A key output of a dual coding strategy is an abundance of open codes. After all, a reasonable set of interviews with experts will accumulate a huge number of different ideas and concepts.
But what on earth do you do with them all? How are you to make sense out of that seeming chaos?
Here in the HIPSTER project, we have been using dual open coding as part of our analytical approach. You can read more about our experiences with dual open coding here in our previous blog.
To get our heads around our own huge list of open codes, we held a focused workshop session. We booked a quiet room, grabbed plenty of Post-it notes and fired up NVivo, our Qualitative Data Analysis tool.
Our first task was to write out the codes onto Post-it notes. You might wonder, what’s the point; weren’t QDA tools designed precisely to take away this kind of labour? Well, yes and no. Never underestimate the power of the Post-it! By freeing your codes from your computer screen and having a physical representation of them that you can tangibly move around, you will find, as we did, that it brings your codes to life in a new way. After becoming accustomed to seeing your codes in a particular order for so long – usually alphabetically or through one of the other filters your QDA tool allows – it can be hard to see the wood for the trees. Putting your codes on paper gives you much more freedom and allows you to see your codes differently. You can begin to move them around, group them into categories, and get a deeper sense of understanding about what your codes are actually saying. You might find, as we did, that even at the point of writing them out, you naturally start to identify some potential groupings or themes.
Our next step was to revisit our main research question to remind ourselves of what we were trying to address; only then were we in a good position to find out how our codes and groupings might help us to do that. Our research question seeks to explore how developers make decisions about security in the health IoT sector. We identified that a certain portion of our open codes could be grouped together and described as factors or considerations related to why security decisions are made (or not made). This group of codes was of interest to us, but it was not the most pertinent to exploring how security decisions are made. Another portion of our codes were more relevant to the how of security decision making, and these were grouped together in a different category. The remaining codes were interesting but not relevant to the research question; we grouped these and set them aside. Grouping our codes in this way allowed us to make some initial sense of the codes we had created and to see which group held the most promise for our analysis.
Throughout this process, we had our project open in NVivo so that we could go back to it for reference. This allowed us to open a code of interest and remind ourselves of the content from our interviews that we had coded to it. We were also able to look at which of our open codes had the most case files associated with them, and thus potentially held the most analytical power. We put an asterisk on those particular Post-its so that we could come back to look at them in finer detail.
We found this type of focused session very helpful for making sense of our open codes. We left with a much deeper understanding of our codes, of how they relate to one another, and thus of how to synthesise meaning from all our many hours of interview transcripts. In particular, we identified the core group of codes that held the most analytical power for addressing our research question. And that was what we wanted.
If you are using QDA software for coding qualitative data, why not give the Post-it approach a try?
- Anna Dyson
Post-it is a trademark of 3M.