Objective: This study compared two group treatments for aggressive behavior in adjudicated male adolescents: a cognitive-behavioral program (low process) and a program integrating group interaction with cognitive-behavioral training (high process). Method: The court-ordered adolescents were divided into two age groups, younger and older, and randomly assigned to one of the two treatment conditions for 10 weekly 2-hour sessions. Data were collected at pretest, posttest, and 9-month follow-up. Dependent variables included recidivism, behavior ratings, parent-teen conflict, anger control, social problem-solving, and social skills. The analysis was ANOVA on the ranks of adjusted gain scores. Results: High process was more effective for the older boys, and low process was more effective for the younger boys on conduct disorder and parent-adolescent conflict. Conclusions: The older boys appeared more cognitively and socially able to take advantage of the self-determined, interactive format, whereas the younger boys seemed to respond better to the leader-guided, highly structured format. (Contains 6 tables.)