In lieu of an abstract, here is a brief excerpt of the content:

Reviewed by:
  • Fuzzy-Set Social Science
  • Tim Futing Liao
Fuzzy-Set Social Science. By Charles C. Ragin. University of Chicago Press, 2000. 352 pp. Cloth, $48.00; paper, $18.95.

Whilst in the 1980s applying fuzzy logic to American Catholic membership in my doctoral research on fertility differentials, I kept a vigil on my university library bookshelves for new arrivals on fuzzy set theory and applications. All I read were books in engineering and computer science. The only social science entry I could locate in the Library of Congress catalogue was Michael Smithson’s Fuzzy Set Analysis for Behavioral and Social Science. I quickly purchased a copy but put it [End Page 354] aside almost as quickly for it was not terribly useful in dealing with the wide range of possible social science application of fuzzy set theory.

The need for a general book on applying fuzzy logic in the social sciences, unfilled by the Smithson book, remained so for the remainder of the century. Has Charles Ragin’s Fuzzy-Set Social Science filled the void? The answer is “No” though the book otherwise has extremely important messages for empirical researchers.

The cardinal component of fuzzy set theory is its grade of membership (GoM) concept. Simply put, in classical logic a set is crisp with a true or false membership whereas in fuzzy logic a set is noncrisp (or fuzzy) with a membership varying between null and full and its GoM ranging from 0 to 1. For example, conventionally Catholics are conceived as a crisp set, one either is or is not a Catholic. Using fuzzy logic, one may be a non-Catholic (0), a Catholic (1), or a partial Catholic with a GoM value between 0 and 1. Fuzzy memberships can be developed and estimated using a spectrum of methods. I used latent class models for the estimation; Max Woodbury, Ken Manton and associates developed a GoM model (with software) for multivariate analysis in social and biomedical research. Manton and colleagues published widely, including a 1992 chapter in Sociological Methodology on using GoM techniques in regression. My initial disappointment by the absence of any such references in Ragin’s book was replaced by the discovery that the book purports to introduce not fuzzy logic and its potential or accomplished social science applications, but its application to a particular method known as qualitative comparative analysis (QCA) that Ragin developed over the years beginning in his 1987 book, The Comparative Method.

Ragin’s 2000 book has two parts. In the first half of the book he presents the basic principles of diversity-oriented research (or QCA for the new millennium), and in the second half, he applies fuzzy logic to QCA, which has previously relied on conventional logic. The diversity-oriented approach is pitted against the quantitative, variable-oriented approach. In doing so, Ragin makes some valid and valuable criticisms of the conventional quantitative approach.

One such criticism is that the conventional approach is insensitive to the definition of population since researchers in this convention work with a predetermined population. However, the definition of a population is more complicated and must use case-based knowledge. In addition, the conventional quantitative approach is ignorant about causal complexity and incapable of dealing with it. In contrast, QCA can analyze necessary and sufficient conditions and distinguish between them. The advice to establish a dialogue between ideas (substantive theory) and evidence (data) is also crucial for conducting meaningful sociological research. These and other well-taken points characterize diversity-oriented research.

In theory, fuzzy logic elevates QCA. But how much does fuzzy logic add to QCA in practice? The book has two empirical examples, one analyzing IMF protests and the other studying welfare state generosity. However, it is not clear how much difference fuzzy logic makes as no comparison is made with other methods, except [End Page 355] some brief comments. For instance, certain findings are similar to those from a previous analysis using conventional quantitative techniques, and the method based on fuzzy sets offers a different approach to the problem of multicollinearity. The reader would benefit from a comparison of the results based on three methods: conventional quantitative analysis, QCA, and QCA with fuzzy logic...

Additional Information

Print ISSN
pp. 354-356
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.