Crystal Biruk on her book, Cooking Data

Interview by Sheng Long

https://www.dukeupress.edu/cooking-data

Sheng Long: In your book, you quoted a Chichewa phrase “kuphika madata” (cooking data) to analyze the global health’s quantification projects in Malawi. Can you briefly unpack what it means by cooking data? How does your ethnography of cooking data give rise to a new theorization of data?

Crystal Biruk: In the opening pages of my book, I discuss how I often heard Malawian fieldworkers use this vernacular phrase to playfully, mostly jokingly, comment on their colleagues’ job performance, invoking it to accuse them of collecting data in a sloppy or lazy way, or of outright ‘faking’ data by writing down numbers into blank spaces on the questionnaire without bothering to ask the questions of research participants. For example, fieldworkers would make this accusation if someone finished an interview more quickly than expected (which signaled, perhaps, that they skipped some questions or made up answers for expediency). In Chichewa, kuphika means to cook or cooking (as in the kitchen); a cooking pot is called mphika. The connotations of –phika signal the practices and processes of cooking food: food, not unlike data, must be prepared in a specific way and under specific conditions in order to be good or edible. We can take this analogy further by thinking of the questionnaire—with its boxes waiting to be filled in and empty household rosters yet unpopulated by names—as a kind of ‘recipe’, a means toward an effective end or desired outcome (whether a perfect key lime pie or pristine dataset!). Before collecting any data from rural households, fieldworkers attended intensive trainings, led by American or European demographers, that spanned 7-10 days and aimed to inculcate good habits (writing neatly, not missing any questions, probing for answers if respondents said, “I don’t know”, double checking the internal consistency of a completed questionnaire, etc). But most of all, fieldworkers were cautioned against “cooking data.” The Malawian fieldworkers thus appropriated this phrase from their employers (demographers), which I read as a commentary on how often and forcefully they were told “Never cook data in the field”: to accuse someone of cooking data was funny precisely because ‘cooked data’ was so feared by their bosses.

From the colonial period to the present, African fieldworkers involved in knowledge production projects ranging from censuses to global health projects have been suspected of fabricating data, and cast as a threat to data quality—these anxieties, of course, embed racist assumptions and stereotypes. Yet, this fear of human error or deliberate mucking up of data in the field relies on a taken-for-granted fundamental difference between raw and cooked data, a binary that my book (building on work by others, such as Lisa Gitelman’s important edited volume “Raw Data” is an Oxymoron) destabilizes. Demographers use the term ‘cooked data’ in two main ways. In the first sense, cooked data are raw data (say, numbers, codes, or other information written onto questionnaires by fieldworkers) that have been processed, ‘cleaned,’ and analyzed according to demographic standards and norms; this kind of cooking is necessary to ensure that data are high quality. In the second moralized sense of “cooking”, however, raw data are deformed, dirty, or made useless as a result of bad data practices, fabrication, or human error. My book contributes to conversations in critical data studies and STS by drawing on long-term ethnographic experiences within survey research projects in Malawi to closely track the social lives of data, in the process illuminating how fieldworkers improvise, reinvent and improve upon top-down standards (a form of ‘cooking’) for data collection as they implement them in the field. From some of the ethnographic vignettes and scenes in the book, I hope readers glean that making good data requires creativity, tinkering, and improvisation as much as it does harmonization and consistency. I hope that my book prompts readers to step away for a moment from arbitrating whether any given data set is good or bad, weak or strong, and toward thinking about how data are constituted by their everyday processes of production and circulation, including social and political contexts. Field research may appear to be simply the systematic collection of information from respondents, but in reality it necessitates a complex and flexible infrastructure of people, equipment, technical and logistical know-how, and so on. Seeing this firsthand can denaturalize the assumptions that data are neutral, objective, or ‘clean’.

Sheng Long: As you say in the book, you are a fieldworker among fieldworkers. Doing research about researchers and fieldworkers, how did you handle your relationship with them? Were there any conflicting moments that led to a compromise of each other?

