The purposes of this study are (1) to find out the levels of smoking, drinking, and using illicit drugs (inhalants, marijuana/hashish, amphetamine-type stimulants, and MDMA) among Japanese university students between gender (G; male & female), age group (AG; non-adult & adult), and part-time work experience (W; student with/without experience), (2) to determine the optimal classified subgroups for predicting cigarette and alcohol use according to these three variables and estimating the risk in each subgroup, and (3) to identify the group with the highest risk for substance abuse among university students. Data were obtained from the Japanese Youth Survey Project on Alcohol and Other Drugs (JYPAD) conducted in 2007. The sample was 672 respondents, composed of 661 undergraduate and 11 graduate students at 4-year universities from randomly selected residents aged 18-22 in the Kanto region of Japan. The results indicated (1) that the risks for cigarette smoking were significant (p ＜ .01) for gender, age group, or work experience, and (2) that the risks for alcohol drinking were significant (p ＜ .001) for age group or work experience. (3) A logistic regression model identified gender and work experience (G ＊ W) as the optimal model for predicting cigarette use. According to this model, the highest risk group was male students who worked part-time jobs during the past year. (4) Additionally, the optimal model for predicting alcohol use was age group and work experience (AG ＊ W). There was a much smaller gender difference for university alcohol use. The highest risk group for alcohol use was university students aged 20―22 who worked part-time jobs. These results suggest that a survey for substance abuse among high school graduates should be conducted with a larger representative sample.
