CHAPTER 3. RESEARCH METHODOLOGY
3.3. Research methodology and methods
3.3.3. Quantitative and qualitative data collection
As indicated above, this was a mixed methods study with quantitative and qualitative data being used to assess different aspects of the groups and their operations and to complement each other.
First, quantitative data was necessary to understand the financial performance of groups and member’s socio-economic characteristics. A panel survey of member and group level financial performance was collected from group records over multiple rounds of data collection. The
77 instrument collected data on member’s savings and loans at the individual level and also collected the performance data of the group as a whole (e.g. Return on Savings). This data collection was in fact highly complicated and time consuming as data records were rarely complete or well kept. Nevertheless, persistence with this paid off because it did much to reveal the reality of how money was actually being managed as will become evident below.
Second, a questionnaire survey was administered to all group members on socio-demographic and livelihoods data. In particular, the survey used some similar questions as are contained in the national FinAccess surveys in order to provide data that could be compared with those datasets.
The financial data and socio-economic data on group members were analysed in two main ways using the SPSS statistical package for the recording and data analysis. First, the socio-economic data was used to produce descriptive statistics of the profile of members to compare the sample with available national level data so that the sample can be put into a wider context.
Second, the financial performance data have been used alongside scores regarding rule following (see below) to construct a number of measures of financial and rule following performance. These have then been used to statistically test the correlation of these with socio-economic variables of gender, poverty, wealth differentials and education in order to assess whether these appear to be influencing performance in any systematic way.
For the in-depth analysis of rules and norms and the influence of social relations on these, four approaches to qualitative data collection were used. First, semi-structured interviews were conducted with group members. We endeavoured to interview the two or three main leaders of the group - chairman, secretary and treasurer; and three members who were randomly selected. The data were collected both by directly discussing the issues or rules and norms with them but also indirectly by collecting accounts of what had happened in the groups and the nature and extent of rule following that could be understood from these. For example, if we knew there were delays, default or fraud in the groups, we asked what the group had done in order to recover the funds. We also used critical empirical incidents in the group history as a focus of such discussions with group members and leaders to examine whether and how the rules were actually practiced.
Second, key informant interviews were carried out with trainers themselves and other key informants such as project managers, local chiefs and so on, on a range of issues, including the training procedures, critical empirical incidents in particular groups, and issues around rule following and enforcement that arose as the research proceeded.
78 Third, during the third round of visits the groups were observed actually undertaking their group meetings in order to directly assess how the rules that the groups had agreed to follow were implemented. This was done by myself and two research assistants using a checklist, taking notes and triangulating our findings afterwards. These results were also turned into quantitative scores of rule following and used in the quantitative analysis. Qualitatively, the analysis also then uses these observations with the training methodology, and trainers and member’s accounts of the training.
Finally, in order to understand the composition of groups in terms of relatedness of members to each other, relationship maps were drawn in the smaller sample of eight groups. After a group meeting all the group members were invited to assist in the process of drawing the map on a flipchart that everybody was able to see. First the names of the leaders were written in the middle and group members names were then added one by one and all the members were encouraged to contribute how the members were related to leaders. Different colours were used to mark the following relationships between the leaders and members: from the same nuclear family, aunts, uncles, cousins, grandparents, in-laws and neighbours or friends. If some of the members were not related to any of the leaders, we inquired whether they were related to other members. Also, if they were not related to anybody, they were asked who had brought them to the group.
In practice, four visits were made over a period of just over two years: the first field visit took place in April-May 2011, the second in October-November 2011, the third in April-May 2012 and the fourth in May-June 2013. All of the qualitative data was stored in Nvivo and analysed with it.
During the first field visit two research assistants with Masters degrees from the Institute of Development Studies at the University of Nairobi assisted in the collection of the quantitative data. Since they were not available during the second and subsequent field visits, another female research assistant assisted in the collection of the qualitative data. She had been working for several years in a project that had done research on SGs and other similar groups and hence had a good understanding of the issues. She undertook interviews in Swahili, rather than the local languages. Whenever the respondents were not fluent in Swahili, I interviewed them with the assistance of local translators, one from Nyamira who was a native Kisii speaker and another from Rachuonyo who was a native Luo speaker. The local translators also organised field visits and interviews with the group members. The translator from Nyamira was a university graduate who had previously worked for a large donor funded health project in
79 Kisii, and the translator from Rachuonyo was a young trained English teacher who had not yet obtained a permanent teaching position.
Regarding the analysis and the coding of the material, I had some prior aspects I was interested in on the basis of the literature; for example gender relations, age and other forms of power dynamics around institutionalisation of rules. Further, the literature had also indicated that groups fail so I was particularly attentive to the reasons why they fail. The main reasons for group failures, default and fraud, came out over time as I observed delays and other problems. My research was oriented towards understanding that these dynamics were likely to have problems. These themes proved to be important features of how the systems were working that I needed to examine and pay attention to as the COSALO programme seemed to suggest that increased transparency and accountability of the SG methodology will avoid these.
I therefore analysed material and undertook coding around these themes across all the 204 interviews, of which 171 were SG participant, 12 trainer and 21 key informant interviews.
However, not all the material in all the interviews was coded but that that related to above mentioned themes. In coding I further concentrated on the eight groups on which I had more in depth material and presented them as case studies in chapter 7.
However, I did not just encode data based on my understanding from the literature. These themes also partly emerged from observation. I did analysis between the different stages of fieldwork and this analysis then informed my next round of fieldwork. For example after collecting financial data during the first and second field trips the substantial challenges with repayment, quality of recordkeeping and misappropriation of funds became clearer and led me asking more questions on these themes during the consequent field visits and hence deepening the focus of the analysis on these areas.