Data Use, Technology, and Educational Leadership
Almost 2 years ago, I compared some the qualitative coding software out there. That included Atlas.ti and Dedoose. Recently, I received an incredibly insightful and helpful email from Julia Scherba de Valenzuela at the University of New Mexico about her recommendations regarding Dedoose. She responded to some of my critiques and suggested some updates.
Pro: Collaboration. Our big agreement is that Dedoose is really unmatched when it comes to sharing and collaborating around data sets. It’s for this reason that I suggested that BC’s latest batch of doc students in ed leadership considering using Dedoose — they need to do a lot of collaborating and sharing of data as a part of their dissertations in practice. Plus, it’s pretty intuitive, without so many odd terms or “special” features no one really uses.
After experiencing some outrageously costly troubles involving NVivo (both in terms of dollars and lost time), Julia’s research team made the move to Dedoose. In her words:
In contrast, I can add folks to my projects in dedoose quickly and easily and we can both code on the same project. As far as I know, most of the other programs don’t allow that kind of collaboration, or at least not so easily or cheaply. And perhaps more importantly, I can have my doc students add me to their projects in dedoose. Then, we can chat about their data even if we aren’t meeting face-to-face. We aren’t tied to the one computer where they have their project loaded.
An aside: although I’m shifted my own work (which tends to be less geographically distributed) back to Atlas.ti, I’m at this very moment “locked out” of my own Atlas files because I think an assistant has the files open.
Pro: The Cooccurence Matrix. I hesitate to share this one. I’m sometimes interested in when codes co-occur. For example, if studying Twitter, I might be curious about the extent to which a certain topic is an original tweet or retweet. I’m a fan of the Dedoose’s matrix for this (see image above). It’s easy to read and seems to make sense. My caveat is that my doc students saw this and tried to make inferences about their data that maybe shouldn’t have been made.
Cons: Who owns the data? Why am I still back to using Atlas? I like “owning” my data on my university server/laptops. This is more about $$$ and practicality. I go through spells where I get caught up with other teaching, research, department, or life obligations that prevent me from digging too much into the data. I was on the month-to-month plan with Dedoose, which I had originally thought would be cheaper. But with this kind of schedule, I ended up paying as much, if not more than a traditional package. I wanted the ability to come back again and again, perhaps over several years, without having to pull out my credit card (and run through the reimbursement hassle) and without wondering what will happen to the cloud in years to come.
I was also a little irked that although Dedoose would accept data from other packages, it couldn’t export into those formats. For example, when I made the switch back to Atlas, I finally just printed out a bunch of Dedoose output, then had an undergrad assistant recode them it all. On that note, the print outs of Dedoose data seemed a little clunky and hard to read to me. I was not one of those people that lost data when Dedoose did some updates a few weeks ago, but I am one of those people who has corrupted/lost data probably due to his own mistakes in handling Atlas’s VERY byzantine file structure.
Other clunkiness. Julia also responded to my (now dated) comments about clunkiness when coding with Dedoose. Her take is below — I’m not sure what the browser cache thing is all about.
You mentioned two problems, both of which I believe have easy solutions. First, you wrote that “before tagging selected text with a code, you have to reach across the screen to the “excerpt” button. Then you have to drag all the way across the screen to assign the codes.” While that is one way to create an except (and a number of my colleagues insist on doing it this way). I find it much easier to select the text I want, then right click to create the excerpt. You do then have to move your cursor from the excerpt to the code, but I think that’s pretty typical. Secondly, you noted that “a ‘child code’ doesn’t automatically mark the ‘parent.'” I don’t know how it was in 2012, as they update the software fairly often, but you can currently select that excerpts in child codes are automatically coded in the parent or not. I appreciate having the option of selecting which way I want it to be coded.
I do agree with you that it is clunky to have to delete and then re-create and recode excerpts if you want to make a small change in your starting or stopping place for an excerpt. This is one of my biggest frustrations with dedoose, aside from having to clear out my browser cache frequently.
Bottom line? Dedoose might be worth a try, especially if you are working as part of a geographically distributed research team. That definitely fits the description of my EdD program. Our students work hard and constantly throughout their dissertation time.