by Christina Silver, 1st October 2016
In an earlier post on CAQDAS critics and advocates I promised to provide evidence for my position that CAQDAS packages are not distancing, de-contextualising, and homogenising, as is sometimes claimed. So I’m starting a series of posts. First I'm taking the suggestion that the use of CAQDAS distances us from our qualitative data and illustrate why I believe the converse to be true. Here I outline my position, and I'll illustrate my argument with examples in subsequent posts.
THE IMPORTANCE OF BEING CLOSE TO OUR DATA
Being close to data is essential when undertaking qualitative analysis, whatever its form, the sources it comes from, or the strategies we use to analyse it. I have discussed this in relation to transcription as an analytic act, elsewhere, arguing that just like when children learn to read they must repeatedly revisit letter-sounds, words, sentences and stories to consolidate their learning, we as qualitative analysts must read and view our data multiple times in order to become sufficiently familiar with it to develop a nuanced, accurate and authoritative interpretation. Being “close to data” is therefore of utmost importance.
THE SUGGESTION THAT TECHNOLOGY DISTANCES
I'm not best-placed to outline the position that the use of CAQDAS packages distances us from our data, seeing as I don't believe they do. However, this suggestion has something to do with the idea that we need tactile and direct contact with data in order to reach the level of familiarity required for analysis. It is argued that this is achieved by working with hard-copy print-outs, marking interesting passages using highlighter pens and writing notes in the margin. We can flick back-and-forth through the pages, and visually annotate the associations we see.
This is indeed what I did for my first two QDAs, my A-Level sociology project and undergraduate dissertation. In those days there were sadly only four highlighter pen colors, but I had other tactics at my disposal: scissors, post-it notes, blue-tac, and thread.
I photocopied transcript printouts, highlighted interesting segments, cut them out and constructed a matrix of themes by respondents on my bedroom wall. Most of the time I remembered to scribble on the back of each cut-out segment approximately where in the transcript it had come from and who had said it and I used thread to make links between related cut-outs. This is what I ended up with...
It was fun and it sparked my interest in qualitative analysis. And I was certainly close to my data. It surrounded me in fact, I couldn't get away from it in my tiny student bedroom. But I faced a few challenges in managing the data, accessing the source context reliably and handling the fact that many of my cut-out segments related to more than one theme, and therefore belonged in more than one matrix cell.
The argument is that the use of CAQDAS puts some kind of barrier between us as analysts and our data, such that we cannot interact with it in the same way as when working manually, we cannot be close enough to it to reach the necessary level of familiarity.
What about non-textual data?
Many researchers work entirely or predominately with textual forms of data, but what about when audio, video or graphic data are our sources? We have to use technology to capture and view these data. We could print graphics out and analyse them by hand, but that's not possible with audio and video. I'm not sure where the close-to-data-critics stand on this point...
CAQDAS CAN BRING US WAY, WAY CLOSER
I believe that CAQDAS tools, when harnessed appropriately, can bring us way, way closer to data than is possible when working with hard-copy or using non-dedicated software (such as word processing and spreadsheet applications).
IT’S ALL ABOUT ACCESS
For me, it all has to do with the reliability and speed of access. As soon as we use a dedicated CAQDAS package we are able to create links within and between all of our materials and all of the ideas we have about them. Once we start doing that - and there are several ways we can do it - we can access anything and everything related to any bit of material and any part of our analysis, whenever we want, however we want, for whatever purpose we want. Reliably, quickly and repeatedly. And that means we can review it, revise it, build on what we have done before. And so it continues. And we can do it with any form of data our CAQDAS package supports.
Flicking back-and-forth takes on a whole new meaning.
CLOSER TO MORE THAN JUST OUR "DATA"
And its not just our “data", but all the materials that are relevant to our study. Grant proposals, previous literature, supplementary information, internet content, notes we have made about our participants, about our data, about our processes, about the ideas we have had about about anything during the process that we have recorded within our CAQDAS project or linked to from it. We can access earlier work we have done, review it, alter it, write about it. Build on it.
OUR IDEAS ARE "DATA" TOO
The suggestion that CAQDAS packages distance us from our data is mystifying to me in part because it ignores the status of our interpretations, somehow prioritizing "data" above and beyond our description, analysis and interpretation of it. Sure, the datasets we construct in order to answer our research questions are important, fundamental even, but what we produce as the result of the work we do is always some form of account of that data, not the data in its raw format. We do of course illustrate our accounts with extacts of data, but the account itself is some sort of structured synthesis, summary, representation, pathway through, or reduction of it. Depending on our objectives and guiding methodology the account might be presented as a theory or model, it might be presented as a narrative, it might be presented as visually or dynamically.
Using dedicated CAQDAS packages enables us to keep all the insights we have about our data, throughout the process, stored together with our data. Moreover, we can integrate our insights with our data, linking the reflections we make with the segments of data - in whatever form, from whatever source - that prompted them. If you asked me what the main benefit is of using a dedicated CAQDAS package this would be it. Being able to treat our insights about data as data, and access them independently of, as well as integrated with, all our materials is a level of closeness that simply isn't possible when the tactics we use for analysis are highlighter pens, scissors, blu-tac, and thread.
In my next post I'll take a specific analytic task and illustrate how CAQDAS tools can bring us closer...