Implementing “insider” ethnography: lessons from the Public Conversations about HIV/AIDS project in rural South Africa

Abstract

We describe the conceptualization and implementation of a research methodology in which “insider” community members work with “outsider” investigators as participant observers to document everyday conversations taking place in public settings in their communities. Our study took place in a resource-poor area of rural South Africa and focused on HIV/AIDS, yet we aim here to provide a road map for those interested in implementing this approach in other contexts for various empirical ends. Because this approach is unusual, we highlight considerations in selecting a team of ethnographers, describe the training process, and offer ways to ensure the data collected are trustworthy and confidential. We describe the advantages and limitations of utilizing “insider ethnography” in contexts where being indigenous to the study site provides access to perspectives that cannot be obtained through other methods. Finally, we examine how mutuality and the positionality of the research team affect data collection and quality.

This paper aims to serve as a road map for scholars, researchers, and practitioners interested in collecting contextualized ethnographic data in foreign communities by working with “insider” ethnographers, individuals indigenous to the study site of interest and thus immersed in the workings of everyday life. We begin by defining what we mean by “insider” in light of the extant theoretical and emic understandings of the concept and then describe the conceptualization of our research method, “insider ethnography” (IE). Next, we highlight our methodological approach’s similarities to and differences from other forms of ethnography. We then introduce our study, “Public Conversations about HIV/AIDS”, and describe our recruitment of a team of insider ethnographers to document conversations about AIDS taking place in public settings in their communities. This is followed by an in-depth description of the training process, including how we adapted the theoretical underpinnings of ethnography into practical use, as well as the processes we implemented to ensure that fieldwork produces trustworthy data where complete confidentiality is observed. We conclude with a discussion of the strengths and limitations of our approach as well as how to adapt IE to other research situations where having an insider status can provide unique access to local social networks and the conversations taking place within them. In this study, we focus on a rural, resource-poor area in South Africa where HIV prevalence is high.

‘Insider’ and ‘Insider Ethnography’ (IE) Defined

It is important to first address what we mean by “insider.” The notion of “insider”, as distinguished from “outsider” or “stranger”, has been challenged by feminist and postmodernist theorists in recent decades on the grounds that these binary distinctions are over-simplified and presume that social positions are static (e.g., Griffith, 1998; Naples, 1996; Obasi, 2012; Turgo, 2012). We do not wish to reify uncritically these distinctions in our use of the term “insider”. We recognize, following Merton (1972), that insider status is variable, context-dependent, and shaped by one’s own status within the community. For our purposes here we conceptualize “insiders” as those who are indigenous to the communities of interest and “outsiders” as those who are foreign to them. Variations of this distinction have roots in classical sociological theory (see for example Simmel, 1950) as well as in emic understandings of outsiders in our study site, a former apartheid homeland area in South Africa where mulungus (a local reference for “white people”) are viewed as foreign to the community (also see Weinreb, 2006). To attend practically to theoretical considerations about the variability of insider status, we selected a diverse group of insider ethnographers representing a variety of standpoints (see Fawcett and Hearn, 2004) for our research team, a discussion we pursue further in the sections below.

Our project – “Public Conversations about HIV/AIDS” or simply, “Conversations” – substitutes “insiders” (locals) for “outsider” (foreign) researchers on the premise that insiders have unique, intimate, and regular access to everyday conversations that those who are foreign to the research setting do not, and could not, unless they were prepared to spend years learning the language of the community and living within it, as classical ethnographers do. Even so, insiders, as native speakers of the local language and by virtue of their embeddedness in daily life in the community, have privileged access to the social processes and interactions taking place within it. This is not to suggest, however, that there is no bias from insiders. What insiders observe depends on their social networks and daily routines, which will differ based on their “status set” (Merton, 1972), including attributes such as age, gender, and social and relationship status. We also do not presuppose that what people say to one another in social situations is unmediated or unfiltered simply because the researchers are insiders. Indeed how people behave in social settings and appear to others is influenced by their social positions vis-à-vis others (Goffman, 1959). Yet, outsider researchers are also subject to these concerns, and often face additional obstacles such as having to rely on sabbaticals and grants to conduct research. Working with a team of insider ethnographers can thus facilitate access to research populations, increase efficiency in the data collection process, and enhance insights (Schatz et al., 2014). For these reasons, we believe that “insider ethnography” serves as an important addition to the current arsenal of methodological strategies and can provide added value for studies where having an insider status may be of particular benefit.

