This article reports the results of a meta-analysis of technology-based intervention studies for children with autism spectrum disorders. We conducted a systematic review of research that used a pre-post design to assess innovative technology interventions, including computer programs, virtual reality, and robotics. The selected studies provided interventions via a desktop computer, interactive DVD, shared active surface, and virtual reality. None employed robotics. The results provide evidence for the overall effectiveness of technology-based training. The overall mean effect size for posttests of controlled studies of children with autism spectrum disorders who received technology-based interventions was significantly different from zero and approached the medium magnitude, "d" = 0.47 (confidence interval: 0.08-0.86). The influence of age and IQ was not significant. Differences in training procedures are discussed in the light of the negative correlation that was found between the intervention durations and the studies' effect sizes. The results of this meta-analysis provide support for the continuing development, evaluation, and clinical usage of technology-based intervention for individuals with autism spectrum disorders.