In a recent report, positive outcomes were described for use of a prevention program serving students with behavioral and emotional disturbance and those at risk for emotional disturbance (ED) in urban elementary schools. This primarily school-based prevention program consisted of social skills activities, peer tutoring, and individual and classwide behavior management components. Improved behaviors and decreased aggression were noted for the experimental group in treatment for approximately I½ school years, compared to a control-wait group.This report is a further analysis of the prevention program investigation, including (a) findings from implementation across multiple school years with longitudinal student outcome data; (b) replication effects for the prevention program for the second cohort; (c) effects of variations in program implementation; specifically, strength of treatment and classroom structure; and (d) individual cases of treatment successes and failures. Results for two cohorts indicated that inappropriate behaviors decreased (i.e., aggression, out of seat, negative verbal behaviors) and positive behaviors increased (academic engagement, behavioral compliance) under conditions of strong implementation of program components and in the context of high classroom structure. Individual case studies further confirmed that students exposed to more treatment in each year of the program and in classrooms with more structure had better outcomes. Implications of these findings supporting early, multiyear, high-quality prevention and the adverse effects of weak treatment efforts and low structure are discussed in the context of current practice and the need for additional research.