32. Estimate causes, not just correlations
Use DAGs, confounders, colliders, mediators, randomized trials, natural experiments, matching, weighting, difference-in-differences, instrumental variables, and regression discontinuity. This chapter gives a practical language for cause-and-effect claims.