Historical Antecedents

Our methodological approach is not without historical precedent. The “Mass Observation” studies of mid-century working class life in England utilized a volunteer team of observers and writers to record their day-to-day lives in diaries to learn more about the lives of ordinary Britons during wartime (The Mass-Observation Archive and the Centre for Continuing Education, 1991). Another example is the work of the Rhodes Livingstone Institute (the first social science research facility in Africa) and the Manchester School, which, during the colonial period of the mid-twentieth century, had teams of anthropologists and their African research assistants collect field observations of conversations people had during the day and their stories of the past (see Schumaker, 2001). In more recent years, researchers have employed “indigenous fieldworkers”, defined as current and former commercial sex workers (Power, 1994) and injecting drug users (Elliott et al., 2002), to study covert communities; others have worked with “peer ethnographers” to study young peoples’ sexual and reproductive behavior (Price and Hawkins, 2002) and substance use among gay and bisexual men of color (Mutchler et al., 2013); and one study included adolescent women as “research partners” to learn more about young people’s school participation (Akerstrom and Brunnberg, 2013).

We are also not the first researchers to work with local ethnographers for AIDS research in a rural African setting. In 1998, Susan Cotts Watkins and colleagues asked several village residents in rural Malawi to go about their daily routines, make mental notes of conversations they overheard about AIDS, and write their recollections in a journal (Watkins and Swidler, 2009). Close to fifteen years of data from the Malawi Journals Project (www.investinknowledge.org) have produced a unique longitudinal archive, and multiple studies of various social dimensions of the AIDS epidemic (Kaler et al., 2014), including the formulation of innovative strategies of prevention (Watkins, 2004), resistance to condom use (Kaler, 2004; Tavory and Swidler, 2009), the social context of transactional sex (Swidler and Watkins, 2007), local attitudes towards HIV testing (Angotti, 2010, 2012; Angotti and Kaler, 2013; Angotti et al., 2011; Kaler and Watkins, 2010) and anti-retroviral treatment (Conroy et al., 2013), and emic interpretations of death and disease (Ashforth and Watkins, 2013).

Similar to the Malawi Journals Project, our study aims to understand what Hammel (1990) has called the “clouds of commentary” that accompany social change, in contrast to examining the underlying structures, trends, and norms of communities or larger societies, a strategy that characterizes the other group ethnography projects cited above. These “commentaries” include those emanating from the government, religious organizations, or local tribal authorities as well as informal social networks, and they circulate in various spaces, such as on the radio, in funeral sermons, or at community village meetings. Everyday conversations among those living amidst a severe AIDS epidemic thus provide a distinctive perspective on intelletual, emotional, and social responses to the epidemic: they tell us what people say to each other rather than what they say to interviewers or in focus group discussions, the two primary conventional sources of local understandings of the AIDS epidemic.

Background to Conversations Project

The Conversations project took place in the Agincourt Health and Socio-Demographic Surveillance System (“Agincourt HDSS”; “Agincourt”) in the Bushbuckridge sub-district of Mpumalanga Province, a remote, former apartheid-era homeland in rural South Africa, a setting in which the project’s investigators have been working for several years. This area has been studied extensively using several standard methods of research, including: annual demographic and health censuses, periodic surveys addressing topics such as child health and food security, and qualitative studies using focus group discussions and interviews (see Kahn et al., 2012). More recently, a community-based HIV prevalence and sexual behavior survey was conducted in Agincourt to characterize the HIV epidemic, estimating HIV prevalence at 20% and as high as 50% among those in their reproductive years (Gómez-Olivé et al., 2013).

