A quantitative review of social skills interventions for children and adolescents with autism spectrum disorders.

摘要

Children and adolescents with autism spectrum disorders (ASDs), also known as pervasive developmental disorders (PDDs), experience severe social skills impairments throughout their development, including deficits in nonverbal social skills (e.g., eye gaze, joint attention, affect recognition and production, gestural recognition and production), social communication skills (e.g., initiations, responses, engagement, conversations, requests, compliments, greetings), and social play skills (e.g., initiations, responses, engagement, sharing, turn taking). This problem is of particular importance, given the recent and expected increases in estimated prevalence rates, mandated special education services, and costs of ASDs to society. Due to the variability of ASDs, many specific types of social skills interventions have been developed. Previous qualitative reviews highlighted promising interventions. However, they lacked a universal, quantitative metric of comparison. Bellini, Peters, Benner, and Hopf (2007) used the percentage of nonoverlapping data (PND) statistic (Scruggs, Mastropieri, & Casto, 1987b) to quantitatively examine treatment effects. However, their study was restricted by a number of key limitations. The present review builds upon the foundation of previous reviews, in an effort to conduct the largest, most comprehensive quantitative review of social skills interventions for children and adolescents with ASDs to date. Searches through the end of 2009 identified 168 single-subject design target studies with 482 eligible, school age participants. A median intervention PND score of 78% ( n = 167) indicated effective interventions, as treatments produced positive effects between the baseline and intervention phases. A median maintenance PND score of 98% (n = 63) indicated very effective maintenance of positive treatment effects between the baseline and maintenance phases. A median generalization PND score of 75% (n = 60) indicated effective generalization of treatment effects across different settings, social skills, and treatment providers. Differences across study characteristics (i.e., publication types), participant characteristics (i.e., diagnosis types, age groups, intelligence levels), and treatment characteristics (i.e., specific types, settings, formats, locations) are compared. Strengths of the present review include methodological improvements in research design and reporting characteristics. Remaining limitations include a significant publication bias and a small, but expanding sample of identified target studies. Practical, training, and research implications for the field of school psychology are discussed.