Meta-analysis finds that peer-mediated and video-modeling social skills interventions are effective for children with autism, but caution is urged in interpreting the results

摘要

Design: Meta-analysis of single subject research using HLM methodology. Data sources: The authors searched the following online databases: Ovid Medline, PsyclNFO, ERIC, Web of Science, and TOC Premier databases. In all, 64 articles were identified for further selection and coding. Study selection and assessment: The authors identified studies for the meta-analysis using the following inclusion criteria: (a) the research focused on peer-mediated or video modeling as the main intervention/ independent variable, (b) the research was published between 1994 and January 2008; (cj at least half or more of the participants were children diagnosed with ASD; (d) the dependent variables of the study were related to social skills; and (e) the researchers used a single-case research design to evaluate the effects of the intervention on the dependent variables. Studies that included mixed methods (such as a combination of single-case and group designs) were included only if the individual information and outcome graphs for the participants in the single-case part of the research study were provided. Studies that included peer mediated or video-modeling intervention techniques as part of a multicomponent intervention were not included due to concerns about the confounding effects of multiple treatments. The first author identified 14 studies for further analysis using the criteria listed above. The reliability of study inclusion decisions was not assessed or discussed. The second part of "Study selection and assessment" is concerned with data extraction. Specifically, what data did the researchers extract from each study (e.g., participant diagnosis, age, setting, etc.), including an appraisal of the quality of the individual studies. Again, there was no information provided on the reliability of data extraction decisions-that is, whether the variables were identified by two separate coders and, if so, how reliable the coders were in the data extraction process. Outcomes: The results of the studies were quantified in terms of effect sizes to describe individual intervention effects. Additionally, an HLM analysis was conducted to determine the significant effects of peer-mediated/video modeling interventions with children with ASD and the moderating effects of age and type of intervention on intervention outcomes.