Quality Meets Quantity: Case Studies, Conditional Probability, and Counterfactuals

Publisher Website
Google Scholar
Abstract
In contrast to statistical methods, a number of case study methods—collectively referred to as Mill's methods, used by generations of social science researchers—only consider deterministic relationships. They do so to their detriment because heeding the basic lessons of statistical inference can prevent serious inferential errors. Of particular importance is the use of conditional probabilities to compare relevant counterfactuals. A prominent example of work using Mill's methods is Theda Skocpol's States and Social Revolutions. Barbara Geddes's widely assigned critique of Skocpol's claim of a causal relationship between foreign threat and social revolution is valid if this relationship is considered to be deterministic. If, however, we interpret Skocpol's hypothesized causal relationship to be probabilistic, Geddes's data support Skocpol's hypothesis. But Skocpol, unlike Geddes, failed to provide the data necessary to compare conditional probabilities. Also problematic for Skocpol is the fact that when one makes causal inferences, conditional probabilities are of interest only insofar as they provide information about relevant counterfactuals.Jasjeet S. Sekhon thanks Walter R. Mebane Jr., Henry Brady, Bear Braumoeller, Shigeo Hirano, Gary King, John Londregan, Bruce Rusk, Theda Skocpol, Suzanne M. Smith, Jonathan N. Wand, the editors of Perspectives on Politics, and three anonymous reviewers for valuable comments and advice.

This publication has 28 references indexed in Scilit: