A quantitative synthesis of research findings on the effects of three major activity-based elementary science programs developed with National Science Foundation support was conducted. Controlled evaluation studies of the Elementary Science Study (ESS), Science-A Process Approach (SAPA), or The Science Curriculum Improvement Study (SCIS) were used to examine effects on student outcomes and effects on classroom practices. For the effects on student outcomes, a meta-analysis of 57 studies resulted in 400 separate comparisons involving 13,000 students from 1,000 classrooms. An average improvement of 20 percentile units on science process tests for students in activity-based programs over those in other programs was found. This improvement was significantly greater for disadvantaged students and less for advantaged students. Effects on classroom practices were synthesized from 12 studies in which teaching practices were systematically observed during 1,800 science lessons presented in activity-based and non-activity-based classrooms. Results indicate that, on the average, the use of student hands-on activities increased by nearly 10% in the classrooms of the new programs. Talk decreased by 9% and lecturing by 7%. (DC)