On Coding an Interview

I am looking at a pile of interviews on my desk. The phase of conducting interviews is nearly over, I have to start analyzing them. After some postponing (a grant proposal had to be written, a book had to be read) I finally take a deep breath and start.

I take the first part of the first interview, which I transcribed in full. Line by line I read it and attach codes to lines, to alineas, and sometimes to single words – codes being words, concepts, short phrases that capture what is being said in the interview on a slightly more abstract level. The idea is that through coding and recoding, constant comparison of passages, writing memos on the process during the process, eventually a picture arises of a possible interpretation of the interviews - or of several possible interpretations. Codes get more and more abstract during the process, and eventually you end up with some sort of theoretical model.

But I am not there yet. After having worked through the first part of the first interview, I have a list of 170 codes. Way too much, but I should probably not bother so I start fiddling around with them – putting them in families (uses, functions, developments, narration traits…), subdividing the families, trying to make sense of the relationship between the families, trying to find out if I should not skip one of the families because it still is too much for a useful analysis, and asking myself where those families come from. Am I theoretically too prejudiced, do the families really come out of the interview or did I  put them in? I know “interviews don’t speak”, texts do not turn into theory by themselves, a researcher is needed. But I don’t want to end up with something I knew beforehand – if I would like to write up what I think I should have become a philosopher, an essayist, or a theorist; but I want to stick to empiry,  be surprised by it, tell a plausible story about it.

The 170 initial codes seem to be transferable into a set of about 40 slightly more general codes. Some of them are still empty but I know there are interviews which will fill those codes. Some of the families seem unbalanced still – the most important family (at least that’s what I think now), the one of functions, will probably shrink from 11 to something like 5 or 6 codes. But I don’t know how yet.

And then there is literature, and theory. I am dying to connect my codes to existing theory, to literature. I know for some of the codes and code families what to read, where to look; for others I am not sure yet. But I have decided to do a lot more analysing first, trying to keep the theory a bit out, for as far as that is possible. Interviews don’t speak, but if the theories shut up, occasionally you hear them whisper.

So tomorrow it’s back to coding. I don’t know where things come from, and I don’t know where I am heading. That makes me uncertain. But in writing this down, I recognize that that is something described in all the books, and that I have to take my time. Dust will settle gradually. The book will be written eventually. And after that – more coding…