A research design for generalizing from multiple case studies

Many evaluations would be more useful to decision makers if their results were both explanatory and generalizable. In view of the limitations of traditional methods in meeting both these goals at the same time, the authors describe a research design explicitly intended for generalizing about the range of applicability of “explanatory patterns” (within a designated “target” population of cases). The paper presents an overview of the main features of a “multiple case study design,” shorthand for a multiple site, structured case study design. It then discusses the nature of explanatory patterns, how case study investigators pursue and recognize valid patterns, and how an analyst can apply the same logic to cross-site analysis in order to make valid inferences about the limiting conditions under which particular explanatory patterns apply to particular populations.