I am currently working through the interviews of my first study, using Nvivo 10 to help me code the transcripts. These interviews explored how people at a wedding engaged with photography during and after the event. I will use this case to develop a framework for exploring personal photography practices that includes the purposes for which people engage in these practices and the technological and cultural drivers that influence them.
I have plenty of reservations about Nvivo, in particular the way that it seems to coerce me into treating the data in particular ways that I might not otherwise have wanted to do. An example is its quantification of everything – it’s easy to get sucked into comparing the number of instances of different codes and then attaching an inappropriate value to this comparison. I am keeping thorough notes on these observations as well as any other methodological issues that come up.
I have decided to use Framework Analysis (NatCen, the originators of this approach, have good information on their site about this) because of the clarity it encourages around reporting on research process and how conclusions are arrived at.
I have six transcripts to analyse and, having familiarised myself with the data, I am currently doing an initial coding (called “indexing” in Framework Analysis terms) to refine a thematic framework that I developed through a literature review and some pilot work. Once I am satisfied that I have a framework that fits the data, I will do a second round of coding, then “chart” coded excerpts according to each theme in the framework which should allow a thematic view across interviews. I will then conduct the “mapping and interpretation” phase, which is essentially an analysis of secondary data where excerpts and themes are visualised through diagrams, patterns are searched for, and theories and typologies are generated and defined. Any concepts emerging from this step will be checked against the original data by returning to the close readings of the transcript sections from which the excerpts were extracted, and, potentially, by validating conclusions with the participants themselves.