Randomized Comparison of a Peer-Mediated Social Intervention to ‘Business as Usual' Social Programming for Young Elementary Students with Autism

Background:  This proposal presents results of a randomized control trial comparing a Peer-Mediated Intervention (PMI) to a business as usual social curriculum on children’s social communication in early primary grades. Research over several decades has shown positive effects of PMI’s for improving social communication and interactions between children with autism and peers without autism (Chan et al., 2009; Strain et al., 1979). The majority of research to date has been single-subject designs. Exemplary models exist for (a) teaching social-communication using modeling, picture cues, and scripts; (b) use of naturalistic strategies such as joint action routines; and (c) the inclusion of peers as change agents (Goldstein, 2002; Goldstein, Schneider & Thiemann, 2007; Kamps et al., 2002). Unfortunately, a large number of children with ASD continue to be served in segregated special education settings; and when they do have access to typically developing peers it is generally during settings that are less structured and therefore more difficult to socially navigate (e.g., recess, lunch, and PE).

Objectives:  

(1)  To conduct a randomized trial to determine if students with ASD who experience a PMI have higher levels of social-communication with peers than a group of children with ASD who receive the usual school curriculum.

(2)  To measure generalization of social communication to non-treatment settings.

Methods:  

Fifteen students with autism, and 34 peers were enrolled in year 1. The PMI includes: (1) direct teaching of communication targets (e.g., ask and share, comments); (2) child-adult practice; (3) child-peer practice with adult feedback; and (4) play activities with peer and adult prompts. Each lesson includes written and picture cues of social language targets. Groups meet for 30 min, 3 times/ week, and include one child with ASD and 2 peers. School staff is trained on implementation of the PMI. Fidelity measures are used to monitor teacher implementation across schools. Rates of behaviors in treatment sessions and generalization settings are collected through direct observation three times per year.

Results:  

Nineteen of the 21 school staff trained as implementers had treatment fidelity averages above 80%. Data analyzed from year 1 showed the children receiving the PMI increased their rates of initiations and responses from an average of 10 to 22 during 10 min probes, and peer rates increased from 9 to 24. Generalization data in non-treatment settings was less compelling. Higher functioning students showed steady improvement in skill use during treatment; adaptations for lower functioning students were necessary. Social validity ratings measuring teacher perceptions of changes in social behaviors improved for 10 of 13 children in the PMI group. The ‘business as usual’ group showed minimal changes in social communication skills across the pre- mid- and post-social probes.

Conclusions:  

The PMI was effective in increasing communication between children with autism and peers without disabilities. Modifications are necessary for some students with lower language. Generalization to unstructured free-play or center activities was limited. Solutions for this challenge have included (a) mini- whole class social lessons, (b) decreasing group structure, (c) posting scripts in non-treatment settings.