The Kansas City School Behavior Project was an experiment intended to enhance the social-emotional development of individual pupils through treatment of mild behavioral disturbances and primary prevention programs, utilizing the teachers in their roles as group leaders. The teachers in the experimental group received summer training in the methods and techniques of small group interaction, and were given support through the school year by project staff in weekly meetings. An epidemiologic follow-up indicated more positive teacher ratings of work habits and fewer absences. Data from family files demonstrated prevention effects, whereas school data varied, showing limited positive effects and some negative effects. Community data from police and juvenile court tend to confirm the prevention effect, though police data's positive effect wore off in the final year. Recommendations for future implementation and evaluation of this approach are suggested. (KS)