For this review, approximately 85 studies were identified. The majority were merely descriptions of training programs. There were few studies of teaching assistants' cognitive or affective gains. A number of studies reported significant effects on observed teaching behavior. Whereas quasi-experimental studies generally found significant effects on student variables, true experimental studies generally did not. However, the two types of studies tended to assess different types of training programs. Conclusions, further research questions, and methodological suggestions are offered. (Author)