Published online by Cambridge University Press: 04 January 2017
Gordon and Smith (2004) do a great service by introducing innovative and creative quantitative methods that incorporate information from qualitative sources. It is nevertheless important to examine the conditions under which the proposed estimators will be useful in practice. These conditions prove to be surprisingly restrictive: with the possible exception of extremely low-information settings, virtually all of the cases of discernible causation must be coded as such, those codings must contain virtually no errors, and the process by which qualitative researchers produce evaluations of discernibility must conform to the authors' model of the qualitative data-generating process (QDGP) if the procedures are to retain any comparative advantage.