Davis.

Berry, J. W. (1993). Ethnic identity in plural societies. In M. E.Bernal & G. P.Knight (Eds.), Ethnic identity: Formation and transmission among Hispanics and other minorities (pp. 271–296). Albany, NY: SUNY Press.

Berry, J. W., Kim, U., Minde, T., & Mok, D. (1987). Comparative studies of acculturative stress. International Migration Review, 21, 491–511. doi:10.2307/2546607

*Breslau, J., & Chang, D. F. (2006). Psychiatric disorders among foreign-born and US-born Asian-Americans in a US national survey. Social Psychiatry and Psychiatric Epidemiology, 41, 943–950. doi:10.1007/s00127-006-0119-2

Brown, T. A., DiNardo, P., & Barlow, D. M. (1994). Anxiety Disorders Interview Schedule for DSM–IV: Client interview schedule. Oxford, UK: Oxford University Press.

Chan, S. (1998). Families with Asian roots. In E. W.Lynch & M. J.Hanson (Eds.), Developing cross-cultural competence: A guide for working with children and their families (pp. 251–344). Baltimore, MD: Brookes.

Chen, X., & Rubin, K. H. (2011). Social, emotional, and personality development in context. In X.Chen & K. H.Rubin (Eds.), Socioemotional development in cultural context (pp. 99–127, New York, NY: Guilford Press.

Cheung, M. W. L. (2014a). Fixed- and random-effects meta-analytic structural equation modeling: Examples and analyses in R. Behavior Research Methods, 46, 29–40. doi:10.3758/s13428-013-0361-y

Cheung, M. W. L. (2014b). metaSEM: An R package for meta-analysis using structural equation modeling. Unpublished manuscript. Retrieved from http://courses.nus.edu.sg/course/psycwlm/internet/metaSEM/metasem.pdf

Cheung, M. W. L., & Chan, W. (2005). Meta-analytic structural equation modeling: A two-stage approach. Psychological Methods, 10, 40–64. doi:10.1037/1082-989X.10.1.40

Cheung, M. W. L., & Chan, W. (2009). A two-stage approach to synthesizing covariance matrices in meta-analytic structural equation modeling. Structural Equation Modeling, 16, 28–53. doi:10.1080/10705510802561295

Chin, D., & Kameoka, V. A. (2006). Sociocultural influences. In F.Andrasik (Ed.), Comprehensive handbook of personality and psychopathology: Adult psychopathology (pp. 67–84). Hoboken, NJ: Wiley.

*Choy, Y., Schneier, F. R., Heimberg, R. G., Kang-Seob, O., & Liebowitz, M. R. (2008). Features of the offensive subtype of taijin-kyofu-sho in US and Korean patients with DSM–IV social anxiety disorder. Depression and Anxiety, 25, 230–240. doi:10.1002/da.20295

Clark, D. M. (2001). A cognitive perspective on social phobia. In W. R.Crozier & L. E.Alden (Eds.), International handbook of social anxiety: Concepts, research, and interventions relating to the self and shyness (pp. 405–430). New York, NY: Wiley.

Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In R. G.Heimberg, M. R.Liebowitz, D. A.Hope, & F. A.Schneier (Eds.), Social phobia: Diagnosis, assessment, and treatment (pp. 69–93). New York, NY: Guilford Press.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

De Los Reyes, A., Aldao, A., Thomas, S. A., Daruwala, S., Swan, A. J., Van Wie, M., . . .Lechner, W. V. (2012). Adolescent self-reports of social anxiety: Can they disagree with objective psychophysiological measures and still be valid?Journal of Psychopathology and Behavioral Assessment, 34, 308–322. doi:10.1007/s10862-012-9289-2

*Dinnel, D. L., Kleinknecht, R. A., & Tanaka-Matsumi, J. (2002). A cross-cultural comparison of social phobia symptoms. Journal of Psychopathology and Behavioral Assessment, 24, 75–84. doi:10.1023/A:1015316223631

Duval, S., & Tweedie, R. (2000). Trim and fit: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysisBiometrics, 56, 455–463. doi:10.1111/j.0006-341X.2000.00455.x

Gagnier, J. J., Moher, D., Boone, H., Bombardier, C., & Beyene, J. (2012). An empirical study using permutation-based resampling in meta-regression. Systematic Reviews, 1, 18.

