Aim: Assessing social competence is important for clinical and preventive interventions of depression. The aim of the present paper was to examine the factor structure of the Japanese Interpersonal Competence Scale (JICS).
Methods: Exploratory and confirmatory factor analysis was performed on the survey responses of 730 participants. Simultaneous multigroup analyses were conducted to confirm factor stability across psychological health status and sex differences.
Results: Two factors, which represent Perceptive Ability and Self-Restraint, were confirmed to show a moderate correlation. Perceptive Ability involves a more cognitive aspect of social competence, while Self-Restraint involves a more behavioral aspect, both of which are considered to reflect the emotion-based relating style specific to the Japanese people: indulgent dependence (amae) and harmony (wa). In addition, Self-Restraint may be linked to social functioning. Both constructs may confound a respondent's perceived confidence.
Conclusion: Despite its shortcomings, the JICS is a unique measure of social competence in the Japanese cultural context.
The relationship between social competence and depression has been well established.1 Depressed people are likely to evaluate themselves as less socially competent2–4 and often experience interpersonal difficulties.5 Such difficulties generate further interpersonal conflicts, which are linked to negative life events (e.g. marital disruption), and which result in deterioration of their depressive symptoms.6,7 Moreover, deficient social competence can undermine the acquisition and utilization of an effective social support system. Zeiss and Lewinsohn reported that depressive symptoms in patients with low social skills were less improved by cognitive–behavioral therapy, and suggested that therapeutic relationships with therapists were insufficiently established due to their low social skills.8 In line with this, depression prevention programs, which include a method to enhance social competence, have been reported as substantially effective.9
Depression and anxiety are often comorbid.10–15 But previous studies investigating the relationship between social competence and anxiety are limited because they have mainly focused on social anxiety and phobia in children and adolescents.16–19 Smari et al. have indicated that social anxiety is strongly related to a low perceived social competence and a high threat of social appraisal by peers.20 Although this knowledge may be useful for anxiety disorders in adulthood, several important points such as the continuity of clinical and non-clinical anxiety and developmental changes by age have yet to be identified.
Several cross-cultural studies have noted social competence as a protective factor against psychological maladaptation. People who involve cultural transition inevitably experience acculturative stress during their relocation,21–23 leading to the onset of depressive and anxiety disorders. Social competence significantly predicted the extent of psychological stress and the effectiveness of coping strategies in the international students.24 In addition, social competence in intercultural communication can be developed through psychoeducational training programs.25 Thus, assessment of social competence is important to researchers and mental health practitioners who are interested in providing effective interventions to immigrants and international students.
Measures of social competence
The concept of social competence involves the ability of an individual to effectively and appropriately interact with other people.26 A number of inventories have been developed to assess social competence.27–31 These instruments assess various aspects of social competence: assertiveness, empathy, role taking, ambiguity tolerance, behavioral flexibility, and interaction management. But caution should be exercised when using such measures in non-Western contexts. Sociocultural factors partly define one's relating style and methods to control the social environment.32,33 A relating style that may be effective and appropriate in one cultural setting, may not be appropriate in another. Some investigators have noted the importance of social competence in a specific cultural setting.34,35 Therefore indigenous cultural factors involving social competence should be carefully considered to assess social competence in cross-cultural setting.
Social relationship in the Japanese culture
Social relationship in the Japanese culture dates back to 1946 when Benedict wrote The Chrysanthemum and the Sword, in which she noted that many characteristics of Japanese behaviors contradict each other.36 She wrote, ‘All these contradictions, however, are the warp and woof of books on Japan (p. 2)’. An anthropologist, Nakane, noted, ‘The vertical relation, which we predicted in theory from the ideals of social group formation in Japan, becomes the actuating principle in creating cohesion among group members (p. 26)’.37 In a vertical society, group membership is homogeneous and requires harmony (wa) in the group based on the respect of a superior by a subordinate. This is accompanied by the hierarchical consciousness of the group members. The superior is expected to know without explicit communication what the subordinate wishes to express or desire. Expressing his or her own ideas and wishes, particularly for a subordinate, may be seen as arrogant or a violation of the societal norm.
