Because of the pressure towards immediate implementation of innovative educational programs, evaluation emphasis has been put on the overall effectiveness of these programs. This type of research yields only yes-no answers about general effectiveness without probing into the process of education itself. More research is needed on the factors that make programs effective; this research can be undertaken without sacrificing action-orientation, by carrying out the research and the program simultaneously. Evaluative research should include assessment of both possible and intended outcomes, measurements of antecedent conditions and consequences of intervention. If this alteration in research orientation takes place, it could be seen as a shift from the engineering model of evaluation studies (input-output differences relative to cost) to the medical model (concern for specific processes). A major advantage of the medical model is that goals, side effects and program by-products receive increased attention. Finally, this type of evaluative research can serve to advance science as well as social welfare. (MH)