A peer-mediated intervention with clinic-referred socially isolated girls. Generalization, maintenance, and social validation.

Peer-mediated interventions have been widely employed with seriously impaired autistic and handicapped children to modify social withdrawal. This study extends the literature by examining the impact of a peer intervention on the interactions of developmentally normal, socially isolated girls. Two to three actual classroom peers were trained to serve as helpers to initiate and maintain interactions with the subjects. The intervention was introduced sequentially across 2 girls in a multiple-baseline design, and a within-subject ABA withdrawal design was used to assess maintenance. Behavioral observations during recess periods indicated that both children's positive social interactions with peer helpers and other classmates were increased significantly during intervention and were maintained in return-to-baseline conditions and at 4-month follow-up. Increases in positive social interactions of both subjects generalized to a second recess setting, in which intervention was not introduced. Subjects' interactions in both recess settings reached levels comparable to those of social-comparison groups of peers. Teacher and self-report ratings suggested that both girls had fewer social problems and experienced less loneliness and dysphoria as a result of the intervention.