Image via WikipediaYesterday I spent time in the bowels of my Journal Project...to be specific the bowels of the project are located in NVivo, my Qualitative Data Analysis Software package of preference. I was working on specific codes. "Coding on" as we say in the trade, that is, taking a larger bin of references and hacking it up into smaller identified chunks as well as reorganizing codes--merging, deleting, etc.
As I did this I found myself so satisfied with the visual process of how the tool was showing me the changes that were being made. Over and over I would create a new child node, drop materials into it and see the label of the sub-node pop up underneath its 'parent'. It was a visually satisfying process to see the shrinkage or collapse of materials, along with the relief of organization appearing. It put me very much in mind of Bruno Latour's discussion of "cascades" of paper, graphs, charts, etc. that emerge in modern times.
I couldn't help but think back to the coding processes of yore. Actually, I liked the old cut and paste, too. (At the time I didn't know there were alternatives.) But what I appreciated was different, visually. I remember the importance of the wide right margin (or sometimes you put the column of text right in the center to given you a wide left and right margin)...and using the line numbering function so that you would be able to create a more exact reference. As you can see I am describing a computer word processor based form of data analysis...which is, itself, represents a brief period at the end of the 20th century.
I would write my codes by hand in the margins, xerox the pages later (making more copies for those pages with overlapping codes), and then cut them up, making sure to put a page # on each fragment, and then file in a file drawer full of files with little fragments of text. I thought this was highly efficient, and I found it to be visual and satisfying. Just as now, I feel satisfaction watching a list of codes emerge in a nice little line.
I'm not arguing about which is best here. Rather what has caught my attention is how, in both cases, there is a deep reliance on visualization in the qualitative research analysis process. This is a kind of visualization that is not captured by pictures or in photos--it's not a drawing.
I think back to Anselem Strauss' book on analysis (the white book I have always called it--because of the color of the jacket)...and Miles and Huberman (the large gray book)...and I think about the ways they describe data representation. They are talking about text, but text that mixes the meaning of the text with visual formats that help us to move the blocks of text around in our heads: A table with codes organized; a schematic map of social relationships or connections between events. These qualitative researchers honed in on the ways that the work of analysis was reliant on visual constructions.
Both manual and computer based methods of qualitative research analysis are highly visual. It's not that manual was not and QDAS is. QDAS, however, does have new visual affordances. As I was working on my "dream" code, I found myself in the source bin exploring the three ways that references can be viewed: summary, text, and icons. I found the summary helped me to understand how much and where it was located. I worked in the full text of the code and the view with multiple icons to choose from and single text.
I do have to say that working in QDAS felt more efficient. If I'd never experienced it, I wouldn't know. But I do know, and I know it is easier to move my materials around and that the visualization of the text through terms and positioning is richer than what I had available to me before.
Now my question is--how is this form of visualization different than the kind of visual that I was talking about earlier in the blog--drawings, collage, and other non-text creations? It's something to think about!