A survey of 312 teachers and families of students with severe disabilities in California, Colorado, Kentucky, Utah, and Virginia was designed to measure the extent of implementation of 12 educational "best practices" and the level of demonstration of 7 student outcome variables. The educational practices included: integration into school and family activities, integration into community recreational activities, functional educational program, data-based instructional programs, community-based instruction, transdisciplinary model, teacher integration, program reflecting respect for students, teacher inservice training, teacher education, teacher experience, and principal involvement. Additionally, three demographic variables--the type of community, family socioeconomic status, and the age of the student--were included as factors which may be strongly associated with student performance. Student outcome variables included social skills development, communication skills development, positive affective demeanor, percent of Individualized Education Program objectives achieved, level of independence, positive nondisabled peer attitude, and positive parental expectations for their child's future. Results of a series of multiple regression analyses indicated that a single program variable, the degree to which students were integrated into school and family activities, was significantly associated with each one of the high priority student outcome variables. (Contains 41 references.) (Author/JDD)