Crystal Biruk: I am grateful to the demographers and Malawian fieldworkers who were kind enough to let an anthropologist tag along in the field (notably, ‘the field’ carries many of the same connotations for them as it does for anthropologists). I knew that in order to understand the particular cultures of science that make up projects like those I spent time with, it would be important to be part of the everyday and repetitive work of fieldwork. Most days, we woke up very early, got into minibuses or SUVs (approximately 6 fieldworkers per vehicle) and headed to rural areas of a given district in Malawi. Once we arrived to a central point near the households from which we would be collecting data that day, a Malawian field supervisor would hand out questionnaires and give each fieldworker a hand-drawn map—crafted by a fieldworker in a prior iteration of the survey—to help them find their assigned household. At the household, the fieldworker would identify their respondent and engage in the interview/questionnaire for anywhere from 1-3 hours. When finished, they would check the questionnaire for completion or internal inconsistencies and submit it to their supervisor for another check. I helped with these checks (and with other tasks in the office and the field) and, thus, became familiar with the kind of ‘assembly line’ on which statistics are manufactured. Some of the statistical claims—the numbers and percentages—that now appear in published papers in demography journals are made up of data points (say, a number written in pencil on a questionnaire page by a fieldworker) that I myself may have checked! This experience helped me come to see statistics not as abstract, free-floating numbers, but, rather, as social artifacts born of hundreds of relations and exchanges. A lot of my fieldwork involved trying my best to ‘touch’ data, to make the abstract tangible, so as to better see and understand them. The Malawian fieldworkers saw me as a novice fieldworker—I hadn’t, in the beginning, accumulated the many years of expertise and knowledge they had, as most of them spend much of the year working for projects like the ones I describe in the book. This made it harder for me, for example, to ‘check’ questionnaires (e.g., to know whether a response about how many kgs of a crop a household had grown last year was viable or not) rapidly and efficiently earlier on. I was, at the time, of a similar age to the fieldworkers, and because I was helping out in small but useful ways, it was quite easy to build rapport and trust.

As a kind of honorary fieldworker, I was privy to many things that happened in our long days in the field—the kinds of deviations from schedules or set plans that occur in all workplaces—but my allegiance was primarily with the fieldworkers. I think in the beginning they were a bit surprised that I kept coming back every day, and over time, they shared with me their grievances and concerns and gripes about what they called ‘living project to project,’ about the lack of employment in Malawi (most fieldworkers had skills and educational credentials that exceeded those needed to do the kind of data collection work they were engaged in), and about the difficulties of being ‘middle-men’ between foreign projects and rural Malawians. Demographers would sometimes inquire with me why the field teams hadn’t visited as many households as required in a given day, and I would have to navigate these encounters carefully. However, as a kind of ‘middle-person’ myself (between project management and the fieldworkers), I also brought suggestions or critiques from the fieldworkers—say, about the wording of a question on the questionnaire that was not working, for example—to the demographers. I reported back to them some of the challenges faced by fieldworkers, especially pertaining to their struggles in dealing with complicated situations or reluctant participants. Fieldworkers were typically too busy and exhausted from the work of fieldwork to make these suggestions or reports themselves. I thought a lot about reciprocity in doing this work, and did my best to not be ‘dead weight’ by contributing to the labor of data collection. As I built meaningful relationships with the fieldworkers around me, of course, we also became entangled in relations of mutual support and drew on our respective knowledge, networks, and resources to help each other achieve our diverse interests.

Sheng Long: Instead of being mere “respondents” in surveys, you have found that Malawians actively participated in the HIV/AIDS research. On the one hand, their cultural practices were interwoven with the research (such as the gift); On the other, they showed great comprehension of these projects, beyond the mystification of indigenous notions (such as the taboo of blood). How do anthropological methods contribute to a different understanding of the Malawians’ agency in these research projects?

