CausalitIES in Policy Studies and Beyond: to Each its Own and Conditionally Matching

The paper integrates different layers of knowledge about causal analysis in policy and political studies – from the philosophical to the methodological one – around non-randomized strategies. The first section sets the ontological foundations: it embraces a realist interpretation of causation and maintains that causal claims involve necessary connection and counterfactual assertions. The second section compares non-randomized strategies to ascertain causation by following a macro-to-micro rationale that ‘zooms in’ causation, from quasi-experimental designs, selection correction and structural equation models, to Qualitative Comparative Analysis, Bayesian case study and process tracing. It provides a map of the assumptions that such strategies embark when they establish the credibility of a causal claim, thus illuminating the special ground upon which they build their language and design. Finally, the third section discusses the threats to the credibility of causal claims that arise in any research strategy due to undiscounted heterogeneity and provides guidelines to research designs that combine techniques to compensate their blind spots instead of summating their weaknesses, thus going beyond ‘naïve interpretations’ of mixed- and multi-method research.