Causal inference and American political development: common challenges and opportunities

This sense of humility about what we can learn from political science is by no means shared universally, but strikes me as a key lesson that one gains from taking the CI framework seriously.

Shared concepts do not guarantee cumulation. Fields can instead get bogged down in intractable disputes, as arguably eventually happened with the “party effects” literature discussed below.

It is worth noting that APD also had origins in American political thought and in political theory more generally. A prominent debate in APD concerned the relative importance of “ideas” and “institutions” in shaping outcomes; advocates of a more “institutional” approach tended to define the subfield over time, though much work incorporated a role for “ideas.”

Failure to do so is not necessarily a sign of trouble. APD’s initial theoretical coherence in part reflected its isolation from and opposition to mainstream American politics scholarship. Arguably, the declining coherence of the subfield is an understandable consequence of younger APD scholars’ less oppositional and more integrated relationship with mainstream American politics. I thank Devin Caughey for this suggestion.

Kelly (2019) highlights another important potential issue with persistence studies: spatially correlated errors can lead to false positives. Kelly offers an approach to tease out when that error is likely to occur.

An example of a literature that has combined theory-development with rich historical research and (growing) efforts at causal inference comes from recent work on political parties. The UCLA “School”—responding to and building off of earlier theoretical work by Aldrich (1995) and others—has elaborated a theoretical perspective that views political parties as coalitions of intense policy demanders that coordinate their efforts to achieve often non-centrist goals (Bawn et al. 2012; Cohen et al. 2008). Several scholars associated with that perspective have undertaken major historical studies drawing on and refining its insights (see, e.g., Karol 2009; Noel 2014; Masket 2009; Baylor 2013). At the same time, scholars have used experimental evidence and observational analyses to assess the operations of some of the specific mechanisms—e.g., endorsements—that the UCLA school argues are crucial to party coordination and effectiveness (Kousser et al. 2015; Hassell 2016). Other scholars, though departing from the claim that parties are best viewed as coalitions of policy demanders, have nonetheless engaged closely with the UCLA perspective in developing their own theoretical and historical accounts (see, e.g., Schlozman 2015; Grossmann and Hopkins 2016; see also McCarty and Schickler 2018). Collectively, that body of work has offered important insights into our understanding of political parties.