The Rewards and Risks of Rapid Ethnography: Researching How Microfinance Entrepreneurs in Sri Lanka Make Business Decisions

Project Overview and Context

This project, which led to my PhD, examined how microfinance borrowers in a developing country make business decisions and acquire entrepreneurial expertise. Entrepreneurs who had borrowed from a microfinance institution (MFI) in Sri Lanka completed a survey, were interviewed individually and in groups, completed diaries about their business activities, and were observed working in their businesses and during meetings about loan repayments, all within 2 months.

The types of techniques and the short time span are characteristic of rapid ethnography (RE), “a form of multi-method ethnography involving data collection from numerous sources over a relatively short period of time including interviews, participant observations, document review and sometimes surveys and focus groups” (Baines & Cunningham, 2013, p. 74). Although the fundamental principles of RE are similar to those of conventional ethnography, a number of its elements, for example, planning and preparation, and socialization and rapport-building with participants, are done before researchers go into the field (Baines & Cunningham, 2013; Handwerker, 2001; Millen, 2000). When planning the research, rapid ethnographers collect background data and secondary information to narrow down the research questions (Handwerker, 2001). This information is also used to determine the time period to visit the field so that the researchers can observe relevant activities. In addition, rapid ethnographers use technology-assisted communication such as regular emails, phone calls, and instant messages to build rapport before going into the field (Baines & Cunningham, 2013). Although the actual time spent in the field is often less than 3 months, researchers typically use time-deepening strategies such as multiple and parallel data collection methods and multiple observers (Baines & Cunningham, 2013; Handwerker, 2001; Millen, 2000) to compensate for the limited time. Thus, RE is not an inferior way of collecting data, but an alternative way of studying the research issues within a natural context (Pink & Morgan, 2013).

I became interested in studying how microfinance borrowers make business decisions when I was employed in an MFI in Sri Lanka. Sri Lanka is a lower middle-income developing country in South Asia where around 77% of people live in rural areas, and where very small service-based, manufacturing and agriculture/animal husbandry businesses are common. These businesses are frequently financed by very small loans from a range of financial service providers, including MFIs, which cater to people, particularly women, who often cannot get finance any other way. MFI borrowers form a small, “solidarity” group of peers whose members are responsible for each other’s loans. Microfinance is not a “magic bullet” to reduce poverty or create fast-growing businesses (Augsburg, De Haas, Harmgart, & Meghir, 2015; Banerjee & Duflo, 2011; Bateman, 2010; Karlan & Zinman, 2011). Indeed, solidarity groups can discourage individual borrowers from initiating risky business ventures (Armendáriz & Morduch, 2010). Peer pressure used to ensure loan repayment can also hinder business growth, and some group repayment tactics have been described as abusive and ethically unacceptable (Ali, 2014; Ghatak & Guinnane, 1999). Similarly, because their main concern is loan repayment, MFIs are sometimes said to encourage entrepreneurs to “play safe” (Banerjee & Duflo, 2011). Nevertheless, many people take out microfinance loans and create sustainable businesses. I wanted to understand how these people make decisions at business start-up and later, including the effect of a microfinance loan on their decisions.

Research Design

It made sense to take advantage of my knowledge of Sri Lanka and my contacts with MFIs and through them, microfinance borrowers, when designing the research. An early decision was to use a survey to ask microfinance borrowers about their business thinking at the business start-up and development stages of their organizations, to see whether and how their thinking changed over time, the extent to which they practiced specific business skills, and so on. The theories used were causation and effectuation in entrepreneurship (Sarasvathy, 2001) and deliberate practice (Ericsson, 2006). Then, I planned to use additional, qualitative data gathered from the following:

• Interviews with and observation of MFI staff (to understand the MFI’s processes);
• Analysis of MFI documents;
• On-site interviews with business owners;
• Observations of business owners at solidarity meetings and elsewhere;
• Diaries kept by business owners;
• My field notes.

Research Practicalities

My PhD needed to be completed within a limited time: 3 to 4 years. This made RE an attractive technique because it would allow me to gather good data fairly quickly. A number of factors in the research situation made RE feasible for me, including the following.

