Currently, formative evaluation strategies are used in special education programs to the near exclusion of impact or outcome evaluations. Fast practices, misconceptions concerning group impact evaluation designs, and philosophical orientations have all contributed to this trend. Impact evaluations of special education programs, however, can be designed and conducted so that they produce information useful for program improvement-regardless of whether the intervention is massive or results are statistically significant. Illustrative case studies of recent impact evaluations support this position. Assessment of program implementation and measurement of outcome, two critical areas that can greatly improve the usefulness of evaluation results, are emphasized in the concluding discussion.T VO DECADES AGO THE following passage appeared in a newly published book on research on teaching:Claims made for the rate and degree of progress which would result from experiment were grandiosely overoptimistic and were accompanied by unjustified depreciation of non-experimental wisdom. The initial advocates assumed that progress in the technology of teaching had been slow just because scientific method had not been employed . . . When, in fact, experiments often proved to be tedious, equivocal, of undependable replicability and to confirm pre-scientific wisdom, the overoptimistic grounds upon which experimentation had been justified were undercut and a disillusioned rejection or neglect took place. (Campbell & Stanley, 1963, pp. 2-3) This encapsulated history of 50 years of educational research served as Campbell and Stanley's point of departure in their classic text, which revolutionized the field of applied educational research.
Ironically, though not unpredictably, experimental (and quasi-experimental) designs for evaluation in special education have become the subject of growing disillusionment. Few would deny that findings from impact evaluations in special education have often been equivocal, failing to support the theories (and hopes) of the researchers.Recent evaluations of mainstreaming are a case in point (e.g., Hersh & Walker, in press). Due at least in part to this state of affairs, we have witnessed a growing reluctance to engage in impact evaluations. Instead, formative evaluation procedures-often involving little more than an examination of whether legal and administrative guidelines are met (see Kennedy, 1980)-have captured the attention of thf*majority of special education administrators and evaluators. Only rarely have researchers aimed recent evaluations at determining how much students learn, or whether maximal learning can be linked to a particular educational approach. When such evaluations have been carried out, most have been of the single subject variety. But even in these instances, Richard Jones' (1979) appeal has been largely ignored. Jones advocated using single subject designs for formative evaluation of 16 RASE 5(2), 16-24 (1984) 0741-9325/84/0052-0016$2.00©PRO-ED Inc.