This course explores the statistical methods used for causal inference in the social sciences within the potential outcomes framework. Using this perspective puts the logic of statistical inference for both experimental and non-experimental studies within the same framework. Though randomized experiments serve as the gold standard for causal inference, the course also outlines how it may sometimes be reasonable to treat non-experimental data as if it had been drawn from an experiment. Usually, this involves a set of assumptions or substantive factual information about how the natural world produced the data. Research designs and methods covered include randomized experiments, matching, instrumental variables, difference-in-differences, synthetic control, and regression discontinuity designs.