A quasi-experimental approach was used to study the effect of different types of summer programs in Alberta (Canada) on the skills and behaviors of severely handicapped children ages 2 1/2 to 18 years. The experimental group, composed of 98 students, were enrolled in 6-week summer programs characterized as high structure (emphasizing specific educational skill and goal-oriented behavioral change), medium structure (mixed educational-recreational), or low structure (recreational). The control group (103 children) were enrolled in no program at all, or in short (2-4 week), primarily recreational summer programs. Data were collected on program characteristics and on child-child and staff-child interactions. The Inappropriate Behaviors Observation Checklist was developed to assess student behavior, and the Teachers' Interview Skills Checklist to assess changes in skill levels. Among the reported results were that programs designated as high structure and medium structure produced the greatest improvement in skills and behavior, with little difference between the two types in terms of results. Recommendations included: (1) the provision of full 6-week summer programs for all severely handicapped children; (2) increased staff preparation; (3) a greater degree of normalization; and (4) a follow-up study to measure retention of skills and behaviors over a longer period of time. (Author/JW)