Causal Inference and Comparative Methods

Abstract

Various ways are considered to infer causality from a relatively small number of cases that can be selected but not manipulated. The development of the “comparable-cases strategy” is examined first, together with the claim that it constitutes the comparative method. Mill's “method of agreement” is then presented as an alternative method of comparison, a method that not only can and has been used quite effectively with survey research in comparative politics, but one that is completely free from the methodological short-comings attributed to it. Cases, in short, may be selected for their similarity or their contrast. But because both of these qualitative methods of comparison (even when used jointly) are considerably less powerful than the “method of concomitant variation,” a third strategy is proffered to comparativists. It is a strategy employed extensively by Durkheim, but apparently lost sight of in attempts to reduce it—as well as the method of agreement—to a single comparative method based upon comparable cases.