The purpose of this study was to clarify the subgroup differences in the association between participation in school-based extracurricular activities and exercise and levels of cigarette, alcohol, or marijuana use during one's lifetime and in the past year. This study also sought to determine the optimal classification of subgroups based on four variables (participation in school-based extracurricular activities, exercise, gender, and school year). Data consisted of a combined sample (aggregate sample) from the Japanese School Survey Project on Alcohol and Other Drugs (JSPAD) conducted in 2004, 2006, and 2009. The aggregate sample consisted of 75,726 1st-, 2nd-, and 3rd-year students (37,697 males and 38,029 females) at full-time high schools nationwide that were selected by stratified, single-stage cluster sampling during each survey. The aggregate sample combined randomly selected samples of the same size from each year the survey was conducted. Results indicated that 63.4% of the entire sample participated in school-based extracurricular activities, 64.8% of males did so, and 61.9% of females did so. Results also indicated that 66.4% of the entire sample exercised outside of PE