CHAPTER 4: METHODS: CONCEPTUAL AND METHODOLOGICAL FRAMEWORKS
4.2 Methodological framework
4.2.6 Ethics considerations
Ethics permissions
Ethics permissions were obtained from the national ethics committee in Burundi for both rounds (see Appendices 2, 3, 9).
In addition, data collection for the first round was undertaken as a sub-study of a project which was already approved by UWC Ethics Committee, entitled ‗Experience of three African countries with Global Health Initiatives‘ – though further approval was obtained for round 2 (Appendices 8 and 10).
All study participants were informed about the purpose of the study in the language of their choice and their rights and were asked to sign a consent form (Appendices 4, 5, 6, 7, 11 and 12).
The interviews, audio recordings and the notes were kept anonymous and stored in a secure place. A confidentiality clause was included in the employment contract of the research assistants.
Informed consent
The study was clearly explained to each participant until they understood it fully. Written informed consents were obtained before each semi-structured interview. Participants were informed that they could withdraw from the interview process at any time during and after the completion of the study – and that if they withdrew, they would not be negatively affected in any way.
Initially, participants were given the choice of language – Kirundi and French. Where participants were fluent, interviews were conducted in French, which was often the case.
Confidentiality
Confidentiality was guaranteed to every participant interviewed. Interview notes and transcripts, and audio taped materials were all stored in a secure place and access was limited to the co-researchers and to research assistants.
When quoting anyone, interviewees are referred to through the use of broad categories such as ‗donor institution‘ or ‗government institution‘, to make sure that the identity of the interviewee is not revealed.
Responsibilities to research participants
Researchers ensured that participants did not feel uncomfortable by having them choose the venue of their interview at their convenience. Researchers did not make use of their position to force participants to participate or to disclose information, especially on sensitive issues.
Practical limitations in methods
There are several limitations in this thesis, however.
With regard to coordination analysis, time-based trends could not be respected: some policies were formulated before others, so influences of one policy over another, including those that were formulated later were not distinguishable. The document analysis was not exhaustive as some were not available or were difficult to access.
‗Coordination‘ was a relatively new concept globally. In our findings, it seemed it was used rhetorically, rather than coming from a real understanding of what coordination implied. Nonetheless, rhetoric may become reality after a while, since time is key to the ownership and incorporation of ideas. Our sets of findings could not capture any trend in this sense, given the short time span.
With regard to language, all interviewees were fluent in French, even for Burundese. All the transcripts were translated into English for triangulation. Some meaning could have been lost during the translation process or some cultural specificity
misinterpreted, but the use of local researchers contributed to minimise the extent of these bias. The fact that the main researcher was French added to the quality of the
the second commonest language and is extensively used in official documents and meetings. Higher education is provided in French.
With regard to sampling, all organizations involved in the health sector could not be exhaustively included in the study, given limited time and funding. Some donors were not major actors in the immediate post-conflict period and were not taken into account in this organizational analysis, even though their influence might have grown over the years. A brief overview of excluded organizations is provided in Appendix 18. However, the relatively large size of the sample is likely to counteract the possible sampling bias.
With regard to the researcher‘s position, interviews were easy to get, including ‗elite‘ interviews, either as being a foreigner, and being working for first the WHO and then for a foreign academic institution gave me symbolic power and put me in a higher position, or as a return on investment was expected.
Some interviewees might not have expressed their true opinions, for fear of
repression or of a breaking of confidentiality. In general, however, local people had more confidence in foreigners, as Burundese believed foreigners to be trustworthy, compared with Burundese in general. In addition when interviewee was concerned about repressible content, they asked not to be audio recorded, or asked to stop the recording. Caution was exercised when interpreting sensitive content, by
triangulating findings with local research assistants, and by taking into account the potential interests/fears of interviewees. Any case of reference to embezzlement from the part of the interviewee was treated with special caution, so as to not
perpetuate what could be a simple rumour. But somehow surprisingly, interviewees were quite open about this topic, and pointed out at one person in particular, holding this person responsible for his act, and generally not at one institution (in case of embezzlement only). These cases were reported only when a triangulation was possible.