Good, B., & Kleinman, A. (1985). Culture and anxiety: Cross-cultural evidence for patterning of anxiety disorders. In A. H.Tuma & J. D.Maser (Eds.), Anxiety and the anxiety disorders (pp. 297–324). Hillsdale, NJ: Erlbaum.

*Gordon, T. L., & Teachman, B. A. (2008). Ethnic group differences in affective, behavioral, and cognitive markers of anxiety. Journal of Cross-Cultural Psychology, 39, 424–446. doi:10.1177/0022022108318224

*Hambrick, J. P., Rodebaugh, T. L., Balsis, S., Woods, C. M., Mendez, J. L., & Heimberg, R. G. (2010). Cross-ethnic measurement equivalence of measures of depression, social anxiety, and worry. Assessment, 17, 155–171. doi:10.1177/1073191109350158

*Hardin, E. E. (2002). Depression and social anxiety among Asian and European Americans: The roles of self-discrepancy, optimism, and pessimism. Dissertation Abstracts International: Section B: The Sciences and Engineering, 63(4-B), 2056.

*Hardin, E. E., & Leong, F. T. L. (2005). Optimism and pessimism as mediators of the relations between self-discrepancies and distress among Asian and European Americans. Journal of Counseling Psychology, 52, 25–35. doi:10.1037/0022-0167.52.1.25

Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, FL: Academic Press.

Heimberg, R. G., Brozovich, F. A., & Rapee, R. M. (2010). A cognitive-behavioral model of social anxiety disorder: Update and extension. In S. G.Hofmann & P. M.DiBartolo (Eds.), Social anxiety: Clinical, developmental, and social perspectives (pp. 395–422). San Diego, CA: Elsevier Academic Press.

Heine, S. J., Lehman, D. R., Peng, K., & Greenholtz, J. (2002). What’s wrong with cross-cultural comparisons of subjective Likert scales? The reference-group effect. Journal of Personality and Social Psychology, 82, 903–918. doi:10.1037/0022-3514.82.6.903

Heinrichs, N., Rapee, R. M., Alden, L. A., Bögels, S. M., Hofmann, S. G., Oh, K. J., & Sakano, Y. (2006). Cultural differences in perceived social norms and social anxiety. Behaviour Research and Therapy, 44, 1187–1197. doi:10.1016/j.brat.2005.09.006

*Ho, L. Y., & Lau, A. S. (2011). Do self-report measures of social anxiety reflect cultural bias or real difficulties for Asian American college students?Cultural Diversity & Ethnic Minority Psychology, 17, 52–58. doi:10.1037/a0022533

*Hong, J. J., & Woody, S. R. (2007). Cultural mediators of self-reported social anxiety. Behaviour Research and Therapy, 45, 1779–1789. doi:10.1016/j.brat.2007.01.011

Honjo, K. (2004). Social epidemiology: Definition, history, and research examples. Environmental Health Prevention Medicine, 9, 193–199. doi:10.1007/BF02898100

*Hsu, L. (2004). Patterns of social anxiety and Chinese and European Canadian students(Master’s thesis), Vancouver, B.C., Canada: University of British Columbia. Retrieved from circle.ubc.ca/handle/2429/15575

*Hsu, L., & Alden, L. (2007). Social anxiety in Chinese- and European-heritage students: The effect of assessment format and judgments of impairment. Behavior Therapy, 38, 120–131. doi:10.1016/j.beth.2006.06.006

