The effects of a conflict resolution program on school climate, student self-image, and the use of conflict resolution skills were studied in urban schools in Canada. Benefits and limitations of conflict resolution in comparison with other types of violence prevention programs and methods of evaluating violence prevention programs were also studied. The context was grade-5 classrooms of 3 French and 4 English elementary schools in 2 urban areas of a large Canadian city, for a total of 140 students and their teachers. The evaluation design was based on the Key Features Model of J. S. Renzulli (1975) using a quasi-experimental pretest-posttest control group design. Data sources included student and teacher questionnaires and structured interviews with students and teachers. The conflict resolution program, which featured accepting and respecting difference and skills for self-control and communication, was delivered over 10 weeks. Observations supported the positive effects of the program on school climate. Students reported increased self-confidence and an increase in the use of skills specific to conflict resolution. Teachers reported a decrease in interference with teaching, but the value perceived by teachers for the program did not change with the intervention. Implications for future evaluations of violence prevention programs are discussed. (Contains 5 tables and 39 references.) (SLD)