This paper focuses on a review of data collected from the second wave of a three wave longitudinal study (Head Start Planned Variations-HSPV) designed to assess the relative impacts of a variety of preschool curricula. Head Start children formerly in the HSPV program were tested when they entered selected Follow Through schools and throughout their early school years until they completed Follow Through at the end of the third grade. The study design is reviewed and tests and measures used during the 1970-71 evaluation period are briefly described. Problems in designing assessment instruments for this age group in both cognitive and "non-cognitive" areas are identified. A consideration of the general methodological implications of these design problems is followed by a summary of conclusions on the three main effect questions: (1) overall effects of the Head Start experience; (2) differences between Planned Variation and conventional Head Start programs; and (3) differential effects of the eleven Planned Variation Models (brief descriptions of these are included). Also considered is the question of whether the effects of the various preschool curricula are sensitive to differences in child characteristics. Conclusions address the problem of making curricular evaluations attend more to goals than to psychometric characteristics. (ED)