Generalizability

Conceptual Overview and Discussion

The notion of generalizability has long been a concern for survey researchers. When, for matters of feasibility, they need to confine the study to a sample, they want some assurance that the conclusions remain valid for the entire population. Many statistical tests have been designed to certify the possibility of the generalization of their results. But because a precondition for ensuring generalizability is a large sample size, which is unlikely when undertaking case study research, these statistical tests are not always feasible for case study research.

From a positivist approach, the researcher looks for correlations between phenomena or variables. The more general are the findings, the more valuable is the research. In this endeavor, a case study can represent only a very small sample, exploring a new context, giving the “flesh” and the circumstances for a better interpretation of the correlation. But in order to ensure the generalizability of case study findings, further studies that rely on the same logics as survey research are required. A case study provides the exploratory phase of research upon which more extensive quantitative research can be based.

Most of the various strategies in case study research that act to ensure generalizability rest on realist ontological assumptions. This means that a reality is presupposed as existing underneath and underpinning the apparent world. A case study is one of the best of manners in which to access this underlying reality. Once this reality has been rendered visible and explainable thanks to one case study, the researcher seeks to verify the existence of that reality in other cases to assess the extent to which the findings are generalizable to a specified population. This process has been undertaken in research informed by realist ontological assumptions, including that of Barney Glaser and Anselm Strauss, Robert Yin, and Haridimos Tsoukas.

...