Estimate causal effects from existing data. Use when fitting or interpreting DiD, ITS, synthetic control, regression discontinuity, or other treatment-effect analyses, including robustness checks and counterfactual plots. For choosing a study design before analysis, use designing-experiments instead.