Chapter 5. The research
5.6 Data collection and analysis
5.6.7 Transcription, coding and analysis
I have yet to see a complicated problem not become more complicated (O’Leary 2005, p.11).
Each interview was transcribed by a professional transcription service and reviewed immediately. Interview analysis and documentary analysis were undertaken at the same time and each informed the other in what Silverman (2005) describes as an ongoing reflexive process (p.152). As detailed above, this research was always intended to be an iterative project. In keeping with the advice of Noaks and Wincup (2004), I kept an open mind and flexible approach. In the ‘analytic focus’ there was no pre-determined agenda: I wanted the data to present me with new information and to challenge the way I saw issues (p. 131). I did not leave the analysis of the data till the end of the project. I read and re-read the interview transcripts as they came in coding, revising, re-coding, forming ideas and making notes as I went. There were a number of unexpected themes that emerged from the interviews, which meant that I had to examine new ideas and follow up new areas of research. One of the most unexpected aspects of the interview material was the emphasis that interviewees placed on individual personalities, emotions and their influences on the policy process. This is reflected in the attention given to subjectivity and affective factors throughout the thesis and especially in chapter ten. It also led me to review
5. The research my theoretical approach to ensure that the role of agency was adequately incorporated into my analysis. As outlined above, when interviewees mentioned key reports, cases, important media events or debates in Hansard, they were followed up and then included in the data. Again, the iterative, reflexive nature of the qualitative research process proved to be flexible enough to accommodate these kinds of shifts. As the interviews were undertaken, I began to see clearly how important the interviewee’s background, character, professional position played in their perspective of the policy process. They were not objective data sets, or extracts separated out from the person but discursive subjects, whose views were part of the whole person (Hollway & Jefferson 2000, p.68). Today, when I read their words in the manuscript I can see the person, or hear their voice and for me that is important, I feel that my own immersion in the data and my manual coding brings a level of authenticity and integrity to the findings and brings the data alive.
Initially interviews were coded thematically, from themes identified in the research questions and then from any other consistent themes that emerged from the interviews as they progressed and major gaps were noted. During a second sweep major themes were broken down into sub groups (axial coding), including concepts, ideas, issues, silences. In a third pass through the data, I selected quotations and case examples that seemed to exemplify the key themes of the thesis and which appeared to closely represent the particular viewpoints of the interviewees. As transcripts came in and further issues were raised, the initial transcripts were revisited. If something new or unexpected appeared in one transcript, the others were reviewed again in the light of the new information, to see if there was any resonance with previous material and to assess its significance. During writing, the transcripts were re-checked to ensure that nothing was missed. By the end of the study period each transcript had been coded and recoded at least four times, each transcript separately dated and each manuscript was covered in notes, stickies and different coloured marker pens. The analytical chapters of this thesis reflect the coding categories and sub categories that were established during this process. In the way that O’Leary advises, the documents were also treated like the interviews themselves, themes that resonated with the interview schedules were highlighted
and included in the review process (2004, p.180). In addition anything that stood out as different or unexpected in the documents was noted.
5.7 Writing
Writing each chapter, although distinct from the data analysis itself, was also part of this analytical, reflexive process. As Coffey and Atkinson (1996) state:
Writing and representing is a vital way of thinking about one’s data. Writing makes us think about data in new and different ways. Thinking about how to represent our data also forces us to think about meanings and understandings, voices and experiences present in the data. As such, writing actually deepens our level of academic endeavour (p.109 cited in Noaks and Wincup at p.134). As the data came in and transcripts were coded they formed the basis of draft written materials, draft chapter sections, or conference papers and presentations. Writing helped to tighten ideas and inform the analysis. The final structure of the thesis was formed through this process.
As I was writing I also considered how I was going to cite the interview material since, from a review of the relevant literature, there appears to be no clearly defined rules about using quotations (see Corden & Sainsbury 2012). Throughout the thesis, quotations are used to illustrate key points. Participants often expressed issues far more succinctly that I ever could and I considered that examples from the field provide the reader with great insights on the reasons why and how policy decisions take place. I have tidied up the verbatim quotations - taking away the hesitations and inappropriate language (there was some swearing!) to make them more readable and to make sure that they reflect the overall quality of the information that I was given.
The analytical process described in this section recognises the kind of integrated interpretative process described by Neuman - it combines a first order interpretation of the interviewee material, a second order interpretation of the researcher and a third order interpretation which integrates writing and reflection with theory (Neuman 2006, p.160).
5. The research