Where to Begin

The problem of selecting independent variables for qualitiative comparative analysis (QCA) is addressed. This is a different problem for QCA than for inferential statistical methods, for both technical and epistemological reasons. Technically, QCA can manipulate only a few variables at one time. Epistemologically, QCA expects causation to work in a combinatorial fashion. The authors isolate and reject four ways of choosing independent variables for QCA and advocate a fifth method, the conjunctural theories approach, which is more compatible with the characteristics of QCA. Their decision is supported by way of discussion and an empirical analysis based on theories of the welfare state and U.S. social spending in the Great Depression.