Redesigning Social Inquiry: Fuzzy Sets and Beyond (review)

In Redesigning Social Inquiry: Fuzzy Sets and Beyond, Charles Ragin has offered to the social sciences a robust appeal for renewed commitment to small-N comparative research, pursued with a toolkit of analysis techniques that he has developed. The book contains three arguments that are developed simultaneously: (1. Case-oriented explanation remains vital to the social sciences; (2. Fuzzy-set-based Qualitative Comparative Analysis offers tools that enable the construction and evaluation of complex causal explanations; (3. Prevailing forms of quantitative analysis cannot elucidate causal complexity and are therefore ill-suited to small-N research.

I do not know of a single sociologist who would disagree with Ragin's position that case-oriented explanation is central to the sociological enterprise. This widespread agreement is, to no small degree, the result of Ragin's 30 years of eloquent writing on this point, complemented by longstanding admiration for the highest quality ethnographic and historical work in the sociology cannon. In Redesigning Social Inquiry, Ragin reminds all of us why we feel this way.

Although the specific details of fsQCA are well beyond a short review of this type, Redesigning Social Inquiry succeeds in its goal of explaining that the utility of fsQCA is "to explore evidence descriptively and configurationally, with an eye toward the different ways causally relevant conditions may combine to produce a given out-come."(141) The demonstration proceeds in orderly fashion through all 11 chapters, with a sufficient number of clear, detailed examples.

In Redesigning Social Inquiry, Ragin also moves QCA and fsQCA away from reliance on notions of sufficiency and necessity in defining causal effects and instead embraces counterfactual dependence as well. This was a major weakness (by my reading) of the foundational literature on QCA. Too many researchers came to believe that simple notions of sufficiency and necessity could be inferred from near-invariant associations in small-N research, which thereby cut them off from dominant currents of thought on causality in both philosophy and quantitative methodology. Ragin and co-author John Sonnett show how counterfactuals fit nicely within both QCA and fsQCA, and this is an important development.

I expect that Redesigning Social Inquiry will also increase the usage of fsQCA among comparativist researchers. One often hears the relevance of fsQCA challenged among those most at risk of deploying it. Is it worth the effort? Does the Boolean algebra kill off the scholarly impulse? Does the formality constrain creativity? Does it suppress narrative detail? Does it give sufficient scope to complex narrative time? In Redesigning Social Inquiry, Ragin takes on these whispered critiques will full force, not by direct confrontation, but through convincing demonstrations of the power of fsQCA. I came away from the book with a deeper appreciation for the power of fsQCA to help build subtle causal explanations, and I suspect this will be true for all readers.

No review is complete without at least some critical content, and in this regard there is low hanging fruit in Redesigning Social Inquiry. It is perhaps common in pioneering methodological work to conjure up an opposition, but the enmity in this book is excessive. According to Ragin, "conventional quantitative research" is (1. obsessed with correlations to the exclusion of more fundamental joint distributions, (2. ignores set relations, relying only on linear associations between interval-scaled variables, (3. universally condemns any examination of cases selected on the outcome variable, (4. adopts linear additive models of causation, only paying lip service to nonlinear and interactive effects, and (5. has no interest in understanding the joint distributions of the causal variables that it implores should be included in "net effects" regression models.(cf., 16, 22, 29, 149, 157) Even when moving beyond such "conventional" practice, with references to the work in counterfactual causal analysis that has developed since the 1990s, the commentary is both out of date and incorrect. This less conventional, but increasingly important, literature gives substantial attention to modeling of causal effect heterogeneity, interactions and mechanisms, all...