I was running a workshop today, and talking about the context in which I use a text search query in NVivo, so thought I would share it here in case others find it useful. The caveat here, as usual is that this is not for ALL research, just the majority that I see when I do consultancies, and even for my own analysis projects. All research is different, so this may not apply to you. However, I would say it applies to lots of research projects. Also, this is for analysis, NOT for a literature review. I have very different advice for a literature review, and I think the two are very different processes.
Don’t start with a text search query
I always show how to do a text search query towards the end of my workshops, and this is for an important reason. I think a text search query should come towards the end of the analysis for most projects. Why? Because it truncates the data, and you miss the story, you miss what people are really trying to say. Qualitative analysis is not about picking out some keywords and seeing what people have said about them. It is about really understanding people, what they are telling you, what is important to them. If you start with a text search query, you miss this. In the rare occasions I’ve started with a text search query, without first really getting a depth of understanding of the data, I’ve felt I’ve just wasted my time. It has no context, no story, and I just don’t “get it”. Ultimately, nothing can save you from having to read the data, and really get to understand it.
Even having recently working with very structured positivistic data, I thought I could get away with starting with a text search query, but it didn’t help. I found narratives within the short responses people gave. There was a message people had for me in this data, and it was getting lost in mere text searches.
Use it as a checking tool towards the end of analysis
There are times in my analysis process, where I worry that I haven’t captured everything said about one of my codes/topics/themes etc. It is usually because I start to understand things when I’m a couple of interviews into the analysis, maybe more. So I often have this niggling doubt, have I really understood everything about this? Is there something I’m missing, especially early on? It is here, where I use the text search query. I use it as a way to check that I have captured everything about a topic/theme/idea (or at least done my best to capture it), and to understand it. It gives me the confidence to know that things haven’t been missed.
Find the silences in the data
A text search query is excellent for helping find silences in the data, and something you can’t really do manually. What do I mean by “finding the silence”? You may have created a code/node that you thought was important for your research, only to find very little coded in it. It may because it wasn’t important to your participants, but the silence around it may be important for your research (it may not). But, by doing a text search, you can find if the words were used, if the questions were asked, and what as said or avoided.
I did this as part of a team in one of the very first research projects I worked on as a research assistant. We had a main theme, we asked about it, but had little to nothing coded. We decided to do a text search query. We found that we had asked about in all our interviews, but that everyone had avoided answering the question, saying they didn’t want to think about it. We probed further in our analysis on this topic, and found some really important insights. The findings from this analysis really had an impact on me, and is also one of the most cited papers from that project.
These are just some thoughts and ideas to get you started. I think the main thing to take away, and I will repeat this often, is to be clear about the purpose of your analysis. If you keep this in mind, when you start or continue with your analysis, keep asking yourself “How will this help me answer the research questions?” or “How will this help me understand what people are trying to say?”
Anuja Cabraal (Phd) 