Academic journal article Journal of Comparative Politics

Pitfalls in Qca's Consistency Measure

Academic journal article Journal of Comparative Politics

Pitfalls in Qca's Consistency Measure

Article excerpt

(ProQuest: ... denotes formulae omitted.)


In the almost thirty years since the publication of Charles Ragin's "The Comparative Method", Qualitative Comparative Analysis (QCA) has developed into a widely-used analytical technique in political science. The number of QCA-related articles published in peer-reviewed journals is increasing exponentially, from forty-five in 2012 to no fewer than ninety-nine in 2013 (Marx, Rihoux and Ragin 2014, 115; Rihoux 2014). Over the years, the technique went through numerous modifications and adjustments. One of the most important developments was the introduction of the consistency measure, which eventually became QCA's single most important parameter for assessing sufficiency and necessity (Wagemann and Schneider 2010, 289). Strikingly however, this formula does not meet the requirements Ragin (2006) formulated when he introduced the measure in its current form. Contrary to the latter's assertions, small disconfirming cases have greater bearing on the consistency score than large disconfirming cases. Consequentially, the standard consistency measure does not adequately express the degree to which the empirical data is in line with statements of sufficiency and/or necessity.

After revealing this flaw in QCA's most important parameter, this article demonstrates how it leads to the misinterpretation of empirical evidence for sufficiency and necessity and introduces a formula that more accurately assesses the evidence for sufficiency and/or necessity. The article is structured around three main parts. First, the general purpose of calculating consistency is described. Subsequently, I demonstrate that the standard formula does not meet all requirements Ragin deemed necessary to achieve this purpose and introduce a new formula, which more accurately assesses the evidence for sufficiency and/or necessity. Finally, two recent applications of fuzzy set QCA, Mello (2014) and Schneider and Makszin (2014), are used to illustrate the impact of the flaw on empirical research and the benefits of using the alternative consistency measure.


QCA is generally used to establish set-theoretic connections between one case property, defined as the outcome, and other properties, defined as the causal conditions (Wagemann and Schneider 2010, 380). As extensively demonstrated in Ragin (2000, 203-260; 2008, 13-28) and Schneider and Wagemann (2012, 56-91), such subset relations are intimately linked to the notions of sufficiency and necessity. Since a condition is sufficient if the outcome is always present when this condition is present, the set defined by a sufficient condition constitutes a subset of the set defined by the outcome. Inversely, a condition is necessary if it is always present when the outcome is present. Therefore, the set defined by a necessary condition constitutes a superset of the set defined by the outcome.

The assessment of set-theoretic connections is straightforward in the original crisp set version of QCA. Cases are either present or absent in a crisp set, respectively indicated by a value of 1 and 0. In consequence, establishing a setrelation solely requires examining whether each case with a score of 1 in the alleged subset also has a score of 1 on the outcome. This straightforward procedure cannot be duplicated in the more sophisticated fuzzy set QCA, in which membership scores can vary between full membership (value of 1) and full non-membership (value of 0). In fuzzy sets, assessing subset relations requires examining whether each case's membership score in subset X is consistently equal or less than its corresponding score in superset Y, thus whether X < Y.

Perfect subset relations and fully necessary or sufficient conditions are relatively rare in social science (Ragin 2000, 108). This inspired Ragin to introduce the consistency-parameter, which provides a descriptive measure of the degree a perfect set relation is approximated (Ragin 2006, 292; Wagemann and Schneider 2010, 389). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.