Ihre Browserversion ist veraltet. Wir empfehlen, Ihren Browser auf die neueste Version zu aktualisieren.

It's sets all the way down (and up)...

Posted 19/7/2017

At the 7th Annual General Conference of the European Political Science Association, which took place in Milan (Italy) from 22-24 June 2017, the QCA panel featured a particularly interesting paper by Kevin Clarke, Associate Professor at the Department of Political Science of the University of Rochester (USA).

The paper was titled "It's Sets All the Way Down (and Up)". Its main argument was that QCA's label as a "set-theoretic method" does not distinguish it from traditional quantitative methods because quantitative methods are also grounded in sets; the theory of probability makes extensive use of sets of events. In other words, when proponents of QCA argue that the set-theoretic orientation of QCA accords closer to the logic of social science theory than the correlational net effects logic of quantitative methods, they demonstrate they have not understood what quantitative methods are. The conclusion of the presentation was that QCA has one fewer leg to stand on in consequence.

As Clarke was also citing my work, I reminded him that I have long switched to the term "configurational comparative methods" because I also found the term "set-theoretic" too general and misleading, but certainly not disqualifying. His reply was that it's only me who uses this term. I then reminded him that a very prominent edited book by Rihoux and Ragin is called Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques).

In addition, why should a certain label diminish the utility of a method? Surely, "set-theoretic" implicitly involves more than the notions of union and intersection to compute the probability of two independent or dependent events. It involves the whole apparatus of Boolean algebra, Boolean minimization in particular. All this is lacking in the group of methods Clarke referred to.

The paper goes to quite some lengths to revisit the essentials of random variables and probability theory, but these are pages completely wasted if the goal was to demonstrate that QCA has no value beyond what traditional quantitative methods are capable of doing, just as William Clark, Michael Gilligan and Matt Golder had already tried in a paper published in Political Analysis in 2006. Thus, to me at least, Clarke's anti-QCA paper merely demonstrates that the dogmatic critics of this method start to run out of ammunition.

(if you'd like to read the paper by Clarke, just drop me a line)