Yet these data provide only a partial understanding of the AIDS epidemic in rural South Africa. For example, while the sexual behavior survey data can tell us about the types of people who report using condoms (e.g., teens, older adults) and which relationships appear to be more conducive to regular condom use (e.g., casual, marital), they cannot tell us about the multiple meanings associated with condom use that are constructed and maintained through everyday conversations among friends chatting casually on a bus or at a bar (see for examples Kaler, 2004; Tavory and Swidler, 2009). The Conversations project was thus envisaged to complement and build upon the HIV prevalence and sexual behavior survey by gaining insiders’ perspectives on the AIDS epidemic in their communities. Indeed insider ethnographers, who are engaged in and surrounded by conversations about AIDS throughout the course of their daily lives, provide invaluable perspectives that can give greater texture and meaning to survey results, and even more, offer rich description about and multiple understandings of AIDS circulating in everyday life (Ashforth and Watkins, 2013).

Recruiting a Conversations field team

We recruited a team of community members to work as ethnographers and keep field notes of what people in their communities were saying to one another about the AIDS epidemic. We focused on conversations occurring in public settings that we knew were frequented by locals, such as shared water taps, health clinics, churches and taverns; and community events, like village meetings, pension days¹, and funerals. We recruited our team from among a pool of individuals already working in various research capacities in the Agincourt study site, and who consequently, had previous training in research ethics, had completed secondary school, and were proficient in English (the language of the foreign investigators). Working for the Conversations project was thus a second job – and supplemental income – for most of the members of our team.

Since we both have lived and worked in the Agincourt site, we knew personally, and have worked previously, with several individuals we thought would be suitable candidates for the Conversations field team. Our priorities were to have candidates who performed well in other jobs with the Agincourt research unit, who were particularly sociable, and as a team, were diverse in their demographic characteristics (age, gender) and social interests. We suggested a list of candidates to Agincourt management, and Agincourt management recommended additional people based on these criteria as well as availability. The Conversations field team was initially comprised of five men and five women ranging in ages from early twenties to early forties. We started with a team of 10 because we expected some attrition over the course of the project, as some individuals were offered more attractive and/or remunerative opportunities or were otherwise too busy with other jobs or responsibilities. Our final team included five women and two men of varying ages and with diverse sets of interests: some were heavily involved in their churches, others liked to spend their free time hanging out with friends at music festivals and bars, some were devoted sports fans, and others spent much of their time taking care of loved ones.

One important difference between the ethnographers and the larger community is that the ethnographers were all employed through the Agincourt HDSS, whereas unemployment in the population at large is a significant problem: the unemployment rate in Agincourt in 2008 was around 25% for men and 48% for women (Blalock, 2014). Despite this difference in social location, the ethnographers endure many of the same hardships, daily tasks, and pleasures as their local relatives, friends, and neighbors. Both groups frequently collect water for their homes; attend local events, such as church, village meetings, and funerals; and make use of public services, such as health clinics and local transportation. In many important ways, then, the ethnographers’ daily lives are similar to the lives of those around them.

Teaching the “tricks of the trade”

Our primary focus in training was on developing practical and experiential knowledge of the method – the "tricks of the trade" in Becker's (1998) terminology – rather than emphasizing the theoretical underpinnings of ethnography. Though formal training only lasted five days, our on-going engagement with the team once fieldwork commenced continued throughout the course of the project, as described further below. Formal training was organized around several themes, which we address in turn below: (1) introduction to ethnography; (2) methods for capturing a conversation; (3) practice and feedback; (4) introduction to the project’s subject of interest; and (5) ethics in research.

Introduction to ethnography

Our training began with a brief introduction to ethnography, emphasizing the similarities and especially the differences between this method and others with which the team was already familiar: surveys, interviews, and focus group discussions. We described ethnography as more “natural” than other research methods insofar as it captures what people say as they say it in organic conversations, rather than through responding to a pre-determined set of questions and, in the case of surveys, responses. We emphasized that ethnography encompasses observation and close listening, unsolicited information, and abundant detail, including who was present, where and when the conversation occurred, what stimulated the conversation, and what exactly was said. We encouraged the ethnographers to imagine that they were describing a recent event to a close friend. When recounting the story, what would they say? And if they were the one listening to the story, what sorts of information and details would they want to know?

