The aim here is to examine the factorial structure, internal consistency, and concurrent validity of the Korean version of the Anxiety Sensitivity Index-3 (K-ASI-3) in student samples in Korea. Also, we investigated the cross-cultural differences in the Social Concerns factor.
K-ASI-3 was administered to non clinical samples in Korea. Internal consistency, exploratory factor analysis (EFA), and confirmatory factor analysis (CFA) were undertaken to examine the factorial structure and reliability of the K-ASI-3.
Results from CFA comparing our data to factor solutions commonly reported as representative of European-American samples indicated an adequate fit. The K-ASI-3 showed good performance on the indices of internal consistency and concurrent validity. In addition, using regression analyses, we found the Social Concerns factor is most strongly related to life satisfaction and worry. However, we found no evidence that Korean college students express more Social Concerns than their European Caucasian counterparts.
The authors demonstrate that the K-ASI-3 has highly internally consistent and psychometrically sound items, and that it reliably measures three lower-order domains assessing Physical, Social, and Cognitive Concerns.
Anxiety sensitivity (AS) refers to the fear of anxiety-related sensations that are interpreted as having potentially undesirable somatic, psychological, or social consequences.
Researchers have identified three lower-order domains of concerns related to AS. The lower-order dimensions represent Physical Concerns, Cognitive Concerns, and Social Concerns. Some people are more apt to fear the physical symptoms of AS, believing that these symptoms are signs of physical illness. For example, a person might be worried by shortness of breath, thinking that it may lead to fainting or suffocating. Other people are more likely to fear cognitive symptoms of AS, due to beliefs that anxiety symptoms like derealization (an alteration in the perception or experience of the external world so that it seems strange or unreal) are signals of mental disorder. For example, some people are more concerned about the difficulty to focus their attention on task believing that they are going crazy. Still others may fear publicly observable symptoms due to beliefs that displaying anxiety will result in embarrassment, social criticism, and public ridicule. These people would be perplexed if others took notice of their nervousness or anxiety.
Research on these three factors has elucidated the nature of the relationship between AS and several types of anxiety-related disorders. For example, Physical Concerns are most strongly associated with panic disorder. Cognitive Concerns are more apt to be related to depression, whereas Social Concerns are most strongly associated with the fear of negative evaluation and to social phobia.
Culture may exert important influences on the experience and expression of AS.
According to the previous studies, East Asians in general report higher social anxiety than do Western people.
Although we expect that Koreans would have more Social Concerns, Social Concerns should not be immediately interpreted to psychopathology. The doorstep for negative implications associated with Social Concerns might be higher in Korea. Thus, it is requisite to examine if Social Concerns dimension may contribute more to the subjective well-being and general distress than the other dimensions of AS do in Koreans.
In order to assess AS, Reiss, Peterson, Gursky, and McNally
A possible reason for inconsistencies in factor structure of the ASI might be associated with its small number of items. Majority of its 16 items are related to Physical Concerns whereas few items assess Cognitive and Social concerns. In addition, some of them do not target any specific domain. For example, it is ambiguous which factor is related with the item: "It is important for me to stay in control of my emotions".
It was in an endeavor to find a solution with the problems in the ASI that the Anxiety Sensitivity Index-Revised (ASI-R)
Based on the existing research findings, it was first necessary to devise an AS scale with stable factor structure across different cultures, nations, and languages. Taylor et al.
The aims of this study are 1) to evaluate the factorial validity of the Korean version of the ASI-3 (K-ASI-3), 2) to evaluate the concurrent validity of the K-ASI-3, 3) to evaluate the role of the Social Concern subscale of the K-ASI-3 in predicting life satisfaction and worry, 4) to evaluate the internal consistency of the K-ASI-3, 5) to examine if Koreans express more Social Concerns than their European-American counterparts.
Though it is necessary to investigate samples with clinically significant AS, researches using non-clinical samples are also requisite. The appropriateness of and need for using non-clinical samples when studying AS is bolstered by study indicating that AS has a dimensional latent structure.
Participants consisted of 761 college students recruited from introductory psychology courses at two Universities in Seoul. All of the participants were between 17 and 32 years of age (Mean age=21.43 years, SD=2.59). 55% of the participants were female. No data are reported on the clinical background of these participants.
356 of the 761 college students participated in the exploratory factor analysis. The participants were between 18 to 30 years of age, and 67% of them were female (Mean age=21.50 years, SD=2.29). Of 761 participants, 264 college students participated in the confirmatory factor analysis. The participants were between 18 to 42 years of age (Mean age=22.97 years, SD=3.01). Approximately 70% of them were female. Participants consisted of 142 college students recruited for investigating the concurrent validity and testing the role of the Social Concerns subscale of the K-ASI-3 in predicting life satisfaction and worry. We hypothesized that Social Concerns subscale would explain more variance of life satisfaction and worry than the other subscales of the K-ASI-3. All of the participants were between 17 and 22 years of age (Mean age=18.29 years, SD=0.81). 35% of the participants were female.