Crystal Biruk: For generations, development and global health’s dominant discourse and practices have presumed African lack, neediness, backwardness, or ignorance. Jemima Pierre (2020) discusses this in great detail in her analysis of the racial vernaculars of development in Ghana. In global health worlds, when Africans refuse to participate in research or other projects, dominant explanations always rest on cultural misunderstandings. When frictions arise between health or development projects and participants, it is often assumed that they can be fixed or ameliorated by implementing more culturally sensitive protocols, translating project materials into local languages, or interfacing with locally influential figures. If researchers properly and thoroughly explain a project’s intent and secure truly informed consent, so the story goes, Africans will willingly and enthusiastically participate. In fact, informed consent is often the gold standard for arbitrating whether research is ethical or not. Yet, I found that in Malawi, even when informed consent was secured in proper fashion—that is, rural Malawians were well-informed of the particulars of participating in the survey project and heartily agreed to do so—tensions and conflicts still arose, primarily around the question of reciprocity between research projects and respondents.

In Chapter 3, I discuss how the bars of soap given by research projects to research participants as a token of thanks for their time and energy became a site of contestation and debate in the field (I also consider the historical resonances of ‘soap’ as quintessential colonial commodity). Even those who gratefully received the soap would comment on the ‘smallness’ of this gift, or suggest they deserved more for the work (using the Chichewa word for labor) they did answering questions. Many suggested they should receive money instead of soap. Further, in commenting on what they saw as lopsided and uneven relations between themselves and researchers, they sometimes accused researchers of being bloodsuckers, or sucking their blood. This accusation fits into a larger transhistorical genre that demonizes dangerous others (colonial officials, researchers, politicians, doctors) who are said to steal or accumulate blood for nefarious ends or to ‘do business.’ These accusations that researchers are something akin to vampires are easily dismissible as conspiracy theories, and often explained by reference to culture, as in “Malawians believe the strangest things” or “Rural Malawians do not understand the value or intent of research”. For me, the potential of ethnography in global health and development worlds is its ability to reveal moments of friction or conflict such as these, and to analyze them outside the narrow frame of a single encounter between one project and a population. In documenting and analyzing critiques of the soap-gift made by those who received it, I observed that participants situated present day projects in historical experiences and legacies of the extraction of data and other resources, pointing out that the benefits that researchers accrue are far greater than those they receive. At first glance, the fraught soap-for-information exchange might seem to highlight cultural tensions between ‘African’ and ‘western’ codes of giving. But upon closer examination, it becomes clear that participants conceive of research as a realm of negotiations over proper distributions of past, present, and future benefits. Rather than cultural misunderstanding or unfounded conspiracy theory, the accusation that researchers are sucking blood or the complaint that a bar of soap is a too small gift are historically informed critiques of a social and political order where some people are always giving and others are always taking. Anthropology of/in global health projects demonstrates that the dynamics that play out in any given project cannot be understood without attending to broader histories wherein data, labor, land, and other resources have been stolen amid broken promises of future gain or benefit.

Sheng Long: Following international demographic research projects, you have traced the traveling of numbers among multiple actors in hierarchical relationships. What are different forms of labor, visible or invisible, involved in the data collection, analyses, and dissemination?

Crystal Biruk:  Indeed, my book tries very much to answer the deceptively simple question, What’s in a number? In so doing, I tried to think at every turn about the materiality and relationality of data, that is, about how any given data point (say, a respondent’s response to the question “How frequently do you use condoms in sexual encounters?”) only comes to exist through nested and complex relations between people, things, and ways of knowing. This is why I arranged the book roughly around what I call the life course of data, from survey design to the circulation of polished statistics in policy forums or published work. Part of my project as I wrote the book was to emphasize the invisible labor of Malawian fieldworkers, whose expertise, as discussed above, was crucial to the smooth running of research projects, yet is rarely credited or acknowledged in publications or policy that result from the data they so painstakingly collect. Since the earliest surveys and research endeavors carried out in Africa, fieldworkers have appeared in archived accounts and discourse as individuals whose ‘menial’ and ‘unskilled’ labor is necessary to field research. Yet, they are framed as instrumentalized as cogs in a larger machinery and homogenized or mentioned offhand as “native assistants” or “data collectors”. I hope my book challenges such depictions and prompts all of us, including anthropologists, to think deeply about how we know what we know, and about how to ‘cite’ and acknowledge all the labor that goes into producing knowledge.

