The twofold purpose of this study was to investigate the reliability and construct validity of scores on the Japanese version of an academic self-concept scale titled the Dimensions of Self-Concept (DOSC) Form H and ascertain any relationships between scores on the DOSC scale and selected demographic variables, including class, gender, and self-reporting academic achievement. Confirmatory factor analysis served as a method of investigation. The original English version of the DOSC scale is a 96-item self-report questionnaire comprising 6 subscales, each representing 1 of the 6 hypothesized dimensions of academic self-concept: (1) Level of Aspiration (ASP); (2) Anxiety (ANX); (3) Academic Interest and Satisfaction (AIAS); (4) Leadership and Initiative (LAI); (5) Identification versus Alienation (IA); and (6) Stress (STR). A total of 177 students at a small private college in Kanazawa, Japan completed the Japanese version of the measure. Internal consistency factor analysis revealed that among several alternative models, the a priori six-factor model afforded the best fit. Comparisons of scores by demographic variables revealed a significant difference between males and females on scores of the LAI subscale, between high achievers and low achievers on scores of the ASP, AIAS, and LAI subscales, and between freshmen and sophomores on scores of the ASP, AIAS, IA, and STR subscales. (Contains 3 tables and 22 references.) (Author/SLD)