社交技巧訓練對高功能自閉症及亞斯柏格症者介入成效之後設分析

A Meta-Analysis of Social Skills Intervention for Individuals with High-Functioning Autism and Asperger Syndrome

本文探討2000至2011年國外33篇高功能自閉症及亞斯柏格症者社交技巧訓練單一個案研究的現況，並以不重疊比率，後設分析在不同參與者特徵、介入作法和評量情境上成效是否有差異。結果發現多數參與者為男性、國小學生，學前及高中職以上者較少，無國中生。學前的效果高於國小，高中職以上者成效很好；女性的成效不如男性。單一介入者的篇數和效果皆高於多重；成人中介（家長最少）多於同儕及兩者結合，成人及同儕中介成效中等。大多採個別介入，團體加個別介入效果極佳，個別為中等效果，團體介入則稍有效。模擬情境和其中包含自然情境成分的介入最多，效果中等，而自然情境介入無效。採認知行為取向最多，它和行為取向皆為中等效果，認知取向稍有效。多數介入一種技巧，其中社交認知者較少，且僅稍有效；沒有單獨介入非語言溝通，多將之結合其他三種（社交起始、互惠及認知）；介入三種效果極佳，其次是一種，兩和四種僅稍有效。模擬情境評量優於自然，分屬中等和稍有效果。習得成效中等，短期類化稍有效；而短、長期維持及長期類化無效。由此結果建議結合認知和行為取向優點的社交技巧訓練，促進其維持與類化成效；加強社交技巧介入信實度和社會效度的實施，團體結合個別的社交技巧訓練研究，並探究在自然情境實施有效介入和提升自然情境評量效果的方式。

Purpose: This study used meta-analysis to explore the application and effect of 33 social skill interventions on individuals with high functioning autism and Asperger syndrome. Methods: Meta-analysis was adopted to analyze studies published in peer-reviewed journals in English between 2000 and 2011. The studies used single case designs. Percentage of non-overlapping (PND) was employed to calculate effect sizes. Findings: Most participants were males from elementary school. Fewer participants came from preschool, high school, or postsecondary school and none came from junior high school. Social skill interventions were more effective on participants from preschool than from elementary school. The effects on high school and postsecondary students were high; however, the effect on female students was less than on males. Single interventions were more frequent than were multiple ones, and the effect of the former was better than was the latter. Adult interventions (less parental participation) were more frequent than peer interventions and more frequent than a combination of both types of interventions. The effects of adult and peer mediated interventions were moderate. The individual format was used widely. The effect of an intervention delivered by combining the group format with the individual format was excellent; the effect of the individual format was moderate, and the group format alone was mild in effectiveness. Simulated settings with authentic components were employed and had moderate effects; natural settings had no effect. A cognitive behavioral approach was most often employed and was moderately effective. A behavioral approach was also effective. However, a cognitive approach had only a mild effect. Most studies targeted only one type of social skill; a few studies targeted social cognition with mild effect No study targeted nonverbal communication alone. However, certain studies targeted nonverbal communication combined with other types of social skills. Studies that targeted any 3 types of social skills were highly effective. Studies that targeted only 1 type of social skill were the next most effective. Studies targeting 2 or 4 types of social skills were mildly effective. Interventions had a greater effect when the assessment was held in simulated settings versus natural settings. Simulated settings were moderately effective, whereas natural settings had no effect. Finally, the effect of acquisition was moderate, and the effect of short-term generalization was mild; however, short-term and longterm maintenance and long-term generalization had no effect. Those studies that collected treatment integrity and social validity data had effects that were better than studies that did not collect these data. Conclusions/Implications: first, increasing social skills studies of intervening preschool, secondary, postsecondary, and female individuals with high functioning autism and Asperger syndrome, social cognition, by group format and combining group with individual format, with peers and families’ participation. Second, implementing social skills studies which emphasize treatment integrity and social validity, analyze maintenance and generalization effects, inquire into the methods of advancing intervention effects at natural settings.