*Hsu, L., & Alden, L. E. (2008). Cultural influences on willingness to seek treatment for social anxiety in Chinese- and European-heritage students. Cultural Diversity & Ethnic Minority Psychology, 14, 215–223. doi:10.1037/1099-9809.14.3.215

*Hsu, L., Woody, S. R., Lee, H., Peng, Y., Zhou, X., & Ryder, A. G. (2012). Social anxiety among East Asians in North America: East Asian socialization or the challenge of acculturation?Cultural Diversity & Ethnic Minority Psychology, 18, 181–191. doi:10.1037/a0027690

Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis (2nd ed.). Thousand Oaks, CA: Sage.

*Ingman, K. A. (2003). An examination of social anxiety, social skills, social adjustment, and self-construal in Chinese and American students at an American university. Dissertation Abstracts International: Section B: The Sciences and Engineering, 63(9-B), 4374.

Iwawaki, S., Eysenck, S. B., & Eysenck, H. J. (1980). Japanese and English personality structure: A cross-cultural study. Psychologia: An International Journal of Psychology in the Orient, 23, 195–205.

Kessler, R. C., McGonagle, K. A., & Zhao, S. (1994). Lifetime and 12-month prevalence of DSM–III—R psychiatric disorders in the United States: Results from the National Comorbidity Survey. Archives of General Psychiatry, 51, 8–19. doi:10.1001/archpsyc.1994.03950010008002

Kessler, R. C., Stein, M. B., & Berglund, P. (1998). Social phobia subtypes in the National Comorbidity Survey. The American Journal of Psychiatry, 155, 613–619.

*Kleinknecht, R. A., Dinnel, D. L., Kleinknecht, E. E., Hiruma, N., & Harada, N. (1997). Cultural factors in social anxiety: A comparison of social phobia symptoms and taijin kyofusho. Journal of Anxiety Disorders, 11, 157–177. doi:10.1016/S0887-6185(97)00004-2

Kotov, R., Gamez, W., Schmidt, F., & Watson, D. (2010). Linking “big” personality traits to anxiety, depressive, and substance abuse disorders: A meta-analysis. Psychological Bulletin, 136, 768–821. doi:10.1037/a0020327

*Lampert, M. D., Isaacson, K. L., & Lyttle, J. (2010). Cross-cultural variation in gelotophobia with the United States. Psychological Test and Assessment Modeling, 52, 202–216.

*Lau, A. S., Wang, S., Fung, J., & Namikoshi, M. (in press). What happens when you can’t read the air? Cultural fit and aptitude by values interactions on social distress. Journal of Cross-Cultural Psychology.

*Lau, A. S., Fung, J., Wang, S., & Kang, S. (2009). Explaining elevated social anxiety among Asian Americans: Emotional attunement and a cultural double bind. Cultural Diversity and Ethnic Minority Psychology, 15, 77–85. doi:10.1037/a0012819

*Lee, M. R., Okazaki, S., & Yoo, H. C. (2006). Frequency and intensity of social anxiety in Asian Americans and European Americans. Cultural Diversity & Ethnic Minority Psychology, 12, 291–305. doi:10.1037/1099-9809.12.2.291

*Leung, A. W., Heimberg, R. G., Holt, C. S., & Bruch, M. A. (1994). Social anxiety and perception of early parenting among American, Chinese American, and social phobic samples. Anxiety, 1, 80–89. doi:10.1002/anxi.3070010207

Little, T. D. (1997). Mean and covariance structures (MACS) analyses of cross-cultural data: Practical and theoretical issues. Multivariate Behavioral Research, 32, 53–76. doi:10.1207/s15327906mbr3201_3

Loo, R., & Shiomi, K. (1982). The Eysenck personality scores of Japanese and Canadian undergraduates. The Journal of Social Psychology, 118, 3–9. doi:10.1080/00224545.1982.9924411

Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224–253. doi:10.1037/0033-295X.98.2.224

Mesquita, B., & Frijda, N. (1992). Cultural variation in emotions: A review. Psychological Bulletin, 112, 179–204. doi:10.1037/0033-2909.112.2.179

*Miwa, R. (2005). Understanding symptoms of social phobia and taijin kyofusho: A cross cultural comparison of Japanese and United States university students(Dissertation). Available from ProQuest Dissertations and Theses database, 171.

