• No results found

Data Instruments Design and Administration

Economic Condition

4.8 Data Instruments Design and Administration

The data collection instrument for this study comprises of an in-depth personal interview guide, a focus group discussion guide and an impact survey questionnaire because the mix of tools allows the researcher to have a wider picture of an issue (Pawlak and Szubert, 2004). The impact survey questionnaire instrument is one of the five tools (three qualitative and two quantitative) that were created by Nelson (2000) and which have been useful in assessing the impact of informal finance programs. The instruments have been recognised as one of the few influential materials that have been very useful in carrying out informal finance impact assessments (Barnes and Sebstad, 2000). It is sometimes referred to as AIMS/SEEP impact tools. Moreover, the instruments have been used by other studies (Edgcomg and Garber, 1998; Falaiye, 2002; Larocque et al., 2002; Sharma et al., 2005). Different criteria were taken into consideration in determining whether to use these instruments or not. To make this decision, the researcher reviewed previous works –

Page | 104 including those cited above – that had used the instruments and also took special cognisance of Nelson’s (2000: 1-2) suggestion that users should “choose the tools and adapt them to meet your need because each tool must be adapted to the specific circumstances in which it will be applied”. The comment above led to the modification of the impact survey questionnaires for this study to capture the peculiar conditions of the respondents with special emphasis on the very low educational background of the rural dwellers and the definition of rural areas in section 4.5.1 above in determining the study areas.

The issue of confidentiality was taken very seriously during the field work since the majority of the interviewees are illiterate and all of them reside in the community. As a result, the personal integrity of the interviewee has to be protected to avoid any leakage of information to a third party. For this reason, the interviews were conducted on a one-on-one basis in line with Copestake et al.’s (2002) suggestion because the presence of a third party could influence the interviewee’s response in various undesirable ways. This one to one approach enables interviewees to feel free in responding to the interview questions based on their own personal opinions and experience with the cooperative society. The benefit of this is that it assures the interviewees of the confidentiality of their responses. Data for the study was sourced using survey questionnaires, personal in-depth interview and focus group discussion as shown in table 4.1 above. These required the translation of the questionnaires, interview and focus group discussion guides into the local language (Yoruba) of the respondents in order to gather accurate responses to the questions. For example, cooperative societies are called “egbe alaje seku” in the local language of the respondents. This approach has been suggested by Imp-act (2005: 5) to emphasise that a researcher “needs to think about how to translate the questions into local languages and make sure the meaning remains the same”. A similar process was used in Kenya where “all questions were printed in Swahili as well as English” (Copestake et al, 2002: 9) and in Mali where impact survey questionnaires were written in French and Bambara (Nelson, 2000).

Page | 105 The interviews were conducted in Yoruba, the local language of the respondents. Likewise the focus group discussions (FGDs) were conducted in Yoruba to ensure full and maximum involvement of the participants since the majority of group discussants were illiterate. This is made possible because the researcher is equally fluent in the local language of the respondents and discussants. The interview comprised structured open ended questions, to allow participants to describe their personal experiences with the cooperative and to enable them to speak without any restriction using their own words. Both the interviews and the FGDs were jointly documented using a tape recorder and interview guide record sheets. The recorded discussions from the interviews and the FGDs were transcribed and later translated into English. These transcriptions and translations were then compared and reviewed with the interview notes. The interview and FGD is to enable the members to respond to questions relating to why and how of challenges and successes in the program that cannot be reflected directly through quantitative data with limited options which may not capture all there is in the program from the participants perspective.

The impact survey questionnaire tests the hypotheses on household and enterprises levels. It was divided into seven sections comprising of twelve groups of questions covering individual basic information, household level basic information, income and assets, enterprise income, profit and assets. The data collection instrument was translated into Yoruba - the local language of the study areas and respondents - because “translation of both quantitative and qualitative tools into local languages spoken by the clients is critically important for some program sites” (Nelson, 2000: 3-21). Moreover, successful “questionnaires also require translation into local languages and then reverse translation to cross check the accuracy of the translation” (Sebstad, 1998: 15). The translation was undertaken due to the respondents’ low level of education and in order to enable respondents to communicate fluently with the researcher and for them to feel comfortable in providing answers to the questions asked without being easily distracted in the course of the research.

