Congruence Analysis

Both are within-case, theory-oriented methods, but they ask different questions. Congruence analysis compares the predictions of competing theories with observed values in the case and ranks the theories by how well their expectations match the evidence; it focuses on correspondence between predicted and observed patterns. Process tracing instead reconstructs the causal chain step by step, looking for diagnostic evidence (straw-in-the-wind, hoop, smoking-gun, doubly-decisive tests) that a specific mechanism operated. Blatter and Haverland treat them as distinct approaches, and they are often combined — congruence to adjudicate theories, process tracing to verify the mechanism.

Because an observation that all rival theories predict cannot help decide between them; only observations on which the theories make different or competing predictions carry inferential weight. Blatter and Haverland advise concentrating on these high-discrimination observations: congruence on a unique prediction of one theory strongly supports it, whereas congruence on a shared prediction is nearly uninformative. Weighting evidence by discriminating power is what allows a single case to adjudicate among theories rather than appearing consistent with all of them at once.

Not on its own. Demonstrating that a theory's predictions are congruent with the observations shows that the theory fits the case better than its rivals, which builds the theory's relevance and credibility, but congruence is a relationship between expectations and evidence rather than a demonstrated causal mechanism. To strengthen causal claims, researchers typically combine congruence analysis with causal-process tracing within the case, and where possible with cross-case comparison, so that fit, mechanism, and comparison together support the inference.