The question of how best to account for the multidimensional character of welfare states has become an integral part of discussions on the so-called dependent variable problem within comparative welfare state research. In this paper, we discuss challenges from an attempt to capture productive and protective welfare state dimensions by means of several methodological techniques, namely Z-score standardisation, cluster analysis, factor analysis and fuzzy-set ideal type analysis. While we illustrate that a decision to use any one of these techniques has a substantial bearing on the produced findings, we specifically argue that fuzzy-set ideal type analysis offers considerable advantages over more traditional, statistically rooted approaches. This is particularly true if the observed dimensions are conceptually distinct and ‘antithetical’.
Copyright © Cambridge University Press 2010
Access to external full text or publisher's version may require subscription.