Sheng Long: Citing James Scott’s Seeing Like a State, you similarly frame the medical projects as “seeing like a research project.” What do you observe as the major consequences of “seeing like a research project” in global heath? How do you envision your research as academic and public interventions to these issues and limitations?

Crystal Biruk: Demography, or the quantitative study of human populations, is generally a positivist science rooted in the assumption that reality can be observed, measured, and counted. The questionnaires discussed in my book are tools that aim to do just that. Whereas anthropologists are number averse and harbor suspicions of quantification as a mode of knowing and governance, demographers fetishize numbers. For Scott, seeing like a state means aspiring to make something yet unseen or unknown legible and visible: this transformation from illegible to legible always involves standardization, reduction, and abstraction. The methods used to calculate or organize or map a population or a terrain or a phenomenon always bake in assumptions about who or what is important, about who or what counts, about who or what is valuable (according to Scott, 19th century German foresters’ conception of ‘the forest’ reduced forests—as complex and diverse ecologies and relations—to metricized concepts like revenue yield). In a similar way, demographers build into the design of survey questionnaires assumptions about what is worth knowing, assumptions about what constitutes even the basic units of analysis that underlie the validity of surveys (such as ‘the household’ or ‘poverty’, concepts that, of course, may be defined or imagined differently by, say, an anthropologist!), and assumptions about what good data are. All of these determine, then, what is ‘seen’ or what it is possible to know from the tools, technologies, and epistemologies that constitute quantitative/demographic data. This is why what is seen by an anthropologist may differ from what is seen by a demographic survey project, even if they are looking at the same thing. Yet, the point of my book is not necessarily to elevate anthropological ways of knowing or seeing above other disciplinary modes of knowing. Rather, I hope that my book, in attending closely to the micro-interactions and tacit assumptions and value judgments baked into demographic datasets, might prompt all of us who make, consume, or think about data of diverse kinds to understand them as contingent and partial artifacts of social processes and normalization (‘disciplines discipline us’, as I tell my students!)

As for the consequences of seeing like a research project for global health in particular: as anthropologists and others working in critical global health studies have shown, global health (as assemblage of actors, resources, geographies, and priorities) very much prizes and desires quantitative data such as the numbers discussed in my book. Numbers travel and translate easily and carry an aura of objectivity and self-evident truth that make them a convincing stand in for reality. Yet, when we identify a problem with or through numbers, say high rates of malnutrition, it narrows our imagination of solutions or interventions for that problem. In my current work on the Global Fund’s efforts to ‘end AIDS’ in Africa, for example, I have observed that donors are interested in counting vulnerable populations (such as men who have sex with men, MSM) so as to have a denominator of ‘people at highest risk of HIV’ against which their efforts and interventions can operate. This denominator is the starting point for measuring how many vulnerable people are ‘reached,’ ‘tested,’ or ‘treated.’ Achieving impressive numbers in this regard stand in as proof of success or efficacy, yet, such metrics overlook or exclude from measurement the forms of vulnerability that impinge on, say, an LGBTI-identified person’s life in Malawi. These genres of vulnerability far exceed the viral load or HIV status of that individual. Global health is successful at demonstrating success in targeting narrow, easily quantifiable problems, but, of course, less successful at ‘seeing’ health or other problems as hopelessly entangled in complex contexts and as a product of structural or political failings. In my book, I try to move beyond the anthropological impulse to show, using ethnographic methods, what numbers get wrong or what they overlook. To do so, I examine not so much the finished numbers themselves and what they elide, but rather the criteria and metrics that underscore and determine data’s production and consumption, and measure whether or not they come to ‘count’ as data in the first place. A granular analysis of research worlds in a particular place at a particular time encourages us to more critically engage with the kinds of evidence we too often take for granted, whether inside or outside our discipline or training.

References

Gitelman, Lisa, ed. 2013. “Raw Data” is an Oxymoron. Boston:MIT Press.

Pierre, Jemima. 2020. “The racial vernaculars of development: A view from West Africa.” American Anthropologist 122(1):86-98.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s