The participants included 791 males and females in 1974 and 791 males and females in 1975. They completed a self-administered questionnaire on their psychosocial adjustment at least 1 year before leaving school. Results revealed that, as expected, school, family, behavioral, social, and personality variables could all predict dropping out of school in the two samples. Furthermore, these predictors were quite stable over time. However, statistical improvement measures in logistic regression analyses indicated that the best screening variables for potential dropouts were school experience variables, including grade retention, school achievement, and school commitment. Other psychosocial variables made significant contributions but did not much improve the capacity to identify who will drop out of school. Findings indicated that an extensive questionnaire is not necessary for identifying potential dropouts. Instead, a short questionnaire that asks about grade retention, school achievement, and commitment is sufficient to distinguish most of the at-risk students. This finding is of particular interest for those who want to implement a general secondary prevention program in their school and for time and money are limited. Tables and 45 references (Author abstract modified)