• No results found

Method and Research Design

4.2 Research design

A research design is an overall plan or strategy for how and why the research is to be carried out – however, the design of a qualitative study rarely begins from a predetermined starting point, with fixed steps to steer the process. Instead, a qualitative research design is constructed and reconstructed throughout the research process (Maxwell, 2013). The central components in a research design are goals, theories, research questions, methods, and validity, and the connection between these are continuously reconsidered during the study. The purpose of the research design is to show the link between these components. As presented in the

introduction chapter, the research question guiding this thesis is: How did the TRACKS knowledge mobilisation process impact on climate change adaptation in northeast Bangladesh, evaluated as governance? The theoretical framework of this study has been

61

presented in Chapter 2 and 3. In this section, the choice of method will be explained with reference to the research question, goals and theories.

4.2.1 Method: Qualitative interviews

Before designing a research project, it is essential to ask “what do I want to learn?”. The answer to this question guides the choice of method. According to Aberbach and Rockman (2002), qualitative interviews are a suitable method when attempting to understand how people think or how they interpret an event. Rathbun (2008) argues that qualitative interviews are best fitted to gather information about the characteristics about the social world that are different from those of the natural world, and when wanting to include factors such as culture, norms, ethics, perceptions, learning and cognition.

Qualitative interviews were considered a useful method for this thesis for a number of reasons. As mentioned in the methodological discussion above, quantitative methods are usually chosen when breadth is considered more relevant than depth, and when the goal is to be able to make general statements that can be applied to society or people. The case being studied in this thesis, is highly context specific: because climate change affects regions of the world very differently, with large, local variations, adaptation must be tailored to each

community – there is no ‘one fits all’ adaptation strategy. As such, the aim is not to apply the results to the wider society, but rather to understand climate change adaptation relative to the local context in which it is embedded. A second reason for choosing qualitative interviews is the approach: applying a participatory evaluation approach made qualitative interviews a natural choice of method, as it allows the respondents to answer according to their own knowledge systems. This is a crucial criterion when the evaluation criteria are based on the respondents’ opinions and points of view. In settings like this, when the aim is to investigate people’s conceptions of a specific topic or social phenomena, Stenbacka (2001) and Aberbach and Rockman (2002) state that open questions are important in order for the informants to be able to organise their answers in accordance with their own references and knowledge

structures. The choice is also influenced by my methodological perspective as a researcher: in identifying mostly with the constructivist perspective, the focus of this thesis is not to uncover a single truth, but rather to understand what is true for the respondents in this study. This also relates to the necessity of tailoring evaluation to its specific context, as discussed in Chapter 2 and 3. Rather than evaluating the impacts of TRACKS based on predetermined criteria derived from theory, qualitative interviews that allow the respondents to answer relative to their own perceptions, was considered the best suited method in this respect.

62 4.2.2 How to ensure quality in qualitative interviews

Because qualitative and quantitative methods are essentially different research approaches, the quality of the research must also be assessed relative to different concepts of quality.

According to Golafshani (2003), many of the most common concepts of quality, such as reliability, validity and generalisability, were originally developed for the assessment of quantitative methods. As such, they need clarifying and redefining before they can be applied to qualitative research. Stenbacka (2001) states that the most common definitions of

reliability, validity and generalisability are not appropriate for qualitative research because they aim to explain, while the aim of qualitative research is to generate understanding.

Without redefining them for use in qualitative research, these quality concepts can lead to unnecessary restrictions and erroneous conclusions, according to Stenbacka (2001). With this criticism as a backdrop, I will now discuss how concepts of quality can be understood in terms of qualitative research.

4.2.2.1 Validity

A common understanding of validity is that it is a quality concept for the researcher to make sure that he is measuring what he wants to measure (Ringdal, 2013). However, because the aim of qualitative research is not to measure, this understanding is not well suited to assess the validity of qualitative methods, according to Eneroth (1984). Stenbacka (2001) states that it is the purpose of the research, that primarily decides if the data is valid. Based on this, she upholds that it easy to answer how validity can be ensured in qualitative research. When the aim is to generate an understanding of a social phenomenon, one is interested in

understanding the respondent’s conception of reality about a specific topic. This means that the understanding of the phenomenon is valid if the respondent is involved in the topic and gets the opportunity to speak freely according to her own knowledge structures (Stenbacka, 2001).

4.2.2.2 Reliability and replicability

Ringdal (2013) defines reliability as to what extent repeated measurements with the same tools will lead to the same results, in other words, if the results can be replicated. According to Stenbacka (2001), this understanding of reliability is irrelevant for qualitative research – again, because the aim of qualitative methods is not to measure. If qualitative interviews were to be repeated with the same people, a few years later, would the answers be the same? Quite certainly not, as people and contexts change over time. Stenbacka (2001) therefore suggests

63

alternative concepts to ensure reliability in qualitative research: if the researcher makes the whole research process transparent, from planning to collection of data and analysis, the study will be reliable.

4.2.2.3 Generalisability

Statistical generalisability is a common quality concept in quantitative research – meaning that it can generate general claims about a population or society. Yin (1989) distinguishes between analytical and statistical generalisation, and upholds that only analytical

generalisation is relevant for qualitative research. This means that general, theoretical conclusions can be drawn from the empirical material. In other words, the aim of qualitative research is not to make universal claims about a population, but rather about a theory (Yin, 1989).

4.2.2.4 Ensuring quality in this thesis

Based on the discussion above, and in recognising that qualitative research requires different quality concepts than quantitative research, this thesis aims at ensuring validity, reliability and analytical generalisability as defined by Stenbacka (2001) and Yin (1989): validity is

understood in terms of to what extent the study fulfils its goals, reliability as transparency, and analytical generalisability as the possibility of drawing general, theoretical conclusions based on the findings. In the following section I will now present how the data collection was carried out, before I discuss strengths and weaknesses with the approach.