THE FACTOR STRUCTURE OF THE 28-ITEM GENERAL HEALTH QUESTIONNAIRE WHEN USED IN JAPANESE EARLY ADOLESCENTS AND ADULT EMPLOYEES - AGE-CULTURAL AND CROSS-CULTURAL COMPARISONS

The principal-component analysis (PCA) with varimax rotation on the 28-item General Health Questionnaire (GHQ-28) in Japanese translation was performed separately for two samples of the Japanese population: early adolescents (junior high school students) and adult employees. The initial PCA yielded different number of components across sub-samples. The later PCA, however, with restriction of the number of components four generally produced a similar structure across sub-samples by means of visual inspection and the coefficient of factor similarity between the components calculated by using the loadings matrices, while a slight different structure emerged for the employees aged 40-49 years. Then, based on these two ways of comparison, the factor structure of GHQ-28 was compared between the present Japanese samples and the European and Turkish school-aged adolescent populations with three ethnic backgrounds. The results showed that its structure was highly stable across age (generation) and several cultural backgrounds, at least among these nations. The internal consistency reliability of the GHQ-28 was at a high level among the present sample of Japanese.