Building on Solid Ground: Robust Case Selection in Multi‐Method Research

The social sciences are currently witnessing a trend toward multi‐method research (MMR). However, many important issues have not been sufficiently addressed so far. The focus of this paper is case selection for process tracing on the basis of regression results, which is the main point of intersection between the two methods. Based on a review, we first show that the current empirical and methodological literature does not fully appreciate the implications of modeling uncertainty and non‐robust quantitative results. The major problem is that non‐robust regression results may lead to invalid choices and faulty inferences. We develop a novel selection procedure that takes these issues into account and improves causal inference in MMR.