An overview of data collection methods

I adopted three strategies of data collection to achieve a multifaceted look at the health situation of Roma communities in England. They were: 1) a review of health policy documents; 2) participant observation in Roma community organisations; and 3) interviews with community members, community support workers and health professionals. Data collection methods were selected with the aim of developing a holistic view of the environment in which Roma community members interact with health professionals, as well as to capture broader interactions with UK public institutions (focusing specifically on health-related benefits and immigration insecurity).

This table provides an overview of my three modes of data collection, outlining the specific activities involved in each method and defining their connections to my research questions.

Table 3: Data collection methods

Method Research questions Data collection activities

1) Policy analysis To what extent do the Health and Social Care Act of 2012 (and

2) Participant observation To what extent do language and communication barriers

conceptualisations of their position within UK social institutions?

3) Interviews To what extent do language and communication barriers

An analysis of policies related to GRT health (Chapter 5) offered a view of current levels of health service provision to Roma populations, revealing striking local variations in the extent of local decision makers’

recognition of Roma needs. The policy analysis provided a foundation for collection and analysis of field data. Comparative analysis of policy documents furthermore indicated the current state of thinking about Roma needs within UK health care institutions, suggesting the power relations implicit in decision makers’ efforts to define Roma health needs (often without community consultation).

Where the policy analysis responded to my research questions by elucidating aspects of the relationship between Roma communities and health system institutions, it did not provide insight into Roma experiences of health services, nor does it facilitate description of day-to-day life in Roma communities.

My field research investigated Roma experiences of accessing services, with a view inter alia of discovering whether there are discrepancies between governmental accounts of service provision to Roma communities and their experiences ‘on the ground’. Achieving this aim required a robust strategy for developing an understanding of Roma participants’ beliefs, opinions and experiences, and I adopted a combination of participant observation and interview strategies to capture multiple dimensions of Roma health experiences.

Participant observation occurred throughout my fieldwork and was key to facilitating a smooth entry into an unfamiliar research field, as well as for honing my impressions as I moved into data analysis. This method enabled me to develop initial familiarity with the Roma community and to formulate impressions of the ways in which Roma interact with health systems. Furthermore, by beginning data collection as an observer and conducting interviews only after I had spent a number of months in the field, I was able to build relationships with participants before beginning formal recruitment for interviews. This proved to be essential to gaining the trust of prospective participants and ensuring that they engaged in my project with an understanding of my research goals and activities. Participant observation was not only vital in offering an introduction to a novel social setting, but it also aided in contextualisation of data gathered via other channels. I actively collected data between April 2015 to June 2017, during which I spent approximately seven hours per day, four days per week primarily in Roma community centres, though there were also opportunities to accompany participants to meetings with health care providers and local councils. Once I had concluded my formal period of data collection, my ongoing employment with Roma Support Group allowed me to remain partially entrenched in the field as I analysed and wrote up my results. Spending this substantial amount of time in community spaces – and also observing how participants interacted in more official institutional settings – increased my sensitivity to beliefs and lifestyle preferences that may impact health-related perceptions and behaviours, and thus allowed me to refine the questions for interviews.

As participant observation largely yields data informed by the views and impressions of the researcher, interviews constituted a key component of my data collection in giving voice to Roma community

members, community advocates and health professionals. Interviews validated and further explored concepts derived from participant observation, and the interview topic guide was regularly updated based on concepts emerging from interviews and participant observation (in accordance with guidelines for developing sensitivity to concepts outlined in Strauss and Corbin’s (1998) text on grounded theory methodology). While I loosely followed the topic guide, I found it vital to remain flexible in the interview process and allow participants to discuss areas of personal significance in-depth. This strategy was intended to minimise the unequal power relationships that can arise in the researcher-respondent relationship, thus subverting the typically dominant role of the researcher in determining the direction of the conversation (Beck, 2011; Harvey, 1949; Peller, 1995).

In some cases, interviews diverged substantially from the topic guide, taking on the form of a non-directive narrative interview (Brinkman, 2018). These interviews centred on pressing and at-times overwhelming concerns in participants’ lives; while these concerns did not necessarily address topic guide questions, I ultimately deemed it most appropriate to afford participants the space to discuss issues that were of greatest relevance to their lives. This resulted in a set of nine interviews that were largely narrative in character – incorporating stories of traumatic health events or distressing experiences of claiming health-related benefits – which formed the basis of my narrative re-analysis of this subset of interview data.

Adopting this three-pronged approach to data collection allowed for consistent interrogation of multiple dimensions of Roma health experiences. I sought to holistically explore the concepts arising through observation and analysis by drawing connections across the results from different modes of data collection, thus revealing the relationships between Roma cultural beliefs, current and historical social positions of Roma communities, health system policies, attitudes of health professionals and the role of non-Roma community advocates.