Most policy changes—in education as in every other field-- do not work as intended. Human behavior is complex; it is just too hard to anticipate whether students, parents and teachers will react in the intended way. Thus, policymakers must recognize opportunities to learn: to identify comparison groups for their initiatives, to implement new programs in ways which lend themselves to evaluation and to be consumers of efficacy research done in other settings. In this class, we will discuss the key controversies in education policy over the last 30 years and the evidence which has been collected thus far on each of them. Along the way, we will discuss alternative evaluation designs—randomized trials, difference-in-differences, regression discontinuity and value-added (covariate adjustment)—and the conditions conducive to each. No single policy idea in vogue today could make a bigger difference two decades from now than creating the infrastructure for testing a series of evolving ideas. We will discuss how other fields—from retail sales to pharmaceuticals—have created systems to generate sustained improvement, and explore ways similar progress could be made in education and other fields of social policy.
Prerequisites: Successful completion of S-040 (HGSE), API-202 (ÌÇÐÄvlog¹ÙÍø), or prior equivalent training in multiple regression.
Also offered by the Graduate School of Education as A164.