Forty second and third grade children identified by their teachers as aggressive were randomly assigned to a child problem solving skills training (PSST) or teacher consultation. PSST consisted of 20 45-minute sessions of three to five children, held twice weekly at school. Teacher consultation involved four 30-minute individual sessions. Posttreatment effects were investigated in a 2 (group) by 2 (time) factorial design. Dependent measures included classroom observations, parent and teacher behavioral ratings, sociometric measures, measures of social-cognitive skills, and self-concept. Both interventions resulted in improvement at posttreatment in peer ratings of social competence, teacher ratings of aggression, classroom on-task behavior, and parent ratings of externalizing problems. Improvement in social-cognitive skills mediated behavior improvement for PSST children only. Children whose mothers reported more accepting attachment histories showed the most behavioral improvement. Follow-up data consisting of teacher and parent behavior ratings were completed 12 months post-intervention and found no group differences. (Author)