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

Rejoinder to Hug's reply in Qualitative & Multi-Method Research 12(2):24-27

Posted 7/2/2015

I thank Simon Hug for his reply (Hug 2014) to my comment (Thiem 2014) on Hug (2013).

I am also very grateful to Robert Adcock, the outgoing editor of Qualitative & Multi-Method Research, for having made the publication of this comment possible after Political Analysis had rejected it. Since the deck is always stacked in favor of that scholar who makes the first move because she gets the last word, I use this blog as a means to redress the balance in a final rejoinder.

In the first part "Mill's Methods and QCA", Hug (2014:25) eschews a confrontation with my argument that QCA is a much more advanced technique than Mill's methods of agreement or difference for the configurational analysis of observational data in the social sciences. Instead, he merely references "prominent QCA scholars" (2014:25) who have drawn some connections between Mill's methods and QCA.

Subsequently, Hug digresses to topics of maximum likelihood estimation and split-sample designs in the second section "Deductive Use of QCA", instead of addressing my point that it makes no sense to conceive of QCA as an originally "deductive" method. As argued in my comment, the truth or falsehood of an implicational hypothesis can be tested by simple data parsing in search of empirically falsifying facts. But this is not the primary objective of QCA. Instead, it is to identify causally interpretable models by analyzing configurational data. The testing of hypotheses can be part of this process at various stages, but this is not its main element.

In the third section on "Measurement Error and Missing Observations", Hug criticizes that I described it as awkward to run a simple QCA loop over a collection of data sets and call it grandiloquently "Monte Carlo simulation", although there was no stochastic component whatsoever, and simple combinatorics that are mathematically faithful to QCA could have done the job at a tiny fraction of the effort (2014:26). Of course, exhaustive loops to derive QCA solutions as used by Hug will always work, but they will also always mask what is going on behind the scenes, something Hug seems not to be interested in. In addition, he again remains silent about the errors and problems I have identified in his original article.

To sum up, Hug has not engaged on a substantive level in his reply with any of the points raised in my original comment, but filled three pages with methodological digressions, spontaneous heaps of praise on Collier & Co., and thin-skinned complaints about unadorned language that merely challenged his confrontational assertions.

Hug calls for an open discussion of QCA's limitations in the title of his reply - a goal I would always support without hesitation - but as long as Political Analysis, on whose editorial board he sits and which has just declared his article to be one of its Greatest Hits, continues to blatantly favour QCA-disparaging pieces that are anything but cutting-edge methodology while rejecting all other submissions, such calls reduce to lip service. If QCA's critics go so far as to suggest that the method is useless, they should not be that afraid of having their own arguments properly scrutinized.


Fields marked with * are required.