Items were selected out of the pool of 36 K-ASI-R
The satisfaction with life scale (SWLS), developed by Diener et al.
The Center for Epidemiologic Studies-Depression Scale (CES-D)
The Penn State Worry Questionnaire (PSWQ) is a 16-item questionnaire that assesses excessive and uncontrollable worry.
The Brief Fear of Negative Evaluation Scale
Informed consent was obtained first, and then students completed the battery in a classroom setting, during class time. Researchers were available during questionnaire completion to answer any questions. The questionnaires took approximately 20 minutes to complete via pen and pencil.
Prior to analysis, data were examined for departures from normality. The Kolomogorov-Smirnov Test was applied to examine the skewness and kurtosis of the distribution. Since data departed from normal distribution (e.g., for the K-ASI-3 item 10, skewness=3.05 and kurtosis=10.02), and consequently, the latent variable analyses were performed using robust maximum likelihood (MLM) in Mplus 2.02.
To explore the relationship between the K-ASI-3 subscales and the other variables, we used Spearman ρ correlations. Given the number of correlations being tested, p values were set at 0.006 to control for experiment-wise error (the Bonferroni approach to multiple significance tests was utilized, so an initial alpha of 0.05 was divided by the number of measures or 0.05/8). To determine the internal consistency reliability of the K-ASI-3 total scale and subscales, we used Cronbach's alpha, with thresholds of alpha at or above 0.70.
Three subscales of the original ASI-3 are moderately correlated with each other. Therefore an oblique (promax) rotation was used. The number of factors to retain was evaluated using (1) Kaiser's eigenvalue greater than 1 factor extraction rule,
Four factors possessed eigenvalues greater than one (6.00, 1.95, 1.79, 1.01). From the scree analysis, we estimated that one to three factors were necessary to explain the data, but the one- and two-factor models were obviously not enough to explain the data fully (
CFA was performed to investigate the construct validity of the K-ASI-3 in a Korean student sample. Based on the EFA results, our hypothetical factor model contained three latent factors labeled Physical, Cognitive, and Social Concerns. The goodness of fit of the CFA models was evaluated using the Root Mean Square Error of Approximation (RMSEA),
The analysis examined the three-factor model of the K-ASI-3. An acceptable model fit was obtained for a three-factor solution [χ2(132)=215.430, TLI=0.932, CFI=0.941, RMSEA=0.049]. Further, both the standardized residuals and the modification indices suggested that no significant improvement to the structure could be proposed. Thus, the CFA finding cross-validated the three-factor structure suggested in the EFA.
In order to examine the relationship between the K-ASI-3 domains and the other variables, Spearman's correlations were assessed (
Regression analyses were conducted to test the role of the Social Concerns subscale of the K-ASI-3 in predicting life satisfaction and worry. We hypothesized that Social Concerns subscale would explain more variance of life satisfaction and worry than the other subscales of the K-ASI-3.
Separate hierarchical regression analyses were conducted, predicting scores on the K-PSWQ and the K-SWLS. During each analysis, Social Concerns was entered in the second step, and Physical Concerns and Cognitive Concerns were added in the third step. We also conducted regression analyses with the order of entry of the K-ASI-3 factors reversed.