We emphasized several qualities inherent in a “good” ethnographer, including being eager to interact with others, a good listener, and attentive to detail. We asked the ethnographers not to summarize a conversation but rather to think about themselves as having a video camera on their heads recording exactly what they saw and heard. We outlined the types of details they should include in their field notes, including: gender; age; number of people present; what people were wearing (expensive or shabby clothes, a blue jacket, a red hat) and doing (eating sugar cane, waiting for a bus); the time of day; the mood of the setting (joyous, somber); the event (pension day, on a minibus); and most importantly, what people were saying, as expressed in words and nonverbal communication.

During the final part of the first day of formal training, we discussed what an ethnographer should not do, notably: interview, conduct a focus group, steer the conversation towards the topics of interest, use recording devices, or take notes in public. We also did not want them to change their daily routines, since the project was interested in the ordinary rather than the unusual. Finally, we emphasized how publicly displaying their identities as researchers would temper the advantages of the method because people would likely censor themselves, as evaluations of survey data quality have shown (Miller et al., 2001).

Capturing a conversation: “JWR”

We established a three-step process for writing field notes, which we termed “JWR”: Jot, Write, Review. This provided the team with a concrete template for collecting the data and highlighted the systematic, accurate, and detailed nature of taking good field notes.

The first step is “Jotting”, writing down key words (or triggers) directly after observing a conversation. Jots, written while the ethnographer is still in the field (in pocket-sized notebooks we provided to them), are basic descriptions that can be used to later remember the details of a conversation. For example, the jottings from a conversation among a group of young women discussing the infidelity of a boyfriend might be: “cheating”, “playboy”, “sms” (text message), and “makhwapeni” (secret lover, literally “under the armpit”). We modeled how jottings should be used as personal triggers by having the team observe the same setting for 10 minutes and then return and share their (different) jottings with the entire group.

The second step in the process is “Writing”. Writing should occur within 24 hours of a conversation so that details are not forgotten. During the writing process, ethnographers should expand on their jottings and record in detail what they actually observed and heard. We emphasized the importance of properly ordering the events observed and including everything that was said during the course of a conversation, even the parts that might seem trivial at the time. Conversational topics do not arise from nowhere, and may be triggered by something someone observed or by small talk, information that could prove important during data analysis. We also stressed the value of including information about nonverbal communication: if someone says “You should use a condom” and then laughs, the statement has a different meaning than if the injunction is said in a serious tone of voice. We thus asked the team to try to appreciate the meaning of the scene by imagining what behaviors, actions, and facial expressions might mean to others.

The final step is “Reviewing” field notes before submitting them. In reviewing, we encouraged the team to read through their field notes, expand on any details that might be missing or unclear, and write their interpretation of what transpired. We encouraged them to include local sayings in Shangaan (local language) if they were difficult to translate into English, but to follow this with a translation or explanation that would clarify the meaning. Finally, we asked the team to use brackets to set apart their own interpretations or explanations to distinguish them from what they actually observed. Thus, in many regards, insider ethnographers – who are familiar with local symbols and idioms – are also key informants and cultural interpreters.

Practice and feedback

The bulk of the training was experiential, providing the team with ample opportunities to practice before they went into the field. After the first training session, for example, we gave the ethnographers an assignment: to return the next day with a written description about one conversation that heard that evening following the training session. The next morning, they read their field notes aloud to the group so that they could provide one another with feedback and constructive critique, applying what had been learned the previous day. This active sharing process allowed the ethnographers to improve their field note techniques from the get-go.

On the second day of training we provided the team with two examples of field notes from the Malawi Journals Project. One excerpt exemplified several useful techniques, including how to capture the flow of a conversation, use direct quotations in field notes, and identify individuals (e.g., “the old man”) while maintaining confidentiality. The other excerpt illustrated how to write about words or phrases that were spoken during a conversation but were not easily translatable (in the Malawi case, from Chichewa to English). We also took turns reading a long field note excerpt aloud, which the ethnographers compared to the field notes they had written the night before.

On the third training day the team dispersed to nearby public places for observation. They observed for 30 minutes and then returned to the training room to write jottings, followed by a full field note. After reading them aloud, the team provided one another with feedback on both the jottings and completed field notes, a process that enabled the ethnographers to play an active role in the training process and to work on improving their field notes as the training progressed.

