I have been struggling lately to find the data I would like to use for my next CA-related project. To this end, I’ve had to start thinking about this very early piece in the analytic process – finding and/or generating relevant data – moreso than ever before. I think in our training, quite often it is assumed that the object, sites, and interactions of interest are pre-existing. This is ironic given how much we are supposed to align our data procedures to the questions we ask; we often go the opposite way (“what questions can we ask and answer with this data we already have”).
I am interested in how macro- and popular discourses around issues of (in)justice. Following Kitzinger & Frith, there is, in my view, quite a large collection of instances wherein policy, practice, or discourse move contradictory to our understanding of every day talk. For example, when we hear “critical methods are revealing,” rarely does an actant orient to this by responding, “well, all methods are revealing.” We do not hear an assessment of the relative importance of a particular category as implying assessments of other categories. And yet, “black lives matter” is responded to with, “well, all lives matter.” This is one of many potential critical CA questions that I think CA is well positioned to answer.
But now, I am “looking” in quite a motivated way. We have not really belabored Sacks’s approach of “unmotivated looking,” but it seems to me that the call is more to not assume that a particular turn is performing a particular action, or even that a particular action is being performed at all. Rather, we take the “data” as is and see what’s what. I don’t hear this as the opposite of choosing data specifically.
What of looking for a lexical action? It seems natural to search a corpus for all instances of refusals and then to further analyze how these are done. This may indeed be a type of coding. But it’s really just combing through the data and narrowing down the analytic focus. What of looking for a particular term, etc, to see how it is taken up differently in different places? All of this feels like “fair game” and not necessarily in contrast to unmotivated looking. That said, I could imagine a critique suggesting that by narrowing down the lexical term of interest, you may miss other lexical terms that are performing the same function (and vice versa – by looking for a particular function right off the bat, you might miss deviant cases quite easily).
Back to coding, then. As we try to narrow the data set, coding is natural. Isn’t it in all/most qualitative work? Memoing, narrowing, and seeking emergent patterns – that which is similar across the data – seems quite natural. But in my read I cannot fully grasp if Stivers is suggesting the kind of coding that is quantitative for the purposes of statistical analysis. It does not seem that she ever gets as far as t-tests and whatnot. In this case, coding is then just for the purposes of saying “44% of cases did x” or what have you, and this makes coding part of the storytelling rather than of claim-making.
I certainly appreciate the idea that coding might address demographic-related information in a way that CA alone cannot. This in turn suggests that coding might be quite useful to a critical agenda. I am not totally convinced that I would use this approach, but I find it quite productive – the medical example was compelling. It reminds me of Samantha’s question some weeks ago about whether or not CA could see if two different racial groups were oriented to differently by a university officer.
But I am thinking about her second stated boon a bit more. She writes, “The second advantage of this approach and thus the second account for the mixed methods expansion is that it allows conversation analysts to reach audiences who would otherwise be inaccessible. In some cases these are professionals—physicians, medical educators, journalists, or journalism instructors, etc. CA research can offer powerful implications for professionals” (p, 12). This prompts a question from me. When research cannot be communicated effectively, is it appropriate to change the methodology? Or is it just an extra challenge for the researcher to explain why this methodology as is is appropriate, or why the findings are valid, etc. Should we change how we talk about the method, or change the method itself (which, in my view, is essentially what Stivers suggests)? It is tempting to say, no, of course we cannot sacrifice methodological purity. But purity to what end? If you want to make, say, schooling better, your priority really can’t be theoretical fidelity or methodological purity. It must be practical advancement. Isn’t it possible that lots of really powerful research never got taken up in meaningful ways because the researcher insisted on purity tests? I’m not sure what I would do in this situation. I think we have our theoretical, epistemological, and methodological commitments for a reason and that they’re not all pedantic – sometimes they are really important! So I suppose it’s a question of picking your battles. I would be quick to critique the idea that we should change our CA methods because we’re not able to talk about the current ones in a compelling way, but I wanted to stop and think through this critique.
This blog post took a sharp left turn halfway through but the moral of the story is I still don’t know what data I’m using for class whoops, and also hm, coding.