Let’s not conflate APD with political history, and other reflections on “Causal Inference and American Political Development”

Abstract

American political development (APD) is a distinctive field of research that should not be conflated with, or flattened into a caricature of, historical research that uses historical data to make flawed causal inferences. It is a problem-driven inquiry into the dynamics of American politics, a substantive and theoretical exploration of how American politics has changed over time. APD research uses diverse types of data from a wide range of sources and employs multiple methodologies and analytical approaches, as appropriate. Because APD is a substantive and theoretical inquiry and not a method per se, there is no a priori reason to think that design-based causal inference cannot play a valuable role in studies of America’s political development, just as advanced quantitative methods have. However, while APD research does often seek to explain outcomes and establish causal relationships, that is not its only goal, and its orientation toward causality, causes, and theory tends to differ from much of the work in the causal inference tradition. This essay endeavors to clear up some of the confusion by offering the author’s perspective on what APD does well and how it does it. It also suggests how experimental research and APD research might be brought into more fruitful intellectual exchange and concludes with some thoughts on the value of methodological and intellectual pluralism.