Chapter 11: Problems in the Estimation of School Effects: Insights from Improved Models

Understanding the effects of school interventions on children’s learning outcomes is of obvious importance for education policy, although identifying program effects is far more complicated than most policymakers and many researchers realize. For example, students and their families self-select into different educational settings in part on the basis of student characteristics that affect achievement and are quite difficult to measure. The methods for addressing this self-selection problem that are most commonly used in educational research appear to be susceptible to bias. More sophisticated methodological approaches appear to paint a more optimistic picture than the previous literature about the promise of various educational programs.