Nonexperimental Replications of Social Experiments: A Systematic Review. Interim Report/Discussion Paper.

Reports initial findings from a synthesis and reanalysis of all available evidence on whether quasi-experimental evaluation designs produced unbiased estimates of the impacts of education, training, and welfare programs. This methodological paper tries to understand conditions under which quasi-experimental methods can lead to the same conclusion as a well-designed and well-executed randomized experiment. It has practical implications for the design of new evaluation research and synthesis of existing evidence on the effectiveness of social programs.