Invited to submit at Annnual Review of Political Science
(with Mike Findley and Kyosuke Kikuta)
External validity refers to how the inferences drawn from a given study apply outside of the immediate sample to a broader population or other target populations. The concept of external validity is so common in social science that one might mistakenly assume it has been addressed at length. Unfortunately, scholars do little more than pay lip service to external validity. In this article, we clarify the conceptual dimensions of external validity, formally demonstrate why external validity matters, and then advance three evaluative criteria: Model Utility, Scope Plausibility, and Specification Credibility. In addition to articulating the components of these criteria, we conclude with proposed evaluation and reporting standards for external validity. We hope that these evaluation and reporting standards re-balance scientific inquiry, such that the current obsession with causal inference is complemented with an equal interest in generalized knowledge.