Another key issue in understanding Japanese people is amae, which means dependency on others. Doi explained that amae is the noun form of the verb ‘amaeru’, which means ‘to depend and presume upon another’s benevolence'.38 These two words, amae and amaeru, have the same root as amai, which is an adjective meaning ‘sweet’. Thus, amae is accompanied by the sweet feeling between the two, which implies a strong attachment to the other.39Amae and amaeru are ubiquitous in Japanese interpersonal relationships from children to adults. These words are colloquial expressions used to describe interpersonal relationships and personality. Resistance to separate from a parent-figure is strong and a sense of unity with such a figure gives unutterable delight, which makes amae an important psychic organization in the Japanese. With this feeling one expects that others (whom he or she depends on) will guess what he or she feels and thinks. Consequently, Japanese interpersonal communication is more ambiguous and based more on emotion than interpersonal communication compared to Western societies. Persons who are depended upon should have the intuitive capacity to infer what is inside their subordinate's minds.40 These specific faculties of social competence are difficult to measure using an instrument developed specifically for people with a Western cultural background.
Assessment of social competence in the Japanese cultural setting
The assessment of social competence in a Japanese population must be considered in a cultural context. The Japanese Interpersonal Competence Scale (JICS) was developed to assess social competence in the Japanese cultural setting.41 The JICS consists of 22 items rated on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree), and generate the five subscales: Perceptive Ability, Self-Restraint, Hierarchical Relationship Management, Interpersonal Sensitivity, and Ambiguity Tolerance (Appendix I). The alpha coefficients were 0.64–0.80 and the 6-week test–retest reliabilities were 0.64–0.81. The JICS indicated moderate correlations to other social competence scales, which were developed for Japanese and non-Japanese individuals (rs = 0.42–0.66).41,42 One crucial concern is the factor structure. The five subscales yielded by an exploratory factor analysis (EFA) are assumed to have orthogonal dimensions but, given that the JICS assesses each dimension of the social competence in a person, such constructs should be correlated to some extent. Therefore, the factor structure of the JICS should be examined in terms of dimensional correlations.
The JICS has a great potential to be used in cross-cultural setting. The JICS may provide a marker of psychological adaptation of immigrants and international students. Social competence specific to Japanese culture may be linked to acculturation of culturally diverse immigrants in Japan, and in a similar way, of Japanese people outside Japan. But, before application to a cross-cultural matter, the construct of the JICS should be examined in a Japanese population. Such procedure provides a basis for the validity of the JICS across cultures. Moreover, this procedure allows us emic and etic perspectives to understand relationships between social competence and specific health issues in cross-cultural setting. Murphy noted that researchers who study cultural phenomenon should be required to have insightful knowledge of their own culture.43 Such knowledge can reduce each individual's own cultural bias.
We collected data from two groups: psychiatric outpatients and university students. This sampling allowed us to investigate the construct and magnitude of social competence assessed by the JICS. Given that social competence is linked to depression, the outpatient group was expected to have lower social competence than the student counterpart. But the construct of the JICS was expected to indicate invariant factor structure between the two groups because the core of Japanese cultural values in the JICS (wa and amae) appears to be shared among Japanese people to some extent. Thus we treated the data from the two groups as a single dataset, except during subgroup analysis.
The purpose of the present study was to test the factor structure of the JICS using exploratory and confirmatory factor analysis (CFA) in a Japanese population.
