This paper summarizes research in which the public health models of epidemiological assessment and primary prevention were used to (1) determine the prevalence of ethnic social segregation among school children in Toronto, Canada; and (2) alter ethnically segregated friendship patterns in Toronto schools. A preliminary study of children aged 12 to 14 years and belonging to Anglo majority or to Chinese or Italian minority groups, documented ethnic segregation as a fact of life for many Toronto school children. In a second study, an experimental group of sixth graders was divided into heterogeneous learning teams which worked cooperatively to complete assignments. Pre- and post-experiment measures were taken from a sociometric self report and an attitude to different others scale for the experimental group and a control group. Results indicated the positive effects of using small cooperative learning teams to increase cross ethnic friendship and to prevent racist behavior. (Author/MJL)