Advancing the Value of Ethnography

Creating Ethnography


What is an anthropologist? What does an ethnographer actually do? I used to believe that my own answers to these questions were sufficient. In reality, however, the existential dilemma at the foundation of any institution—academic, professional, or otherwise—is a socially constructed affair. In other words, whether I want to admit it or not, my answers are partially your answers—for that matter, all of our collective answers.

Indeed, the very existence of a mutually shared set of practical assumptions about ethnography is what makes these questions so important. Meanings are contested, negotiated, and (if you believe Berger and Luckmann1) thereby constitutive of the agreed-upon fiction we call “reality.” Most of us might agree that we are, more or less, the biographers of that fiction. We are interested in how it comes into being, what sustains it, what motivates it, and how it responds to challenges. We pluck assumptions from reality and sell them to clients. Together, this activity constitutes its own reality—a reality of what ethnography in industry is.

Sam Ladner (Microsoft) at the EPIC2014 salon Debating Ethnographic Epistemology: Should Corporate Ethnography Even Attempt To Avoid Reductionism?

Sam Ladner (Microsoft) at the EPIC2014 salon Debating Ethnographic Epistemology:
Should Corporate Ethnography Even Attempt To Avoid Reductionism?

So what is it? I wanted to find out. Last year in New York, Christian Madsbjerg and I co-facilitated an EPIC Salon session, asking over fifty EPIC participants how they do ethnography. It was a nuanced, satisfying conversation. In addition to making me feel even more grateful to be a part of this community, listening to other researchers was useful fieldwork. It was an opportunity to unpack the constructed reality of our research practice and reflect on the professional ethnographer as subject.

There is an abundance of conversation about ethnographic projects and ethnography in this or that context, but comparatively little discussion of what exactly is meant by ethnography. People are tired of that conversation. It’s messy. It’s challenging. We don’t like to be challenged. Not where we live anyway. So we treat it as a foregone conclusion, a means to an end, or a form of performance—ethnography is “whatever I want it to be” or “what I make of it.” If my ethnographic practice is a form of self-expression, then questioning my practice means questioning me. And we just don’t do that in polite society. It’s a sensitive subject.

Lively discussion at the salon.

Lively discussion at the salon.

In the spirit of disclosure, I have made no secret of how horrifying I find the notion that ethnography is “whatever.” Nor have I veiled my objection to the idea that each ethnographer’s methods are an exponent of their individual personality or predilection. But am I just an academic curmudgeon? Clinging to his slide-rules and journal articles?

On my best days, I regard myself as a scientist. On other days, I might describe myself as a dilettante, but an honest dilettante. One of the ways I try to maintain that honesty is through transparency. I try to make my methods and assumptions transparent. I source the origins of my concepts and ideas. My expectation is that these references will make my work replicable and, if necessary, falsifiable.

I am certainly not alone in such endeavors—these are the ritual practices that constitute the reality of science. For example, in Pandora’s Hope2, Bruno Latour demonstrates that the scientific enterprise, however socially constructed, establishes its validity through chains of reference. For example, let’s look at maps: a two-by-two square meter of dirt becomes an intersecting point of latitude and longitude, that point on the map becomes part of a vector, that vector becomes a grid, the grid becomes a projection, and that projection, a map. And the map? It becomes “objective truth.”

But there is no objective truth in science, Latour contends, just reference. The “truth” of map is only as valid as the chains of reference, back to the projection, back to the vector—yes, all the way back to the dirt. Each chain in that reference can be questioned, interrogated, falsified, or improved. Together, these intersecting chains form a delicate lattice that we regard as truth. Latour’s proposition is not altogether different from what Wittgenstein3 wrote about language and meaning.

But beyond that, Latour uncovers the hidden strength of science: it lends itself to interrogation and falsification—in fact, questioning strengthens it. But this is only possible if our methods and findings come out of our head and into reality, where they can be questioned, interrogated, and, if warranted, they can be falsified.

Therefore, what I am proposing is a conversation about our methods, and delineation between scientifically valid and invalid methods. That is to say, we should develop ethnographic research that can be tested, replicated, and questioned. Ethnography should be a form of research in which everyone can participate critically, as part of a dialogue. That dialogue is over the moment ethnography becomes “whatever I make of it.” At that point, it becomes something else: a stifling conversation about neither the data, nor the analysis, but the ethnographer and their self-representation.

Specifically, by “tested, replicated, and questioned” I mean that when we make claims to have “predicted” a socio-cultural trend, or product, or technology, we should become responsible for explaining the “goodness of fit” of our model to the data, and what proportion of the variation in our population—be it a socio-cultural group, or group of consumers, or volunteer subjects—our model explains. We should be able to apply the same ethnographic methods to actually predict something using new data. Moreover, chains of reference from the individual subject to the larger society must be considered and elucidated. It should be our obligation to ask if a given person is a sufficient example of the assumptions that make a certain form of society possible. The meaning encoded in phenomenological artifacts like cultural practices or product experiences must be decoded, and then compiled, and tested on representative cases.