Introduction to the subject of interest: HIV/AIDS

We introduced the project’s substantive focus on the afternoon of day three: conversations about HIV/AIDS. We kept the definition of a conversation about HIV/AIDS wide open by including both direct and indirect references to AIDS. A direct reference would be someone actually saying “HIV” or “AIDS”, while an indirect reference was something that might be more broadly associated with AIDS, such as a story of someone who suffered from the “disease of nowadays” (“mavabyi ya masiku lawa” in Shangaan) or someone who was “swallowing pills” (referring to HIV treatment).

We asked the team to construct a list of the ways people talk about HIV/AIDS. As a group they identified 10, including sign language implying HIV by placing “three fingers” over one’s head, as well as euphemisms in the local vernacular, such as “xinghunghumana” (something frightening but invisible, like the “boogie man”). This process gave us insight into the universe of colloquial words and phrases to which we may not have otherwise been privy, as well as assured the ethnographers of the abundant content to which they would be exposed during fieldwork, assuaging concerns they had voiced that there would be little to write about if a conversation had to directly invoke “AIDS”.

Ethics in research

We obtained ethical approval for Conversations from Institutional Review Boards in the United States and South Africa as well as from the Department of Health in the study site. We obtained community consent for conducting the research in each village from its community leaders – the Induna (village chief) and the community development forum (village-based political representatives). With institutional approval we were able to forgo gaining formal individual consent because we were focusing on conversations occurring in public settings (and thus available for others to hear) and the data would be fully anonymized and de-identified. Gaining formal consent from the individuals participating in the conversations might actually have done more harm than good in that it would have provided a link to participants’ identities. It would also have potentially undermined the value of the method because individuals may have altered their conversations in the presence of a known researcher (see for example Spano, 2006).

We took steps to ensure the confidentiality and anonymity of both the ethnographers and the individuals participating in the conversations throughout the processes of data collection, data cleaning, and dissemination (Kaiser, 2009). During data collection the ethnographers anonymized their field notes. We offered suggestions for pseudonyms or markers for individuals participating in conversations, such as “an older man” or “the woman wearing the red dress”, and the ethnographers offered their own, such as “Buti [Brother] M” or “Gogo [Grandma] A”, reflecting the familial nature with which people commonly refer to one another in the setting. We assigned a pseudonym to each ethnographer, which they wrote on their field notes when they submitted them. We acknowledge that we cannot know for certain that the ethnographers kept their field notes and pseudonyms confidential, but this was emphasized in the training and we have no reason to believe otherwise. In data cleaning and dissemination, we assigned the ethnographers a second set of pseudonyms to further mask their identities. We also changed unessential details of the conversations in the field notes, such as the date on which a funeral occurred or the name of a particular village bar, to make the conversations unidentifiable to others (Weiss, 1994, as cited in Kaiser, 2009). Despite our concerted efforts to maintain the anonymity of the ethnographers during fieldwork, we suspect that others – such as spouses – may have known that an ethnographer was working on a research project even if they were unaware of the specifics of the project (and it was certainly well-known to those who knew the ethnographers that they were employed with the Agincourt HDSS in a research capacity). We have no way of ascertaining how this might have affected the data our team had access to, and whether or not some people altered their speech while in the presence of the ethnographer(s). To the best of our knowledge, however, no breaches of confidentiality or anonymity occurred, nor were any issues of community conflict or distrust raised.

Ensuring data quality

It is important to strike a balance between encouraging productivity and ensuring that ethnographers do not feel undue pressure to meet an unreasonable field note page limit. The ethnographers were paid for each complete set of field notes that they submitted and were hired on a flexible month-to-month basis, which allowed them to take time off during fieldwork if need be. We paid the ethnographers ZAR 500 (about $60 USD) a month for 30 pages of handwritten field notes, an amount we determined in concert with Agincourt management to be consistent with the site’s salary pay scale. Although the team expressed doubt that they would be able to meet the page requirement, it was clear after the first month of fieldwork that they were not short of notes. As Audrey, one of the ethnographers remarked to us, “People are talking [about AIDS] … If you are with people, you will hear them talk.” Nonetheless, some ethnographers complained about how time consuming it was to write detailed field notes at the end of a long work day and vis-à-vis other domestic responsibilities. We thus settled on a range of 23–27 pages of field notes per month to account for this concern.

