Purpose. The hypothesis tested was that experimental subjects, relative to controls, would demonstrate significant increases in school performance and decreases in drug involvement at program exit (5 months) and at follow-up (10 months). Design. A two-group, repeated-measures, intervention trial was the design used. Setting. The study involved four urban Northwest high schools. Subjects. Participants included 259 youth at high risk of potential school dropout, 101 in the experimental group and 158 in the control group. Intervention. The Personal Growth Class experimental condition was a one-semester, five-month elective course taken as one of five or six regular classes. It had a 1:12 teacher-student ratio, and integrated group support and life-skills training interventions. The control condition included a regular school schedule. Measures. School performance measures (semester GPA, class absences) came from school records. Drug use progression, drug control, and adverse consequences were measured by the Drug Involvement Scale for Adolescents. Self-esteem, school bonding, and deviant peer bonding were measured using the High School Questionnaire: Inventory of Experiences. All multi-item scales had acceptable reliability and validity. Results. As predicted, trend analyses revealed significantly different patterns of change over time between groups in drug control problems and consequences; in GPA (but not attendance); and in self-esteem, deviant peer bonding, and school bonding. The program appeared to stem the progression of drug use, but group differences only approached significance. Conclusion. Program efficacy was demonstrated particularly for decreasing drug control problems and consequences; increasing GPA and school bonding; and desired changes in self-esteem and deviant peer bonding. Program effects on progression of drug use were less definitive.