Applied Econometrics for Public Policy

Catalog description

This class is an applied introduction to multivariate statistical inference that is aimed at graduate students with little prior statistical experience, and satisfies the Quantitative Methods and Analytics requirement in the MPA Program. We will begin with a brief introduction to basic statistical concepts and probability theory before introducing the linear regression model. We then review several tools for diagnosing violations of statistical assumptions, including how to deal with outliers, missing data, omitted variables, and weighting. We will next consider situations in which linear regression will yield biased estimates of the population parameters of interest, with particular attention paid to measurement error, selection on unobservables, and omitted variables. The course will end with an introduction to extensions of the linear regression model, including models for binary and categorical outcomes. While statistical modeling is the focus of the course, we proceed with the assumption that models are only as good as the theoretical and substantive knowledge behind them. Thus, in covering the technical material, we will spend considerable time discussing the link between substantive knowledge and statistical practice. The course is designed primarily for professional masters students.