Qualitative Research Resources for Journal Coding

I’ve recently started my dissertation project research. The dissertation has several moving parts including two national surveys, follow-up interviews, and classroom observations. Currently I’m waiting on my IRB approval to get started on collecting data from real people, so while I’m in a holding pattern I’ve started to work on my first chapter. I’m looking at how “community” has been used historically as a term in the journal College Composition and Communication in order to understand (1) what compositionists mean when we say “community”, (2) how this definition has shifted over the past ~70 years, and (3) what that means for “community” and composition today.

I started my research with a quick search in JSTOR and other databases to see how many times “community” shows up in the 64 volumes of CCC, and I got a manageable 1700 hits. Can you imagine if my term had been even more generic like process? Needless to say, even 1700 instances of a term can be a lot for one researcher to handle. Luckily, there are some great resources that currently exist to organize the research process, facilitate coding, and analyze and visualize data.

The Coding Manual for Qualitative Researchers by Johnny Saldaña (2012)

I cannot emphasize how important, useful, and accessible this book is for anyone who is interested in coding data. Becky Rickly introduced me to it a couple of years ago, but I had no idea what coding was (“Wait, it’s not about HTML?”) or that I would eventually use it in my dissertation. Like many compositionists, my background in research methods is rather limited. I took one graduate course that was useful but had to cram so many methods into a relatively short period of time. We went over coding but stopped with grounded theory methods. Saldaña’s text actually gets into the nitty-gritty of qualitative coding and offers many models for how to code. Developing categories of codes can be very intimidating especially when you don’t know the best way to approach your data (yet), so his breakdown of different ways to look at data makes coding more accessible to scholars. I haven’t gotten to the sections on analysis or visualization yet, but I’m looking forward to those explanations! The text also includes some screenshots of a variety of programs that can be used for qualitative research, and he also discusses hand coding for smaller data sets.

I’ve also tried out a few qualitative research programs and wanted to share those experiences.

HyperResearch from ResearchWare, Inc.

We used HyperResearch in my graduate research methods class, and it’s a good intro to qualitative research programs. It allows researchers to code data in text, audio, and video format. It’s very easy to highlight text and create a new code or apply a previous code to the text.

Screenshot of HyperRESEARCH

Example of HyperRESEARCH

It also allows some limited data visualizations such as word clouds and word frequency reports. There are some drawbacks. The main issue for me is that all text has to be a RTF file. I used the program to code discussion forums from the Rhetorical Composing MOOC, and the formatting in RTF was all messed up. I had a limited data set (only 10 albeit extensive discussion threads), and I wasn’t sure if there was a way to get HyperResearch to autocode a much larger data set that I could then check. This could simply be researcher error. Overall I would recommend HyperResearch for small data sets and new coders. It’s available as a free download, but the free version limits the number of codes you can create and the number of documents you can include in one study. There’s an educational rate for the program, and I think it was about $200.

Dedoose

Dedoose markets itself as a qualitative and mixed-methods program created by qualitative researchers. It’s an online app, which means that you don’t clutter your computer and limited hard drive space with large study files, and you can access the files from any computer without needing to install software. It also makes Dedoose open to collaborative work-flow.

Screenshot of Dedoose study home page

Dedoose study home page

Like HyperResearch, Dedoose can work with multiple data types including text, audio, and visuals. Even better, it can take text files in multiple formats including RTF, .doc and .docx, and HTML (unfortunately not PDF). I was really impressed by Dedoose’s options for collaboration. There is a training center option that allows for inter-reader reliability training and testing on dummy data before coding the real data set. It also offers more data visualization and analysis options. It’s also much cheaper than the other options I’ve used: it’s based on a monthly subscription model, and for a student rate it only costs $10 a month.

However, I ended up not using Dedoose. For me, the ability to be able to upload PDFs was very important (all my documents are PDFs, and to export in a new format was adding a significant amount of time per article). Dedoose was very buggy for me. I tried uploading text in multiple formats, and it would often freeze, which required me to refresh the page and have to log back into the app. I also did not find the program very intuitive for the coding process; it seemed better suited to importing spreadsheets with coding results. I think Dedoose has tremendous potential, and the collaborative work is very exciting, but at this point it’s not what I need (a PDF reader, easy coding workspace).

NVivo by QSR International

My final program is NVivo, which is probably one of the best known qualitative data programs out there. It’s another desktop program, and luckily it works for both Windows and Macs (unlike QDA Miner, which looks good but unfortunately I can’t run). So no collaborative work available in this program, but it offers a much better coding workspace than Dedoose (as in, I can actually figure out how to code in it!).

Screenshot of NVivo study home screen

NVivo study home screen

Once again, NVivo works with text, audio, and video data. The best feature is that it accepts text in PDF format, although it is buggy for the non-OCR PDFs that I’m currently working with (has issues highlighting). Coding isn’t as intuitive as it is in HyperResearch; you can highlight text, but codes are called “nodes,” which took me a while to figure out. You can create in vivo codes easily or code using recently used nodes, but to create a new node or a node you haven’t used for a while requires several clicks. NVivo allows for pretty good source  organization, and there’s an awesome query feature that allows for quick data analysis or selections to code (great for me because I’m coding one term). There are some autocode options that I haven’t used yet, but I have noticed that there is a feature that will create a graph based on the term “community” and show me all of the words that surround it. I also like that it allows you to create analytical memos and keep those organized with your materials. NVivo doesn’t make it as obvious that a particular section has been coded as HyperResearch does; there’s a small view that shows coding stripes, but it isn’t as glaringly obvious as HR.

Overall, I like NVivo the best simply for it’s ability to work with PDFs, and I’m sure the Query and Analyze functions will come more in handy as I work through this project. There is a free 30 day trial, and the Mac educational rate is only $90, which makes it the cheaper option.

Update: I ended up switching back to HyperResearch. NVivo for Mac had serious limitations (the Windows version was much more robust), and my file became corrupted so I was unable to do some of the limited functions available on Mac. It was easier to convert all my PDFs to TXT files for Hyperresearch. I also found that Hyperresearch’s customer suport was phenomenal; they responded to emails and tweets at all hours to help me troubleshoot. They also provided me beta access to some tools. There are still some frustrations with Hyperresearch (difficulty in organizing articles, needing to select individual articles for coding, etc.), but the HR folks assure me these issues are on the list of updates in the future.