The Logic & Best Practices of Process Tracing (QMMR A)
Andrew Bennett, Jeffrey T. Checkel, and Tasha Fairfield
Half Day, 9:00 AM – 1:00 PM
Los Angeles Convention Center, 406A
This short course covers the underlying logic and best practices of process tracing, which is a within-case method of developing and testing causal explanations of individual cases. We begin by exploring the philosophies of science behind process tracing: scientific realist and interpretive. Next, we highlight, define and provide examples of the central concept process tracing measures – causal mechanisms – noting their difference from causal effects and interpretive understandings of causation.
The core of the short course is then an introduction to the logic and best practices of process tracing, both its ‘front end’ data collection and ‘back end’ data analysis. For data collection, we consider the typical ways in which process tracing gathers evidence on the observable implications of causal mechanisms, including archival work, document analysis of secondary sources, various field methods (interviews, political ethnography, ethnography), and surveys. In reviewing these methods, we consider the inferential and ethical challenges each raises when accessing process-tracing data. On data analysis and process tracing, we begin by considering the informal manner in which many scholars proceed; more important, we survey the growing number of techniques (e.g., Bayesian logic, directed acyclic graphs) that allow us to conduct the process tracing analysis more formally and transparently. We finish this part of the course by articulating a set of best practices for conducting process tracing.
After this overview of the philosophical, causal and data logics of process tracing, the course introduces participants to two different types. We begin with Bayesian process tracing—comparing rival hypotheses; evaluating the inferential weight of evidence by “mentally inhabiting” the world of each hypothesis and asking which one makes the evidence more expected; updating prior views about which hypothesis is more plausible; and fostering transparency through systemization. We then turn to interpretive process tracing—inductive approach; practice logic; establishing local causation; transparency through ethical self-reflection. Full details on each approach will be offered in separate afternoon short courses: “Bayesian Reasoning” (QMMR B) led by Tasha Fairfield and Andy Bennett; and “Interpretive Process Tracing & Practice Tracing” (QMMR C) led by Jeff Checkel and Vincent Pouliot.
Throughout the course we will emphasize best practices and applications to exemplars of process tracing research. While the examples are primarily drawn from international relations and comparative politics, the methods we discuss are applicable to all the subfields of political science, to sociology, economics, history, business studies, public policy, and many other fields.
The course’s final section is devoted to small-group breakout sessions, where participants workshop how they plan to use process tracing in their research. Are there data access, data collection, data analysis or ethical issues with which they are grappling? Instructors and fellow students will offer constructive advice on how best to address such issues.