Understanding mentoring relationships with disaffected young people (Helen Colley)
My research project (Colley, 2001a) investigated mentoring relationships between ‘disaffected’ 16 and 17 year olds on a pre-vocational training scheme, and volunteer mentors who were all university undergraduates. Most of the volunteers were student teachers or following a degree in applied social sciences. I was particularly interested in the power dynamics of those relationships. Most research on such mentoring consists either of anecdotal accounts that have been criticised for their favourable accounts despite a lack of substantial evidence; or of quantitative psychological studies which use questionnaire surveys measuring before-and-after outcomes, focus on individuals to the exclusion of context, and fail to explore the process of mentoring. Issues of power have almost exclusively been considered in terms of the mentors’ superior power over the mentee. Moreover, few studies of youth mentoring actually gave any voice to the views of the young people themselves.
I wanted to conduct a qualitative study that might generate new insights into the development of mentoring relationships with socially excluded young people, reveal some of the processes that took place over time, and placed these relationships within their wider context, including power dynamics that might affect both members of the mentoring partnership and the practice of mentoring itself. From my own socialist feminist perspective, I felt there were three neglected questions about mentoring which I wanted to try and answer:
Are mentors subject to external sources of power through control and surveillance, including self-surveillance?
How are mentoring dyads situated in relation to wider power relations, through their overt institutional setting as well as more covert aspects of power such as dominant discourse and structural forms of oppression?
Because I wanted to make sense of the way the young people and their student mentors experienced mentoring and the meanings it had for them, recognising that these are inevitably mediated by contextual factors, my main method for generating data was to carry out in-depth, semi-structured one-to-one interviews with mentors and mentees. I selected members of matched pairs who had established an on-going relationship, and carried out follow-up interviews towards or after the ending of the relationship, in most cases six months to a year later. This resulted in the generation of data from nine mentoring relationships.
The interviews produced a large quantity of very rich data, and when the generation stage was complete, I faced the daunting task of analysing it and making sense of what I had found. (A more detailed account of this process is given in Colley, 2001b.) I turned to some of the research method textbooks reviewed above for guidance, and accordingly set about constructing categories from mind-maps I made from each interview transcript. I knew I was not genuinely using grounded data theory as Kim Diment has outlined it above, because the relationship between my analysis and data generation was not premised on the evolving alternation of the formation of hypotheses and their verification in the field (Strauss and Corbin, 1998a). But like many other researchers I drew on its ethos, trying to ensure that the analysis emerged from the data, that I had ‘saturated’ all my categories and that I had not glossed over relevant data (Bryman and Burgess, 1994a,b). I hoped that I would be able to discern relationships of similarity and difference, both within each group of interviewees, and between them, and I worked extremely hard at the laborious task of trawling the data. I was pleased that the categories which emerged both reflected my original research questions, but also raised some surprises – I had not expected the mentors’ very strongly-felt sense of being under surveillance, nor their self-censorship of discussions with their mentee. I then began to code the data in order to produce a written account of the findings.
It was during this process that I began to encounter a number of problems. Firstly, no matter how hard I tried to concentrate as I cut-and-pasted passages from the interviews into the various categories, and despite the assurances of the textbooks authors that with care this would not happen, I found myself constantly drifting into an automatic mode. My very familiarity with the data was decontextualising it – an error which was particularly disappointing given the way I wanted to locate mentoring through my research.
Secondly, the coding process led me into an unintentional prioritisation of the data from the mentors over that generated with the young people. The students were highly articulate, and gave much more theoretically constructed accounts that often linked their mentoring experiences with their academic studies. Although some of the young people were extremely talkative, their accounts were often intuitive, and the data from those who were extremely shy, or had learning disabilities, was simpler far smaller in volume. There was so much more to cut-and-paste from the mentors, and this data was far easier to categorise. The voices of the young people were becoming overwhelmed.
Thirdly, my written accounts of the findings appeared increasingly embarrassing. My attempts to come up with a typology or spectrum of mentors’ understandings of mentoring seemed absurd when it was only drawing on the views of 9 mentors. I was missing the point of in-depth but small-scale case study research, which cannot produce comparisons or generalisations, but has value because it considers each case as singular in its own particularities, revealing deeper insights that large-scale generalisations obscure.
Finally, the technique of ‘slicing’ data according to categories fragmented my representations of the mentoring relationships themselves – although to create insights into such relationships was one of the main purposes of my research. Categorising the data led to considering the respondents in groups (mentors, mentees and scheme staff). All my efforts seemed to be propelling me away from the very ambitions I had for my research.
