Data obtained from the 1988 and 1990 student, parent, and administrator surveys of the National Education Longitudinal Study yielded a sample of 8,931 students in 693 schools. The dependent variable was the student's report on whether students were able to make friends with students of other racial and ethnic groups in the school. This variable was measured while students were in the 10th grade. The individual-level variables were intended to control for individual-level factors that might have an impact on attitudes about interracial relations in schools. In an effort to measure student academic status within the school, the study included a set of dummy variables that contained information on the student's course of study or track. School-level independent variables focused on four features of the formal and informal social organization of schools that may influence the quality of interracial relations in the school: tracking differentiation, ethnic diversity, school size, and consolidated inequality at the neighborhood level. The latter variable quantified the extent to which Whites were advantaged over non-Whites in employment, BA attainment, high school completion, having two-parent families, and living above the poverty line. This variable was included because the conflation of race and resource differences increases the strength of the boundaries that students must cross to be friendly or at least tolerant of one another. The study conducted four sets of analyses. The first was a within-school analysis that included only student-level variables. The second analysis was a between-school analysis, which examined the effects of contextual characteristics on the school mean racial climate. The third analysis combined Level 1 and Level 2 equations to examine the effects of school characteristics, controlling for the students who attended those schools, as well as the individual-level effects, controlling for the characteristics of the schools they attended. The fourth analysis explored some cross-level interactions. The study found that students were more likely to observe interracial friendships in schools with high levels of ethnic diversity, which apparently increased opportunities for interracial contact. This effect was not as strong, however, in schools with high enrollments, thus indicating that intergroup contact tends to decrease as relative group size increases. Students were less likely to observe interracial friendships if their school was characterized by a high level of tracking, particularly for those students involved in vocational and technical courses of study. The existence of tracking thus served to decrease opportunities for interracial friendship, particularly for those students in lower status tracks. This practice, which is intended to improve the quality of instruction, has an impact not only on student achievement but also on the social climate of high schools. This has the effect of undermining the aims of desegregation in schools. Suggestions for future research are offered. 4 tables, 1 figure, 12 notes, and 38 references