This course covers core ideas relevant for quantitative data analysis in the social sciences, with a focus on causal and statistical inference. Building on the material in 231A, we cover instrumental-variables analysis, non-linear regression models such as probit and logit, mediation and path analysis, and selection models. We also further discuss natural experiments and regression-discontinuity and difference-in-differences designs. We make frequent use of simulations, including the bootstrap. Students will complete a project using replication data from published work. Throughout the course, we emphasize the role of strong research designs in promoting valid causal inferences, and also the limitations of design. Model specification is a central area of concern.
Discussion scheduled Wednesdays 4-6pm.
Political Science 231A or equivalent. Experience with R is assumed.