Read this before running fsQCA

Just last week I wrote about two papers that examined the validity of QCA. They were by no means the first ones to do so, but that doesn’t make these papers any less important.

Now, QCA isn’t exactly static, even though it remains focused on its founding father. Fuzzyset QCA (fsQCA) is often used these days as it promises to overcome some of the shortcomings of QCA. Unfortunately, even if you buy into the concept and epistemology, the empirics still don’t add up.

Krogslund, Chris, Donghyun Danny Choi, and Mathias Poertner. 2014. “Fuzzy Sets on Shaky Ground: Parameter Sensitivity and Confirmation Bias in fsQCA.” Political Analysis, November, mpu016. doi:10.1093/pan/mpu016.

Krogslund and colleagues used simulations to check how robust fsQCA is. The approach is quite intriguing. Rather than using data generated in the computer as is often done in such situations, they have used three existing studies. After replicating these studies, they modified tiny bits. With a robust method, such tiny changes will not have a substantive impact on the results. With fsQCA, however, the results often changed radically: it is a very sensitive method.

Read this before running QCA

Thinking of using QCA in your work? Read these two papers, and make up your own mind:

Achen, Christopher H. 2005. “Two Cheers for Charles Ragin.” Studies in Comparative International Development 40 (1): 27–32. doi:10.1007/BF02686285.

Achen summarizes the three claims by Ragin: (1) case studies have their own methodology. Nobody disputes this (first cheer). (2) Often quantitative analyses do not match social reality (assumptions, linearity assumptions, etc.). Yes (second cheer), but that’s a question of applying the existing tools appropriately, not a problem with the tools. (3) That statistical analysis cannot cope with social reality. That’s not true, or only true if quantitative methods are applied in a poor way (no cheer here). Achen then outlines nicely that each step in QCA has an equivalent in long-established quantitative analysis, to which many have contributed.

Lucas, Samuel R., and Alisa Szatrowski. 2014. “Qualitative Comparative Analysis in Critical Perspective.” Sociological Methodology 44 (1): 1–79. doi:10.1177/0081175014532763.

Lucas and Szatrowski wanted to know whether QCA is any good. To find out, they used computer simulations where the true causal path was known (the programmed it). Their conclusions are quite clear: QCA correctly identifies only 3 out of 70 causal paths correctly. They are critical of the epistemological underpinnings, too.