Assessing the efficacy of any intervention can be a difficult task. In the case of children with autism who require comprehensive and long-lasting interventions, the task becomes even more difficult. In this paper, intervention studies based on comprehensive interventions for young children with autism are reviewed and examined in reference to elements that are essential to a well-designed treatment study, e.g., randomization/matching procedures, and outcome measures. Given the complexity of these comprehensive intervention programs, and the few data available on interventions with clear empirical validation, a plea is made for studies to consider the active ingredients or component parts of an intervention, e.g., number of hours, type of setting, and teaching approach. Suggestions are further made for improving future studies, including more rigorous designs, and measuring change.