The research was conducted from 17 June 2003 to 30 September 2005 in the Tokyo metropolitan area. The data of the outpatient group were collected in a psychiatric clinic for 1 month. A set of questionnaire was distributed to 800 outpatients. Of those, 220 outpatients responded. They consisted of 96 men, 124 women, and three outpatients who did not report their sex. The mean age was 35.2 ± 9.9 years. Outpatients with dementia, mental retardation, and alcohol or drug abuse were excluded. Correspondingly, a set of questionnaires was distributed to 573 students in a class. Of those students, 12 turned down the participation. Twenty foreign students were excluded due to their poor command of Japanese. Available data were collected from 541 Japanese students, which consisted of 212 men and 329 women. The mean age was 20.0 ± 2.6 years. The mean age of the outpatients was significantly higher than for the student group (t(218) = 23.4, P < 0.001), and the mean age of the men was higher than that of the women (men, 25.4 years; women, 23.1 years, t(257) = 3.3, P < 0.001). Two-way analysis of variance indicated a significant interaction in the mean age (F1,716 = 34.6, P < 0.001), but significant differences were not indicated in the sex ratio between the outpatient and student groups (χ2(1) = 0.153, n.s.). Only the participants with complete JICS data were included; the participants with missing data (eight and 20 outpatients and students, respectively) and three outpatients with undetermined sex were excluded. Thus, a total of 730 participants consisting of 209 outpatients and 521 students were analyzed.
A questionnaire that contained the JICS, items tapping demographic features, and other items not reported in this study, was distributed in a cross-sectional manner to outpatients as they attended a psychiatric clinic. Each outpatient was asked to complete and return his or her questionnaire in a self-addressed stamped envelope. In addition, the questionnaire was distributed to students in psychology classes and returned to the researchers. In both settings, incentives were not provided. Each participant's self-determination to participate in the study and the anonymity of the responses were maintained. This project was approved by the ethical committee of Kumamoto University Graduate School of Medical Sciences (Institutional Review Board: Approved No. 4, epidemiology).
Before beginning a series of factor analyses we randomly divided the sample groups in half (group 1, n = 365; group 2, n = 365). A plausible model was refined by EFA in group 1 and subsequently examined on CFA in group 2. In the EFA, the principal factor method was used to extract factors, which allows a theoretical solution uncontaminated by unique and error variability to be obtained.44 The number of appropriate factors was determined by an eigenvalue above unity,45 the scree test,46 and interpretability of the factors. The substantial threshold of the factor loading in each item was determined as ≥0.40.47 Subsequent CFA were performed to identify the optimal model. Simultaneous multi-group analyses between the outpatients and students and between the two sexes were also conducted to test the factor stability. The maximum likelihood estimation method was adopted to produce standardized parameter estimates. Consistent with common practice, the model fits were evaluated using five indicators: χ2, comparative fit index (CFI),48 standardized root mean square residual (SRMR),49 root mean squared error of approximation (RMSEA),50 and the Akaike information criterion (AIC).51 Although the χ2 is the most common fit test, it is almost always statistically significant for models with large samples. CFI >0.90 is an acceptable fit, while a value >0.95 fits the data well. An SRMR <0.08 and an RMSEA <0.06 indicate good fits. A lower AIC indicates a better fit among a class of competing models but the AIC does not assume a true model, rather it tries to identify the optimal model. All statistical analyses were performed using SPSS 13.0 (SPSS, Chicago, IL, USA) and AMOS 6.0 (Smallwaters, Chicago, IL, USA).
Descriptive statistics of the subscales
Table 1 shows the means, standard deviations, and Cronbach's alpha coefficients in the JICS subscales. All the subscales, except for Self-Restraint, in the subgroup analyses did not indicate significant differences between the outpatient and student groups. The mean score of the Self-Restraint subscale in the students was significantly higher than those in the outpatients (t(743) = 3.5, P < 0.001); gender did not significantly affect any of the subscales.
Table 1. JICS subscale characteristics (n = 730)
JICS, Japanese Interpersonal Competence Scale.
Hierarchical relationship management
A principal factor analysis with promax rotation yielded two factors for group 1. The first five eigenvalues were 4.11, 2.43, 1.20, 0.92, and 0.84, suggesting two factors in the scree test and three factors in the Kaiser criterion. A sample size >200 participants provides a fairly reliable solution for the scree test,52 while the Kaiser criterion is accurate when all communality estimates after extraction are >0.70.47 The present result included a relatively low final communality estimate. Thus we chose a two-factor solution (Table 2). These factors were considered to represent Perceptive Ability and Self-Restraint. Items 1, 2, 16, 17, and 23 were excluded due to the very low final communality estimates (<0.16),53 item 22 was excluded due to a dual loading (0.34 and 0.36 for Perceptive Ability and Self-Restraint, respectively).
