Validity and Circumstances
"Validity is not a commodity that can be purchased with techniques...Rather, validity is like integrity, character, and quality, to be assessed relative to purposes and circumstances"
(Brinberg & McGrath, quoted in Miles & Huberman, p. 39)
I liked this quote; it seemed somehow reassuring. Our quick ramble through various approaches to validity in qualitative research left me feeling like it was a hodge podge of or wish list of "ideal types" that were each in their own way trying to define a multi-dimensional, historically contingent, ideosyncratic process as a codifiable regime of trustworthy inferences. That's a tall order.
The typologies of validity are useful, however, foremost for directing our attention to the many ways in which our research can fall short in one or more dimensions. Thinking through the details of what makes for authetic findings is critical to the research enterprise. Valid inferences are the goal, but the pitfalls loom everywhere (errors of omission or characterization, for example).
I do wonder about the refrain that holds qualitative studies are doomed to have limited generalizability. This overstates the case, don't you think? What kinds of qualitative studies might result in findings that are generalizable to a wider popluation?
posted by Stu @ 6:01 PM
4 comments
4 Comments:
I think the dialogue against the generalizability of qualitative studies occurs because it is easy to overemphasize the differences that exist amongst similar groups, while neglecting the commonalities that also do exist. Qualitative research, it seems, can be generalizable, because the overlying behaviors that affect human processes, such as, for example, reacting to a certain teaching method or how one interacts with web technology, can be, and have been, categorized and classified. The researcher must exercise caution when comparing groups or groupings of data; nor does he or she want to create to rigid of boundaries. But, I believe the overlying commonalties allow for generalizability. Chapter 8 of the Huberman and Miles text does a nice job of providing examples and examining this issue; I particularly found the concept of “fittingness” interesting (Huberman & Miles 178).
I like this quote as well. Sometimes I think of validity in a rigorous, cut and dry, formulaic outcome. But in reality, it can also be the sum of consisent behavior + truth. This fits nicely in with our search for validity in a qualitative world.
I think the discussion about generalizability has to do with the relative difficulty each method has of moving the arguement from cases studied to general population. If you look at one of the recent postings on my blog, you'll see where I tried to define the research space along the dimensions of "aspect of experience" and number of subjects. Qualitative research looks at lots of aspects of a relative small number of subjects. Quantitative looks at a relatively large number of subject, but focuses on a small number of aspects.
If the goal of generalizability is to "predict" or "anticipate" how all subjects will likely experience the same aspect, quantitative research has the advantage. It's already looked at more subjects. Generalizability is more suited to quantitative studies.
I see the analog to generalizabity in qualitative studies as the goal of predicting how an individual subject will experience other aspects than the ones studied.
I don't see generalizabity having the same meaning in the two research types, so we probably need to find some other type of construct that will be the equivalent of it for qualitative research.
Coming from a biomedical/nursing perspective I see qualitative research as the stepping stone to intervention-type studies which would then yield emirical support for changing practice or initiating new practice parameters. It is a stepping stone, but one that is required to elludicate "the problems" in order to move on to additional types of research that affect the health care system in our country (and perhaps internationally). So, while many may say, "it's not generalizable", the subsequent research that evolves from qualitative certainly is/should be generalizable.
Post a Comment
<< Home