Fourteen sites attempting to integrate 245 severely handicapped students (aged 3-22) with non-handicapped children were examined. Trained observers recorded social interaction between handicapped and non-handicapped students, incorporating macroscopic and microscopic views. Adaptive behavior of each S was also measured, as was teacher and school support for integration. Additional data were gathered from Individualized Education Programs (IEPs), measures of non-handicapped students' attitudes, state level information requested by the U.S. Department of Education, analysis of state certification practices, and responses to a previously conducted survey of state support for the severely handicapped. Results indicated that integration defined in terms of social behavior occurred in each of the sites, with the degree of integration statistically related to information about six factors: antecedent factors of the state and local education agency, people present in the environment when integrated, organization of the environment, rate of social input from non-handicapped students, degree of support for interaction from the teacher, and the teacher's educational planning process. The social behavior of severely handicapped Ss differed depending on whether the behavior was directed to other severely handicapped students or to non-handicapped students. The degree of integration was related both to the proportion of objectives achieved on the target students' IEPs and to the attitudes of non-handicapped students. Effects on severely handicapped and non-handicapped students revealed that integration can be a positive part of the educational program for severely handicapped students. (CL)