Instead of pursuing adaptations of this paradigmatic approach to data analysis, I turned to an alternative method often used in life history research – narrative analysis. In a sense, this is not analysis at all, but rather a process of synthesis: Finally, the heuristic researcher develops a creative synthesis, and original integration of the material that reflects the researcher’s intuition, imagination, and personal knowledge of meaning and essences of the experience…In this way the experience as a whole is presented, and, unlike most research studies, the individual persons remain intact (Moustakas, 1990: 50-51, original emphasis).
One of the major shifts that distinguishes narrative analysis from paradigmatic analysis is in its abandonment of the quest to catalogue similarities and differences:
The search is for data that will reveal uniqueness of the individual case or bounded system and provide an understanding of its idiosyncrasy and particular complexity (Polkinghorne, 1995: 15). I used this approach to carry out my final analysis of the data, employing the method of ‘emplotment’ (Polkinghorne, 1995). This meant working backwards chronologically, framing the outcome of each mentoring relationship, then selecting data, including contextual material, according to its contribution to the ‘plot’. I made a point of always beginning with the data from the young person (rather than the mentor) at each stage of the plot, and soon I had constructed case studies of two mentoring relationships which I felt were far more satisfactory. I thought I had found my ‘golden key’ to data analysis, and that the account of the next relationship I wanted to describe would be quickly produced. However, I found myself once again frustrated.
Lisa, the mentee, had been bereaved of her mother three years previously, and felt that she had had to hold her family together emotionally since then. She had a repeated pattern of starting well in the work placements on her pre-vocational training course, then quitting them a few weeks later. Yvonne, her mentor, had long experience of working with disabled children and adults. She helped her mother look after her learning disabled brother at home, and she had been employed full-time and part-time in a respite home for severely disabled children. Yvonne was becoming increasingly frustrated and disappointed by Lisa’s cycle of repeated failure, and it was undermining her own confidence. Both of them spoke about how their relationship was ‘going round in circles’, but how neither of them felt able to progress or bring their relationship to an end, and both were finding it increasingly difficult to communicate in their meetings.
My problem was that, as I tried to use the linear method of emplotment to creative a narrative from this circular data, I found myself going round in circles. On the one hand, the description of repetitious incidents seemed boring. On the other hand, it was producing a reductive diminishment of a story which was far more complex than I can represent in the remit of this paper. The interpretations it produced suggested that Yvonne was bullying Lisa to ‘stick with it’, and was unable to reflect on her own practice sufficiently to break the cycle of repeated failure in Lisa’s placements and in her own role as mentor. It was returning me again to typical representations of the individual power of the mentor over the mentee. In the end, I found myself with a serious and worrying bout of writer’s block.
My supervisors suggested using creative writing to break the block at least (cf. Nelson, 1993), and a senior colleague recommended a further alternative to try as a means of analysing the data: the use of ‘radial narration’. While linear narrative derives from the tradition of Aristotelian logic, radial narration derives from a Celtic tradition which circles about, repeats, and elaborates a central theme (Le Guin, 1981). Using this idea, I engaged in some free writing about Lisa and Yvonne’s relationship. Instead of focusing on the opposition between the two young women, and the tussle in their relationship, something else emerged much more clearly.
I came to see how both Lisa and Yvonne’s identities were closely bound up with a gendered notion of caring for others – a feminine stereotype which oppresses women in a deeply internalised way, through expectations of self-sacrifice, the repression of their own feelings, and the attempt to absorb and neutralise the emotional burden of others. The process of mentoring and being mentored was producing and reproducing this ideological construction of care in both Lisa and Yvonne, and the longer it went on, the less able they were to escape the idealised images each brought to the process, or to admit that truth to each other. These were not opposite or opposing experiences, but parallel. And their parallel nature revealed the way in which mentoring was influenced by wider power dynamics that are so deeply ingrained in our patriarchal society that they have become almost invisible.
The lesson I have drawn from this experience is that there are notechniques, whether conventional or radically non-conventional, to which we can turn with certainty that they will resolve our problems in making sense of qualitative data. If deployed unthinkingly, research techniques may drive our enquiry off course rather than help us gain in understanding. The use of ‘radial narrative’ is no more a guarantee of success than any other method, but each method must be chosen to fit not only the data, but also the interests and values we pursue in any research study, as Kim Diment has so powerfully argued above.