Interview transcripts were anonymised through removal of real names, use of pseudonyms and removal of any contact information, employer details or other significant identifying information.
Anonymised data was loaded into DeDoose[1], a web-based qualitative data-encoding tool. Broad themes were identified in a first round of analysis, with notes taken during interviews used as a source of initial codings. Once all transcripts had been coded a second parsing of the data took place to ensure relative consistency across codings. A further reflexive parse of the transcripts was made with particular consideration of the concept of “transposing data” (Dey 1999), interrogating the “what ifs” of the questioning process. For instance in this research “what if this interview was taking place in a face to face context” was a useful transposing question. Finally extracts from the transcripts were analysed with the Voyant text analysis tool to further interrogate and visualise the most popular or significant terms in texts - particularly clusters of quotes around a particular trend (see Fig. 4).
Figure 4: Interactive embedded Voyant Tools analysis of comments from questionnaires tagged as relating to “dysfunctional learning”. Click on a word in any section of the screen in order to explore the analysis
[1] DeDoose: http://www.dedoose.com/
Previous Section: Questionnaire analysis | Next Section: Presentation of findings