20. Causal evidence when you cannot randomize
Covers natural experiments, difference-in-differences, regression discontinuity, instrumental variables, and matching at a beginner-friendly level. You will see how researchers make stronger causal claims when randomization is not possible.