Research on the effectiveness of school-based programs for preventing or reducing aggressive behavior was synthesized with a meta-analysis. Changes in aggressive behavior between pretest and posttest were analyzed for developmental patterns and characteristics associated with differential effects. Control groups showed little change in aggressive behavior, but there were significant reductions among intervention groups. Most studies were conducted on demonstration programs; the few studies of routine practice programs showed much smaller effects. Among demonstration programs, positive outcomes were associated with a variety of study, subject, and intervention characteristics. Most notably, higher risk youth showed greater reductions in aggressive behavior, poorly implemented programs produced smaller effects, and different types of programs were generally similar in their effectiveness, other things equal.