1. Why are case studies considered particularly appropriate to reveal causal mechanisms?
When comparing the several methodological tools available in social research, case studies seem to be particularly appropriate to contribute to establishing complex causalities in a particular theory or in international politics in general. Using mostly qualitative evidence, case studies complement process-tracing methods to analyse causal inference in politics. From an inductive perspective, historical approach to specific events may test theories through formulating hypotheses. As a result, using process-tracing method in a case study allows the researcher to explain causal mechanisms and outcomes, in particular why several different causal paths lead to the same outcomes or why several different outcomes originate from a single variable (George and Bennett 2005, 9-10).
Moreover, a case must be not only integrated into a wider direction with a comparative purpose, but also investigated internally to identify all the intervening variables, so a case study is a ‘well-defined aspect of a historical episode’ (George and Bennett 2005, 18). Nevertheless, causality and correlations must not be confused as a causal relation aims at generalizing a claim, even though it has limited potential to estimate an average or a recurrence of a variable. To conclude, it must be noted that there is no consensus on the appropriateness of case studies to uncover causal mechanisms, as Gerring asserts that a single-case study is a ‘way of defining cases’ rather than ‘a method of modelling causal relations’ (2004, 353).
2. What are the advantages of a cross-case study approach over a single case study approach (refer to Gerring’s arguments)?
Gerring proposes a classification of strengths and weaknesses of both single-case and cross-case study according to seven considerations. In a cross-case study, the type of inference is causal (that is to say deductive), the scope of proposition is both breadth and boundedness (which refers to the incorporation of spatial and temporal elements), the unit homogeneity is characterised by representativeness, the insight is on causal effect, the causal relationship is probabilistic, the strategy of research aims at testing or confirming theory, and the useful variance concerns many units (2004, 346).
However Guerring highlights his preference for descriptive inductive case studies which are ‘easier to conduct’ (2004, 347). The depth of the unit’s scope, the comparability potential of the single-case study, the causal insight on mechanisms, the invariant causal proposition, the possibility to generate theories and more generally the variance for a single unit are key elements. Indeed, in his methodological discussion, the author exposes the risk for the researcher to lose familiarity with the complexity of cases, although providing contextualised details allow to preserve the uniqueness of the cases selected.
To conclude, Gerring claims that a single-unit case study is generally more useful than a cross-units case study (2004, 352) although he argues for a complementarity of single-unit and cross-unit approaches. In terms of importance, the variations of the theory and the constancy of other variables are determinant in choosing which methodological approach is most appropriate.
George, Alexander L. and Andrew Bennett (2005) Case Studies and Theory Development in the Social Sciences. Cambridge, MA and London, England: MIT Press.
Gerring, John. ‘What is a Case Study and What is it Good for?’. American Political Science Review 98 (2004): 341-354.