Why natural experiments? -- Standard natural experiments -- Regression-discontinuity designs -- Instrumental-variables designs -- Simplicity and transparency: keys to quantitative analysis -- Sampling processes and standard errors -- The central role of qualitative evidence -- How plausible is as-if random? -- How credible is the model? -- How relevant is the intervention? -- Building strong designs through multi-method research

1. Social sciences -- Experiments
2. Social sciences -- Research
3. Experimental design