Methodology: an explanation and a critique

The ideal is equality between researcher and researched, but this is hard to achieve. (Coslett et al., 2000)

Introduction

The aim of this book was always to explore the experiences of women who have risen to the top, or near to the top, of their organisations or professions. The study focuses predominantly on women in the UK public sector, including chief executive officers (CEOs) and women in other senior positions. This chapter sets out the assumptions and perspectives behind our methodology and looks at how we went about selecting our sample, devising our interviews and carrying them out, as well as providing a critical and reflective account of the process.

Quantity or quality?

From the outset of the project, we wanted to find a way of giving a voice to 'successful' women in the public sector. We therefore ruled out a large-scale quantitative survey in favour of a smaller, qualitative, approach. In line with much auto/biographical feminist writing (see, for instance, the collection by Coslett et al., 2000) we were attracted by the idea of talking directly to women in order to tap into their own accounts and subjective experiences of their career progression. We wanted to understand their attitudes and beliefs as well as listening to a descriptive account of their careers. Moreover, we were attempting to avoid reproducing the unequal power relationships which often characterise interviews and in which women have often been cast as passive subjects (Coslett et al., 2000). We aimed for an exploratory approach in which the biographical details of each woman's life and career could unfold in a way largely determined by them. Rather than impose strict criteria or boundaries on the talk, we wanted the interviewees to set the agenda as far as was feasible and to allow for an unfolding of their accounts of their experiences, and their feelings about them, in as 'authentic' a way as possible.

This approach assumes a particular perspective on, and attitude towards, the subjects of the book and the material produced. A recent article by Burr (2009) provides a useful reminder of the basis of much qualitative, small-scale research based on observation and/or an interest in interaction. Talk, including that produced in interviews, is an important aspect of social interaction, comprising crucial 'expressive techniques' where individuals present themselves, like actors on a stage (Erving Goffman's 'dramaturgical model', outlined first in 1959). Such performance of self includes reference to the role of others, in the sense that the presentation of self necessarily includes the ability to take the role of the other (see Burr, 2009; but also Blumer, 1969; Holdsworth and Morgan, 2007) so that the in-depth interview can be seen as a significant interaction between active subjects. When this general approach is applied within feminist scholarship, with its history of linking the personal to the political and with its emphasis on auto/biography (see, for instance, Davies, 1992; Stanley, 1992), the result is a position where 'the subjective element must not be left out of the practice of research methods, such as the interview' (Coslett et al., 2000).

Lumby (2008) takes the notion of 'multiple constructions of [selves] which change and can be understood differently over time and in different contexts'. In asking these women to talk about their careers, we were well aware that they would be (possibly unconsciously) talking from a particularly privileged position of achievement. They would be, as Erving Goffman might have put it, managing the presentation of particular identities and 'selves', and these would undoubtedly be shaped or framed according to the context of the interview itself as well as by the label of 'successful' woman (see Goffman, especially 1959, 1974, 1981). All these elements could undoubtedly influence their perspective on women's success and achievement more generally. Nonetheless, since we were aiming to uncover the women's subjective experiences, bound up with their identities, we took the construction of their narratives to be constitutive of their own realities (see Silverman, 2007, for an outline and critique of this position). Put more simply, the fundamental assumption of this research is that there are multiple realities and multiple truths and these are bound up with social and individual status and identities, including gender. The book aims to present the realities of a small group of high-achieving women as constructed in the context of an interview.

The interview format - some issues

Inevitably, there were some provisos built into our initial aspirations for an open-ended exploratory approach in the interviews. We realised early in the process that the notion of success was problematic and needed to be interrogated and that the interviews would need to incorporate explicit exploration of the concept. We were also looking for the interviewees to provide us with a narrative account of what they considered to be important in their lives and careers and this needed to be in a form that we could work with later in the writing-up phase. Therefore we were already imposing some shape, direction and order on the narratives that would emerge. In order to achieve these broad outcomes, we devised a loose framework within which a conversation could take place. From these parameters, a semi-structured set of questions was devised. The checklist used in the interviews is reproduced in the appendix at the end of the book. The outline of areas to be covered followed a largely chronological format - asking interviewees to tell us about their early lives, work experiences and so on, then moving on to more thematic issues such as notions of success or the advice they would give to other women - and men (see also Kumra and Vinnicombe, 2008).

The sample

Selecting a sample for any research project can be problematic (see, for example, Sara Arber's chapter on sampling in Gilbert, 2001). In this case, there were quite specific issues to deal with. Much discussion took place about the definition of 'success' encompassed by the study and this is outlined in greater detail in chapter linkChapter 4 of this book. Once parameters for selection had been agreed (women occupying the top level post - such as CEO - or other senior positions within an organisation or profession), the potential population narrowed down considerably. Numbers of women at the top are limited, notably in the public sector - indeed this is what prompted our interest in the subject in the first place - so the pool of potential participants was naturally small. We began searching for suitable participants within familiar territory - higher education - and used contacts, including initial interviewees themselves, to suggest a broader range of participants. As already noted, obtaining a reasonably sized sample was not unproblematic, since women in such senior posts are inevitably very busy. In a number of cases, potential candidates for interview agreed but then cancelled, unable to find a time for the interview to be carried out. Thus our sampling method was what might be described as network or 'snowballing' sampling (Arber, 2001) and we do not claim any scientific basis for it. Although we attempted to make our sample representative of the population of 'top women', it is not a probability sample where precise inferences can be made about the population. In a small-scale qualitative study, it is unrealistic to use a probability sample but there are both strengths and weaknesses in utilising a snowballing sample. The main strength is the use of existing networks to find sample members (i.e. women recommending other women). In addition, in this case, participants in the study needed to be assured of the 'legitimacy' of the project and the networking element was useful for this since we began with women known to David in his role as CEO of a university college. The main weakness stems from the same source: women who may be successful but who are not part of existing networks could be left out. We acknowledge that had the composition of our sample been different we may have heard different stories. But our intention was not to produce a scientifically reliable survey but to gain a flavour of the career issues and events seen as significant by a number of senior women. In this respect, we would argue that we were able to collect rich, valid and detailed material which casts light on these women's experiences and subjective truths that might in turn tell us something more general about the gendered experience of success. This approach echoes Inglis' 2010 article, where it is argued that 'sociological forensics' can be used to make valid connections between micro case studies and macro social structures. As Inglis puts it, this assumes 'that the structure of social action reflects in some way the structure of the society and culture of which it is a part' (Inglis, 2010). While we are cautious about making grand claims, it is possible to detect patterns in the interviews which may have more general significance. More detail...