Continuing with my discussion of my recent book and some of the scholarship issues that I worked through, today I'm going to write about some of the tips and pitfalls I discovered actually doing qualitative empirical research. Hopefully some of these can be helpful to others who are embarking on a similar endeavor:
Interview recording: We started the interviews by having a stand-alone mic that we plugged into a laptop computer to record the interviews. We quickly realized this was too cumbersome and bought a small digital recorder, along these lines. This was much better -- easier to use and less intrusive. You can use your phone's voice recorder, but you always run the risk of running out your battery. Having a dedicated digital recorder is a better plan.
Post-interview form: One thing I regret not doing is having a basic post-interview form for us to record our basic thoughts following each interview. Things like - important points, key stories, apparent themes, issues raised, etc. When you do so many interviews, it's easy to forget these things that you might have fresh in your mind right after. Having a form to complete following each interview would have helped.
Transcription: We didn't even try to transcribe the almost 100 hours of interviews ourselves. Instead, we hired someone via Elance who had been recommended to us by a colleague. She was great and charged $40 per interview hour, much less than some other services we looked into.
Coding software: Getting to coding and analysis time, the big question was what software to use. NVivo is the big player here, but it just didn't work well for our purposes. Having two authors working on the coding at the same time presented problems with NVivo - one of us had a PC and the other a Mac, and NVivo was not so easy on a Mac. Also, syncing our updates was not easy on NVivo, involving extra server software that was very expensive.
Dedoose solved these issues and more. Dedoose is internet-based software that runs the same regardless of the computer (since it's on the internet). It updates real-time, so if two people are working separately their updates can be seen almost immediately by the other. Both of the major NVivo issues were solved with Dedoose.
Plus, Dedoose was easier to use in many ways. Its interface was intuitive and very easy to use. Its customer service was very responsive. It was relatively cheap, did everything we would have asked NVivo to do with our data, and had helpful video tutorials. We started with Dedoose almost out of necessity; we continued using it because it was easy and powerful.
Developing codes: We made the mistake when we developed the codes for our interviews of creating too many codes and sub-codes. This seemed to make sense at the time, as we wanted to capture fine distinctions, but it became too unwieldy as we tried to actually apply the codes. Thankfully, Dedoose allowed us to merge the codes without too much trouble, but it would have been better if we had been more judicious with our codes and sub-codes at the start.
Fake names: This is a little thing, but it was hugely helpful. If you're like me, when you write exams, you make up fake names for your fact patters. At some point though, you run out of good names and feel stupid using combinations of family members' and friends' names. For our book, we use almost all fake names (to protect identities). It would have been a bear to come up with so many fake names. But thankfully, there's a website to solve the problem. Actually, there are many websites. The best one I found was the appropriately named Fake Name Generator. It generates all sorts of fake names for you, customizable by sex, country of origin, and ethnicity. It was an invaluable resource for writing.
There are other decisions to make in actually doing qualitative empirical research, but these are the ones that were either the most important or the most valuable in passing on. When in doubt, there's no better advice than this - ask someone else who's done it before!
Comments