(Informal) training continued after the team began submitting their monthly field notes. We met with the ethnographers each month to review their notes, having scheduled these dates in advance on the last day of training. Prior to the meetings, we read through their field notes, making note of what needed clarification, translation, and/or elaboration. Meetings lasted between 30–60 minutes and were structured around offering praise for strong field notes, providing constructive feedback for improving observational and writing techniques, soliciting additional details about the contexts in which conversations occurred, and if time allowed, deeper engagement with the ethnographer about the content of his/her notes. These meetings also provided valuable opportunities to verify our preliminary interpretations of the data and ask for clarification about complicated issues. Thus, oftentimes the ethnographers served as both key informants and cultural interpreters by adding local insights and embedded knowledge about their field notes to the investigators’ understanding of them. For example, meetings with one of our most prolific ethnographers, Audrey, often included ancillary discussions about the topics raised in her field notes: a field note capturing a group of women gossiping at church about the sexual transgressions of their local pastor might segue into other discussions about religion and AIDS in her community, such as local people sojourning to faraway cities to meet with pastors infamous for promising a cure for AIDS in exchange for a fee.

Because neither of the investigators was in the research site full-time once the Conversations project began, we created a “Conversations Coordinator” position. The position was filled by an ethnographer who worked full-time with the Agincourt HDSS and was available at the field office each month to collect field notes. The coordinator was responsible for collecting and accounting for the ethnographers’ field notes, informing us of any issues that arose and, when necessary, communicating any important messages between the team and us. We provided her with sealable manila folders in which to place the completed sets of field notes after they had been turned in and signed by both the ethnographer and the coordinator. The coordinator was also responsible for making sure that her supervisor from the Agincourt HDSS – who visits the field office regularly – transported the field notes to the administrative office (about an hour’s drive away) for scanning and archiving. We followed this process closely by communicating with the coordinator and administrative assistant each month when field notes were submitted (via email, phone, or in person when possible).

To ensure the quality of the scanned field notes, the ethnographers wrote their pseudonym on the top of each page and this identifier was used in the file name of the scanned copy. At the beginning of each unique note or entry, the ethnographers noted the setting where the conversation occurred and the date and time of the observation, as an ethnographer may have heard several conversations in a day. They also differentiated each new observation by skipping a line between observations (but not starting a new page). They included page numbers at the top of each handwritten page of notes and used one-inch margins from all edges of the paper. These small details ensured that the handwritten transcripts scanned successfully the first time, and were essential in producing a set of usable scanned transcripts true to the handwritten version. Once scanned, the field notes were uploaded to a secure digital folder accessible only to the project investigators and later transcribed.

Discussion

Insider ethnography (IE), as defined and described here, is a research method that provides access and insights into everyday forms of expression as they are manifested in daily life, observations that could not be captured by the more standard methods of social research (such as surveys and interviews) and that would not be possible for researchers foreign to the research setting to collect on their own (Schatz et al., 2014; also see Watkins, 2004). In this paper we focused specifically on how we implemented IE in an effort to learn more about an issue of pressing social concern: HIV/AIDS in a high prevalence, rural African setting. Below we consider the strengths and limitations of our approach to recruiting and training a team of insider ethnographers, and to the IE method more generally, and in so doing offer suggestions for how IE might be usefully extended to other settings for other empirical ends.

