Types of Case Studies and Case Selection

Abstract

The goal of generating statements about empirical regularities implies that a population of cases exists to which these statements extend. Unless one is able to examine all cases in the population, something that might be possible when the population is small, one faces the pertinent challenge of case selection. The systematic choice of cases is crucial because the case is not interesting in itself (at least not in the first place), but for learning some-thing about the population of cases from which it is drawn. One central goal of this chapter is to show that different types of cases lend themselves to this purpose. 1 The types of cases are characterized by different features, each type having different implications for causal inference and calling for different case selection strategies. Figure 3.1 gives a snapshot of the types of case studies and basic selection strategies and additionally demonstrates that distinguishing between the three dimensions introduced in Chapter 1 – research goals, levels of analysis, and variants of causal effects – is central for case studies and case selection. 2 On the top level, the discussion follows the distinction between the three theory-centered research goals. 3 In each of the three sections dealing with the generation, testing, and refinement of a hypothesis, I first focus on the cross-case level and turn to the within-case level afterward. For each level of analysis, it is further necessary to differentiate between correlational and set-relational propositions, with the latter in turn being subdivided into statements of necessity and sufficiency. 4 For hypothesis-testing case studies, I further describe case selection for the test of a single hypothesis and then extend the discussion to comparative tests of multiple propositions and the incorporation of interaction effects in the case selection process. An additional source of complexity to consider here is that one can follow two different case selection strategies that are tied to different types of case studies. The first strategy follows the idea of distribution-based selec-tion because cases are chosen with respect to their location in a distribution of cases spanned by the cause or causes and/or the outcome of theoretical interest. Depending on the research goal, the three distribution-related types