Method in Action

After gaining ethics approval from the university, I wrote formally to the Chairperson of the MFI requesting permission to collect data. Approval was granted quickly, perhaps because the Chairperson was interested in the research and because I was a former employee. While still in Australia, I used Skype, email, and telephone calls to let my former colleagues know about my project. I introduced myself to those who had arrived since I was employed, briefly explaining my former role in the organization. These introductions and technology-assisted socialization processes helped me later when I interacted with staff members in person and collected key documents.

Field data collection in Sri Lanka took 2 months, most of which I spent in the MFI’s head office and four of its 18 branches. At head office, interacting with staff and collecting document data was relatively easy as the Chairperson had given approval and most staff already knew me. In contrast, at regional and branch level, I had to rely on other staff, some of whom I had never met and who had no particular reason to be interested in the project. I first gave a formal presentation to regional managers to gain their support. Although regional managers did not have any practical role in collecting data, approaching them first helped me to understand and work appropriately within the MFI’s reporting structures. Via the regional managers, I liaised with branch managers to plan when to visit an individual branch. I spent about a week in each branch location, explaining the research to staff and gaining their insights, and visiting clusters (a cluster comprises 10–12 solidarity groups) to meet and observe individual microfinance borrowers.

At the cluster, solidarity group, and individual borrower levels, I relied on the MFI’s field officers to help me start collecting data. In each branch, I visited selected clusters to observe their meetings and speak with borrowers. At the start of the meeting, the MFI field officers would introduce me to cluster leaders and individual borrowers. Without introductions from the field officer’s “familiar face,” it would have been impossible to gather cluster-level data because people’s activities at meetings focused on loans and other business matters. After the introduction, I observed cluster meetings, held discussions with solidarity groups, and conducted focus group discussions to gather data about the MFI–borrower interface.

Introductions from MFI field staff helped me to gain the confidence of gatekeepers such as cluster leaders, and to understand the operational side of the MFI’s activities at branch level. In addition, the field officers helped me with practicalities such as planning my travel to the next location. Nevertheless, working closely with MFI field officers meant ensuring that I was not seen as part of the MFI because, given that borrowers have obligations to the MFI, this might have made borrowers feel compelled to take part in the research. Both the field officer and I emphasized borrowers’ freedom not to take part, to the point of letting people know hownot to participate if that was their preference, and how to keep this decision private. (Borrowers could simply hand in a folded blank survey form, and not volunteer to be interviewed.) I also needed to observe the field officers themselves, and being able to speak Sinhalese helped with this. Indeed, it would have been impossible to understand the nuances of some of the MFI’s loan collection practices without knowing the language, as some officers were reluctant to explain their tactics, some of which were not listed in the MFI operational manual. A typical example was that field officers controlled the time that meetings started. Some insisted on waiting until all borrowers were present before allowing the meeting to begin, which led to those who had arrived on time expressing annoyance with latecomers. Others started the meeting but it would inconvenience latecomers by keeping them waiting outside for a time. A field note makes clear the effect this had:

Direct observations combined with knowledge of the language allowed me to interpret facial expressions and concerns people had without disturbing the cluster proceedings.

Practical Lessons Learned

Although some traditional ethnographers see RE as a “quick and dirty” method, my project demonstrates that modern technology and time-deepening strategies can be used to collect quality ethnographic data even if the researcher spends less than 3 months in the field, so long as certain factors are managed carefully. First, it is essential to communicate with and get to know the research organization or likely respondents before visiting the field. This can be done using modern technology-assisted communication tools such as email, Skype, or chat sessions. This helps organizations and respondents become familiar with the research and build trust with the researchers. However, in instances where research participants cannot be contacted using modern communication technologies, as with this study’s microfinance borrowers who lived in rural areas, a person whom research participants are familiar with and trust (the MFI field officer in this case) should introduce the researcher.

It also helps if the researchers have at least a basic idea of the selected organization’s internal functions, power relations, and reporting structures. However, this is often difficult in practice as researchers are usually external to the organization. In this study, I had prior knowledge as I had been employed at the selected MFI before beginning my PhD project. Where there is no insider knowledge, the researcher should approach the organization through formal channels. This will facilitate getting approval from the formal gatekeepers.