Social Concerns was entered in the second step of the regression equation predicting the K-PSWQ and was significant, t(133)=4.78, p<0.001. Physical Concerns and Cognitive Concerns did not account for significant additional variance in the K-PSWQ, beyond that accounted for by Social Concerns, when it was entered on the third step, t(131)=-0.45, p=0.64, t(131)=1.88, p=0.06, respectively, and Social Concerns was still significant, t(131)=3.50, p<0.01 (
Cognitive Concerns domain was entered in the second step of the regression equation predicting the K-PSWQ and was significant, t(132)=3.04, p<0.01. Physical Concerns dimension was also entered in the second step of the regression equation predicting the K-PSWQ and was insignificant, t(132)=0.19, p=0.84. Social Concerns accounted for significant additional variance in the K-PSWQ, beyond that accounted for by Physical Concerns and Cognitive Concerns, when it was entered on the third step, t(131)=3.50, p<0.05, and Cognitive Concerns was no longer significant, t(131)=1.88, p=0.06 (
Social Concerns was entered in the second step of the regression equation predicting the K-SWLS and was significant, t(134)=-3.21, p<0.01. Physical Concerns and Cognitive Concerns did not account for significant additional variance in the K-SWLS, beyond that accounted for by Social Concerns, when it was entered on the third step, t(131)=-0.42, p=0.67, t(131)=-1.11, p=0.26, respectively, and Social Concerns was still significant, t(131)=-2.12, p<0.05 (
Cognitive and Physical Concerns were entered in the second step of the regression equation predicting the K-SWLS and was insignificant, t(133)=-1.87, p=0.06, t(133)=-0.82, p=0.41, respectively. Social Concerns accounted for significant additional variance in the K-SWLS, beyond that accounted for by Physical Concerns and Cognitive Concerns, when it was entered on the third step, t(132)=-2.12, p<0.05. In the final model, Physical, Social, and Cognitive Concerns accounted for 12.6% of the variance in the K-SWLS, F(4, 131)=6.84, p<0.01 (
Internal consistency tests gave a Cronbach's alpha of 0.87 for the global scale, with an alpha of 0.73 (6 items) for the Physical Concerns subscale, 0.83 (6 items) for the Social Concerns factor, and 0.86 (6 items) for the Cognitive Concerns domain. All factors exceeded the 0.70 level considered as an acceptable internal consistency. Item-total correlations ranged from 0.37 to 0.63 for the Physical Concerns subscale, from 0.41 to 0.70 for the Social Concerns factor, from 0.54 to 0.72 for the Cognitive Concerns domain, and from 0.25 to 0.64 on the entire scale (
Effect size differences were calculated to examine if Korean college students express more Social Concerns than their U.S. counterparts. The U.S. male sample
The present study represents the first evaluation of the psychometric properties and the latent structure of the Asian (Korean) translation of the ASI-3. Results from our EFA and CFA displayed that a three-factor solution provided a suitable fit for the present data. Consistencies in the factor structure are shown between the current study and the Taylor et al.'s study.
The correlations between the K-ASI-3 and other self-report measures were examined. Scores on the total K-ASI-3 were moderately correlated with measures of depression, worry, life satisfaction, and negative evaluation sensitivity. In addition, the K-ASI-3 Social Concerns factor was more strongly associated with negative evaluation sensitivity than the other factors, while the ASI-3 Cognitive Concerns factor was more related to depression than the other dimensions.
We found the Social Concerns factor is most strongly related to life satisfaction and worry. However, contrary to the expectation, the current findings displayed that the Social Concerns domain scores of the Korean college students were not higher than those of the European American college students. These results are not in line with the previous findings that East Asians generally endorse greater social anxiety than do Westerners and that social anxiety is related to AS
The present study has several important limitations. First, the study was conducted with university students. Therefore, we should be cautious about generalizing these findings to other samples, and they should be replicated with a more representative sample from the general population. Second, only self-reporting data was included, and thus relationships between study variables may have been inflated by questionnaire-specific method variance. Third, the results of this study can be potentially applicable to Koreans residing in the metropolitan area in Korea. The study area should include the other regions in Korea in order to get a more comprehensive picture of the general population. Lastly, socioeconomic status was not evaluated, thus it is possible that variability in our results may be caused by SES differences.
This research suggests that the K-ASI-3 appears to be a promising measure of AS. The K-ASI-3 was determined to be composed of highly internally consistent and psychometrically sound items. It seems to reliably measure three lower-order domains assessing Physical, Social, and Cognitive Concerns. Future studies with this scale could give a more comprehensive picture of the role of AS in the development and maintenance of mood, anxiety, and substance-related disorders in Korea.
Goodness-of-fit indices for the K-ASI-3 models: Exploratory Factor Analysis
K-ASI-3: Korean version of the Anxiety Sensitivity Index-3, RMSEA: root mean square error of approximation, RMR: root mean square residual
Promax rotated loadings (3 factor model: students sample)*
*salient loadings (≥0.30)
Zero-correlations among the study measures for study participants (N=142)
*p<0.05, ***p<0.001. K-ASI-3: The Korean version of the Anxiety Sensitivity Index-3, K-PSWQ: The Korean version of the Penn State Worry Questionnaire, K-BFNE: The Korean version of the Brief Fear of Negative Evaluation Scale, K-SWLS: The Korean version of the Satisfaction with the Life Scale, K-CES-D: The Korean version of the Center for Epidemiologic Studies Depression Scale
Summary of regression analysis for variables predicting K-PSWQ (N=142)
**p<0.01, ***p<0.001. K-PSWQ: The Korean version of the Penn State Worry Questionnaire
Summary of regression analysis for variables predicting K-SWLS (N=142)
*p<0.05, **p<0.01. K-SWLS: The Korean version of the Satisfaction with the Life Scale
Mean, standard deviation, correlation of each K-ASI-3 item with the sum of the other items and internal consistency if the item is deleted
K-ASI-3: Korean version of the Anxiety Sensitivity Index-3
Means and standard deviation for the western (N=462)
*effect size. ASI-3: Anxiety Sensitivity Index-3