Conditional Hypotheses in Comparative Social Science: Mixed-Method Approaches to Middle-Sized Data Analysis

Abstract

This paper discusses under which circumstances and how configurational comparative methods (i.e. QCA) and statistical methods can be combined to provide tests for the ‘quasi’-sufficiency of any given set of combination of causal conditions. When combined, QCA provides the ability to explore causal substitutability (i.e. multiple paths to a given outcome) and the ways in which many multiple causes interact with one another to produce effects, while the statistical elements can provide robust indications of the probable validity of postulated hypotheses. The potential utility of the mixed-method approach for analyzing political phenomena is demonstrated by applying it to cross-national data regarding party positions on European integration and party-based Euroscepticism in Western Europe. The findings show that oppositional stances to European integration are partly associated with non-governmental ideological fringe parties on both the left and right. The empirical example presented in this paper demonstrates that configurational methods can be successfully combined with statistical methods and supplement the QCA-framework by providing statistical tests of ‘almost sufficient’ claims. However, combining QCA with statistical methods can sometimes be problematic in middle-sized data analysis, especially as the latter usually cannot handle limited diversity (i.e. insufficient information) in the data and/or overtly complex relationships (i.e. having a large number of conjunctional conditions or interacting variables).