In this study, we conducted a multilevel meta-analysis to determine whether the self-regulated strategy development (SRSD) approach to teaching writing to students with autism spectrum disorder (ASD) improves significantly the number of words written and overall quality of writing, whether the effects of SRSD were consistent or variable across studies, and whether the moderator of age could explain the potential variability in the number of words written and overall quality of writing. Results indicated that SRSD had large, statistically significant and meaningful effects on the number of words written and overall quality of writing. There was variability in the effectiveness of SRSD, both between cases and between studies, and older students had benefitted more from the intervention, but only for number of words. Implications and directions for future research are provided.