Applied Econometrics: Mostly Harmless Big Data
A graduate-level econometrics course from MIT OpenCourseWare covering the empirical toolkit used in applied microeconomics research. Lectures work through regression and matching methods, instrumental variables, differences-in-differences designs, regression discontinuity, and the proper construction of standard errors. A dedicated block of eight to nine lectures turns to the analysis of high-dimensional data sets, the methods now grouped under the label Big Data, showing how machine learning tools intersect with causal inference. Materials include lecture notes and course readings drawn from published applied micro papers, following MIT's standard OCW format of free downloadable content without live instruction or a certificate. The course assumes prior exposure to statistics and econometrics and is aimed at students who already want to run these designs on real data, not at newcomers looking for a first introduction to regression.