Page | 106 The study made use of two field assistants in addition to the researcher in each local government area in administering the questionnaire because of the magnitude of the study and the need to cover all designated research areas within the available time without compromising the quality of the study. The researcher and a field assistant conducted the personal interviews and focus group discussions. The questionnaires were personally administered by the researcher and the field assistants directly to the respondents by reading out the questions to the respondents and they in turn provided their answer which was then recorded in the appropriate space on the questionnaire. Moreover, the field assistants were restricted to one local government area within a senatorial district for their services, meaning that, a field assistant who had participated in administering the questionnaire in one local government area was then excluded from assisting during the interview and focus group discussions in the same local government area, though he could still function in that capacity in another local government area. The researcher therefore used twelve field research assistants who are university graduates with either a first or second degree1 for the survey questionnaire, while six field research assistants were employed for the interviews and focus group discussions. Edgcomb and Garber (1998) made use of a highly qualified post graduate interpreter2. Prior to the collection of data, the field assistants were trained by the researcher as suggested by Hulme (2000) and Imp-act (2005) and they were required to administer the instruments to the researcher and other field assistants on a one-on-one basis. The result of this exercise led to a further briefing and guidance on areas that required additional understanding and input from them. This training eventually proved useful during the data gathering process as it helped in the proper administration of the instruments and documentation of responses.

1

Appendix 2 page 1 of Edgcomb and Garber (1998) shows that 7 out of the 11 personnel used for microfinance impact assessment in Honduras holds a second degree. They were recommended by a colleague based on their previous experience in similar field work.

2

Page | 107 The researcher was introduced to the cooperative members, in most cases during their meetings by the executives, because “motivation can be enhanced by having interviewers introduced by program officers” (Hulme, 2000: 90). This allows for acceptance which enables the researcher to introduce himself, his mission, the data to be collected, its purpose and its usage. In a few communities, the researcher was first introduced to the village head and thereafter sought and received the permission of the village head in compliance with their culture and as a seal of authority to conduct the study because such villages had never had visits from researchers in the past. To avoid bias in completing the questionnaires, officials of the cooperative societies were excluded from working as field assistants to the researcher because “using program staff introduces the risk of biased responses and of compromising the validity and reliability of the data” (Barnes and Sebstad, 2000: 44 and 45)1. Moreover, “their links with the institution may influence the way respondents answer the questions” (Pawlak and Szubert, 2004: 6). The officials were neither allowed to respond to the survey questionnaire nor participate in the interviews and focus group discussions. These precautions were undertaken to avoid bias and to enhance the validity and reliability of the study by reducing subjectivity and third party interference.

The researcher encounters some challenges during the field work that are worth mentioning. The cost of data collection was higher than the budget because of shortage in the supply of fuel which led to increase in the prices of petroleum products. The fuel scarcity also led to changes in data collection period in few locations. More days were therefore used for data collection. The response rate in some locations was slow in terms of availability of the respondents and response to the research questions. This also increased the number of days for data collection. In few cases, the participants became emotionally aggrieved of the perceived neglect of their community by the government and its agencies.

1

In a 1993 survey of microfinance clients in two programs in Egypt, half of the enumerators were program staff and half of the enumerators were not. In analyzing the results, the research team established through statistical analysis that there was a clear bias in the responses given to the program staff (Barnes and Sebstad, 2000: 45).

Page | 108 Such people were allowed to decide if they wanted to continue or withdraw their participation. Most of the participants affected decided to continue with the survey based on their further conviction that the researcher does not represent the government or it agencies.

The field work in rural areas of Ogun State took eight months - October 2009 to May 2010 – instead of the six months earlier projected because of the above challenges. During this period, eight personal in-depth interviews and one focus group discussion were held in each local government area. In all, forty eight interviews were conducted (table 4.3 above) and six focus group discussions were held in the six local government areas. The minimum size of each focus group was six with the largest being eight with each group comprising of both loan and no-loan members (table 4.4 above). During the field work, 223 questionnaires were administered to loan members and 79 to no-loan members. 42 people participated in the six focus group discussions comprising 12 no-loan members and 30 loan members. Out of the 48 participants that took part in the in-depth interviews, 38 were loan members while 10 were no-loan members. In all, 392 individuals from 54 cooperative societies participated in this study.