*Norasakkunkit, V., & Kalick, S. M. (2002). Culture, ethnicity, and emotional distress measures: The role of self-construal and self-enhancement. Journal of Cross-Cultural Psychology, 33, 56–70. doi:10.1177/0022022102033001004

*Norasakkunkit, V., & Kalick, S. M. (2009). Experimentally detecting how cultural differences on social anxiety measures misrepresent cultural differences in emotional well-being. Journal of Happiness Studies, 10, 313–327. doi:10.1007/s10902-007-9082-1

*Norasakkunkit, V., Kitayama, S., & Uchida, Y. (2012). Social anxiety and holistic cognition: Self-focused social anxiety in the United States and other-focused social anxiety in Japan. Journal of Cross-Cultural Psychology, 43, 742–757. doi:10.1177/0022022111405658

*Norton, P. J., & Weeks, J. W. (2009). A multi-ethnic examination of socioevaluative fears. Journal of Anxiety Disorders, 23, 904–908. doi:10.1016/j.janxdis.2009.05.008

*Okazaki, S. (1997). Sources of ethnic differences between Asian American and white American college students on measures of depression and social anxiety. Journal of Abnormal Psychology, 106, 52–60. doi:10.1037/0021-843X.106.1.52

*Okazaki, S. (2000). Asian American and white American differences on affective distress symptoms: Do symptom reports differ across reporting methods?Journal of Cross-Cultural Psychology, 31, 603–625. doi:10.1177/0022022100031005004

*Okazaki, S. (2002). Self-other agreement on affective distress scales in Asian Americans and white Americans. Journal of Counseling Psychology, 49, 428–437. doi:10.1037/0022-0167.49.4.428

*Okazaki, S., & Kallivayalil, D. (2002). Cultural norms and subjective disability as predictors of symptom reports among Asian Americans and white Americans. Journal of Cross-Cultural Psychology, 33, 482–491. doi:10.1177/0022022102033005004

*Okazaki, S., Liu, J. F., Longworth, S. L., & Minn, J. Y. (2002). Asian American-white American differences in expressions of social anxiety: A replication and extension. Cultural Diversity & Ethnic Minority Psychology, 8, 234–247. doi:10.1037/1099-9809.8.3.234

Paulhus, D. L. (2002). Socially desirable responding: The evolution of a construct. In H.Braun, D.Jackson, & D.Wiley (Eds.), The role of constructs in psychological and educational measurement (pp. 49–69). Mahwah, NJ: Erlbaum.

R Development Core Team. (2012). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from http://www.R-project.org/

Ruscio, A. M., Brown, T. A., Chiu, W. T., Sareen, J., Stein, M. B., & Kessler, R. C. (2008). Social fears and social phobia in the USA: Results from the National Comorbidity Survey Replication. Psychological Medicine, 38, 15–28. doi:10.1017/S0033291707001699

*Saw, A., & Okazaki, S. (2010). Family emotion socialization and affective distress in Asian American and white American college students. Asian American Journal of Psychology, 1, 81–92. doi:10.1037/a0019638

*Schreier, S. S., Heinrichs, N., Alden, L., Rapee, R. M., Hofmann, S. G., Chen, J., . . .Bogels, S. (2010). Social anxiety and social norms in individualistic and collectivistic countries. Depression and Anxiety, 27, 1128–1134. doi:10.1002/da.20746

Singelis, T. M. (1994). The measurement of independent and interdependent self-construals. Personality and Social Psychology Bulletin, 20, 580–591. doi:10.1177/0146167294205014

