I embarked on a doctoral study in 2006 to assess two things: first, which categories of the poor are being served by microfinance institutions: are they the very poor, middle poor, or less poor ones? Second, whether household access to microfinance reduces poverty, and if so, to what extent and across which dimensions of well-being. The study was based on direct observations and primary data gathered from semi-structured interviews of 1,132 households across 11 districts in the rural areas of the province of Punjab in Pakistan. Methodologically, the research employed a quasi-experimental research design which employed data collected by interviewing both borrower (treatment) and non-borrower (control) households. Sample selection biases were controlled by matching propensity scores.
This case discusses the limitations of using either qualitative or quantitative data alone to interpret survey responses and observations from the field. It explains how and why the mixed-methods approach offers a solution to this predicament. The case study develops further to provide a detailed account of how the mixed-methods approach was used to collect, analyze, and interpret qualitative and quantitative data. Within mixed methods, the case goes on to explain the triangulation approach to analyzing data. It also discusses, among other aspects, the significance of pilot testing a survey instrument prior to launching a full-scale survey.