Generalizing From Qualitative Data as a Moral Activity: An Example From Regulatory Decision Making for Medical Products
The generalizability of quality research is discussed in terms of the generalizing actions of scientists and policymakers. Using decision making in regulatory drug trials as a case study, I illustrate how questions of generalizability arise when specifying outcome measures to use in clinical trials. Specifically, qualitative concept elicitation interviews are required to infer the relevant symptoms and functional impacts of a particular disease (the Source). Whether such qualitative data and the measure developed from them can be extended to a different disease (the Target) is an important question faced often by researchers and makers of medical products. I suggest that the logical steps in considering such inferences are (1) identifying the salient attributes of the settings that should be compared between the Source and Target; (2) assessing the fit between the Source and the Target with respect to the salient characteristics; and (3) deciding whether the degree of fit is sufficient to justify making an inference about the Target using data from the Source. I underscore the importance of moral considerations when deciding to engage in a generalization from a Target to a Source and call for greater transparency regarding the values upon which people’s generalizations are based.