Recruiting a research team

Recruiting our team of insider ethnographers from the staff of an on-going research site (the Agincourt Health and Socio-Demographic Surveillance System in South Africa) had both scientific and logistical advantages. It allowed us to select individuals who were comfortable communicating and writing in English, who were familiar with ethics of research, and who could be expected to be reliable, discreet, and conscientious. Recruiting a field team from the Agincourt HDSS also abated many of the challenges associated with access (see for example Gokah, 2006): the insider ethnographers lived in the area of interest and did not change their daily routines to conduct their work, effectively providing the project seamless entrée to the community. We were also able to work with the Agincourt HDSS administrative unit for logistical tasks like arranging space for training and meetings and processing the ethnographers’ payment, which was essential to getting the project off the ground in a matter of weeks rather than months. Being employed with the Agincourt HDSS, however, also meant that the ethnographers had less time to devote to Conversations. This likely contributed to some of the attrition we experienced throughout the course of fieldwork. Researchers employing insider ethnographers who have full-time work elsewhere may need to explore incentives or different ways to motivate them to continue with the project over time.

We were able to leverage our affiliation with the Agincourt HDSS as an intermediary to help recruit ethnographers. For other researchers interested in utilizing insider ethnography, other sorts of intermediaries – such as a school administrator, local pastor, or nongovernmental organization (NGO) – might be helpful in identifying interested, particularly talented, and trustworthy individuals to hire for a local research team. The most suitable intermediary, of course, will depend on a researcher’s analytic goals: school teachers might be best able to identify competent adolescents for such a task (if one’s interest, for example, is in the effects of technology on young peoples’ lives). A local NGO may be best able to identify a target population with specific knowledge or experience from which insiders may be drawn, such as an organization that works with those receiving home-based care or those using particular medications for a chronic disease (if one’s interest, for example, is in the experience of aging).

Training techniques

We used excerpts from the Malawi Journals Project to illustrate how to recount who said what to whom as well as how to apply some of the more technical lessons from the training to a set of field notes. Examples help make the art of writing ethnographic field notes more concrete. Other researchers might opt for different types of examples, such as those that help their team see how to accurately report the colloquial speech they hear. In South Africa and elsewhere in Africa, for example, those who are familiar with AIDS research or with NGOs working in AIDS-related advocacy may be more likely to use terms propagated by the global AIDS community, such as “HIV status” (referring to one’s diagnosis) or “ART” (referring to treatment), rather than the more commonly used local ways of referring to the epidemic. Field note examples also help address the question of “how much” an ethnographer should write. If the circumstances in which discussions arise are important, researchers might encourage their team to include as much detail as possible about the context – such as what people were saying before and after the focal discussion arose, facial expressions and hand gestures used, and so forth – rather than settling on shorter entries that target only the topic of interest. We also offer the recommendation of one of our ethnographers, Zanele, who told us that it was only after reading a popular novel that the Conversations training really “clicked”. Zanele’s remark suggests that utilizing excerpts of good novels might be another approach to teaching ethnographic writing to those not academically trained in the logics of qualitative research.

In addition to providing formal training, we also remained actively engaged with the ethnographers throughout the course of fieldwork to help them further develop their observational and field writing techniques and to vet our understandings of the data. This face-time was critical to the production of high-quality data. Importantly, it also helped mitigate the sense of isolation that this type of data collection can lead to, as IE is much less structured than other forms of qualitative research. While we met individually with the members of our team at the end of each month when field notes were due (a routine we chose because neither of us could be in the study site full-time during the fieldwork), other studies using insider ethnographers may opt for more frequent meetings. Mutchler and colleagues (2013), for example, met with their peer ethnographers each week, which was possible because they were in the same city. Meeting with the team as a group in addition to individually might be another strategy to encourage consistent participation in the project and reduce isolation during fieldwork. Whatever the routine may be, we believe IE requires consistent engagement with the field team. Indeed, in other projects in Southern Africa that have used this type of ethnography but where consistent contact was not possible, data quality suffered (Personal Communication with Susan Watkins).

Methodological Considerations

Insider ethnography, like any method of research, also raises important methodological considerations, ones that are both specific to the Conversations project as well as to social science research more generally.