The success of the RE technique depends on time-deepening techniques such as the use of multiple and parallel data collection methods. In this study, I used a range of techniques: discussions, formal and informal observations, interviews, and daily activity journals and field notes to collect data simultaneously, and all were used to flesh out and interpret the results of my survey. Some rapid ethnographers recommend having multiple researchers at sites, so that the same phenomena can be observed from different perspectives. Furthermore, technology-assisted methods, such as video recordings, email, and social media postings, can be used as supplementary tools to collect data.

The success of the interviews and discussions depends on the ability of the researcher to explain their research and introduce themselves in a manner which participants can understand and which creates trust. This is specially required when conducting research in rural areas or other situations where researcher and participants may initially appear very different from each other. For example, I mentioned my background, education, and hometown details, before explaining I was a researcher in an Australian university. This allowed the participants to link me to things they already knew, and to better understand who I was and what I was doing. The ability to communicate in one of the local languages was an advantage.

Obtaining the support of local agents, such as the MFI field-level staff, even collaborating with them, helps in data collection, particularly in locating and approaching information-rich participants, arranging logistics, and understanding the “research terrain.” Although there is some potential for response bias being created when researchers are introduced by representatives of a particular organization, this can be minimized by asking probing questions and carefully observing the context.

Even in a high-speed situation like RE, researchers should take time to reflect during data collection. Taking a day or two between interviews and discussions allowed me to improve my interview and discussion technique. I had time to identify better ways of asking questions, and think about whether additional information was required to understand a certain phenomenon.

It is also recommended to start translating and transcribing interviews and discussions as soon as a reasonable body of data has been collected. These “pre-analysis” tasks will show quickly whether additional information is needed, if there are missing data, and so on. Starting the data analysis quickly also takes account of the fact that data analysis is an iterative process which evolves with time: Comparing early analyses with later ones reveals where the researcher’s thinking has changed and why—an important aspect of maintaining reflexivity. Finally, starting data analysis early gave me time later to recode a sample of interviews a year after the initial coding, which assured me that my qualitative coding was reliable.

Nevertheless, RE has its challenges. The fundamental principles of RE include (a) reducing the time spent at the field using planning, (b) using technology-assisted communication methods to begin building rapport and becoming familiar with the research site before visiting it, and (c) using multiple methods or multiple observers to compensate for limited time spent in the field and allow quality data to be collected. (I did not use multiple observers, but did use multiple methods: survey, diaries, interviews, focus groups, document analysis, field notes.) As I was already familiar with the organization and the microfinance sector, it was fairly easy to adhere to these principles. Nevertheless, it was challenging not to become complacent as result of familiarity, or to remember that changes had taken place since I worked in the case organization and to take nothing for granted. For example, as I had previously worked with some of the MFI staff, they assumed I was familiar with recent MFI developments and changes and tended to skip over information. On those occasions, I had to remind them I was not familiar with the new initiatives.

It is essential to reflect on your position as a researcher within a familiar context. Reflexivity is about explaining how interpretations and findings are influenced by the researcher’s background, experience, and position in relation to the context (Creswell, 2014). As my supervisors questioned me, clarified issues, and asked for explanations of what was not clear to them, I become aware that not all I knew—or thought I knew—was known to others. I needed to make my tacit assumptions explicit in the interests of in-depth analysis and discussion.

Conclusion

This case study explained my experience of a complex research problem encountered and solved: trying to understand how female microentrepreneurs in Sri Lanka make business decisions and gain entrepreneurial expertise. I assessed the advantages and disadvantages of RE in this process, including when my own ethnic background made me familiar with the cultural context and the organization. In addition, the case described practical strategies that can be used to collect ethnographic data during a short data collection time and showed how to collect quality data without violating ethnographic research principles. Furthermore, this case outlined the potential for RE in entrepreneurship and management research by showing how RE data could be used to explain quantitative findings in a mixed-methods study, lend support to findings that might otherwise seemed more susceptible to “same source” bias, and help develop new theoretical models.