ABSTRACT
This chapter discusses the tools to use regressions for selecting case studies. These tools are useful when doing mixed methods research. When one works with observational data (as opposed to experimental), linear regressions cannot, by themselves, answer causal inference questions. That is, although they can unveil the existence of a relation between the independent and dependent variables, the investigation would be incomplete if one is not able to demonstrate, through other methods, how these variables are causally connected. The chapter discusses a method called nested analysis, which is the combination of statistical analysis of a large sample with in-depth study of one or more cases contained in that sample. The selection of extreme cases involves identifying observations that are located far away from the mean of the distribution of the independent or dependent variable. The selection of similar cases involves identifying two cases that are similar.