Table 2. Factor loadings of the JICS for the two equally divided subgroups (group 1, n = 365)
Final communality estimates
Bold, factor loadings with absolute values ≥0.40.
Exploratory factor analysis was performed for the first group of participants who were randomly assigned into two groups.
Explained variance (%)
Correlation of factors
Using the data from group 2, CFA was used to examine the possible models. The two-factor model refined in the EFA indicated an acceptable fit for group 2, but the modification indices strongly suggested error covariance between items 14 and 31. When the error covariance was added, the initial model was greatly improved and produced CFI, SRMR and RMSEA of 0.92, 0.06, and 0.05, respectively. Upon assuming the third factor, which consisted of items 14 and 31, the model fits deteriorated slightly (0.91, 0.07, and 0.06 for the CFI, SRMR, and RMSEA, respectively). Figure 1 shows the factor loadings of the final model. All the factor loadings were significant (Ps < 0.001). Hence, we cross-validated this model to group 1. Acceptable fits were obtained; CFI, SRMR, and RMSEA were 0.91, 0.06, and 0.06, respectively. Table 3 shows the goodness of fits for the competing models in group 2. Of those, the two-factor model had the best fit to the present data. The AIC was 272.6, which was lowest among the models. Cronbach's alpha of Perceptive Ability was 0.81 for both groups 1 and 2, while Self-Restraint was 0.73 and 0.75 for group 1 and group 2, respectively.
Table 3. Fit indexes of the proposed models for the two subgroups (group 2, n = 365)
P < 0.001.
All factors are independent.
A model proposed by exploratory factor analysis.
AIC, Akaike information criterion; CFI, comparative fit index; RMSEA, root mean squared error of approximation; SRMR, standardized root mean square residual.
Confirmatory factor analysis was performed for the second group of participants who were randomly assigned into two groups.
Simultaneous multi-group analysis provided substantial factor stability (Table 4). Model A was the baseline model used to test the common factor pattern, and the magnitude of the factor loadings was allowed to vary. Model B assumed that the corresponding factor loadings between the two groups were equal. Model C was the same as model B except that the respective common factor variance for the two groups was assumed to be equal. All the χ2 in the outpatient and student groups comparison did not indicate significant increments between model A and B (χ2(16) = 24.428, n.s.), or between A and C (χ2(1) = 0.254, n.s.). All the χ2 in the men and women groups comparison did not indicate significant increments between model A and B (χ2(16) = 21.365, n.s.), or between A and C (χ2(1) = 0.012, n.s.).
Table 4. Fit indexes of the invariance of the JICS across the subgroups
P < 0.001.
Model A is factor pattern invariance; model B is factor loading invariance; model C is strong factorial invariance. AIC, Akaike information criterion; CFI, comparative fit index; RMSEA, root mean squared error of approximation; SRMR, standardized root mean square residual.
The mean scores of the two subscales in the present study were evaluated in terms of health status, sex, and age (Table 5). The Perceptive Ability subscale did not show significant differences across the subgroups. In contrast, the mean score of the Self-Restraint subscale in the outpatients was significantly lower than that of the students. Significant interactions between sex and group were not indicated. Moreover, the mean scores of the Self-Restraint subscale significantly differed by age. Multiple comparison on Dunnett's test found that the mean score for participants who were ≤24 years was significantly higher than the mean score for participants between 35 and 44 years (mean square of error = 26.88, P < 0.05). Other differences by age groups were not indicated.
The aim of the present study was to examine the factor structure of the JICS. A series of factor analyses identified two factor structures, representing Perceptive Ability and Self-Restraint with a moderate correlation.
