PSYCHOMETRIC PROPERTIES OF THE 30-ITEM VERSION GENERAL HEALTH QUESTIONNAIRE IN JAPANESE

Psychometric properties of the 30-item General Health Questionnaire (G HQ-30) were explored using the Japanese data collected in various sett ings. A six- or seven-factor structure emerged for the entire sample a nd each gender and age-group divided into a 10-year interval. The fact ors were highly identical across the subsamples. The main four factors showed a high agreement with those of UK and Hong Kong studies: appro ximately 80% of the items were similarly allocated to factors. However , the response endorsement of two ''loss of positive attitude'' items was much common regardless of the score level. These items counted up approximately one-fourths of the score, so that these items made the s core spuriously higher. Accordingly, a question arose as to whether th ese items should be included in the scale for its use in Japanese.