Susan Athey
This talk will provide an overview of recent research combining ideas from machine learning and econometrics, with a focus on estimating causal effects and doing counterfactual policy evaluation in applications with big data, such as internet search. We briefly review a series of papers that cover structural modeling in economics; hypothesis testing in network experiments; robustness of causal estimates using regression tree-based methods; and novel methods for hypothesis-testing when analyzing causal effects. The talk will then go further in depth on a study of the problem of estimating heterogeneity in causal effects in experimental or observational studies and conducting inference about the magnitude of the differences in treatment effects across subsets of the population. In applications, our method provides a data-driven approach to determine which subpopulations have large or small treatment effects and to test hypotheses about the differences in these effects. For experiments, our method allows researchers to identify heterogeneity in treatment effects that was not specified in a pre-analysis plan, without concern about invalidating inference due to multiple testing. New methods for cross-validation in this context are highlighted.