Improving Research on Necessary Conditions: Formalized Case Selection for Process Tracing after QCA

Abstract

AbstractThis paper aims at strengthening causal inference in necessary condition research. We demonstrate how process tracing based on purposefully selected cases can complement findings on cross-case patterns identified with Qualitative Comparative Analysis (QCA). Using an empirical example, we discuss the meaning of typical and deviant cases in analyses of necessity, develop formulas for identifying the most typical and most deviant cases, and detail the implications of so-called SUIN conditions for meaningful case selection. In addition, we clarify various viable variants of comparative process tracing and formulas for identifying the best-matching pairs of cases.IntroductionNecessary conditions play a more prominent role in social science theory than is often acknowledged. Moreover, this type of condition has unique causal qualities: if it is absent, the outcome cannot occur but, conversely, when it is present, the outcome does not always occur (Harvey and Starr 1989: chap. 3). In recognition of their substantive relevance and vexing properties, the methodological literature has made considerable progress in improving the analysis of necessary conditions. Braumoeller and Goertz (2000) introduce statistical tools for assessing whether a given distribution of cases that is not fully in line with a pattern of necessity is still good enough to claim that the condition in question is necessary. Similarly, Ragin (2006) proposes the descriptive measure of consistency for necessary conditions, which expresses the extent to which a distribution of cases is in accord with the claim that a condition is necessary.1 Once a condition is declared to be necessary for an outcome, the follow-up question is how relevant or trivial it is, a task taken up in proposals by Ragin (2006), Goertz (2006), and Schneider and Wagemann (2012). Further conceptual improvements in the analysis of necessity include Ragin's (2000) extension of this concept from classic crisp sets to fuzzy sets, and Mahoney, Kimball, and Koivu's (2009) notion of SUIN causes. SUIN causes denote attributes of cases that are sufficient but unnecessary for a condition that is insufficient but necessary for an outcome.These developments in necessary condition research are valuable contributions to the improvement of Qualitative Comparative Analysis (QCA) as the most formalized technique for set-relational research (see also Bol and Luppi in this issue). Yet, all developments are primarily, if not exclusively, concerned with assessing necessity based on a crosscase perspective. The underlying causal mechanisms and processes operating within cases are largely neglected. This is problematic because is a case-based method that is most effective when researchers engage in a dialogue between cross-case and within-case analysis. In light of this, the goal of our paper is to improve the current state of necessary condition research via by developing principles for the choice of cases after set-theoretic results have been generated on the cross-case level.We deem it analytically fruitful to decompose the constant dialogue between ideas and evidence (Ragin 2000, 317) into a pre-QCA and a post-QCA phase.2 In this context, QCA refers to as a technique (Berg-Schlosser, De Meur, Ragin, & Rihoux 2008), that is, the analytic moment of deriving set-relational cross-case patterns from a truth table. The distinction between pre- and post-QCA is important because case studies performed in these different phases have different goals. The main purpose of pre-QCA case studies is to delineate the universe of cases (Ragin, 2000, chap. 2), to search for the sources of contradictory truth table rows (Ragin 1987, chap. 6), and to increase measurement validity (Rihoux & Lobe 2009). The main purpose of post-QCA process tracing, instead, is to improve the theory underlying the model.3 There are two ways of doing so. By studying typical cases, existing hypotheses on causal mechanisms can be tested or new ones developed. …