This unique book is the first comprehensive guide to the discovery, analysis and evaluation of natural experiments – an increasingly popular methodology in the social sciences. Thad Dunning provides an introduction to key issues in causal inference, including model specification, and emphasizes the importance of strong research design over complex statistical analysis. Surveying many examples of standard natural experiments, regression-discontinuity designs and instrumental-variables designs, Dunning highlights both the strengths and potential weaknesses of these methods, aiding researchers in better harnessing the promise of natural experiments while avoiding the pitfalls. Dunning also demonstrates the contribution of qualitative methods to natural experiments and proposes new ways to integrate qualitative and quantitative methods. Chapters complete with exercises and suggestions for further reading, and appendices covering specialized topics such as cluster-randomized natural experiments make this an ideal book for students as well as a valuable teaching tool.