Evaluations were generally of poor quality and did not provide evidence that substance abuse education reduces drug use. This study used meta-analysis to review a carefully selected sample of evaluations in this area. Outcomes of 33 evaluations were statistically integrated. The meta-analysis showed that typical substance abuse education had its most positive effects on knowledge and attitudes, but was unsuccessful in changing the drug-using behaviors of students. Attitudinal effects were significantly higher when peers were used as instructional leaders and when group discussion was part of the instructional method. Furthermore, students who volunteered for substance abuse education reported lower drug use after treatment than did students who were required to participate in such programs. (Author abstract modified)