We all know that correlation does not imply causality. However, in most cases in economics and other social and natural sciences it is causality that we are interested in, not just correlation. We want to know what the causal effects of educational interventions are on outcomes later in life. We want to know whether medical interventions are effective. We want to know what the effect on the economy is of decisions on interest rates by the Federal Reserve Bank, or the effect of a merger of two companies. In natural sciences we are often able to do randomized experiments, and these form the basis of decisions on drug approvals by the Food and Drug Administration. In social sciences it is often impossible to do randomized experiments; they may not be feasible conceptually or not ethical. We therefore rely in many cases on observational studies to infer causal effects. Econometrics has, since its early days in the 1920s, been focused on drawing credible causal inferences in such settings. In this lecture Guido Imbens will discuss some of the history of these methods and some recent examples, with a particular focus on the recent "credibility revolution" in economics.

A conversation moderated by Provost Richard M. Locke, Schreiber Family Professor of Political Science and International and Public Affairs, will follow the talk.