Qualitative data redux

I've been asked to chair a session at the annual Big Conference for Data Library People in May. That is good; it makes it look like I'm going there to network, be professionally productive and pad my scholarly resume rather than because it's a chance at a subsidized vacation with a bunch of interestingly odd people who like data and tell dirty jokes.

However, the session they've asked me to chair is the one on qualitative data. Pretty much everything that I know about qualitative data was summed up in an earlier blog post where I noted that, as far as I could tell, qualitative analysis software doesn't actually do anything, or at least nothing that I'd consider analysis. I'm still waiting to be enlightened on that.

I wonder if this invitation is a result of one of those bits of cosmic irony, or if one of the conference organisers has been reading my blog.

Books

I love that bug-eyed look students get when they come in asking for data on something and I go away then come back with an armful of books. In this case, musty, dusty, thick books. Sorry kids, but no, not everything is online.

Not that I don't wish that Statistics Canada would digitize its back publications. But I take my entertainment where I find it.

Software Paradigm

Last month I took a one-day class on a qualitative data analysis software package, NVivo. Qualitative data has long been something of a problem for me as a person who is expected to deal with data. I've played with the software, gone through the tutorials, and even helped a student with file import problems, but none of that got me past this basic question: what, exactly, does qualitative analysis software do?

The answer came to me during this class. It doesn't do anything.

Let me unpack that a little.

With statistical software, the quantitative stuff like SAS, SPSS and Stata I'm familiar with, there are all sorts of tools for editing and manipulating and looking at your data, but the point of all this data manipulating is to get it into shape so you can use the software to run a statistical procedure or two. At that point you're done: you inspect your results and go write them up. (At least in theory. My husband the former grad student discovered that one can spend years trying to perfect a model.) But the point of the exercise, the whole reason the software exists, is to run that model. All the other functionality is subservient to that.

With qualitative software, there are likewise all sorts of functions for editing and manipulating and looking at data, many of which I'd noticed previously and some of which I'd even figured out how to use. My problem was that I couldn't figure out what the point of it was: I'm manipulating the data to get it into shape so that I can use the software to do what, exactly?

But the software doesn't do anything - there is no final step, no reason for editing and manipulating the data. Or rather, editing and manipulating the data is the whole point, at least as far as the software is concerned. The final step, making sense of the data, test ing theories and drawing conclusions, has to take place in the qualitative researcher's own head. It's a curious and initially somewhat worrisome concept.

Back at work

And a question: if I tell a patron who is looking for data on some topic (actually, in this case a rather exhaustive list of topics) that I need to know more specifically what s/he is looking for so that I can actually find something relevant, and the patron responds "Just send me everything," is it appropriate for me to reply with a link to Google and a map to the library?