Read on Twitter
You see this a lot, and the problem is, ironically, when the training/underlying philosophy of science is too focused on making qual methods look as much like quant as possible. The point of qualitative isn't representativeness; it's revealing mechanisms, processes, and meaning. https://twitter.com/JessicaCalarco/status/1213801274677874692
This is why as I get older and plan new projects I get less and less interested in comparison for comparison's sake. I mean, it's cool, and I'll always have a bit of a comparative angle, but the point is really to flesh out mechanisms through deep dives, not to cover every angle.
though to be fair, there are often qualitative papers that try to use qualitative data to make quantitative, representative-style arguments, and in that case, reviewers are right to say the data cannot support the claims.
here's an example I run into a lot. Let's say you're looking at an interview transcript, and there's some degree of fishing, or priming, or any other way to get an answer that isn't "natural": that's only a problem if you're trying to find out if that answer would happen anyway.
But if the point is to determine how X or Y answer is described or experience or narrated then it's not the priming that matters-that can actually be intentional!-but rather what the priming reveals in terms of how X or Y answer is experienced in certain contexts/institutions.
there's still the representativeness question: just how common is X or Y meaning or experience anyway? And that's why the best qualitative and quantitative work is iterative, using qual insights to build better survey questions, then using quant work to contextualize qual work.
of course, qualitative work can provide context to other qualitative work too, and usually does. In my work with creationist Muslims and creationist Evangelicals, I draw on both quantitative and (mostly) qualitative studies to contextualize what is a study of just four schools.
I *do* make a causal argument, as I think qualitative scholars are entirely free to do, and I make an argument I posit is representative based on my reading of other work, but I leave it to my quantitative colleagues if that's worth testing more thoroughly.
