Causes, theories, and the past in political science

Abstract

A theoretically grounded approach to causal questions illuminates both the utility and limitations of the potential outcomes (PO) framework as a model for historically-focused, quantitative empirical research. While some causal questions are immediately reconcilable with the PO framework, for others, theoretical guidance is valuable in ascertaining relevant comparisons or characterizing the generalizability of findings to different contexts. A third category of important causal relationships feature strategic or information-based interactions, or multiple or unobservable mechanisms, many of which cannot be directly tested using the PO framework. Here, theory is critical in elucidating additional, observable implications that may be tested empirically. In all three categories, historical research promises special benefits: it expands the set of cases on which to test causal claims, may provide counterfactuals not available in contemporary contexts, and can feature institutional transformations that function as plausibly exogenous modifier variables. We clarify this classification of causal questions using examples from our own historical research.