Chapter 2. Methods and methodology
2.3 Qualitative Interviews
In total I conducted 55 semi-structured interviews (including the 21 fieldwork interviews) to better understand views and collect anecdotal evidence (Appendix 2 outlines the interviews in more detail). I started with an initial set of informants that I chose to study 'innovation approaches to NTDs'. However, during the course of the interviews it became apparent that a different story was emerging. These first interviews constituted an early fact-finding and overview of the NTD community. In keeping with a 'grounded theory' approach to theoretical sampling I conducted an NVivo analysis and a new set of conceptual cores emerged around 'measurement' and 'advocacy'. These cores guided my selection of further informants, whose current institutional affiliations spanned across academic, policy and charitable fields.
The next set of interviewees were sampled using an element of 'snowballing' and 'chain-referring' to choose interviewees once I had made contact with a key actor in the field or completed an interview. The strengths of such an approach were to find interviewees within networks that were not so obvious through an outside search. The limitation was being constrained to those original networks and so not including more oppositional viewpoints. As Ruane points out, relying completely on snowballing may risk omitting individuals who are not well 'networked' (2005, p. 118). Therefore, snowballing was in complement to a purposeful sample of interviewees, generally following the criteria of policy awareness or engagement.
The NTD community itself is small and could be regarded as an elite, with high-profile scientists and politicians as policy influencers and participants in policy change. To clarify elite interviewing, I adopted Hochschild's definition of "...discussions with people who are chosen because of who they are or what position they occupy... the term indicates a person who is chosen by name or position for a particular reason, rather than randomly or anonymously"
(2009, p. 124). By the end I had interviewed a mixture of policy actors through semi-structured qualitative interviews. As Clifford et al. describe this type of verbal interchange allows the interviewer to elicit information through prepared questions but it is done so in a conversational manner, with the chance for participants to "...explore issues they feel are important" (2016, p.
143). Therefore, it is a more of a two-way process as opposed to survey or structured interview.
These interviews were predominantly with research institutes and universities in Europe, the US and Australia, as well as my fieldwork countries Brazil and China (totaling 35). I also identified five 'activist scientists' in the US, UK, and Australia as instrumental in developing policy for NTDs. Finally for a rounded policy picture, the remaining 15 actors I interviewed were from governance institutions (the WHO being especially important), NGOs, and Pharma companies.
In addition, five of the interviews formed part of a podcast series with colleague Erman Sozudogru at UCL. This was an online podcast and workshop series on NTDs called 'Challenging Neglect'. We interviewed five high-profile academics about their work on NTDs in different disciplines ranging from history, philosophy, and the biomedical sciences and held workshops to discuss disciplinary perspectives.17
For the fieldwork interviews, where I wanted to meet actors in their work settings, I did not employ a formal translator. I found that as most of my interviewees were engaged in global policy-making and forums for NTDs, English was the modus operandi. When I did need assistance with translation this was provided on several occasions at Tsinghua University and the National Institute for Parasitic Diseases by the colleagues of the interviewee. Translation by colleagues aided the conversation, which can sometime be stifled and formalized by the presence of an outside translator. Still, there was is a restriction in using English searches of the literature and for interviewees, which means I may not have included some non-English sources of information and insight.
The typical interview time varied from forty-five minutes at the minimum to up to two hours in some cases. Most interviews were held in the participant’s office but in cases where this was not possible a Skype or telephone interview was conducted instead. As well as taking detailed notes, I recorded all interviews and transcribed them. The exceptions to recording, included once when the interviewee did not agree and another when the setting was particularly noisy.
The interview questions were centred on six main lines of questioning:
1. Reason for involvement in NTDs: Can you tell me about your background, what led you to the topic?
2. Problem and solution understanding and framings: What are the causes of NTDs and what are the proposed ways of solving them?
3. Approach being taken to address NTDs: Which particular strategies are being pursued and why?
4. Reflections on the change of term: What is your opinion of the use of the term 'NTDs'?
5. What is the relationship between actors: Who are the main partners you work with?
6. Positioning of Brazil and China within global policies for NTDs: What is being done differently in these countries, what are the challenges, what are the strategies being deployed? (The last question I tweaked the most depending on the interviewee).
(See Appendix 3 for a more detailed list of interview questions).
As the interviews progressed I did aim at greater specificity or directness, although it struck me that interviewees still will want to say what they want, rather than what you want, and this is as it should be. I avoided steering the conversation too heavily, only moving to next questions when
17See www.challengingneglect.com.
ground was repeated or asking qualifying questions when answers were unclear, rather than due to my being 'unsatisfied' with answers. It may have been because of the type of interviewees, most of whom were senior scientists engaged in policy and public roles, that opinions and accounts of events had already been formed in their minds and some experienced difficulty in looking at these from a new light. When I did ask questions prompting reflexivity, I often had the response of "you tell me" or "I hadn't thought about that before" or a deflecting answer in order to avoid going into too much detail.
