This paper reviews both the content and the methodological issues involved in the small but growing number of empirical studies on the use of Direct Instruction programs and teaching procedures with special populations. The research reviewed includes experimental and quasi-experimental work, summative evaluations of entire programs, and fine-grained evaluations of single instructional components. These studies evaluate the effects of Direct Instruction reading and language programs among subjects ranging from those with mild academic deficits to the severely handicapped. Studies are organized for discussion according to the type of evaluation design utilized in order to acquaint the reader with the benefits and drawbacks associated with each type of design. For each study reviewed, the major findings are presented and methodological problems and measurement issues analyzed. Directions for future research are suggested. In its final section, the paper turns to the problem of adequately defining independent variables, noting that precision in these definitions is particularly important. Suggestions are drawn from the research for resolving this problem. (Author/PGD)