Subjects were drawn from 300 students identified by school officials as being at substantial risk of dropping out. Independent variables included treatment and sex (and for one analysis, age). Dependent variables were teacher-assigned class grades, teacher-reported work habits, cooperation, class attendance and punctuality, persistence in school, and student report of attitudes and self-perceptions. Participating students also completed a program evaluation questionnaire at the end of the workshop. The remaining members of the group of 300 were used as nontreatment comparisons to the treatment group. Premeasures and postmeasures of school performance and participation variables were probed with analyses of covariance. The treatment effects were generally negative but not significant. In addition to school performances, measures of attitudes and self-perceptions were obtained using the Wisconsin Youth Survey. These scales suggest plausible explanations for the lackluster outcomes of the intervention, chiefly substantial bonding of group members to each other and their ensuing alienation from regular school activities. 3 tables and 33 references. (Author abstract modified)