Process Tracing: Best Practices Flashcards
What are some characteristics of research communities that use process tracing?
- Start research with puzzle-driven questions
– Questions that are theoretically or practically important regardless of whether they fit particular research methods. They drive the research process and should guide the choice of methods rather than the other way around. - Aim for contextualized explanations
– They combine structure (social and institutional factors) and agency (actors’ decision-making). They also specify scope conditions, explaining when and why the explanation applies. - Focus on “effects of causes” questions
What are the two forms of analysis in process tracing?
- Inductive: theory development
– By starting with an outcome of interest and identifying the cause and mechanism from the case study information. This process often involves comparing the developed explanation to alternative explanations. - Deductive: theory testing
– To examine the observable implications of hypothesised causal mechanisms within a case to assess the mechanism’s plausibility. This assumes the researcher has already identified the mechanism through an inductive phase.
What are the benefits of process tracing as a research method?
- More systematic and transparent research
- Methodologically plural
List some best practices in process tracing.
- Cast the net widely for alternative explanations
- Be equally tough on alternative explanations
- Consider potential biases of evidentiary sources
- Take into account how critical the case is for alternative explanations
- Make a justifiable decision on when to start
- Be open to inductive insights
- Be relentless in gathering diverse and relevant evidence
- Combine process tracing with case comparisons
- Think of empirical implications deductively
- Acknowledge that analysis does not need to be fully conclusive
Where can I find alternative explanations?
- Case-specific sources: regional specialists, topic experts, historians
- Non-scholarly sources: participants, journalists
- Scholarly publications
What are the four types of tests for evaluating explanations in process tracing?
- Hoop tests
- Smoking-gun tests
- Doubly decisive tests
- Straw-in-the-wind tests
What are some things to keep in mind when considering primary sources?
- Agents may have ulterior motives, so adjust credibility accordingly
- Consider the context in which statements are made (private, public, spontaneous, rehearsed)
- Be aware of selection or availability bias
- Absence of evidence is not evidence of absence
What is a most-likely case?
A case that is highly likely to be explained by a particular theory. If the theory fails to explain this case, the theory may need to be revised or rejected
What is a least-likely case?
A case that is unlikely to be explained by a particular theory. If the theory can explain this case, it strengthens the theory’s validity
When should I stop gathering evidence?
- When financial and temporal constraints are reached
- When data repetition suggests additional data will not change findings
How can process tracing and case comparisons be combined?
- Starting with process tracing: use comparative design to test the causal mechanism or situate the case in comparative perspective
- Starting with comparative design: use process tracing to systematically examine cases or refine case selection
Why should empirical implications be defined prior to analysis?
To minimize the risk of ad hoc operationalizations and ensure a rigorous research design
Can a process tracing analysis have multiple or inconclusive outcomes?
- Yes. It is acceptable to conclude that there is not enough data to rule out alternative explanations
- It is also possible to conclude that more than one explanation is valid, potentially explaining different aspects of the outcome
Hoop tests
These tests identify necessary evidence for an explanation to be considered relevant. They are useful for quickly eliminating weak explanations. Passing a hoop test doesn’t necessarily mean an explanation is strong, only that it cannot be immediately dismissed.
Smoking-gun tests
These tests provide sufficient evidence for an explanation to be considered strong relative to others. However, failing a smoking-gun test doesn’t automatically disqualify an explanation. This type of evidence is difficult to find and should be treated with caution, especially when dealing with sources from opaque autocracies or non-state actors
Doubly decisive tests
These tests offer evidence that simultaneously confirms the relevance of one explanation and eliminates all alternatives.
Straw-in-the-wind tests
These tests provide evidence that, while not definitive, suggests an explanation’s plausibility. The accumulation of multiple pieces of straw-in-the-wind evidence can increase confidence in an explanation.
