This meta-analysis of 12 dependent variables from 9 quantitative studies comparing music to no-music conditions during treatment of children and adolescents with autism resulted in an overall effect size of d =.77 and a mean weighted correlation of r =.36 (p =.00). Since the confidence interval did not include 0, results were considered to be significant. All effects were in a positive direction, indicating benefits of the use of music in intervention. The homogeneity Q value was not significant (p =.83); therefore, results of included studies are considered to be homogeneous and explained by the overall effect size. The significant effect size, combined with the homogeneity of the studies, leads to the conclusion that all music intervention, regardless of purpose or implementation, has been effective for children and adolescents with autism. Included studies are described in terms of type of dependent variables measured; theoretical approach; number of subjects in treatment sessions; participation in and use, selection, and presentation of music; researcher discipline; published or unpublished source; and subject age. Clinical implications as well as recommendations for future research are discussed.