Convincing explanations with QCA : the contribution of “essential” configurational models

Since long, the scientific discourse maintains that sound models are a necessary requisite to convincing explanations. How to design them so that they suit configurational thinking is the question that Amenta and Poulsen first have explicitly put in the methodological agenda of Qualitative Comparative Analysis. The article contributes to the substantive and to the technical side of the answer. It reasons that the cogency of the starting hypothesis requires factors supporting the expectation that, were they jointly given, the outcome would certainly obtain; and considers Ostrom’s “action situation” as a fruitful framework for guidance in selection. It then addresses the related risk of overly rich models by adapting Baumgartner’s difference-making principle for specification tests to be run on candidate factors before minimizations. Exercises in replication will show the extent to which these criteria provide a useful diagnostics of existing models, and a blueprint for more convincing configurational explanations.