These hurdles may be generally reflective of the realities of interviewing as a research method, disciplinary differences, and the type of interviewee. However what I did find surprising was the level of openness interviewees had about airing disputes and being very frank about dislike for individuals, organizations, and activities, even though many did not want to have anonymity. I have not included all of these references as they were not always relevant to the research but they certainly assisted me in navigating and gaining an understanding of the NTD community.
I gave a varied level of anonymity to all the participants ranging from full disclosure to full anonymity, depending on their preference. After the interviews I transcribed the text and coded the field notes to be classified according to themes, topics, and other relevant categories with the aid of the software package NVivo (See Appendix 4 for node listing). Therefore, my data analysis was supported through creating distinctive categories, which allowed linkages to be made between concepts, and comparisons between cases and events. I then analyzed the coded text and other data inspired by a ‘grounded theory approach’ in order to concentrate on developing theory related to concepts, relationships and explaining actions to identify the perspectives of groups of actors, reflecting on my interpretative role as researcher. Thus following the general steps typical for grounded theorizing, I did not strictly apply grounded theory, concentrating instead on a close and exploratory-oriented reading of the data. I began with the early stage of data collection and continued up until the writing-up stage with the focus of enquiry clarified over the course of the research.
I developed analytical categories to make sense of the data by coding at the level of theoretical codes. The codes were derived from the data rather than being pre-determined and these involved not only substantive codes but the gathering data from various interviews or documentary sources under the same relevant category (Sapsford & Jupp, 2006, p. 251). It is the segmenting and comparison of data from different parts of the data record that
"...distinguishes grounded theorizing from other forms of qualitative data analysis" (ibid.).
I also undertook a small amount of participant observation – described as the observation and participation of the social action being documented (Hume & Mulcock, 2004) – at a number of technical (and semi-technical) conferences and events listed in Appendix 5.
2.3.1 Ethical considerations
The departmental ethics board approved my research plan and I gained informed consent for all my interviews according to the criteria listed by UCL Research Ethics Committee (ethics.grad.ucl.ac.uk) on qualitative research. Interviewees were given an information sheet with a consent form and a copy for them to keep. The sheet covered the following aspects of the research:
• Research background
• Project aims and objectives
• Information about who is interviewing
• What the interview will be used for and how
• How information is stored and protected Table 2 Overview of ethics sheet
As mentioned earlier, interviewees were asked if they wanted to be anonymous, what level of anonymity and if not whether they wanted to be informed of which quotes were used, with the possibility to change anonymity after circulation.
2.3.2 Qualitative secondary analysis
For a subset of interviews in this thesis I undertook a qualitative secondary analysis. I reused qualitative research data in the form of interviews with individuals who, for various reasons, I was not able to interview. Some key figures who were instrumental at the beginning of NTDs developing as a concept were deceased, from both the 'Great Neglected Diseases of Mankind' Program' (GND) and the WHO Special Programme for Research and Training in Tropical Diseases (TDR).
Kenneth Warren (1929 – 1996) was the director of health services the GND. I interviewed historian Conrad Keating who will publish a biography on Warren in 2017 and interviewed Warren's contemporaries.18 At the WHO, Lee Jong-wook (1945 – 2006) was director general in 2003, when he initiated changes in the organization for the control and elimination of NTDs (which have been called a 'paradigm shift'). In addition I used secondary research through interviews with Lee published in policy documentation. There were a number of other interviews that I refer to of directors of the TDR WHO, published on policy websites (the WHO and Genève Internationale) of: John Reeder: (2012 – present); Robert Ridley: (2004 – 2012); and Adetokunbo Lucas: (1976 – 1986).
18Nick White is Professor of Tropical Medicine Mahidol Oxford Research Unit and David Weatherall is the founding Director of the Weatherall Institute of Molecular Medicine.
Other high profile individuals whom I did not interview but relied on newspaper interviews included Bill and Melinda Gates the founders of the Gates Foundation, and Nobel Prize winner Tu Youyou. A second reason why I relied on secondary qualitative research was because I was interested in the public-facing representation of NTDs. I acknowledge there are limitations in using this type of data source. Sarah Irwin defines qualitative secondary analysis as: "...the use of already produced data to develop new social scientific and/or methodological understandings" (Irwin, 2013). Irwin with Mandy Winterton state that the distance from primary data production and the related knowledge of "...proximate contexts of data production" can present methodological, ethical, and practical concerns (2011, p. 2).
The focus of my analysis was on the representation of arguments and ideas, therefore the context where the sources were generated is less important. Contextual and situatedness of the sources are points that I address when quoting the interviews, addressing the ‘problem of data fit’ by providing detail on the nature of the source (Heaton, 2010, p. 4). Irwin and Winterton support this view that: "So long as data analysis is fit for purpose then secondary analysis is no poor relation to primary analysis" (2011, p. 2).