This study indicates that a two-factor model best fits the present data, unlike the original study, which indicated a five-factor structure. The EFA and CFA in the present study assumed an oblique solution, while the EFA in the original study assumed an orthogonal solution. Thus, the discrepancy may be due to the differing methodologies of the two studies. In the present study Perceptive Ability consisted of the original six items plus items 24 and 31. Item 24 involves the ability to perceive how others' feeling can be hurt, while item 31 refers to the ability to perceive when and to whom polite expressions are necessary. Both of these items can involve cognitive judgment in an indirect communication style. Moreover, Self-Restraint consisted of the original six items plus item 14. Item 14 refers to the daily practice of polite expression, and can be linked to social skills used to avoid interpersonal conflicts with a superior or supervisor. The moderate correlation between the two factors extracted in the present study suggests a shared construct, which may reflect core values of Japanese relating styles: indulgent dependence (amae) and harmony (wa). Indulgent dependence expects Japanese individuals to share others' feelings, intentions, and ideas as a part of oneself,54 which may carry forward to develop Perceptive Ability to infer what others are thinking without being told. Harmony expects Japanese individuals to learn social skills such as deference to superiors, use of polite expressions, and forbearance of one's opinion in public, which may contribute to development of Self-Restraint to maintain a cohesive group. Thus, Perceptive Ability involves a more cognitive aspect, while Self-Restraint involves a more behavioral aspect, both of which may reflect an emotion-based relating style.
The items excluded in the present study were used to constitute the Interpersonal Sensitivity and Ambiguity Tolerance factors in the original study. These items had low communality estimates when they were assumed to be correlated, suggesting that the Interpersonal Sensitivity and Ambiguity Tolerance constructs may have an insufficient link to social behaviors specific to Japanese people. Rather, these factors may reflect the perception of interpersonal attractiveness and assertiveness in general communication. But the possibility of a three-factor model remains. The model with the third factor, which consists of items 14 and 31, also has an acceptable fit (data not shown). The error covariance between items 14 and 31 suggests a shared construct discrete from the two factors in the present study. Hence, a further study is needed to validate the JICS.
The two factor structure in the present study indicated no significant difference across the subgroups. This result suggests that the constructs of the JICS are relatively stable according to psychological health status and sex differences. Therefore, Perceptive Ability and Self-Restraint factors may provide a certain part of the basic pattern of Japanese social behavior. The outpatient group had a significantly lower mean Self-Restraint score than the student group, but the mean Perceptive Ability scores were similar for both groups. Hence, Self-Restraint may be linked to social functioning. The social skills utilized in an actual behavior, rather than the social cognition underlying the behavior, may be linked to one's effective and appropriate social behaviors. The mean Self-Restraint and the Perceptive Ability scores decrease with advancing age, although significance was not reached for the Perceptive Ability scores. This contradicts the expectation that these scores would increase with age because social competence is acquired through accumulated interpersonal experiences. One possible explanation is that these constructs in this study may confound respondents' perceived confidence. The developmental task of young adult is to establish intimate relationships with peers and to balance relationships with adults, including their superiors and supervisors. Thus, individuals in the young adult age group may be more sensitive in managing interpersonal relationships than other age groups.
Limitations of the present study should be considered. First, the JICS is evaluated in a self-report manner. Social desirability may bias the subjective social competence. A multi-group CFA using the data from a partner's report may rectify this matter. Second, the JICS was assessed in a cross-sectional design. It remains to be elucidated whether the Japanese culture influences people's attitude and interpersonal competence or vice versa.55 A longitudinal study is required to identify the dynamic formation of social behavior and indigenous culture. Third, the Japanese culture specified in the JICS may be covered by a broader Asian culture.56 A validation study using the data from culturally diverse groups can be helpful. Finally, the present results are limited in representation of the target population. The sampling in the present study focused on the urban area, in which temporary relocation of the population occurs, and is dominated by young adult students, who are assumed to have internalized a more Westernized Japanese culture. The factor structure of the JICS may be swayed when including data from people in rural areas and when adjusting the proportion of age groups. Further replication is needed before reaching a conclusion. This limitation can be addressed by using a more diverse aged sampling covering a broader settled area.
In summary, the JICS in the present study is constituted by Perceptive Ability and Self-Restraint factors with a moderate correlation. These factors stem from the emotion-based relating style specific to the Japanese people. Although further study is needed, the JICS is a unique measure of social competence in the Japanese cultural context.
Japanese Interpersonal Competence Scale
Reverse coded item.
Residual item in the original item pool (not included in the body of JICS).