Children with Autism Spectrum Disorders (ASD) manifest a triad of deficits, in the areas of socialization, communication, and repetitive and stereotyped behaviors (American Psychiatric Association, 2000; Volkmar, 2005). Tailoring interventions to address these issues is important when working with children with ASD. It is critical to remember that although children with ASD share common deficits, each child is entirely unique. Interventions must be tailored to each child's specific needs and interests.

The current study assessed the effectiveness of peer-mediated intervention on increasing social skills in a sample of 6, 5 to 9 year old children with ASD. The intervention was run class-wide, during students' recess time. Although perhaps the most naturalized setting to practice socialization goals, recess remains one of the least utilized limes to address goals during the school day. Peer-mediated intervention addresses goals in the most naturalistic of settings, actual peer interactions, while minimizing cost in teacher time and expense to the district.

Participants were paired and instructed to play with typically developing buddies. Buddy pairs were rotated to promote exposure to multiple peer exemplars. Intervention occurred in 3 weekly sessions over 6 weeks, for a total of 18 sessions. Prior to intervention, 2 weeks of baseline data was collected on participants' social bids. Social bids were adapted from those assessed in Baumringer's (2002) study. Doctoral students were trained in the fidelity of treatment, and once 80% reliable on the operationally defined behaviors, collected data on the target dyads. Data was collected on frequency of social bids and percent of time engaged with peers during baseline, intervention, and at follow-up. Social bids were defined as any behaviors that functioned to maintain or initiate peer interaction (Baumringer, 2002). In addition to direct observational data, teachers and parents completed The Social Skills Rating System (SSRS; Gresham & Elliot, 1990) prior to and following intervention. The study utilized a single subject (N=6), ABA design. Results were assessed via visual inspection of data. Busk and Serlin's (1992) techniques were also used to monitor effect sizes for each participant.

It was hypothesized that following the intervention, participants would demonstrate: (1) an increased frequency of social bids; (2) an increase in percent of time spent engagement with peers; and (3) increased levels of overall social skills from pre to post assessment as measured by the SSRS-Teacher Form. Results demonstrated strong support for Hypothesis 1. Five of the 6 participants demonstrated increases in the frequency of bids. Results also demonstrated strong support for Hypothesis 2. All participants demonstrated significant increases in percent engagement with peers. Hypothesis 3 was not supported; analysis of teacher ratings did not indicate consistent gains in social skills. Implications for future research are discussed.