Coding: An Information Trail without Walking the Prison Trail

Coding: An Information Trail without Walking the Prison Trail

By Sarah Doxat-Pratt

I joined the COMPEN team in February this year, tasked with coding interview data from the Entry-Exit-Post-Release research. The EEPR study involved following the journeys of prisoners through the criminal justice system, interviewing them soon after entry to prison, shortly before release, and in the community a few months after their release. The study aimed to deepen understanding of the experience of prison, but also to focus on the significance of transition moments in and out of prison – and so interviews covered details of the whole sentence, from before the arrest to life post-release. Coding is not the most glamourous aspect of a research project, either to do or to reflect on, but I hope here to provide some reflections from my experiences which will illuminate the process. 

Julie Laursen has written a piece for this blog on the benefits of ‘walking the prison trail’ – that is, of having experiential and ‘visceral’ knowledge of the prisons that interviewees have been in and the journeys they may have taken to, between, and from prisons. The in-depth knowledge, being able to ‘see, smell and hear the prison’ helps ethnographers be ‘better equipped to understand the meaning behind’ what interviewees say. I joined the COMPEN team after completing my own doctoral research, which included significant fieldwork in one English prison, but I did not have the extensive knowledge of the prison trails in England & Wales and Norway that my colleagues did – indeed, when I started I had never even been to Norway (though, fortunately, that has now been rectified). My challenge, in starting as I did as the fieldwork phase drew to a close, was to understand and make meaning of what interviewees said without having walked the prison trail.

The extent to which coding – that is, dividing data up into thematic segments – constitutes analysis is the subject of some debate (see, for example, St Pierre & Jackson, 2014); simply put my view is that coding is primarily an aid to further analysis, but that analysis happens throughout the research process, coding included. Much of the literature about qualitative data coding, however, assumes that the coder has been involved in the fieldwork as well – coding is an ‘opportunity for further reflection’ (Noaks & Wincup, 2011, p130) and a way to ‘refine their reflections’ (Basit, 2003, p143). Those giving guidance to team-based research projects do acknowledge the potential for outsourcing or hiring in extra help for coding, much like transcription, but the concern tends to be practical – how to avoid too much stylistic or interpretative variation (MacQueen et al., 2008). Few people have much to say about the challenge of actually doing the coding without having been there.

There are inevitably practical issues of not knowing any of the research sites or relevant personnel. What does it mean for a Norwegian prisoner to ‘take Luft’? (Ans: go outside for air, but often the word is not translated in transcripts.) Who is the officer they are all describing that everyone seems to love – or are they different officers? When they say, ‘I was sent up here…’, does ‘up here’ mean a different prison somewhere geographically north, or up a hill, or just upstairs in the same prison? Sometimes the answer emerges through the transcript, but not always; asking colleagues of course provides answers, but it is difficult to know when it is worth asking the question in the first place.

But more than not knowing the research sites, I have found it significant that I do not know any of the people in the study. Perhaps much like transcribers, I found myself in the odd position of knowing intimate details of the lives of people I had never met. Interviewees were, of course, made aware at the start that the interview would be handled by others in the research team as well as the interviewer, but it seems this did not prove to be a barrier to candid conversation. It is common to hear of the incredible openness that exists between researcher and participant in prison research projects (or indeed, any research project, as Jason Warr has discussed for this blog). The interview can provide an opportunity for cathartic release, and the interviewer can be a combination of confidante and counsellor, being made privy to all kinds of personal information in the context of a two-way conversation. But in my experience of reading through these conversations, I am struck by the lack of familiarity I have with the interviewees. I could tell you the life narratives of hundreds of people – their childhood trauma, their criminal activity, their hopes for the future, details of their everyday lives in prison. I know what’s been done to them, by whom; I know what they have done to others and how they feel about it. But still, they are nameless and faceless to me, known only by a research code and pseudonym. I sometimes end up caring about them deeply, but I would not know them if I walked past them on the street. I have a seemingly endless supply of the personal narratives of people with whom I have never interacted.

