How do you handle your data? One big file? Hundreds of randomly-lableled files, in odd folders? Or a carefully curated, updated and catalogued easy-retrieval system? Sarah J White set out to find the answer from her fellow Twitter users….
A few weeks ago I started thinking about processes and tools in conversation analysis. This year I have embarked on my biggest CA project since my PhD, so I thought it was time for a refresh to ensure I am keeping up. There are many, many resources available on how to do CA (I even have a methods chapter coming out soon), but that actual processes used to document the analysis seem less well defined.
As conversation analysts on Twitter are quite active and have previously been super helpful, I decided to ask people who do CA a number of questions:
The responses were swift and demonstrated a range of practices. Here are just a couple of helpful replies, from Scott Barnes and Saul Albert:
The ROLSI editor, Charles Antaki, saw what was going on and asked me to put the replies together for a guest blog for ROLSI. To augment that information, I put together a survey based on these responses and was surprised and delighted to receive 61 responses! In addition to asking about processes and about data storage, I also asked about length of CA career, for additional comments, and for recommendations for favourite methods texts. A copy of the survey data can be found in a Dropbox file here.
Tools for CA
Once people had got in touch, I expanded the survey to ask these questions:
- What do you use when creating initial notes?
- What do you use when creating collections?
- What do you use when managing collections?
Each of these included a selection of paper-based and digital processes, with both standard software and specialist software covered (and a free text other option). Respondents could choose more than one option.
Across the three questions, respondents selected a combination of processes or tools. The combinations varied and the survey instrument didn’t allow for a description of how the combinations are used.
- For creating initial notes, the most common tools are: notes on transcripts (45/61), digital logbooks (39/61), and paper logbooks (21/61).
- For creating collections, the most common tools are: text documents (38/61), digital folders (37/61), and spreadsheets (22/61). A concern was raised in the Twitter discussion about using spreadsheets and the potential for over-categorisation. This issue of “binning” was recently discussed at the Language and Interaction Reading Group at Macquarie University, particularly in relation to a recent Enfield and Sidnell paper, so it all felt quite connected for me!
- For managing collections, the most common tools are: digital folders (37/61), text documents (30/61), and spreadsheets (23/61).
Check out the full survey data for more details, including tools identified by respondents such as shared documents for collaboration.
I was quite surprised that the most common tools used are fairly simple – software such as Word and Excel, digital folder systems, and paper notes. This is despite the availability of more specialist software. Are conversation analysts technophobes? Are we too cheap to buy software? Or are we focused on a simplified way of analysing and managing data, which is easily shareable between collaborators who use different systems, that doesn’t create unnecessary categories too early in the analytic process? I suspect the latter.
In the Twitter thread, a few people started commenting on back-ups and the importance of storage, so I included a couple of questions relating to this:
- Do you discard some of this information following publication? If so, what?
- How do you back up data? (you can select more than one answer)
Very few respondents discarded information (6/61) and those that did only discarded initial notes or other information captured elsewhere, or if they were required to by the institution.
The most common back up methods was external hard drives (52/61), which leads me to remind people that these also have a limited lifespan, so make sure you have multiples and that you periodically replace them.
There were a range of texts suggested, though ten Have’s Doing Conversation Analysis, and Stivers and Sidnell’s The Handbook of Conversation Analysis (especially Sidnell’s chapter on Basic Conversation Analytic Methods), received the most mentions. A number of other texts and course notes, particularly those from Manny Schegloff (to see Schegloff’s lectures, join ISCA and get access here), were also suggested a number of times.
Contributors’ final thoughts
Almost half of respondents also shared some other thoughts, mostly about the analytic process rather than the tools. These focused on the importance of the collaborative nature of CA, with the need for a community of practice and the ability to engage in data sessions.
There were also several comments about ensuring that collection building and technology doesn’t get in the way of good quality analysis, with which I wholeheartedly agree. One particular correspondent seemed to sum up my own sentiments in engaging in this brief exploration:
The analytic process, whichever method you choose, has got to be organised. Whether using an online database, physical folders or spreadsheets and word documents, it has to be organised. There’s no pressure for everyone to be using the same standard method, and there may be good reason for different projects and different collections to use slightly different methods. […]. It’s worth putting in the work to set up a good way of managing your collection and analysis – it definitely pays off!
My final thoughts
Finding a method that works for you is the most important thing. I’ve decided to go with a paper log book for my current project (for now). I have so much in digital format from other smaller projects as well as teaching that I can feel almost overwhelmed, so keeping all my initial notes in a paper book is somewhat comforting and I hope will help with my short term goals. I will, however digitise analytic notes as I go along to keep a more permanent record. I am also using Excel and NVivo to manage some of the data as I also have some surveys related to the project. And for now I am storing videos, transcripts, and clips in folders managed by interaction with notes on the spreadsheet to help me find clips so I can ensure all the data from one interaction is housed together. This is important for my data library as some participants agreed to allow their interactions to be used across multiple projects.
This Twitter discussion and survey has taught me how important it is to critique my own processes, to continue to upskill, and to embrace ideas from others in the way I approach analytic processes and documentation. It can be easy to get stuck in your ways. And for those wanting to ask questions, the Twitter #EMCA community is supportive, so don’t be scared to ask what you might consider a simple question.