External Validity and Generalization: Challenges and Strategies

Abstract

The final task of case study research interested in regularities involves the generalization of causal inferences from those cases that one did examine to those cases that are part of the population and have not been analyzed.¹ In qualitative case studies, one generalizes inferences about causal effects and causal mechanisms because a regularities perspective implies the assumption that both are regular (Kühn and Rohlfing 2010). The small-n literature identifies generalization as a major problem for qualitative case studies (Rueschemeyer 2003) as it lacks tools, such as significance testing, for determining the likelihood that the generated results are due to chance or to systematic cause–effect relationships.² The lack of such an instrument gave rise to the assertion that case study researchers have to assume deterministic cause–effect relationships in order to generalize at all (Lieberson 1991; Munck 2005).