Don't blame the tools: researchers de-contextualise data, not CAQDAS

By Christina Silver on Jan 07, 2017 at 06:26 PM in CAQDAS commentary

by Christina Silver, 7th January 2017.

Donít blame the tools: researchers de-contextualise data, not CAQDAS


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. I have already argued that CAQDAS packages actually bring us closer to our data, and given an illustration of how this can happen, so here I consider the de-contextualizing issue.

I’m writing this because I saw a presentation slide on Twitter that listed a disadvantage of using CAQDAS packages as“detach[ing] data from contexts in which they were collected”.

Donít blame the tools: researchers de-contextualise data, not CAQDAS

As discussed previously, I wasn’t at the conference and I have no idea what was said while this slide was being shown. But it got me thinking about why it is that these sorts of criticisms are still made.

 

Why is detachment the fault of technology? 

What does “detaching data from contexts in which they were collected” mean? If it means that data are no longer ‘located’ in their natural context, once they are being used within a software, how is that different from working with data using any of the other tools available for analysis, for example non-dedicated software such as word processing, spreadsheet or note-taking software, or paper and pen? Unless we actually do our analysis physically in the field, together with our research participants, aren’t data always going to be detached from the contexts in which they were collected?

 

Detachment is about power relations in qualitative research

Detachment is about the power relations in qualitative research, not the use of CAQDAS packages, or any other form of technology for that matter. Discussing interview based research Karnieli-Miller et al. say that “with termination of the data collection stage […], formal control and power over the data returns to the researcher. From now on, the story shared with the interviewer is “separated” from the participant, and the researcher becomes the “storyteller” who recasts the story into a “new” historical, political, and cultural context. During this stage, the researcher’s control over the data seems to be absolute” (2009:282). Even in ethnographic studies when researchers’ spend lengthy periods in the field observing real-life situations and events, speaking with participants whilst they go about their lives, and participating in activities, it is common for the analysis to remain the preserve of the researcher. Some forms of participatory research are different, directly involving participants in data analysis, but this is still relatively rare.

 

Computers are portable, CAQDAS is flexible

But not including participants in data analysis has far more to do with epistemology than technology.Those that believe knowledge is constructed rather than an objective truth, for example researchers strongly positioned in interpretivist and constructivist paradigms, may undertake fully participatory research, involving participants directly in analysis. This is possible because of technology. Electronic tools are portable. With the advent of laptops, tablets and smart phones researchers no longer have to be tied to desks in their offices to undertake analysis, or indeed any other type of scholarly work. Just to prove the point I don’t mind telling you that I’m actually in the bath writing this post on my phone! Paper and pen are portable too, you may say, but practically its difficult to box up and cart around a manual analysis. But we can easily take our laptops into the field, show participants what we’re doing and invite them to get involved in analysis. The advent of CAQDAS App versions are currently primarily focused on data collection (e.g. MAXApp and ATLAS.ti mobile for iPad and Android) but there are web-based CAQDAS packages (e.g. Dedoose and WebQDA). Quirkos, one of the more recent CAQDAS packages to be released was designed with participatory research in mind, providing a simple interface and a visual canvas displaying the results of analysis in ways accessible to non-professional or academic researchers. There are many choices out there, together providing a flexible palette of tools to choose from. 

Don’t blame the tools

The suggestion that CAQDAS packages detach data from the contexts in which they were collected reminds me of the well-known proverb “a bad workman blames his tools”, which according to the Oxford Dictionary of proverbs, dates from the late 13th century. We’re not yet at the point in the field of qualitative analysis where the software is doing any of the analysis for us. Technologies and fields that are and whether we should be heading that way is another discussion. But not only is it inaccurate to say that CAQDAS de-contextualizes data, it is important to recognize that technology offers opportunities for involving participants in all phrases of research. Just like easy to use digital cameras and video recorders enable participants to generate their own data, dedicated CAQDAS packages enable analysis to be undertaken wherever, whenever and by whoever. Whether data is detached from the context in which it was collected is to do with research design, and that is the decision of researchers not software developers. There may still be some cost and usability barriers to fully involving participants in analysis, but these are also reducing. The different CAQDAS developers are now also discussing the possibility of a common exchange format, which if successful, will make things easier as projects started in one program will be easily importable into another.

 

Find out for yourself

The best thing to do if you’re unsure if CAQDAS packages are suitable for your needs is to check them out for yourself. For reviews of CAQDAS packages see the CAQDAS Networking Project. For full feature lists see the individual developer websites. For training in the use of any CAQDAS package, project specific consultancy or analysis services, contract us.

 

We have three books being published on the Five-Level QDA method for harnessing CAQDAS packages powerfully in 2017. One each for ATLAS.ti, MAXQDA and NVivo. Sign up to be notified when our books are published.

References

Karnieli-Miller, O., Strier, R., Pessach, L. (2009) Power relations in qualitative research. Qualitative Health Research, Vol. 19. Nr.2, pp279-289

Sep 27, 2018 Arrow1 Down Reply
Eric

I have also heard about this issue previously. Nonetheless, instead of considering de-contextualizing as being analyzed solely by the researcher (i.e., without participants), I have heard that de-contextualizing refers to analyzing data that is fractured. That is, the data we end up analyzing (usually codes) are abstractions taken from different parts of the data based on their similarity and which lose their contextual particularities (the differences) because of this abstraction. However, I believe that this issue mainly emerges because of the facility and strength of QDAS to code (categorize) and work with this abstractions; and not necessarily because the software cannot be used to link (connect) concrete data to generate stories or cases. In other words, I believe that most researchers take softwares as dictating their methodology (tactics guiding strategies as you would say); and since the main functions of this softwares facilitate categorizing strategies (i.e., coding and working with them) they assume that QDAS should be used like that.
This critique further contends that QDAS are good for methodologies that end up with abstractions (e.g., Grounded Theory); and not necessarily good for methodologies that requiere more concrete outcomes (e.g., Case Studies or Narratives). However, I believe that the methodology should guide the use of software (as you also contend) and while it is difficult to use QDAS software to generate outcomes that display, for example, how the data is organized chronologically (as in Life Story Analysis or Case Studies) I believe that it can be done. The question is whether it is the most efficient way to do it with this software, or going back to pen and paper (or other method) is better for organizing the events represented in the data. The latter of course would depend on the researcher, his methodology, and the project. In my personal experience, for example, I have found the Network tool of Atlas.ti as useful for this connecting task. This tool enables you to link quotes to other quotes (which may be fragmented in the text), and then manually organize them chronologically (and also other logical relations). While a daunting method, it ends up serving its purpose, and in my case, it was definitely more efficient that doing it with pen and paper. NVivo, for its parts, enables you to link references, I am still to try how efficient this might be.







Testimonials

So often I have attended training courses and wondered what I was doing there, feeling my life slip away into oblivion while I could be doing other more useful and interesting activities such as watching iron rust or paint flake. However, Christina's was superb. It was exactly what I had hoped for. Pitched at precisely the right level and presented in a very clear, engaging way. Being able to spend time having some real content and context given by someone who really understands the software and researcher's needs, along with a sensible amount of time to try things out and get help was great. I'm much more confident about getting to grips with my data now.
Paul Rause
Interdisciplinary researcher, Southampton University