Informed certainty should be our salesmanship. Not PowerPoint decks, or clever neologisms, or hip gimmicks. Or as close to certainty as the scientific process will legitimately allow.

For those who think this sort of thing is impossible, I leave you with Tom Fricke’s4 demonstration of the scientific merit of ethnographic methods applied in the manner discussed above. Fricke is a demographer, the quantiest of quanty sociologists—a discipline so boring, in fact, that other quanty sociologists find it boring. But it also tackles some of the most fundamental and important questions of our time, and operates at immense timescales. It is truly fascinating. But if you’re looking for a hard-nosed, empirical branch of sociology, this is it. Why I mention all of this will be clear in a moment.

Fricke opens by discussing the absence of an anthropological contribution to the explanation of cultural change observed in a few specific demographic studies, such as the worldwide decline in fertility observed in the last fifty years. There are many theories of intergenerational wealth and resource exchange, and the effects of education. However, the rational processes underlying these theories bear cultural scrutiny.

Fricke demonstrates how the application of ethnographic methods not only explains population change, but challenges prevailing assumptions and moves the scientific enterprise forward. According to Fricke, researchers found women in certain parts of Punjab were giving birth later in life due to higher marital age. They further linked higher marital age to education level: women were pursuing education and delaying marriage. This was, naturally, interpreted as a surefire sign of cultural change within the developing world.

But a team of anthropologists, experts on Punjabi culture, drew very different conclusions after conducting local fieldwork. They discovered local Punjabis indeed valued education, but the cultural value of education effectively made educated women more desirable as wives. Thus, women with even limited education found themselves highly sought after by families seeking to arrange marriages for their sons. The result was that a significant number of women attended school only briefly. This introduces a short delay between secondary education and marriage not observed in previous generations, or in the case of uneducated women, but the fundamental cultural process remained essentially intact. The cultural change story fell apart.

Two things are remarkable to me about this research: first, it demonstrates that researchers cannot understand the entirety of the demographic process under consideration without fieldwork. Second, ethnography informs quantitative analysis in critical ways, generating new, testable hypotheses. This is, fundamentally, the scientific process. It is a possibility inherent in what we do—and one that can be easily communicated to clients.

Currently, many ethnographers situate themselves in a defensive posture when they encounter challenges to the validity and generalizability of qualitative work. Moreover, this tends to be regarded as a political or perceptual problem that should be solved with salesmanship and client management rather than through the application of scientific principles consistent with ethnography itself.

It doesn’t need to be that way. As Fricke’s work demonstrates, our methods are not only capable of, but often do, contribute falsifiable hypotheses that overturn and inform large-scale statistical work and can be verified experimentally. Malinowski, Radcliffe-Brown, and other founders designed ethnography with questions concerning generalization and predictive validity in mind.

The problematic position of the modern industrial ethnographer does not exist because our methods are not intended to, or cannot, answer these questions—it is due to our lack of scientific rigor and theoretical grounding. When one applies ethnography with attention to the scientific standards established by previous scholarly generations, it informs, illuminates, and unveils.

So let’s have our conversations about data, not about ourselves. Let’s debate interpretations, and analysis. Let’s use our work to formulate, test, and invalidate hypotheses. Let’s become responsible for our assumptions. Let’s make it messy, let’s dig in the proverbial dirt. Let’s be scientists. Standards and methods are not the end of anthropological work—they are the foundation, and the promise of what comes next.


1. Berger, Peter L. and Luckmann, Thomas. 1966. The Social Construction of Reality: A Treatise on the Sociology of Knowledge. New York, NY: Anchor Books.

2. Latour, Bruno. 1999. Pandora’s Hope: Essays on the Reality of Science Studies. Cambridge, MA: Harvard University Press.

3. Anderson, Harlene. 2007. “A Postmodern Umbrella: Language and Knowledge as Relational and Generative, and Inherently Transforming.” In Collaborative Therapy: Relationships and Conversations that Make a Difference. Harlene Anderson and Diane R. Gehart, Eds. Pp 7-19. New York, NY: Routledge.

4. Fricke, Tom. 2003. “Culture and Causality: An Anthropological Comment.” Population and Development Review 29(3): 470-479.

Neal Patel working on equations at a white board.



Neal H. Patel, Google

Neal H. Patel leads the Human/Social Dynamics Program at Google’s Advanced Technology and Projects (ATAP) team. His most recent publications include a chapter in The Handbook of Anthropology in Business, available from Left Coast Press.