This paper reviews published observational research that has addressed the impact of race and sex on student peer interactions and describes two studies that focused on the classroom behaviors of black and white sixth grade boys and girls. The studies were conducted in an urban middle school in the northeastern United States, a school that had been advertised as a model of integrated education. The first study consisted of coded observations of classroom behavior. The patterning of peer interactions implied was then analyzed by race and sex. The second study consisted of a sociometric analysis conducted to validate the interaction patterns implied by the behavioral data. Data from both the behavioral and sociometric studies indicated the overriding importance of gender as a grouping variable and the noticeable, though less pronounced, impact of race (or correlated variables) upon interactant choice. This paper concludes with a lengthy discussion of the implications of interaction data for the integration of schools. (GC)