Symposium: Case Selection, Case Studies, and Causal Inference Introduction

For scholars concerned with causal inference, how should cases be selected in case study research? This symposium builds on previously published arguments by James Fearon and David Laitin (2008), who favor random sampling in case study analysis, and by John Gerring (2008), who favors purposive selection. The statistician David Freedman—long an advocate of case studies as an important research tool—comments on these published arguments; responses are offered by Fearon-Laitin and by Gerring; Gary Goertz adds a commentary of his own; and then Freedman offers concluding remarks. In Fearon and Laitin’s (2008) discussion, the goal is to draw insights about causal mechanisms from case studies so as to illuminate the findings from a large-N, regression-type analysis. The idea of random sampling is of course central to the broad literature on statistical inference, and for Fearon and Laitin a key advantage of this approach is to prevent scholars from deliberately selecting cases favorable to their preferred hypotheses, thus engaging in “cherry-picking.” Symposium: Case Selection, Case Studies, and Causal Inference