Since the inception of the Arizona-Mesa Differentiated Staffing Project in June 1970, the project staff was charged with the responsibility of evaluating the effects of the changes brought about as a result of implementation in three pilot schools. It was found that experimental-design research in an ongoing social system was extremely difficult because of inability to control internal and external variables. After reviewing various methods used to evaluate social innovations, the staff decided on a combination of the case-study technique advocated by Weiss and Rein and the quasi-experimental design advocated by Campbell. Factors jeopardizing internal and external validity were identified and tentative research designs were outlined for five questions. The questions involved comparing project and non-project schools in relation to student achievement, school climate, staff attitudes, teacher ability to perform tasks specified in the original proposal, and degree of actual staff differentiation. Each research design specifies a hypothesis, data-gathering instruments, dependent and independent variables, and statistical treatment. (RT)