Myprevious project with Dr Ibrar Bhatt CAQDAS in Science and Technology Studies, sought to explore how software affects, changes and mediates the way qualitative research is done.
The capabilities of computer assisted qualitative data analysis software (CAQDAS) have massively increased since the 1990s, usage is far more pervasive, and the costs and pressures for institutional licensing have grown. Concomitantly the approaches to researching social life have become more complex, digital and mobile. In this context the relative lack of systematic research on CAQDAS use since Fielding, 1998 is striking – there has been a lot of published work, however, the data informing these publications is largely composed of single-person or single-project case-studies.
CAQDAS is often strongly associated with particular approaches to engaging with qualitative data such as coding data and retrieving the codes, abstracting and reducing data to themes and relating data to fixed locations. Books, training courses and marketing materials for these packages actively position and promote them as being supportive but largely transparent in the qualitative analysis process. Such simplistic positioning is challenged by scholarship in science and technology studies (STS) and practices in mobilities research that explore the mediating effects of software on work practices. The methodological concerns of these interdisciplinary research areas also pose significant challenges to conventional CAQDAS approaches – for example the idea of “following the actors” is different from identifying theme, as are explorations and concerns with movement and mobility. We believe that at their intersection there is substantial potential for exploring, reformulating and better understanding what CAQDAS does, can do, and could potentially do. Through exploring the experiences, practices, uses, and the theorisation of the software by researchers working in STS and mobilities research we’re working to explore and intervene in that are.
How do social scientists studying science, technology and mobility in their research consider, engage with and account for the effects of using CAQDAS on their investigations and analysis?
Specific sub-questions then guide the methods and analysis:
What influences choice of software?
Which tools and options within CAQDAS are adopted, and how are they used?
How is the use of software accounted for in public research outputs?
How is the use of software accounted for in private research outputs?
How are breakdowns, limits and workarounds understood, drawn on, incorporated or theorized?
This project uses web-based screen-capture interviews of around an hour with 13 researchers working in STS or mobilities. We’ve tested several different webinar platforms (including testing Webex, Zoom, GoToMeeting and Adobe Connect) to engage with distributed researchers across the UK, Eutirope, Australasia and the Americas. These systems enable recording on-screen software use and an experiential, demonstrative interview approach to explore software use and issues. By using that method we’ve generated a rich view of the layering of research activity and interaction between the software, research work and writing practices. Of course, it’s also substantially reduced fieldwork travel costs and time. Additional data has been collected through participants sharing documents and notes they have made about software use in their written memos, notes, blog posts or similar reflective writing.
The reflexive, iterative, technology-mediated development of our data analysis procedures is also a significant part of the research project. How it is proceeding is being shaped by the practices, suggestions and approaches of the participants. Thematic analysis using code-and-retrieval is forming a one component of this, as are networked and hyperlinked explorations of the connections between elements of the interview accounts, documentary evidence and literature. How, precisely, this continues to develop is therefore a key part of the research process and project rather than just a means to a conclusion.
The transcripts have been synchronized with the screen recordings and initial analysis of that data together with additional documents from participants and relevant literature are being analyzed using several software assemblages. These include the software used by participants – primarily ATLAS.ti on PC and NVivo 11 on PC. The ATLAS.ti iPad app is also being used extensively for working with the literature for the project as well – enabling the development of a literature-informed coding system and working with literature for the project while mobile.
For the project development analysis is also being undertaken in the Mac versions of these packages. Further possibilities and approaches to analysis are also being explored through other software packages that form part of participants practices including tools from the Sciences.po medialab such as Actor-Network Text Analyser (ANTA), as well as programmes including Gephi and the AI-based document organization system DEVONthink.
A core principle of this research is that the data should be re-usable for secondary research and in the training of highly skilled researchers – with a shared dataset across several packages to support better comparison of software rather than a beauty contest between interfaces or disciplinary taste-test in response to the topic of discipline of example projects. By including data which discusses real world researchers experiences with software use selections from this should help build a useful and relevant dataset for those learning to use software.
The research is at an early stage but working comparatively between Mac and PC versions of ATLAS.ti and making live comparisons with the processes in NVivo is proving very enlightening. I am really excited about the new possibilities that will come with version 8 having seen an early beta at the ATLAS.ti user conference and building that into the project.