Findings from the Follow Through Planned Variation Study

T hough Follow Through was originally conceived in 1967 to be a largescale compensatory education program extending Head Start services through third grade, budget problems quickly reduced it to a small program and its goals changed accordingly. Follow Through became known for its Planned Variation of educational approaches, and the hope was that study of these variations would increase knowledge of which kinds of educational programs would have different patterns of effects in various kinds of settings. Nearly twenty different educational models were studied. Each model was implemented in a variety of settings, and children in those settings were exposed to their model from kindergarten through third grade. Though the program and the planned variations are still in place, the study of effects is now completed. The Follow Through study was not a true experiment, or even a well-designed quasi-experiment. For example, school districts were nominated by states to participate in the program, and once accepted, they selected the models they would implement. Only after the school districts had identified the classrooms and children in their districts who would participate in Follow Through were they asked to help identify comparable groups to serve as controls. The study is an experiment in the sense that it systematically varies the treatments, but it is a natural study in the sense that disruptions to the setting were minimized by compromising experimental requirements. Abt Associates, Inc., who were commissioned to analyze third-grade outcomes on some 20,000 children involved in the study as well as their parents and teachers, recently submitted a final report to the Office of Education that stacks nearly six inches high (Abt Associates, Inc., 1977). The results have already been described in newspapers and news magazines as favoring more structured educational approaches. Such a conclusion is important to the educational community, but its validity and generalizability need to be evaluated, since the Follow Through study did not employ strict experimental standards. This paper will present an analysis of the findings from the Follow Through study and evaluate some of their meanings.