TITLE: ‘Statistical Vexations: Causal Inference from Observational Studies‘
A persistent challenge in both policy analysis and social scientific research is to draw credible causal inferences from observational studies. In this presentation I will offer a critical review of some strategies that have been proposed and argue for the essential roles played by longitudinal data (or, at a minimum, pseudo-panel data) and sensitivity analysis in providing an evidence base for such inferences. The relevant issues will be illustrated by example. Connections with the problem of replicability will also be explored.
Henry Braun is the Boisi Professor of Education and Public Policy, and director of the Center for the Study of Testing, Evaluation, and Educational Policy, at the Lynch School of Education Boston College. Professor Braun specializes in testing and education policy; large-scale assessment surveys and achievement gaps; value-added modeling; standard setting, and higher education outcomes. He serves on a number of state, national and international advisory committees He has won awards from the AERA, NCME, is a fellow of both the ASA and AERA and an elected member of the National Academy of Education.