ABSTRACT
In an influential article entitled ‘Case study and situational analysis’, Clyde Mitchell has argued that since the early 1950s there have been major developments in survey research techniques, but that there has been little parallel development in ‘the epistemology of the case study’ (Mitchell 1983: 188). His article is an attempt to remedy this, by outlining the nature of case study analysis and showing that it constitutes a ‘reliable and respectable’ procedure (p.207). Mitchell argues that much criticism of case study research, and in particular the claim that it is invalid because it seeks to generalise from a single case, is based on the misconception that case study and survey research use the same basis for extrapolating from data to theory. This is a misconception, Mitchell claims, because while survey research uses two types of inference, statistical inference from sample to population and logical inference about necessary or logical relationships among features within a case, case study research relies entirely upon logical inference:
The validity of the extrapolation (in case study analysis) depends not on the typicality or representativeness of the case but on the cogency of the theoretical reasoning.
(Mitchell 1983: 207)