*Smith, S. M., Stinson, F. S., Dawson, D. A., Goldstein, R., Huang, B., & Grant, B. (2006). Race/ethnic differences in the prevalence and co-occurrence of substance use disorders and independent mood and anxiety disorders: Results from the national epidemiologic survey on alcohol and related conditions. Psychological Medicine, 36, 987–998. doi:10.1017/S0033291706007690

Stephenson, M. (2000). Development and validation of the Stephenson multigroup acculturation scale (SMAS). Psychological Assessment, 12, 77–88. doi:10.1037/1040-3590.12.1.77

*Sue, D., Ino, S., & Sue, D. M. (1983). Nonassertiveness of Asian Americans: An inaccurate assumption?Journal of Counseling Psychology, 30, 581–588. doi:10.1037/0022-0167.30.4.581

*Sue, D., Sue, D. M., & Ino, S. (1990). Assertiveness and social anxiety in Chinese American women. The Journal of Psychology: Interdisciplinary and Applied, 124, 155–163. doi:10.1080/00223980.1990.10543212

Sutton, A. J., & Higgins, J. P. T. (2008). Recent development in meta-analysis. Statistics in Medicine, 27, 625–650. doi:10.1002/sim.2934

*Swift, A. (2001). Implications of culturally based understandings of the self for the psychological health of Caucasian and Asian American women(Dissertation). Available from Proquest Dissertations and Theses database.

Turner, S. M., Beidel, D. C., Dancu, C. V., & Stanley, M. A. (1996). Scale development and psychometric properties. SPAI: Social Phobia and Anxiety Inventory. North Tonawanda, NY: Multi-Health Systems.

Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of Statistical Software, 36, 1–48.

Watson, D., & Friend, R. (1969). Measurement of social-evaluative anxiety. Journal of Consulting and Clinical Psychology, 33, 448–457. doi:10.1037/h0027806

Wittchen, H. U., & Beloch, E. (1996). The impact of social phobia on quality of life. International Clinical Psychopharmacology, 11(Suppl. 3), 15–23. doi:10.1097/00004850-199606003-00004

*Xie, D., Leong, F. T. L., & Feng, S. (2008). Culture-specific personality correlates of anxiety among Chinese and Caucasian college students. Asian Journal of Social Psychology, 11, 163–174. doi:10.1111/j.1467-839X.2008.00253.x

Xie, Y., & Goyette, K. (2004). A demographic portrait of Asian Americans. Washington DC: Russell Sage Foundation. Retrieved from http://www-personal.umich.edu/~yuxie/Research/brief/Xie-Goyette.pdf

Yee, B. W. K., Huang, L. N., & Lew, A. (1998). Families: Life-span socialization in a cultural context. In L.Lee & N.Zane (Eds.), Handbook of Asian American psychology (pp. 83–135). Thousand Oaks, CA: Sage.

Yoon, E., Chang, C., Kim, S., Clawson, A., Cleary, S. E., Hansen, M., . . .Gomes, A. M. (2013). A meta-analysis of acculturation/enculturation and mental health. Journal of Counseling Psychology, 60, 15–30. doi:10.1037/a0030652

*Zhang, H. (2012). The effect of primary and secondary control on social anxiety in Latino and Asian American college students. Unpublished senior thesis, Trinity College, Hartford, CT.

*Zheng, X., & Berry, J. W. (1991). Psychological adaptation of Chinese sojourners in Canada. International Journal of Psychology, 26, 451–470. doi:10.1080/00207599108247134

*Zhong, J., Wang, A., Qian, M., Zhang, L., Gao, J., Yang, J., & Chen, P. (2008). Shame, personality, and social anxiety symptoms in Chinese and American nonclinical samples: A cross-cultural study. Depression and Anxiety, 25, 449–460. doi:10.1002/da.20358

Submitted: December 19, 2013 Accepted: April 21, 2014