And it does, indeed, seem endless: the EEPR study has over 450 interviews with over 250 individuals. Few people would call the task of coding itself exciting or interesting, but the data gathered and the stories told were exciting and interesting to start with. I made notes diligently, writing down anything I thought particularly noteworthy. After a while, the task became normal and a bit flat. Like some researchers in the field, I found that the excitement wore off and the stories were no longer so affecting. Now, approaching 200 coded interviews, I have to resist the temptation to be blasé about the content – “abuse, injustice, hopeful but maybe a bit unrealistic, same old, same old…”. Research participants can seem nameless, faceless and, 200 interviews in, lacking in original insight. While ethnographic researchers who code and analyse their own data sometimes need emotional or critical distance from the field, I have to work to ensure I do not have too much critical distance, that the task of coding does not become a mere mathematical exercise performed on ‘brute data’ (St. Pierre & Jackson, 2014, p715). I am still engaging with people, if only distantly, making meaning from their words and developing a picture of their reality.

Despite the challenges, though, I have been surprised by the opportunities that have come from this approach to coding, both for myself as a researcher and for the research project. I have learnt much about the research process simply by reading hundreds of interviews: the different styles used by my colleagues have provided me with an insight into interviewing that I imagine few early-career researchers get. Just as few people have such access to the narratives of prisoners without having been inside, it is also likely that few people have such an opportunity to engage with someone else’s research process to this degree.

Opportunities also arise for the project itself, even beyond the practical value of getting lots of coding done quickly. The critical and emotional distance I have from the interviewees brings something different; without wanting to imply an objectivity which I obviously do not have, it is possible that I have an easier job of avoiding confirmation bias in coding and analysis. Engaging with so much primary data over a relatively short space of time has its own immersive quality – there is the feeling of being steeped in the words of the participants. Not that this is the equivalent of ethnographic immersion in a research site, but still it has added a different perspective to the conversation. Discussions about the ‘breadth’ of imprisonment – that is, the reach of the prison sentence beyond the time in prison (see Crewe, 2015) – are perhaps a good example. The three phases of interviews in the EEPR study were carried out with months or sometimes years in between; I read and code them in a day or two. Reading the whole narrative in one go, and coding lots of these narratives in quick succession, make it easier to perceive some of the interwoven accounts of distress and resilience, or of change and continuity in relationships before, in and after prison (and, often, back in again), which shape our understanding of the broader impact of imprisonment.

For more specific findings, coding the data without having some of the expectations that come from doing the primary research can illuminate things, particularly when we realise that what has been seen by researchers and what has been said by participants do not completely align. For instance, strip searching and questioning on entry into prison: those who had observed this process thought one particularly degrading aspect of this was that some forms of invasive searching and personal questioning were done in public, in front of other prisoners and staff; but from what I had read, interviewees rarely, if ever, mention that this was the case, let alone that it was especially shameful or humiliating. This realisation enabled discussion about the nature of entry and what constitutes degrading treatment in prison that might otherwise have been overlooked.

Coding has its challenges, and non-stop coding for a long time especially so. Various of these challenges will be recognised by researchers in other settings and roles as well, whether on individual or team-based projects, but they are still worth thinking through. My experience of this approach to coding is that it can be far more than just the slightly tedious but necessary precursor to or part of analysis. While there is no replacement for ‘being there’ in research, having a lengthy information trail without having ‘walked the prison trail’ does have uses of its own.  

References:

Basit, T. (2003) Manual or electronic? The role of coding in qualitative data analysis. Education Research Vol 45(2): 143-154.

Crewe, B. (2015) Inside the Belly of the Penal Beast: Understanding the Experience of Imprisonment International Journal for Crime, Justice and Social Democracy Vol 4(1): 50-65.

MacQueen K. et al. (2008) Team-based Codebook Development: Structure, Process, and Agreement in Guest. G & MacQueen K. eds. Handbook for Team-based Qualitative Research Plymouth: Rowman Altamira.

Noaks, L. & Wincup, E. (2004) Criminological Research: Understanding Qualitative Methods London: Sage.

St. Pierre, E. & Jackson, A. (2014) Qualitative Data Analysis After Coding. Qualitative Inquiry Vol 20(6): 715-719.