Effects of high-probability requests on the social interactions of young children with severe disabilities.

High-probability requests were used to increase social interactions in 3 young boys with severe disabilities who had been identified as severely socially withdrawn. A multiple baseline design across participants was used to evaluate the effects of high-probability request intervention on (a) social initiations, (b) social responses, (c) continued interactions, and (d) performance of high- and low-probability requests. The students were observed in a second setting to examine generalization effects across peers who did not participate in the training sequence and settings. The results demonstrated that the high-probability requests increased the students' responsiveness to low-probability requests to initiate social behavior. Increases were also found in (a) unprompted initiations and extended interactions to the training peers, (b) unprompted initiations and extended interactions to peers who were not involved in the training procedure, and (c) generalized unprompted initiations and interactions in a second nontraining setting. The students maintained increased levels of initiations and interactions after all prompts were removed from both the training and nontraining settings.