First, there was evidence that the ethnographers sought out places where they expected to hear relevant conversations about AIDS, which we had originally discouraged because of our primary interest in organic conversations taking place in a variety of settings. While our expansive collection of field notes of everyday talk about AIDS includes conversations from various settings – from “boys talk” at a local car wash to gossip among neighbors about the recent death of a mother to AIDS – we also read a large number of conversations from health clinics, obvious venues in which conversations about AIDS are likely to occur. Once it became apparent that this was happening, we opted not to discourage this strategy on the grounds that data from local health centers add empirical value: they provide information on how clinic patrons respond to, modify, and preserve “official health talk” in everyday conversations occurring in clinic waiting areas. Additionally, several of our ethnographers also encountered clinics during the course of their everyday lives, whether they were escorting ill family members there or attending them on their own accord. This indicates what can be learned through insiders’ adaptations to data collection in practice when researchers are able to take a more flexible approach to the implementation of a research method (also see Rogers-Dillon, 2005).

Next, our implementation of IE depended on there being people living interspersed among the population of interest who were proficient in the language of the foreign researcher(s). This may not apply for other research projects, which may thus require the use of an interpreter or additional translation – during training, while meeting with ethnographers, and/or of the ethnographic texts themselves. While this is a possible strategy, it would add a new layer of complexity to the process and may introduce error (see Temple, 2002; Temple and Young, 2004), as the data would have to be translated and transcribed by a third party as opposed to by those collecting the data themselves.

Finally, the social positionality of the foreign investigator(s) and the insider ethnographers, vis-à-vis each other and the research subjects, can produce both advantages and disadvantages in IE research. An advantage, as we have described, is the access allowed via working with locals embedded in the daily life of their communities (Schatz et al., 2014). In rural South Africa, it would not have been possible for us, as the project investigators, to have access to the same types of conversations that the local ethnographers did because of our “foreignness”, notably our race (white), nationality (American), native language (English) and social class. At the same time, we cannot know if the conversations the ethnographers captured in their field notes were filtered by what they believed we wanted to know or hear. Additionally, even though our research team was made up of a diverse group of insiders, we cannot assume that they – or the conversations they hear or participate in – “represent” the entire community (Temple, 2002). In other words, insider ethnographers only have access to the conversations around them by virtue of their daily routines (e.g., whether they walk or take the bus) and the impact of their social attributes (e.g., whether one is an older woman or a younger man) on what is said around them. We know from the Conversations data, for example, that what women say to other women about the source of blame for HIV in their communities is different (in content and tone) from what women say when speaking in the presence of men or from what men say to one another (Sennott et al., 2014). We encourage researchers interested in insider ethnography to consider how the positionality (of both themselves and their research team) will influence the research process, data quality, and team dynamics, as well as how ethnographers’ social locations will affect the conversations to which they are and are not privy.

We believe the success of the IE research method – as we have defined and implemented it – hinges on mutuality between foreign investigators and insider ethnographers: both contribute to the research in specific ways, and each is vital to the research. In our experience, success also depends on the foreign investigators’ prior intimate knowledge of and engagement with the research setting. This familiarity makes it possible to understand idiosyncrasies and colloquial ways of speaking the language of the outsider (in our case, English) as recorded in field notes, as well as evaluate the plausibility of the field notes, and interpret the resultant data. We also believe that working with a team of insider ethnographers to collect contextualized ethnographic data in communities foreign to the investigators is mutually advantageous. As noted, it provides “outsider” researchers with opportunities to garner rich data from a wide variety of social settings, locations, and conversations compared to what they would be able to gather on their own. It also provides “insiders” with advanced research training, opening up prospects for employment with future projects that utilize qualitative research (also see Rogers-Dillon, 2005), which in demographic surveillance sites globally are increasing in number (e.g., Madhavan and Gross, 2013; Schatz and Ogunmefun, 2007).

We encourage researchers whose social locations differ in fundamental ways from the population of interest to consider working with insider ethnographers to collect contextualized ethnographic data, and, as relevant to the project itself, with a team who possess diverse demographic characteristics and social interests to access an array of local social networks and the conversations occurring within them. Indeed, we believe IE is an appropriate methodology for examining a range of topics (e.g., generational attitudes, political change) across a diversity of settings (e.g., workplaces, churches) to explore how groups attribute varying meaning to particular social phenomena (Watkins and Swidler, 2009). While useful on its own, IE also provides complementary knowledge to enhance research findings from other methods of social research. We hope this paper provides a useful framework for thinking about how to implement insider ethnography for a variety of empirical aims.