Process Tracing Methods

This hands-on Master Course aims to provide participants with the methodological tools to refine their use of process-tracing methods in their own substantive research, while also enabling them to embed process-tracing case studies in mixed-methods research design. The course requires VERY active participation.

Morning sessions will be devoted to lectures and discussions about key methodological issues, whereas the afternoon sessions are used to discuss particular aspects of participant projects, including theories of causal mechanisms and how we can develop testable predictions about evidence that the activities associated with parts of mechanisms might leave in a given case.

The promise of process-tracing as a methodological tool is that it enables the researcher to study more-or-less directly the causal mechanism(s) linking a cause (or set of causes) and an outcome, allowing us to open up the ‘black box’ of causality itself. By unpacking causal mechanisms into their constituent parts, composed of entities engaging in activities, and then tracing the empirical manifestations these activities leave in actual cases, we are able to collect what has been termed mechanistic evidence upon which we can make causal inferences about how causal mechanisms actually work (Machamer, Darden and Craver, 2000; Machamer, 2004; Bogen, 2004; Waskan, 2008, 2011). Strong causal inferences about the effect a cause has on an outcome are naturally only possible when we use evidence of difference-making that is produced through experimental manipulation across cases (Woodward, 2003). However, when we use mechanistic evidence to make causal inferences, we are using observational, within-case evidence to make causal inferences about the actual operation of mechanisms in real world cases (Russo and Williamson, 2007; Illari, 2011; Waskan, 2008, 2011). In other words, instead of studying causal effects we are studying how things work. As an example, in medical research the causal mechanism linking smoking and cancer has been broken down into a series of parts that describe the entities and their activities that transmit causal forces to the outcome (Russo and Williamson, 2007: 162). The cause that triggers the process is smoke inhalation. The mechanism then is composed of several steps that are linked together in a causally productive relationship where the causal links are made explicit in the form of activities: smoke inhalation -> hair-like cilia in lungs are destroyed by the smoke -> lungs cannot clean themselves effectively, resulting in cancer-producing agents in the smoke becoming trapped in mucus -> these agents alter cells (especially cell division) in the lungs -> cancer. Mechanistic evidence is the observational data that captures empirically the observables left by the activities of entities (for example smoke destroying cilia) that transmit causal forces to the next part of the causal process, enabling researchers to make evidence-based claims about how smoking produces cancer. 

In the first morning session, we start by discussing how to differentiate process-tracing from other methods; including both large-n quantitative methods, but also other small-n methods such as analytical narratives, comparative case studies etc. Here we define process-tracing by the interest in studying causal mechanisms in single case studies in ways that enable within-case causal inferences to be made. We discuss the four variants of Process Tracing: theory-testing, theory-building, theoretical revision, and explaining outcome process-tracing. This is followed by an in-depth the ontological underpinnings of process-tracing in the second session of day 1, and the full session of day 2, focusing on how to understand causal mechanisms and how they differ from other types of causal theorization.

In day three morning session, we will discuss how inferences can be made using mechanistic evidence, focusing on how to operationalize theories of mechanisms by utilizing informal Bayesian logic. On day four we will discuss challenges relating to the evaluation of evidence in a joint session with the Historical Methods course. The final session on day 5 will turn to questions of case selection and mixed/multi-methods.

The afternoon sessions on days 1 and 2 will be devoted to presentation and discussion of theories of causal mechanisms prepared by each participant. In the afternoon sessions on days 3 and 4, we turn to presentation and discussion of observable manifestations of the activities of parts of mechanisms of each participant, followed on day 5 by a discussion of why participants chose particular cases.

A critical requirement for participation in this course is that your are using in-depth case study methods in your current research project (ph.d., post doc or other), and that you are advanced enough in your research that you have clear theoretical conjectures and ideas about potential empirical observations that can be worked with in the course.

This course also requires that you are familiar with the recent literature on case study methods (post 2010), and assumes familiarity with basic concepts related to Process-tracing. In particular, you should have basic knowledge about debates about causal mechanisms and empirical tests and how they are used in case studies.

The course requires submission BEFORE the course of a theorized causal mechanism and empirical proposition. The instructor will provide information about this well in advance of the course.